* [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement
@ 2011-05-23 21:42 Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef Aurelien Jarno
` (8 more replies)
0 siblings, 9 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
This patch series finally removes softfloat-native as it is not used by
default in any target. In then laters improves the target-i386 code as
it does not need to be compatible with softfloat-native anymore.
v1 -> v2:
fix mips/gdbstub
reorder patches to fix target-i386 bisection
dropped log2 patches
Aurelien Jarno (9):
target-ppc: remove old CONFIG_SOFTFLOAT #ifdef
target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef
target-i386: remove old code handling float64
softfloat-native: remove
softfloat: always enable floatx80 and float128 support
target-i386: use floatx80 constants in helper_fld*_ST0()
softfloat: add float*_is_zero_or_denormal()
target-i386: cleanup helper_fxam_ST0()
target-i386: add support for FPU exceptions
Makefile.target | 3 +-
configure | 2 -
cpu-all.h | 7 +-
fpu/softfloat-native.c | 540 --------------------------------------------
fpu/softfloat-native.h | 531 -------------------------------------------
fpu/softfloat-specialize.h | 7 -
fpu/softfloat.c | 62 -----
fpu/softfloat.h | 76 ++-----
gdbstub.c | 4 -
target-i386/cpu.h | 37 ++--
target-i386/exec.h | 144 +------------
target-i386/helper.c | 41 +++-
target-i386/machine.c | 104 +---------
target-i386/op_helper.c | 329 +++++++++++----------------
target-ppc/helper.h | 2 -
target-ppc/op_helper.c | 11 -
target-ppc/translate.c | 2 -
17 files changed, 208 insertions(+), 1694 deletions(-)
delete mode 100644 fpu/softfloat-native.c
delete mode 100644 fpu/softfloat-native.h
--
1.7.5.1
^ permalink raw reply [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-29 10:58 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef Aurelien Jarno
` (7 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Alexander Graf, Aurelien Jarno
target-ppc has been switched to softfloat only long ago, but a
few #ifdef CONFIG_SOFTFLOAT have been forgotten. Remove them.
Cc: Alexander Graf <agraf@suse.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
target-ppc/helper.h | 2 --
target-ppc/op_helper.c | 11 -----------
target-ppc/translate.c | 2 --
3 files changed, 0 insertions(+), 15 deletions(-)
diff --git a/target-ppc/helper.h b/target-ppc/helper.h
index 51c99c8..470e42f 100644
--- a/target-ppc/helper.h
+++ b/target-ppc/helper.h
@@ -51,9 +51,7 @@ DEF_HELPER_FLAGS_1(cntlzw32, TCG_CALL_CONST | TCG_CALL_PURE, i32, i32)
DEF_HELPER_FLAGS_2(brinc, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
DEF_HELPER_0(float_check_status, void)
-#ifdef CONFIG_SOFTFLOAT
DEF_HELPER_0(reset_fpstatus, void)
-#endif
DEF_HELPER_2(compute_fprf, i32, i64, i32)
DEF_HELPER_2(store_fpscr, void, i64, i32)
DEF_HELPER_1(fpscr_clrbit, void, i32)
diff --git a/target-ppc/op_helper.c b/target-ppc/op_helper.c
index c6cb448..fe66f14 100644
--- a/target-ppc/op_helper.c
+++ b/target-ppc/op_helper.c
@@ -971,7 +971,6 @@ void helper_store_fpscr (uint64_t arg, uint32_t mask)
void helper_float_check_status (void)
{
-#ifdef CONFIG_SOFTFLOAT
if (env->exception_index == POWERPC_EXCP_PROGRAM &&
(env->error_code & POWERPC_EXCP_FP)) {
/* Differred floating-point exception after target FPR update */
@@ -989,22 +988,12 @@ void helper_float_check_status (void)
float_inexact_excp();
}
}
-#else
- if (env->exception_index == POWERPC_EXCP_PROGRAM &&
- (env->error_code & POWERPC_EXCP_FP)) {
- /* Differred floating-point exception after target FPR update */
- if (msr_fe0 != 0 || msr_fe1 != 0)
- helper_raise_exception_err(env->exception_index, env->error_code);
- }
-#endif
}
-#ifdef CONFIG_SOFTFLOAT
void helper_reset_fpstatus (void)
{
set_float_exception_flags(0, &env->fp_status);
}
-#endif
/* fadd - fadd. */
uint64_t helper_fadd (uint64_t arg1, uint64_t arg2)
diff --git a/target-ppc/translate.c b/target-ppc/translate.c
index 9b3f90c..59aef85 100644
--- a/target-ppc/translate.c
+++ b/target-ppc/translate.c
@@ -215,9 +215,7 @@ struct opc_handler_t {
static inline void gen_reset_fpstatus(void)
{
-#ifdef CONFIG_SOFTFLOAT
gen_helper_reset_fpstatus();
-#endif
}
static inline void gen_compute_fprf(TCGv_i64 arg, int set_fprf, int set_rc)
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 22:09 ` Peter Maydell
2011-05-29 11:01 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64 Aurelien Jarno
` (6 subsequent siblings)
8 siblings, 2 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
target-mips has been switched to softfloat only long ago, but
a #ifndef CONFIG_SOFTFLOAT has been forgotten. Remove it.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
gdbstub.c | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)
diff --git a/gdbstub.c b/gdbstub.c
index ae856f9..b9ae30d 100644
--- a/gdbstub.c
+++ b/gdbstub.c
@@ -1105,10 +1105,6 @@ static int cpu_gdb_write_register(CPUState *env, uint8_t *mem_buf, int n)
env->active_fpu.fcr31 = tmp & 0xFF83FFFF;
/* set rounding mode */
RESTORE_ROUNDING_MODE;
-#ifndef CONFIG_SOFTFLOAT
- /* no floating point exception for native float */
- SET_FP_ENABLE(env->active_fpu.fcr31, 0);
-#endif
break;
case 71: env->active_fpu.fcr0 = tmp; break;
}
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-24 15:12 ` Peter Maydell
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 4/9] softfloat-native: remove Aurelien Jarno
` (5 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
Now that target-i386 uses softfloat, floatx80 is always available and
there is no need anymore to have code handling both float64 and floax80.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
target-i386/cpu.h | 22 +---
target-i386/exec.h | 132 +-----------------------
target-i386/helper.c | 5 -
target-i386/machine.c | 101 ------------------
target-i386/op_helper.c | 266 +++++++++++++++++++----------------------------
5 files changed, 115 insertions(+), 411 deletions(-)
diff --git a/target-i386/cpu.h b/target-i386/cpu.h
index 715828f..fe65886 100644
--- a/target-i386/cpu.h
+++ b/target-i386/cpu.h
@@ -532,16 +532,6 @@ enum {
CC_OP_NB,
};
-#ifdef FLOATX80
-#define USE_X86LDOUBLE
-#endif
-
-#ifdef USE_X86LDOUBLE
-typedef floatx80 CPU86_LDouble;
-#else
-typedef float64 CPU86_LDouble;
-#endif
-
typedef struct SegmentCache {
uint32_t selector;
target_ulong base;
@@ -594,11 +584,7 @@ typedef union {
#define MMX_Q(n) q
typedef union {
-#ifdef USE_X86LDOUBLE
- CPU86_LDouble d __attribute__((aligned(16)));
-#else
- CPU86_LDouble d;
-#endif
+ floatx80 d __attribute__((aligned(16)));
MMXReg mmx;
} FPReg;
@@ -654,7 +640,7 @@ typedef struct CPUX86State {
/* emulator internal variables */
float_status fp_status;
- CPU86_LDouble ft0;
+ floatx80 ft0;
float_status mmx_status; /* for 3DNow! float ops */
float_status sse_status;
@@ -865,8 +851,8 @@ static inline void cpu_x86_set_cpl(CPUX86State *s, int cpl)
/* op_helper.c */
/* used for debug or cpu save/restore */
-void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, CPU86_LDouble f);
-CPU86_LDouble cpu_set_fp80(uint64_t mant, uint16_t upper);
+void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, floatx80 f);
+floatx80 cpu_set_fp80(uint64_t mant, uint16_t upper);
/* cpu-exec.c */
/* the following helpers are only usable in user mode simulation as
diff --git a/target-i386/exec.h b/target-i386/exec.h
index ee36a71..9bd080e 100644
--- a/target-i386/exec.h
+++ b/target-i386/exec.h
@@ -98,67 +98,6 @@ static inline void svm_check_intercept(uint32_t type)
#endif /* !defined(CONFIG_USER_ONLY) */
-#ifdef USE_X86LDOUBLE
-/* use long double functions */
-#define floatx_to_int32 floatx80_to_int32
-#define floatx_to_int64 floatx80_to_int64
-#define floatx_to_int32_round_to_zero floatx80_to_int32_round_to_zero
-#define floatx_to_int64_round_to_zero floatx80_to_int64_round_to_zero
-#define int32_to_floatx int32_to_floatx80
-#define int64_to_floatx int64_to_floatx80
-#define float32_to_floatx float32_to_floatx80
-#define float64_to_floatx float64_to_floatx80
-#define floatx_to_float32 floatx80_to_float32
-#define floatx_to_float64 floatx80_to_float64
-#define floatx_add floatx80_add
-#define floatx_div floatx80_div
-#define floatx_mul floatx80_mul
-#define floatx_sub floatx80_sub
-#define floatx_sqrt floatx80_sqrt
-#define floatx_abs floatx80_abs
-#define floatx_chs floatx80_chs
-#define floatx_scalbn floatx80_scalbn
-#define floatx_round_to_int floatx80_round_to_int
-#define floatx_compare floatx80_compare
-#define floatx_compare_quiet floatx80_compare_quiet
-#define floatx_is_any_nan floatx80_is_any_nan
-#define floatx_is_neg floatx80_is_neg
-#define floatx_is_zero floatx80_is_zero
-#define floatx_zero floatx80_zero
-#define floatx_one floatx80_one
-#define floatx_ln2 floatx80_ln2
-#define floatx_pi floatx80_pi
-#else
-#define floatx_to_int32 float64_to_int32
-#define floatx_to_int64 float64_to_int64
-#define floatx_to_int32_round_to_zero float64_to_int32_round_to_zero
-#define floatx_to_int64_round_to_zero float64_to_int64_round_to_zero
-#define int32_to_floatx int32_to_float64
-#define int64_to_floatx int64_to_float64
-#define float32_to_floatx float32_to_float64
-#define float64_to_floatx(x, e) (x)
-#define floatx_to_float32 float64_to_float32
-#define floatx_to_float64(x, e) (x)
-#define floatx_add float64_add
-#define floatx_div float64_div
-#define floatx_mul float64_mul
-#define floatx_sub float64_sub
-#define floatx_sqrt float64_sqrt
-#define floatx_abs float64_abs
-#define floatx_chs float64_chs
-#define floatx_scalbn float64_scalbn
-#define floatx_round_to_int float64_round_to_int
-#define floatx_compare float64_compare
-#define floatx_compare_quiet float64_compare_quiet
-#define floatx_is_any_nan float64_is_any_nan
-#define floatx_is_neg float64_is_neg
-#define floatx_is_zero float64_is_zero
-#define floatx_zero float64_zero
-#define floatx_one float64_one
-#define floatx_ln2 float64_ln2
-#define floatx_pi float64_pi
-#endif
-
#define RC_MASK 0xc00
#define RC_NEAR 0x000
#define RC_DOWN 0x400
@@ -167,11 +106,6 @@ static inline void svm_check_intercept(uint32_t type)
#define MAXTAN 9223372036854775808.0
-#ifdef USE_X86LDOUBLE
-
-/* only for x86 */
-typedef CPU_LDoubleU CPU86_LDoubleU;
-
/* the following deal with x86 long double-precision numbers */
#define MAXEXPD 0x7fff
#define EXPBIAS 16383
@@ -180,23 +114,6 @@ typedef CPU_LDoubleU CPU86_LDoubleU;
#define MANTD(fp) (fp.l.lower)
#define BIASEXPONENT(fp) fp.l.upper = (fp.l.upper & ~(0x7fff)) | EXPBIAS
-#else
-
-typedef CPU_DoubleU CPU86_LDoubleU;
-
-/* the following deal with IEEE double-precision numbers */
-#define MAXEXPD 0x7ff
-#define EXPBIAS 1023
-#define EXPD(fp) (((fp.l.upper) >> 20) & 0x7FF)
-#define SIGND(fp) ((fp.l.upper) & 0x80000000)
-#ifdef __arm__
-#define MANTD(fp) (fp.l.lower | ((uint64_t)(fp.l.upper & ((1 << 20) - 1)) << 32))
-#else
-#define MANTD(fp) (fp.ll & ((1LL << 52) - 1))
-#endif
-#define BIASEXPONENT(fp) fp.l.upper = (fp.l.upper & ~(0x7ff << 20)) | (EXPBIAS << 20)
-#endif
-
static inline void fpush(void)
{
env->fpstt = (env->fpstt - 1) & 7;
@@ -209,65 +126,24 @@ static inline void fpop(void)
env->fpstt = (env->fpstt + 1) & 7;
}
-#ifndef USE_X86LDOUBLE
-static inline CPU86_LDouble helper_fldt(target_ulong ptr)
-{
- CPU86_LDoubleU temp;
- int upper, e;
- uint64_t ll;
-
- /* mantissa */
- upper = lduw(ptr + 8);
- /* XXX: handle overflow ? */
- e = (upper & 0x7fff) - 16383 + EXPBIAS; /* exponent */
- e |= (upper >> 4) & 0x800; /* sign */
- ll = (ldq(ptr) >> 11) & ((1LL << 52) - 1);
-#ifdef __arm__
- temp.l.upper = (e << 20) | (ll >> 32);
- temp.l.lower = ll;
-#else
- temp.ll = ll | ((uint64_t)e << 52);
-#endif
- return temp.d;
-}
-
-static inline void helper_fstt(CPU86_LDouble f, target_ulong ptr)
+static inline floatx80 helper_fldt(target_ulong ptr)
{
- CPU86_LDoubleU temp;
- int e;
-
- temp.d = f;
- /* mantissa */
- stq(ptr, (MANTD(temp) << 11) | (1LL << 63));
- /* exponent + sign */
- e = EXPD(temp) - EXPBIAS + 16383;
- e |= SIGND(temp) >> 16;
- stw(ptr + 8, e);
-}
-#else
-
-/* we use memory access macros */
-
-static inline CPU86_LDouble helper_fldt(target_ulong ptr)
-{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
temp.l.lower = ldq(ptr);
temp.l.upper = lduw(ptr + 8);
return temp.d;
}
-static inline void helper_fstt(CPU86_LDouble f, target_ulong ptr)
+static inline void helper_fstt(floatx80 f, target_ulong ptr)
{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
temp.d = f;
stq(ptr, temp.l.lower);
stw(ptr + 8, temp.l.upper);
}
-#endif /* USE_X86LDOUBLE */
-
#define FPUS_IE (1 << 0)
#define FPUS_DE (1 << 1)
#define FPUS_ZE (1 << 2)
diff --git a/target-i386/helper.c b/target-i386/helper.c
index 89df997..5c4b288 100644
--- a/target-i386/helper.c
+++ b/target-i386/helper.c
@@ -403,15 +403,10 @@ void cpu_dump_state(CPUState *env, FILE *f, fprintf_function cpu_fprintf,
fptag,
env->mxcsr);
for(i=0;i<8;i++) {
-#if defined(USE_X86LDOUBLE)
CPU_LDoubleU u;
u.d = env->fpregs[i].d;
cpu_fprintf(f, "FPR%d=%016" PRIx64 " %04x",
i, u.l.lower, u.l.upper);
-#else
- cpu_fprintf(f, "FPR%d=%016" PRIx64,
- i, env->fpregs[i].mmx.q);
-#endif
if ((i & 1) == 1)
cpu_fprintf(f, "\n");
else
diff --git a/target-i386/machine.c b/target-i386/machine.c
index d78eceb..bbeae88 100644
--- a/target-i386/machine.c
+++ b/target-i386/machine.c
@@ -84,7 +84,6 @@ static void put_fpreg_error(QEMUFile *f, void *opaque, size_t size)
exit(0);
}
-#ifdef USE_X86LDOUBLE
/* XXX: add that in a FPU generic layer */
union x86_longdouble {
uint64_t mant;
@@ -202,102 +201,6 @@ static bool fpregs_is_1_no_mmx(void *opaque, int version_id)
VMSTATE_ARRAY_TEST(_field, _state, _n, fpregs_is_1_mmx, vmstate_fpreg_1_mmx, FPReg), \
VMSTATE_ARRAY_TEST(_field, _state, _n, fpregs_is_1_no_mmx, vmstate_fpreg_1_no_mmx, FPReg)
-#else
-static int get_fpreg(QEMUFile *f, void *opaque, size_t size)
-{
- FPReg *fp_reg = opaque;
-
- qemu_get_be64s(f, &fp_reg->mmx.MMX_Q(0));
- return 0;
-}
-
-static void put_fpreg(QEMUFile *f, void *opaque, size_t size)
-{
- FPReg *fp_reg = opaque;
- /* if we use doubles for float emulation, we save the doubles to
- avoid losing information in case of MMX usage. It can give
- problems if the image is restored on a CPU where long
- doubles are used instead. */
- qemu_put_be64s(f, &fp_reg->mmx.MMX_Q(0));
-}
-
-const VMStateInfo vmstate_fpreg = {
- .name = "fpreg",
- .get = get_fpreg,
- .put = put_fpreg,
-};
-
-static int get_fpreg_0_mmx(QEMUFile *f, void *opaque, size_t size)
-{
- FPReg *fp_reg = opaque;
- uint64_t mant;
- uint16_t exp;
-
- qemu_get_be64s(f, &mant);
- qemu_get_be16s(f, &exp);
- fp_reg->mmx.MMX_Q(0) = mant;
- return 0;
-}
-
-const VMStateInfo vmstate_fpreg_0_mmx = {
- .name = "fpreg_0_mmx",
- .get = get_fpreg_0_mmx,
- .put = put_fpreg_error,
-};
-
-static int get_fpreg_0_no_mmx(QEMUFile *f, void *opaque, size_t size)
-{
- FPReg *fp_reg = opaque;
- uint64_t mant;
- uint16_t exp;
-
- qemu_get_be64s(f, &mant);
- qemu_get_be16s(f, &exp);
-
- fp_reg->d = cpu_set_fp80(mant, exp);
- return 0;
-}
-
-const VMStateInfo vmstate_fpreg_0_no_mmx = {
- .name = "fpreg_0_no_mmx",
- .get = get_fpreg_0_no_mmx,
- .put = put_fpreg_error,
-};
-
-static bool fpregs_is_1(void *opaque, int version_id)
-{
- CPUState *env = opaque;
-
- return env->fpregs_format_vmstate == 1;
-}
-
-static bool fpregs_is_0_mmx(void *opaque, int version_id)
-{
- CPUState *env = opaque;
- int guess_mmx;
-
- guess_mmx = ((env->fptag_vmstate == 0xff) &&
- (env->fpus_vmstate & 0x3800) == 0);
- return guess_mmx && env->fpregs_format_vmstate == 0;
-}
-
-static bool fpregs_is_0_no_mmx(void *opaque, int version_id)
-{
- CPUState *env = opaque;
- int guess_mmx;
-
- guess_mmx = ((env->fptag_vmstate == 0xff) &&
- (env->fpus_vmstate & 0x3800) == 0);
- return !guess_mmx && env->fpregs_format_vmstate == 0;
-}
-
-#define VMSTATE_FP_REGS(_field, _state, _n) \
- VMSTATE_ARRAY_TEST(_field, _state, _n, fpregs_is_1, vmstate_fpreg, FPReg), \
- VMSTATE_ARRAY_TEST(_field, _state, _n, fpregs_is_0_mmx, vmstate_fpreg_0_mmx, FPReg), \
- VMSTATE_ARRAY_TEST(_field, _state, _n, fpregs_is_0_no_mmx, vmstate_fpreg_0_no_mmx, FPReg)
-
-#endif /* USE_X86LDOUBLE */
-
static bool version_is_5(void *opaque, int version_id)
{
return version_id == 5;
@@ -344,11 +247,7 @@ static void cpu_pre_save(void *opaque)
env->fptag_vmstate |= ((!env->fptags[i]) << i);
}
-#ifdef USE_X86LDOUBLE
env->fpregs_format_vmstate = 0;
-#else
- env->fpregs_format_vmstate = 1;
-#endif
}
static int cpu_post_load(void *opaque, int version_id)
diff --git a/target-i386/op_helper.c b/target-i386/op_helper.c
index 3c539f3..4d309ab 100644
--- a/target-i386/op_helper.c
+++ b/target-i386/op_helper.c
@@ -95,25 +95,19 @@ static const uint8_t rclb_table[32] = {
6, 7, 8, 0, 1, 2, 3, 4,
};
-#if defined(CONFIG_SOFTFLOAT)
-# define floatx_lg2 make_floatx80( 0x3ffd, 0x9a209a84fbcff799LL )
-# define floatx_l2e make_floatx80( 0x3fff, 0xb8aa3b295c17f0bcLL )
-# define floatx_l2t make_floatx80( 0x4000, 0xd49a784bcd1b8afeLL )
-#else
-# define floatx_lg2 (0.30102999566398119523L)
-# define floatx_l2e (1.44269504088896340739L)
-# define floatx_l2t (3.32192809488736234781L)
-#endif
-
-static const CPU86_LDouble f15rk[7] =
-{
- floatx_zero,
- floatx_one,
- floatx_pi,
- floatx_lg2,
- floatx_ln2,
- floatx_l2e,
- floatx_l2t,
+#define floatx80_lg2 make_floatx80( 0x3ffd, 0x9a209a84fbcff799LL )
+#define floatx80_l2e make_floatx80( 0x3fff, 0xb8aa3b295c17f0bcLL )
+#define floatx80_l2t make_floatx80( 0x4000, 0xd49a784bcd1b8afeLL )
+
+static const floatx80 f15rk[7] =
+{
+ floatx80_zero,
+ floatx80_one,
+ floatx80_pi,
+ floatx80_lg2,
+ floatx80_ln2,
+ floatx80_l2e,
+ floatx80_l2t,
};
/* broken thread support */
@@ -3442,18 +3436,18 @@ void helper_verw(target_ulong selector1)
/* x87 FPU helpers */
-static inline double CPU86_LDouble_to_double(CPU86_LDouble a)
+static inline double floatx80_to_double(floatx80 a)
{
union {
float64 f64;
double d;
} u;
- u.f64 = floatx_to_float64(a, &env->fp_status);
+ u.f64 = floatx80_to_float64(a, &env->fp_status);
return u.d;
}
-static inline CPU86_LDouble double_to_CPU86_LDouble(double a)
+static inline floatx80 double_to_floatx80(double a)
{
union {
float64 f64;
@@ -3461,7 +3455,7 @@ static inline CPU86_LDouble double_to_CPU86_LDouble(double a)
} u;
u.d = a;
- return float64_to_floatx(u.f64, &env->fp_status);
+ return float64_to_floatx80(u.f64, &env->fp_status);
}
static void fpu_set_exception(int mask)
@@ -3471,12 +3465,12 @@ static void fpu_set_exception(int mask)
env->fpus |= FPUS_SE | FPUS_B;
}
-static inline CPU86_LDouble helper_fdiv(CPU86_LDouble a, CPU86_LDouble b)
+static inline floatx80 helper_fdiv(floatx80 a, floatx80 b)
{
- if (floatx_is_zero(b)) {
+ if (floatx80_is_zero(b)) {
fpu_set_exception(FPUS_ZE);
}
- return floatx_div(a, b, &env->fp_status);
+ return floatx80_div(a, b, &env->fp_status);
}
static void fpu_raise_exception(void)
@@ -3498,7 +3492,7 @@ void helper_flds_FT0(uint32_t val)
uint32_t i;
} u;
u.i = val;
- FT0 = float32_to_floatx(u.f, &env->fp_status);
+ FT0 = float32_to_floatx80(u.f, &env->fp_status);
}
void helper_fldl_FT0(uint64_t val)
@@ -3508,12 +3502,12 @@ void helper_fldl_FT0(uint64_t val)
uint64_t i;
} u;
u.i = val;
- FT0 = float64_to_floatx(u.f, &env->fp_status);
+ FT0 = float64_to_floatx80(u.f, &env->fp_status);
}
void helper_fildl_FT0(int32_t val)
{
- FT0 = int32_to_floatx(val, &env->fp_status);
+ FT0 = int32_to_floatx80(val, &env->fp_status);
}
void helper_flds_ST0(uint32_t val)
@@ -3525,7 +3519,7 @@ void helper_flds_ST0(uint32_t val)
} u;
new_fpstt = (env->fpstt - 1) & 7;
u.i = val;
- env->fpregs[new_fpstt].d = float32_to_floatx(u.f, &env->fp_status);
+ env->fpregs[new_fpstt].d = float32_to_floatx80(u.f, &env->fp_status);
env->fpstt = new_fpstt;
env->fptags[new_fpstt] = 0; /* validate stack entry */
}
@@ -3539,7 +3533,7 @@ void helper_fldl_ST0(uint64_t val)
} u;
new_fpstt = (env->fpstt - 1) & 7;
u.i = val;
- env->fpregs[new_fpstt].d = float64_to_floatx(u.f, &env->fp_status);
+ env->fpregs[new_fpstt].d = float64_to_floatx80(u.f, &env->fp_status);
env->fpstt = new_fpstt;
env->fptags[new_fpstt] = 0; /* validate stack entry */
}
@@ -3548,7 +3542,7 @@ void helper_fildl_ST0(int32_t val)
{
int new_fpstt;
new_fpstt = (env->fpstt - 1) & 7;
- env->fpregs[new_fpstt].d = int32_to_floatx(val, &env->fp_status);
+ env->fpregs[new_fpstt].d = int32_to_floatx80(val, &env->fp_status);
env->fpstt = new_fpstt;
env->fptags[new_fpstt] = 0; /* validate stack entry */
}
@@ -3557,7 +3551,7 @@ void helper_fildll_ST0(int64_t val)
{
int new_fpstt;
new_fpstt = (env->fpstt - 1) & 7;
- env->fpregs[new_fpstt].d = int64_to_floatx(val, &env->fp_status);
+ env->fpregs[new_fpstt].d = int64_to_floatx80(val, &env->fp_status);
env->fpstt = new_fpstt;
env->fptags[new_fpstt] = 0; /* validate stack entry */
}
@@ -3568,7 +3562,7 @@ uint32_t helper_fsts_ST0(void)
float32 f;
uint32_t i;
} u;
- u.f = floatx_to_float32(ST0, &env->fp_status);
+ u.f = floatx80_to_float32(ST0, &env->fp_status);
return u.i;
}
@@ -3578,14 +3572,14 @@ uint64_t helper_fstl_ST0(void)
float64 f;
uint64_t i;
} u;
- u.f = floatx_to_float64(ST0, &env->fp_status);
+ u.f = floatx80_to_float64(ST0, &env->fp_status);
return u.i;
}
int32_t helper_fist_ST0(void)
{
int32_t val;
- val = floatx_to_int32(ST0, &env->fp_status);
+ val = floatx80_to_int32(ST0, &env->fp_status);
if (val != (int16_t)val)
val = -32768;
return val;
@@ -3594,21 +3588,21 @@ int32_t helper_fist_ST0(void)
int32_t helper_fistl_ST0(void)
{
int32_t val;
- val = floatx_to_int32(ST0, &env->fp_status);
+ val = floatx80_to_int32(ST0, &env->fp_status);
return val;
}
int64_t helper_fistll_ST0(void)
{
int64_t val;
- val = floatx_to_int64(ST0, &env->fp_status);
+ val = floatx80_to_int64(ST0, &env->fp_status);
return val;
}
int32_t helper_fistt_ST0(void)
{
int32_t val;
- val = floatx_to_int32_round_to_zero(ST0, &env->fp_status);
+ val = floatx80_to_int32_round_to_zero(ST0, &env->fp_status);
if (val != (int16_t)val)
val = -32768;
return val;
@@ -3617,14 +3611,14 @@ int32_t helper_fistt_ST0(void)
int32_t helper_fisttl_ST0(void)
{
int32_t val;
- val = floatx_to_int32_round_to_zero(ST0, &env->fp_status);
+ val = floatx80_to_int32_round_to_zero(ST0, &env->fp_status);
return val;
}
int64_t helper_fisttll_ST0(void)
{
int64_t val;
- val = floatx_to_int64_round_to_zero(ST0, &env->fp_status);
+ val = floatx80_to_int64_round_to_zero(ST0, &env->fp_status);
return val;
}
@@ -3693,7 +3687,7 @@ void helper_fmov_STN_ST0(int st_index)
void helper_fxchg_ST0_STN(int st_index)
{
- CPU86_LDouble tmp;
+ floatx80 tmp;
tmp = ST(st_index);
ST(st_index) = ST0;
ST0 = tmp;
@@ -3707,7 +3701,7 @@ void helper_fcom_ST0_FT0(void)
{
int ret;
- ret = floatx_compare(ST0, FT0, &env->fp_status);
+ ret = floatx80_compare(ST0, FT0, &env->fp_status);
env->fpus = (env->fpus & ~0x4500) | fcom_ccval[ret + 1];
}
@@ -3715,7 +3709,7 @@ void helper_fucom_ST0_FT0(void)
{
int ret;
- ret = floatx_compare_quiet(ST0, FT0, &env->fp_status);
+ ret = floatx80_compare_quiet(ST0, FT0, &env->fp_status);
env->fpus = (env->fpus & ~0x4500) | fcom_ccval[ret+ 1];
}
@@ -3726,7 +3720,7 @@ void helper_fcomi_ST0_FT0(void)
int eflags;
int ret;
- ret = floatx_compare(ST0, FT0, &env->fp_status);
+ ret = floatx80_compare(ST0, FT0, &env->fp_status);
eflags = helper_cc_compute_all(CC_OP);
eflags = (eflags & ~(CC_Z | CC_P | CC_C)) | fcomi_ccval[ret + 1];
CC_SRC = eflags;
@@ -3737,7 +3731,7 @@ void helper_fucomi_ST0_FT0(void)
int eflags;
int ret;
- ret = floatx_compare_quiet(ST0, FT0, &env->fp_status);
+ ret = floatx80_compare_quiet(ST0, FT0, &env->fp_status);
eflags = helper_cc_compute_all(CC_OP);
eflags = (eflags & ~(CC_Z | CC_P | CC_C)) | fcomi_ccval[ret + 1];
CC_SRC = eflags;
@@ -3745,22 +3739,22 @@ void helper_fucomi_ST0_FT0(void)
void helper_fadd_ST0_FT0(void)
{
- ST0 = floatx_add(ST0, FT0, &env->fp_status);
+ ST0 = floatx80_add(ST0, FT0, &env->fp_status);
}
void helper_fmul_ST0_FT0(void)
{
- ST0 = floatx_mul(ST0, FT0, &env->fp_status);
+ ST0 = floatx80_mul(ST0, FT0, &env->fp_status);
}
void helper_fsub_ST0_FT0(void)
{
- ST0 = floatx_sub(ST0, FT0, &env->fp_status);
+ ST0 = floatx80_sub(ST0, FT0, &env->fp_status);
}
void helper_fsubr_ST0_FT0(void)
{
- ST0 = floatx_sub(FT0, ST0, &env->fp_status);
+ ST0 = floatx80_sub(FT0, ST0, &env->fp_status);
}
void helper_fdiv_ST0_FT0(void)
@@ -3777,34 +3771,34 @@ void helper_fdivr_ST0_FT0(void)
void helper_fadd_STN_ST0(int st_index)
{
- ST(st_index) = floatx_add(ST(st_index), ST0, &env->fp_status);
+ ST(st_index) = floatx80_add(ST(st_index), ST0, &env->fp_status);
}
void helper_fmul_STN_ST0(int st_index)
{
- ST(st_index) = floatx_mul(ST(st_index), ST0, &env->fp_status);
+ ST(st_index) = floatx80_mul(ST(st_index), ST0, &env->fp_status);
}
void helper_fsub_STN_ST0(int st_index)
{
- ST(st_index) = floatx_sub(ST(st_index), ST0, &env->fp_status);
+ ST(st_index) = floatx80_sub(ST(st_index), ST0, &env->fp_status);
}
void helper_fsubr_STN_ST0(int st_index)
{
- ST(st_index) = floatx_sub(ST0, ST(st_index), &env->fp_status);
+ ST(st_index) = floatx80_sub(ST0, ST(st_index), &env->fp_status);
}
void helper_fdiv_STN_ST0(int st_index)
{
- CPU86_LDouble *p;
+ floatx80 *p;
p = &ST(st_index);
*p = helper_fdiv(*p, ST0);
}
void helper_fdivr_STN_ST0(int st_index)
{
- CPU86_LDouble *p;
+ floatx80 *p;
p = &ST(st_index);
*p = helper_fdiv(ST0, *p);
}
@@ -3812,12 +3806,12 @@ void helper_fdivr_STN_ST0(int st_index)
/* misc FPU operations */
void helper_fchs_ST0(void)
{
- ST0 = floatx_chs(ST0);
+ ST0 = floatx80_chs(ST0);
}
void helper_fabs_ST0(void)
{
- ST0 = floatx_abs(ST0);
+ ST0 = floatx80_abs(ST0);
}
void helper_fld1_ST0(void)
@@ -3891,7 +3885,6 @@ static void update_fp_status(void)
break;
}
set_float_rounding_mode(rnd_type, &env->fp_status);
-#ifdef FLOATX80
switch((env->fpuc >> 8) & 3) {
case 0:
rnd_type = 32;
@@ -3905,7 +3898,6 @@ static void update_fp_status(void)
break;
}
set_floatx80_rounding_precision(rnd_type, &env->fp_status);
-#endif
}
void helper_fldcw(uint32_t val)
@@ -3944,7 +3936,7 @@ void helper_fninit(void)
void helper_fbld_ST0(target_ulong ptr)
{
- CPU86_LDouble tmp;
+ floatx80 tmp;
uint64_t val;
unsigned int v;
int i;
@@ -3954,9 +3946,9 @@ void helper_fbld_ST0(target_ulong ptr)
v = ldub(ptr + i);
val = (val * 100) + ((v >> 4) * 10) + (v & 0xf);
}
- tmp = int64_to_floatx(val, &env->fp_status);
+ tmp = int64_to_floatx80(val, &env->fp_status);
if (ldub(ptr + 9) & 0x80) {
- floatx_chs(tmp);
+ floatx80_chs(tmp);
}
fpush();
ST0 = tmp;
@@ -3968,7 +3960,7 @@ void helper_fbst_ST0(target_ulong ptr)
target_ulong mem_ref, mem_end;
int64_t val;
- val = floatx_to_int64(ST0, &env->fp_status);
+ val = floatx80_to_int64(ST0, &env->fp_status);
mem_ref = ptr;
mem_end = mem_ref + 9;
if (val < 0) {
@@ -3992,19 +3984,19 @@ void helper_fbst_ST0(target_ulong ptr)
void helper_f2xm1(void)
{
- double val = CPU86_LDouble_to_double(ST0);
+ double val = floatx80_to_double(ST0);
val = pow(2.0, val) - 1.0;
- ST0 = double_to_CPU86_LDouble(val);
+ ST0 = double_to_floatx80(val);
}
void helper_fyl2x(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if (fptemp>0.0){
fptemp = log(fptemp)/log(2.0); /* log2(ST) */
- fptemp *= CPU86_LDouble_to_double(ST1);
- ST1 = double_to_CPU86_LDouble(fptemp);
+ fptemp *= floatx80_to_double(ST1);
+ ST1 = double_to_floatx80(fptemp);
fpop();
} else {
env->fpus &= (~0x4700);
@@ -4014,15 +4006,15 @@ void helper_fyl2x(void)
void helper_fptan(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if((fptemp > MAXTAN)||(fptemp < -MAXTAN)) {
env->fpus |= 0x400;
} else {
fptemp = tan(fptemp);
- ST0 = double_to_CPU86_LDouble(fptemp);
+ ST0 = double_to_floatx80(fptemp);
fpush();
- ST0 = floatx_one;
+ ST0 = floatx80_one;
env->fpus &= (~0x400); /* C2 <-- 0 */
/* the above code is for |arg| < 2**52 only */
}
@@ -4032,21 +4024,21 @@ void helper_fpatan(void)
{
double fptemp, fpsrcop;
- fpsrcop = CPU86_LDouble_to_double(ST1);
- fptemp = CPU86_LDouble_to_double(ST0);
- ST1 = double_to_CPU86_LDouble(atan2(fpsrcop, fptemp));
+ fpsrcop = floatx80_to_double(ST1);
+ fptemp = floatx80_to_double(ST0);
+ ST1 = double_to_floatx80(atan2(fpsrcop, fptemp));
fpop();
}
void helper_fxtract(void)
{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
temp.d = ST0;
- if (floatx_is_zero(ST0)) {
+ if (floatx80_is_zero(ST0)) {
/* Easy way to generate -inf and raising division by 0 exception */
- ST0 = floatx_div(floatx_chs(floatx_one), floatx_zero, &env->fp_status);
+ ST0 = floatx80_div(floatx80_chs(floatx80_one), floatx80_zero, &env->fp_status);
fpush();
ST0 = temp.d;
} else {
@@ -4054,7 +4046,7 @@ void helper_fxtract(void)
expdif = EXPD(temp) - EXPBIAS;
/*DP exponent bias*/
- ST0 = int32_to_floatx(expdif, &env->fp_status);
+ ST0 = int32_to_floatx80(expdif, &env->fp_status);
fpush();
BIASEXPONENT(temp);
ST0 = temp.d;
@@ -4064,15 +4056,15 @@ void helper_fxtract(void)
void helper_fprem1(void)
{
double st0, st1, dblq, fpsrcop, fptemp;
- CPU86_LDoubleU fpsrcop1, fptemp1;
+ CPU_LDoubleU fpsrcop1, fptemp1;
int expdif;
signed long long int q;
- st0 = CPU86_LDouble_to_double(ST0);
- st1 = CPU86_LDouble_to_double(ST1);
+ st0 = floatx80_to_double(ST0);
+ st1 = floatx80_to_double(ST1);
if (isinf(st0) || isnan(st0) || isnan(st1) || (st1 == 0.0)) {
- ST0 = double_to_CPU86_LDouble(0.0 / 0.0); /* NaN */
+ ST0 = double_to_floatx80(0.0 / 0.0); /* NaN */
env->fpus &= (~0x4700); /* (C3,C2,C1,C0) <-- 0000 */
return;
}
@@ -4116,21 +4108,21 @@ void helper_fprem1(void)
-(floor(fabs(fpsrcop))) : floor(fpsrcop);
st0 -= (st1 * fpsrcop * fptemp);
}
- ST0 = double_to_CPU86_LDouble(st0);
+ ST0 = double_to_floatx80(st0);
}
void helper_fprem(void)
{
double st0, st1, dblq, fpsrcop, fptemp;
- CPU86_LDoubleU fpsrcop1, fptemp1;
+ CPU_LDoubleU fpsrcop1, fptemp1;
int expdif;
signed long long int q;
- st0 = CPU86_LDouble_to_double(ST0);
- st1 = CPU86_LDouble_to_double(ST1);
+ st0 = floatx80_to_double(ST0);
+ st1 = floatx80_to_double(ST1);
if (isinf(st0) || isnan(st0) || isnan(st1) || (st1 == 0.0)) {
- ST0 = double_to_CPU86_LDouble(0.0 / 0.0); /* NaN */
+ ST0 = double_to_floatx80(0.0 / 0.0); /* NaN */
env->fpus &= (~0x4700); /* (C3,C2,C1,C0) <-- 0000 */
return;
}
@@ -4175,17 +4167,17 @@ void helper_fprem(void)
-(floor(fabs(fpsrcop))) : floor(fpsrcop);
st0 -= (st1 * fpsrcop * fptemp);
}
- ST0 = double_to_CPU86_LDouble(st0);
+ ST0 = double_to_floatx80(st0);
}
void helper_fyl2xp1(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if ((fptemp+1.0)>0.0) {
fptemp = log(fptemp+1.0) / log(2.0); /* log2(ST+1.0) */
- fptemp *= CPU86_LDouble_to_double(ST1);
- ST1 = double_to_CPU86_LDouble(fptemp);
+ fptemp *= floatx80_to_double(ST1);
+ ST1 = double_to_floatx80(fptemp);
fpop();
} else {
env->fpus &= (~0x4700);
@@ -4195,23 +4187,23 @@ void helper_fyl2xp1(void)
void helper_fsqrt(void)
{
- if (floatx_is_neg(ST0)) {
+ if (floatx80_is_neg(ST0)) {
env->fpus &= (~0x4700); /* (C3,C2,C1,C0) <-- 0000 */
env->fpus |= 0x400;
}
- ST0 = floatx_sqrt(ST0, &env->fp_status);
+ ST0 = floatx80_sqrt(ST0, &env->fp_status);
}
void helper_fsincos(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if ((fptemp > MAXTAN)||(fptemp < -MAXTAN)) {
env->fpus |= 0x400;
} else {
- ST0 = double_to_CPU86_LDouble(sin(fptemp));
+ ST0 = double_to_floatx80(sin(fptemp));
fpush();
- ST0 = double_to_CPU86_LDouble(cos(fptemp));
+ ST0 = double_to_floatx80(cos(fptemp));
env->fpus &= (~0x400); /* C2 <-- 0 */
/* the above code is for |arg| < 2**63 only */
}
@@ -4219,27 +4211,27 @@ void helper_fsincos(void)
void helper_frndint(void)
{
- ST0 = floatx_round_to_int(ST0, &env->fp_status);
+ ST0 = floatx80_round_to_int(ST0, &env->fp_status);
}
void helper_fscale(void)
{
- if (floatx_is_any_nan(ST1)) {
+ if (floatx80_is_any_nan(ST1)) {
ST0 = ST1;
} else {
- int n = floatx_to_int32_round_to_zero(ST1, &env->fp_status);
- ST0 = floatx_scalbn(ST0, n, &env->fp_status);
+ int n = floatx80_to_int32_round_to_zero(ST1, &env->fp_status);
+ ST0 = floatx80_scalbn(ST0, n, &env->fp_status);
}
}
void helper_fsin(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if ((fptemp > MAXTAN)||(fptemp < -MAXTAN)) {
env->fpus |= 0x400;
} else {
- ST0 = double_to_CPU86_LDouble(sin(fptemp));
+ ST0 = double_to_floatx80(sin(fptemp));
env->fpus &= (~0x400); /* C2 <-- 0 */
/* the above code is for |arg| < 2**53 only */
}
@@ -4247,12 +4239,12 @@ void helper_fsin(void)
void helper_fcos(void)
{
- double fptemp = CPU86_LDouble_to_double(ST0);
+ double fptemp = floatx80_to_double(ST0);
if((fptemp > MAXTAN)||(fptemp < -MAXTAN)) {
env->fpus |= 0x400;
} else {
- ST0 = double_to_CPU86_LDouble(cos(fptemp));
+ ST0 = double_to_floatx80(cos(fptemp));
env->fpus &= (~0x400); /* C2 <-- 0 */
/* the above code is for |arg5 < 2**63 only */
}
@@ -4260,7 +4252,7 @@ void helper_fcos(void)
void helper_fxam_ST0(void)
{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
int expdif;
temp.d = ST0;
@@ -4272,11 +4264,7 @@ void helper_fxam_ST0(void)
/* XXX: test fptags too */
expdif = EXPD(temp);
if (expdif == MAXEXPD) {
-#ifdef USE_X86LDOUBLE
if (MANTD(temp) == 0x8000000000000000ULL)
-#else
- if (MANTD(temp) == 0)
-#endif
env->fpus |= 0x500 /*Infinity*/;
else
env->fpus |= 0x100 /*NaN*/;
@@ -4294,7 +4282,7 @@ void helper_fstenv(target_ulong ptr, int data32)
{
int fpus, fptag, exp, i;
uint64_t mant;
- CPU86_LDoubleU tmp;
+ CPU_LDoubleU tmp;
fpus = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
fptag = 0;
@@ -4310,9 +4298,7 @@ void helper_fstenv(target_ulong ptr, int data32)
/* zero */
fptag |= 1;
} else if (exp == 0 || exp == MAXEXPD
-#ifdef USE_X86LDOUBLE
|| (mant & (1LL << 63)) == 0
-#endif
) {
/* NaNs, infinity, denormal */
fptag |= 2;
@@ -4364,7 +4350,7 @@ void helper_fldenv(target_ulong ptr, int data32)
void helper_fsave(target_ulong ptr, int data32)
{
- CPU86_LDouble tmp;
+ floatx80 tmp;
int i;
helper_fstenv(ptr, data32);
@@ -4392,7 +4378,7 @@ void helper_fsave(target_ulong ptr, int data32)
void helper_frstor(target_ulong ptr, int data32)
{
- CPU86_LDouble tmp;
+ floatx80 tmp;
int i;
helper_fldenv(ptr, data32);
@@ -4408,7 +4394,7 @@ void helper_frstor(target_ulong ptr, int data32)
void helper_fxsave(target_ulong ptr, int data64)
{
int fpus, fptag, i, nb_xmm_regs;
- CPU86_LDouble tmp;
+ floatx80 tmp;
target_ulong addr;
/* The operand must be 16 byte aligned */
@@ -4469,7 +4455,7 @@ void helper_fxsave(target_ulong ptr, int data64)
void helper_fxrstor(target_ulong ptr, int data64)
{
int i, fpus, fptag, nb_xmm_regs;
- CPU86_LDouble tmp;
+ floatx80 tmp;
target_ulong addr;
/* The operand must be 16 byte aligned */
@@ -4516,61 +4502,23 @@ void helper_fxrstor(target_ulong ptr, int data64)
}
}
-#ifndef USE_X86LDOUBLE
-
-void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, CPU86_LDouble f)
+void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, floatx80 f)
{
- CPU86_LDoubleU temp;
- int e;
-
- temp.d = f;
- /* mantissa */
- *pmant = (MANTD(temp) << 11) | (1LL << 63);
- /* exponent + sign */
- e = EXPD(temp) - EXPBIAS + 16383;
- e |= SIGND(temp) >> 16;
- *pexp = e;
-}
-
-CPU86_LDouble cpu_set_fp80(uint64_t mant, uint16_t upper)
-{
- CPU86_LDoubleU temp;
- int e;
- uint64_t ll;
-
- /* XXX: handle overflow ? */
- e = (upper & 0x7fff) - 16383 + EXPBIAS; /* exponent */
- e |= (upper >> 4) & 0x800; /* sign */
- ll = (mant >> 11) & ((1LL << 52) - 1);
-#ifdef __arm__
- temp.l.upper = (e << 20) | (ll >> 32);
- temp.l.lower = ll;
-#else
- temp.ll = ll | ((uint64_t)e << 52);
-#endif
- return temp.d;
-}
-
-#else
-
-void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, CPU86_LDouble f)
-{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
temp.d = f;
*pmant = temp.l.lower;
*pexp = temp.l.upper;
}
-CPU86_LDouble cpu_set_fp80(uint64_t mant, uint16_t upper)
+floatx80 cpu_set_fp80(uint64_t mant, uint16_t upper)
{
- CPU86_LDoubleU temp;
+ CPU_LDoubleU temp;
temp.l.upper = upper;
temp.l.lower = mant;
return temp.d;
}
-#endif
#ifdef TARGET_X86_64
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 4/9] softfloat-native: remove
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (2 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64 Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 5/9] softfloat: always enable floatx80 and float128 support Aurelien Jarno
` (4 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
Remove softfloat-native support, all targets are now using softfloat
instead.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
Makefile.target | 3 +-
configure | 2 -
cpu-all.h | 5 +-
fpu/softfloat-native.c | 540 ------------------------------------------------
fpu/softfloat-native.h | 531 -----------------------------------------------
fpu/softfloat.h | 14 --
6 files changed, 2 insertions(+), 1093 deletions(-)
diff --git a/Makefile.target b/Makefile.target
index 2e281a4..ed3f95a 100644
--- a/Makefile.target
+++ b/Makefile.target
@@ -71,8 +71,7 @@ all: $(PROGS) stap
# cpu emulator library
libobj-y = exec.o translate-all.o cpu-exec.o translate.o
libobj-y += tcg/tcg.o
-libobj-$(CONFIG_SOFTFLOAT) += fpu/softfloat.o
-libobj-$(CONFIG_NOSOFTFLOAT) += fpu/softfloat-native.o
+libobj-y += fpu/softfloat.o
libobj-y += op_helper.o helper.o
ifeq ($(TARGET_BASE_ARCH), i386)
libobj-y += cpuid.o
diff --git a/configure b/configure
index a318d37..0c26a26 100755
--- a/configure
+++ b/configure
@@ -3385,8 +3385,6 @@ if test ! -z "$gdb_xml_files" ; then
echo "TARGET_XML_FILES=$list" >> $config_target_mak
fi
-echo "CONFIG_SOFTFLOAT=y" >> $config_target_mak
-
if test "$target_user_only" = "yes" -a "$bflt" = "yes"; then
echo "TARGET_HAS_BFLT=y" >> $config_target_mak
fi
diff --git a/cpu-all.h b/cpu-all.h
index 54df1d3..fc252ba 100644
--- a/cpu-all.h
+++ b/cpu-all.h
@@ -123,8 +123,7 @@ typedef union {
endian ! */
typedef union {
float64 d;
-#if defined(HOST_WORDS_BIGENDIAN) \
- || (defined(__arm__) && !defined(__VFP_FP__) && !defined(CONFIG_SOFTFLOAT))
+#if defined(HOST_WORDS_BIGENDIAN)
struct {
uint32_t upper;
uint32_t lower;
@@ -148,7 +147,6 @@ typedef union {
} CPU_LDoubleU;
#endif
-#if defined(CONFIG_SOFTFLOAT)
typedef union {
float128 q;
#if defined(HOST_WORDS_BIGENDIAN)
@@ -175,7 +173,6 @@ typedef union {
} ll;
#endif
} CPU_QuadU;
-#endif
/* CPU memory access without any memory or io remapping */
diff --git a/fpu/softfloat-native.c b/fpu/softfloat-native.c
deleted file mode 100644
index 8848651..0000000
--- a/fpu/softfloat-native.c
+++ /dev/null
@@ -1,540 +0,0 @@
-/* Native implementation of soft float functions. Only a single status
- context is supported */
-#include "softfloat.h"
-#include <math.h>
-#if defined(CONFIG_SOLARIS)
-#include <fenv.h>
-#endif
-
-void set_float_rounding_mode(int val STATUS_PARAM)
-{
- STATUS(float_rounding_mode) = val;
-#if (defined(CONFIG_BSD) && !defined(__APPLE__) && !defined(__GLIBC__)) || \
- (defined(CONFIG_SOLARIS) && CONFIG_SOLARIS_VERSION < 10)
- fpsetround(val);
-#else
- fesetround(val);
-#endif
-}
-
-#ifdef FLOATX80
-void set_floatx80_rounding_precision(int val STATUS_PARAM)
-{
- STATUS(floatx80_rounding_precision) = val;
-}
-#endif
-
-#if defined(CONFIG_BSD) || \
- (defined(CONFIG_SOLARIS) && CONFIG_SOLARIS_VERSION < 10)
-#define lrint(d) ((int32_t)rint(d))
-#define llrint(d) ((int64_t)rint(d))
-#define lrintf(f) ((int32_t)rint(f))
-#define llrintf(f) ((int64_t)rint(f))
-#define sqrtf(f) ((float)sqrt(f))
-#define remainderf(fa, fb) ((float)remainder(fa, fb))
-#define rintf(f) ((float)rint(f))
-#if !defined(__sparc__) && \
- (defined(CONFIG_SOLARIS) && CONFIG_SOLARIS_VERSION < 10)
-extern long double rintl(long double);
-extern long double scalbnl(long double, int);
-
-long long
-llrintl(long double x) {
- return ((long long) rintl(x));
-}
-
-long
-lrintl(long double x) {
- return ((long) rintl(x));
-}
-
-long double
-ldexpl(long double x, int n) {
- return (scalbnl(x, n));
-}
-#endif
-#endif
-
-#if defined(_ARCH_PPC)
-
-/* correct (but slow) PowerPC rint() (glibc version is incorrect) */
-static double qemu_rint(double x)
-{
- double y = 4503599627370496.0;
- if (fabs(x) >= y)
- return x;
- if (x < 0)
- y = -y;
- y = (x + y) - y;
- if (y == 0.0)
- y = copysign(y, x);
- return y;
-}
-
-#define rint qemu_rint
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE integer-to-floating-point conversion routines.
-*----------------------------------------------------------------------------*/
-float32 int32_to_float32(int v STATUS_PARAM)
-{
- return (float32)v;
-}
-
-float32 uint32_to_float32(unsigned int v STATUS_PARAM)
-{
- return (float32)v;
-}
-
-float64 int32_to_float64(int v STATUS_PARAM)
-{
- return (float64)v;
-}
-
-float64 uint32_to_float64(unsigned int v STATUS_PARAM)
-{
- return (float64)v;
-}
-
-#ifdef FLOATX80
-floatx80 int32_to_floatx80(int v STATUS_PARAM)
-{
- return (floatx80)v;
-}
-#endif
-float32 int64_to_float32( int64_t v STATUS_PARAM)
-{
- return (float32)v;
-}
-float32 uint64_to_float32( uint64_t v STATUS_PARAM)
-{
- return (float32)v;
-}
-float64 int64_to_float64( int64_t v STATUS_PARAM)
-{
- return (float64)v;
-}
-float64 uint64_to_float64( uint64_t v STATUS_PARAM)
-{
- return (float64)v;
-}
-#ifdef FLOATX80
-floatx80 int64_to_floatx80( int64_t v STATUS_PARAM)
-{
- return (floatx80)v;
-}
-#endif
-
-/* XXX: this code implements the x86 behaviour, not the IEEE one. */
-#if HOST_LONG_BITS == 32
-static inline int long_to_int32(long a)
-{
- return a;
-}
-#else
-static inline int long_to_int32(long a)
-{
- if (a != (int32_t)a)
- a = 0x80000000;
- return a;
-}
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE single-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int float32_to_int32( float32 a STATUS_PARAM)
-{
- return long_to_int32(lrintf(a));
-}
-int float32_to_int32_round_to_zero( float32 a STATUS_PARAM)
-{
- return (int)a;
-}
-int64_t float32_to_int64( float32 a STATUS_PARAM)
-{
- return llrintf(a);
-}
-
-int64_t float32_to_int64_round_to_zero( float32 a STATUS_PARAM)
-{
- return (int64_t)a;
-}
-
-float64 float32_to_float64( float32 a STATUS_PARAM)
-{
- return a;
-}
-#ifdef FLOATX80
-floatx80 float32_to_floatx80( float32 a STATUS_PARAM)
-{
- return a;
-}
-#endif
-
-unsigned int float32_to_uint32( float32 a STATUS_PARAM)
-{
- int64_t v;
- unsigned int res;
-
- v = llrintf(a);
- if (v < 0) {
- res = 0;
- } else if (v > 0xffffffff) {
- res = 0xffffffff;
- } else {
- res = v;
- }
- return res;
-}
-unsigned int float32_to_uint32_round_to_zero( float32 a STATUS_PARAM)
-{
- int64_t v;
- unsigned int res;
-
- v = (int64_t)a;
- if (v < 0) {
- res = 0;
- } else if (v > 0xffffffff) {
- res = 0xffffffff;
- } else {
- res = v;
- }
- return res;
-}
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE single-precision operations.
-*----------------------------------------------------------------------------*/
-float32 float32_round_to_int( float32 a STATUS_PARAM)
-{
- return rintf(a);
-}
-
-float32 float32_rem( float32 a, float32 b STATUS_PARAM)
-{
- return remainderf(a, b);
-}
-
-float32 float32_sqrt( float32 a STATUS_PARAM)
-{
- return sqrtf(a);
-}
-int float32_compare( float32 a, float32 b STATUS_PARAM )
-{
- if (a < b) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (a > b) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int float32_compare_quiet( float32 a, float32 b STATUS_PARAM )
-{
- if (isless(a, b)) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (isgreater(a, b)) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int float32_is_signaling_nan( float32 a1)
-{
- float32u u;
- uint32_t a;
- u.f = a1;
- a = u.i;
- return ( ( ( a>>22 ) & 0x1FF ) == 0x1FE ) && ( a & 0x003FFFFF );
-}
-
-int float32_is_quiet_nan( float32 a1 )
-{
- float32u u;
- uint64_t a;
- u.f = a1;
- a = u.i;
- return ( 0xFF800000 < ( a<<1 ) );
-}
-
-int float32_is_any_nan( float32 a1 )
-{
- float32u u;
- uint32_t a;
- u.f = a1;
- a = u.i;
- return (a & ~(1 << 31)) > 0x7f800000U;
-}
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE double-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int float64_to_int32( float64 a STATUS_PARAM)
-{
- return long_to_int32(lrint(a));
-}
-int float64_to_int32_round_to_zero( float64 a STATUS_PARAM)
-{
- return (int)a;
-}
-int64_t float64_to_int64( float64 a STATUS_PARAM)
-{
- return llrint(a);
-}
-int64_t float64_to_int64_round_to_zero( float64 a STATUS_PARAM)
-{
- return (int64_t)a;
-}
-float32 float64_to_float32( float64 a STATUS_PARAM)
-{
- return a;
-}
-#ifdef FLOATX80
-floatx80 float64_to_floatx80( float64 a STATUS_PARAM)
-{
- return a;
-}
-#endif
-#ifdef FLOAT128
-float128 float64_to_float128( float64 a STATUS_PARAM)
-{
- return a;
-}
-#endif
-
-unsigned int float64_to_uint32( float64 a STATUS_PARAM)
-{
- int64_t v;
- unsigned int res;
-
- v = llrint(a);
- if (v < 0) {
- res = 0;
- } else if (v > 0xffffffff) {
- res = 0xffffffff;
- } else {
- res = v;
- }
- return res;
-}
-unsigned int float64_to_uint32_round_to_zero( float64 a STATUS_PARAM)
-{
- int64_t v;
- unsigned int res;
-
- v = (int64_t)a;
- if (v < 0) {
- res = 0;
- } else if (v > 0xffffffff) {
- res = 0xffffffff;
- } else {
- res = v;
- }
- return res;
-}
-uint64_t float64_to_uint64 (float64 a STATUS_PARAM)
-{
- int64_t v;
-
- v = llrint(a + (float64)INT64_MIN);
-
- return v - INT64_MIN;
-}
-uint64_t float64_to_uint64_round_to_zero (float64 a STATUS_PARAM)
-{
- int64_t v;
-
- v = (int64_t)(a + (float64)INT64_MIN);
-
- return v - INT64_MIN;
-}
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE double-precision operations.
-*----------------------------------------------------------------------------*/
-#if defined(__sun__) && \
- (defined(CONFIG_SOLARIS) && CONFIG_SOLARIS_VERSION < 10)
-static inline float64 trunc(float64 x)
-{
- return x < 0 ? -floor(-x) : floor(x);
-}
-#endif
-float64 float64_trunc_to_int( float64 a STATUS_PARAM )
-{
- return trunc(a);
-}
-
-float64 float64_round_to_int( float64 a STATUS_PARAM )
-{
- return rint(a);
-}
-
-float64 float64_rem( float64 a, float64 b STATUS_PARAM)
-{
- return remainder(a, b);
-}
-
-float64 float64_sqrt( float64 a STATUS_PARAM)
-{
- return sqrt(a);
-}
-int float64_compare( float64 a, float64 b STATUS_PARAM )
-{
- if (a < b) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (a > b) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int float64_compare_quiet( float64 a, float64 b STATUS_PARAM )
-{
- if (isless(a, b)) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (isgreater(a, b)) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int float64_is_signaling_nan( float64 a1)
-{
- float64u u;
- uint64_t a;
- u.f = a1;
- a = u.i;
- return
- ( ( ( a>>51 ) & 0xFFF ) == 0xFFE )
- && ( a & LIT64( 0x0007FFFFFFFFFFFF ) );
-
-}
-
-int float64_is_quiet_nan( float64 a1 )
-{
- float64u u;
- uint64_t a;
- u.f = a1;
- a = u.i;
-
- return ( LIT64( 0xFFF0000000000000 ) < (uint64_t) ( a<<1 ) );
-
-}
-
-int float64_is_any_nan( float64 a1 )
-{
- float64u u;
- uint64_t a;
- u.f = a1;
- a = u.i;
-
- return (a & ~(1ULL << 63)) > LIT64 (0x7FF0000000000000 );
-}
-
-#ifdef FLOATX80
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE extended double-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int floatx80_to_int32( floatx80 a STATUS_PARAM)
-{
- return long_to_int32(lrintl(a));
-}
-int floatx80_to_int32_round_to_zero( floatx80 a STATUS_PARAM)
-{
- return (int)a;
-}
-int64_t floatx80_to_int64( floatx80 a STATUS_PARAM)
-{
- return llrintl(a);
-}
-int64_t floatx80_to_int64_round_to_zero( floatx80 a STATUS_PARAM)
-{
- return (int64_t)a;
-}
-float32 floatx80_to_float32( floatx80 a STATUS_PARAM)
-{
- return a;
-}
-float64 floatx80_to_float64( floatx80 a STATUS_PARAM)
-{
- return a;
-}
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE extended double-precision operations.
-*----------------------------------------------------------------------------*/
-floatx80 floatx80_round_to_int( floatx80 a STATUS_PARAM)
-{
- return rintl(a);
-}
-floatx80 floatx80_rem( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return remainderl(a, b);
-}
-floatx80 floatx80_sqrt( floatx80 a STATUS_PARAM)
-{
- return sqrtl(a);
-}
-int floatx80_compare( floatx80 a, floatx80 b STATUS_PARAM )
-{
- if (a < b) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (a > b) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int floatx80_compare_quiet( floatx80 a, floatx80 b STATUS_PARAM )
-{
- if (isless(a, b)) {
- return float_relation_less;
- } else if (a == b) {
- return float_relation_equal;
- } else if (isgreater(a, b)) {
- return float_relation_greater;
- } else {
- return float_relation_unordered;
- }
-}
-int floatx80_is_signaling_nan( floatx80 a1)
-{
- floatx80u u;
- uint64_t aLow;
- u.f = a1;
-
- aLow = u.i.low & ~ LIT64( 0x4000000000000000 );
- return
- ( ( u.i.high & 0x7FFF ) == 0x7FFF )
- && (uint64_t) ( aLow<<1 )
- && ( u.i.low == aLow );
-}
-
-int floatx80_is_quiet_nan( floatx80 a1 )
-{
- floatx80u u;
- u.f = a1;
- return ( ( u.i.high & 0x7FFF ) == 0x7FFF ) && (uint64_t) ( u.i.low<<1 );
-}
-
-int floatx80_is_any_nan( floatx80 a1 )
-{
- floatx80u u;
- u.f = a1;
- return ((u.i.high & 0x7FFF) == 0x7FFF) && ( u.i.low<<1 );
-}
-
-#endif
diff --git a/fpu/softfloat-native.h b/fpu/softfloat-native.h
deleted file mode 100644
index 6afb74a..0000000
--- a/fpu/softfloat-native.h
+++ /dev/null
@@ -1,531 +0,0 @@
-/* Native implementation of soft float functions */
-#include <math.h>
-
-#if (defined(CONFIG_BSD) && !defined(__APPLE__) && !defined(__GLIBC__)) \
- || defined(CONFIG_SOLARIS)
-#include <ieeefp.h>
-#define fabsf(f) ((float)fabs(f))
-#else
-#include <fenv.h>
-#endif
-
-#if defined(__OpenBSD__) || defined(__NetBSD__)
-#include <sys/param.h>
-#endif
-
-/*
- * Define some C99-7.12.3 classification macros and
- * some C99-.12.4 for Solaris systems OS less than 10,
- * or Solaris 10 systems running GCC 3.x or less.
- * Solaris 10 with GCC4 does not need these macros as they
- * are defined in <iso/math_c99.h> with a compiler directive
- */
-#if defined(CONFIG_SOLARIS) && \
- ((CONFIG_SOLARIS_VERSION <= 9 ) || \
- ((CONFIG_SOLARIS_VERSION == 10) && (__GNUC__ < 4))) \
- || (defined(__OpenBSD__) && (OpenBSD < 200811))
-/*
- * C99 7.12.3 classification macros
- * and
- * C99 7.12.14 comparison macros
- *
- * ... do not work on Solaris 10 using GNU CC 3.4.x.
- * Try to workaround the missing / broken C99 math macros.
- */
-#if defined(__OpenBSD__)
-#define unordered(x, y) (isnan(x) || isnan(y))
-#endif
-
-#ifdef __NetBSD__
-#ifndef isgreater
-#define isgreater(x, y) __builtin_isgreater(x, y)
-#endif
-#ifndef isgreaterequal
-#define isgreaterequal(x, y) __builtin_isgreaterequal(x, y)
-#endif
-#ifndef isless
-#define isless(x, y) __builtin_isless(x, y)
-#endif
-#ifndef islessequal
-#define islessequal(x, y) __builtin_islessequal(x, y)
-#endif
-#ifndef isunordered
-#define isunordered(x, y) __builtin_isunordered(x, y)
-#endif
-#endif
-
-
-#define isnormal(x) (fpclass(x) >= FP_NZERO)
-#define isgreater(x, y) ((!unordered(x, y)) && ((x) > (y)))
-#define isgreaterequal(x, y) ((!unordered(x, y)) && ((x) >= (y)))
-#define isless(x, y) ((!unordered(x, y)) && ((x) < (y)))
-#define islessequal(x, y) ((!unordered(x, y)) && ((x) <= (y)))
-#define isunordered(x,y) unordered(x, y)
-#endif
-
-#if defined(__sun__) && !defined(CONFIG_NEEDS_LIBSUNMATH)
-
-#ifndef isnan
-# define isnan(x) \
- (sizeof (x) == sizeof (long double) ? isnan_ld (x) \
- : sizeof (x) == sizeof (double) ? isnan_d (x) \
- : isnan_f (x))
-static inline int isnan_f (float x) { return x != x; }
-static inline int isnan_d (double x) { return x != x; }
-static inline int isnan_ld (long double x) { return x != x; }
-#endif
-
-#ifndef isinf
-# define isinf(x) \
- (sizeof (x) == sizeof (long double) ? isinf_ld (x) \
- : sizeof (x) == sizeof (double) ? isinf_d (x) \
- : isinf_f (x))
-static inline int isinf_f (float x) { return isnan (x - x); }
-static inline int isinf_d (double x) { return isnan (x - x); }
-static inline int isinf_ld (long double x) { return isnan (x - x); }
-#endif
-#endif
-
-typedef float float32;
-typedef double float64;
-#ifdef FLOATX80
-typedef long double floatx80;
-#endif
-
-typedef union {
- float32 f;
- uint32_t i;
-} float32u;
-typedef union {
- float64 f;
- uint64_t i;
-} float64u;
-#ifdef FLOATX80
-typedef union {
- floatx80 f;
- struct {
- uint64_t low;
- uint16_t high;
- } i;
-} floatx80u;
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE floating-point rounding mode.
-*----------------------------------------------------------------------------*/
-#if (defined(CONFIG_BSD) && !defined(__APPLE__) && !defined(__GLIBC__)) \
- || defined(CONFIG_SOLARIS)
-#if defined(__OpenBSD__)
-#define FE_RM FP_RM
-#define FE_RP FP_RP
-#define FE_RZ FP_RZ
-#endif
-enum {
- float_round_nearest_even = FP_RN,
- float_round_down = FP_RM,
- float_round_up = FP_RP,
- float_round_to_zero = FP_RZ
-};
-#else
-enum {
- float_round_nearest_even = FE_TONEAREST,
- float_round_down = FE_DOWNWARD,
- float_round_up = FE_UPWARD,
- float_round_to_zero = FE_TOWARDZERO
-};
-#endif
-
-typedef struct float_status {
- int float_rounding_mode;
-#ifdef FLOATX80
- int floatx80_rounding_precision;
-#endif
-} float_status;
-
-void set_float_rounding_mode(int val STATUS_PARAM);
-#ifdef FLOATX80
-void set_floatx80_rounding_precision(int val STATUS_PARAM);
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE integer-to-floating-point conversion routines.
-*----------------------------------------------------------------------------*/
-float32 int32_to_float32( int STATUS_PARAM);
-float32 uint32_to_float32( unsigned int STATUS_PARAM);
-float64 int32_to_float64( int STATUS_PARAM);
-float64 uint32_to_float64( unsigned int STATUS_PARAM);
-#ifdef FLOATX80
-floatx80 int32_to_floatx80( int STATUS_PARAM);
-#endif
-#ifdef FLOAT128
-float128 int32_to_float128( int STATUS_PARAM);
-#endif
-float32 int64_to_float32( int64_t STATUS_PARAM);
-float32 uint64_to_float32( uint64_t STATUS_PARAM);
-float64 int64_to_float64( int64_t STATUS_PARAM);
-float64 uint64_to_float64( uint64_t v STATUS_PARAM);
-#ifdef FLOATX80
-floatx80 int64_to_floatx80( int64_t STATUS_PARAM);
-#endif
-#ifdef FLOAT128
-float128 int64_to_float128( int64_t STATUS_PARAM);
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE single-precision conversion constants.
-*----------------------------------------------------------------------------*/
-#define float32_zero (0.0)
-#define float32_one (1.0)
-#define float32_ln2 (0.6931471)
-#define float32_pi (3.1415926)
-#define float32_half (0.5)
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE single-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int float32_to_int32( float32 STATUS_PARAM);
-int float32_to_int32_round_to_zero( float32 STATUS_PARAM);
-unsigned int float32_to_uint32( float32 a STATUS_PARAM);
-unsigned int float32_to_uint32_round_to_zero( float32 a STATUS_PARAM);
-int64_t float32_to_int64( float32 STATUS_PARAM);
-int64_t float32_to_int64_round_to_zero( float32 STATUS_PARAM);
-float64 float32_to_float64( float32 STATUS_PARAM);
-#ifdef FLOATX80
-floatx80 float32_to_floatx80( float32 STATUS_PARAM);
-#endif
-#ifdef FLOAT128
-float128 float32_to_float128( float32 STATUS_PARAM);
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE single-precision operations.
-*----------------------------------------------------------------------------*/
-float32 float32_round_to_int( float32 STATUS_PARAM);
-INLINE float32 float32_add( float32 a, float32 b STATUS_PARAM)
-{
- return a + b;
-}
-INLINE float32 float32_sub( float32 a, float32 b STATUS_PARAM)
-{
- return a - b;
-}
-INLINE float32 float32_mul( float32 a, float32 b STATUS_PARAM)
-{
- return a * b;
-}
-INLINE float32 float32_div( float32 a, float32 b STATUS_PARAM)
-{
- return a / b;
-}
-float32 float32_rem( float32, float32 STATUS_PARAM);
-float32 float32_sqrt( float32 STATUS_PARAM);
-INLINE int float32_eq_quiet( float32 a, float32 b STATUS_PARAM)
-{
- return a == b;
-}
-INLINE int float32_le( float32 a, float32 b STATUS_PARAM)
-{
- return a <= b;
-}
-INLINE int float32_lt( float32 a, float32 b STATUS_PARAM)
-{
- return a < b;
-}
-INLINE int float32_eq( float32 a, float32 b STATUS_PARAM)
-{
- return a <= b && a >= b;
-}
-INLINE int float32_le_quiet( float32 a, float32 b STATUS_PARAM)
-{
- return islessequal(a, b);
-}
-INLINE int float32_lt_quiet( float32 a, float32 b STATUS_PARAM)
-{
- return isless(a, b);
-}
-INLINE int float32_unordered( float32 a, float32 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-INLINE int float32_unordered_quiet( float32 a, float32 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-int float32_compare( float32, float32 STATUS_PARAM );
-int float32_compare_quiet( float32, float32 STATUS_PARAM );
-int float32_is_signaling_nan( float32 );
-int float32_is_quiet_nan( float32 );
-int float32_is_any_nan( float32 );
-
-INLINE float32 float32_abs(float32 a)
-{
- return fabsf(a);
-}
-
-INLINE float32 float32_chs(float32 a)
-{
- return -a;
-}
-
-INLINE float32 float32_is_infinity(float32 a)
-{
- return fpclassify(a) == FP_INFINITE;
-}
-
-INLINE float32 float32_is_neg(float32 a)
-{
- float32u u;
- u.f = a;
- return u.i >> 31;
-}
-
-INLINE float32 float32_is_zero(float32 a)
-{
- return fpclassify(a) == FP_ZERO;
-}
-
-INLINE float32 float32_scalbn(float32 a, int n STATUS_PARAM)
-{
- return scalbnf(a, n);
-}
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE double-precision conversion constants.
-*----------------------------------------------------------------------------*/
-#define float64_zero (0.0)
-#define float64_one (1.0)
-#define float64_ln2 (0.693147180559945)
-#define float64_pi (3.141592653589793)
-#define float64_half (0.5)
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE double-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int float64_to_int32( float64 STATUS_PARAM );
-int float64_to_int32_round_to_zero( float64 STATUS_PARAM );
-unsigned int float64_to_uint32( float64 STATUS_PARAM );
-unsigned int float64_to_uint32_round_to_zero( float64 STATUS_PARAM );
-int64_t float64_to_int64( float64 STATUS_PARAM );
-int64_t float64_to_int64_round_to_zero( float64 STATUS_PARAM );
-uint64_t float64_to_uint64( float64 STATUS_PARAM );
-uint64_t float64_to_uint64_round_to_zero( float64 STATUS_PARAM );
-float32 float64_to_float32( float64 STATUS_PARAM );
-#ifdef FLOATX80
-floatx80 float64_to_floatx80( float64 STATUS_PARAM );
-#endif
-#ifdef FLOAT128
-float128 float64_to_float128( float64 STATUS_PARAM );
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE double-precision operations.
-*----------------------------------------------------------------------------*/
-float64 float64_round_to_int( float64 STATUS_PARAM );
-float64 float64_trunc_to_int( float64 STATUS_PARAM );
-INLINE float64 float64_add( float64 a, float64 b STATUS_PARAM)
-{
- return a + b;
-}
-INLINE float64 float64_sub( float64 a, float64 b STATUS_PARAM)
-{
- return a - b;
-}
-INLINE float64 float64_mul( float64 a, float64 b STATUS_PARAM)
-{
- return a * b;
-}
-INLINE float64 float64_div( float64 a, float64 b STATUS_PARAM)
-{
- return a / b;
-}
-float64 float64_rem( float64, float64 STATUS_PARAM );
-float64 float64_sqrt( float64 STATUS_PARAM );
-INLINE int float64_eq_quiet( float64 a, float64 b STATUS_PARAM)
-{
- return a == b;
-}
-INLINE int float64_le( float64 a, float64 b STATUS_PARAM)
-{
- return a <= b;
-}
-INLINE int float64_lt( float64 a, float64 b STATUS_PARAM)
-{
- return a < b;
-}
-INLINE int float64_eq( float64 a, float64 b STATUS_PARAM)
-{
- return a <= b && a >= b;
-}
-INLINE int float64_le_quiet( float64 a, float64 b STATUS_PARAM)
-{
- return islessequal(a, b);
-}
-INLINE int float64_lt_quiet( float64 a, float64 b STATUS_PARAM)
-{
- return isless(a, b);
-
-}
-INLINE int float64_unordered( float64 a, float64 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-INLINE int float64_unordered_quiet( float64 a, float64 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-int float64_compare( float64, float64 STATUS_PARAM );
-int float64_compare_quiet( float64, float64 STATUS_PARAM );
-int float64_is_signaling_nan( float64 );
-int float64_is_any_nan( float64 );
-int float64_is_quiet_nan( float64 );
-
-INLINE float64 float64_abs(float64 a)
-{
- return fabs(a);
-}
-
-INLINE float64 float64_chs(float64 a)
-{
- return -a;
-}
-
-INLINE float64 float64_is_infinity(float64 a)
-{
- return fpclassify(a) == FP_INFINITE;
-}
-
-INLINE float64 float64_is_neg(float64 a)
-{
- float64u u;
- u.f = a;
- return u.i >> 63;
-}
-
-INLINE float64 float64_is_zero(float64 a)
-{
- return fpclassify(a) == FP_ZERO;
-}
-
-INLINE float64 float64_scalbn(float64 a, int n STATUS_PARAM)
-{
- return scalbn(a, n);
-}
-
-#ifdef FLOATX80
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE extended double-precision conversion constants.
-*----------------------------------------------------------------------------*/
-#define floatx80_zero (0.0L)
-#define floatx80_one (1.0L)
-#define floatx80_ln2 (0.69314718055994530943L)
-#define floatx80_pi (3.14159265358979323851L)
-#define floatx80_half (0.5L)
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE extended double-precision conversion routines.
-*----------------------------------------------------------------------------*/
-int floatx80_to_int32( floatx80 STATUS_PARAM );
-int floatx80_to_int32_round_to_zero( floatx80 STATUS_PARAM );
-int64_t floatx80_to_int64( floatx80 STATUS_PARAM);
-int64_t floatx80_to_int64_round_to_zero( floatx80 STATUS_PARAM);
-float32 floatx80_to_float32( floatx80 STATUS_PARAM );
-float64 floatx80_to_float64( floatx80 STATUS_PARAM );
-#ifdef FLOAT128
-float128 floatx80_to_float128( floatx80 STATUS_PARAM );
-#endif
-
-/*----------------------------------------------------------------------------
-| Software IEC/IEEE extended double-precision operations.
-*----------------------------------------------------------------------------*/
-floatx80 floatx80_round_to_int( floatx80 STATUS_PARAM );
-INLINE floatx80 floatx80_add( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a + b;
-}
-INLINE floatx80 floatx80_sub( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a - b;
-}
-INLINE floatx80 floatx80_mul( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a * b;
-}
-INLINE floatx80 floatx80_div( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a / b;
-}
-floatx80 floatx80_rem( floatx80, floatx80 STATUS_PARAM );
-floatx80 floatx80_sqrt( floatx80 STATUS_PARAM );
-INLINE int floatx80_eq_quiet( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a == b;
-}
-INLINE int floatx80_le( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a <= b;
-}
-INLINE int floatx80_lt( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a < b;
-}
-INLINE int floatx80_eq( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return a <= b && a >= b;
-}
-INLINE int floatx80_le_quiet( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return islessequal(a, b);
-}
-INLINE int floatx80_lt_quiet( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return isless(a, b);
-
-}
-INLINE int floatx80_unordered( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-INLINE int floatx80_unordered_quiet( floatx80 a, floatx80 b STATUS_PARAM)
-{
- return isunordered(a, b);
-}
-int floatx80_compare( floatx80, floatx80 STATUS_PARAM );
-int floatx80_compare_quiet( floatx80, floatx80 STATUS_PARAM );
-int floatx80_is_signaling_nan( floatx80 );
-int floatx80_is_quiet_nan( floatx80 );
-int floatx80_is_any_nan( floatx80 );
-
-INLINE floatx80 floatx80_abs(floatx80 a)
-{
- return fabsl(a);
-}
-
-INLINE floatx80 floatx80_chs(floatx80 a)
-{
- return -a;
-}
-
-INLINE floatx80 floatx80_is_infinity(floatx80 a)
-{
- return fpclassify(a) == FP_INFINITE;
-}
-
-INLINE floatx80 floatx80_is_neg(floatx80 a)
-{
- floatx80u u;
- u.f = a;
- return u.i.high >> 15;
-}
-
-INLINE floatx80 floatx80_is_zero(floatx80 a)
-{
- return fpclassify(a) == FP_ZERO;
-}
-
-INLINE floatx80 floatx80_scalbn(floatx80 a, int n STATUS_PARAM)
-{
- return scalbnl(a, n);
-}
-
-#endif
diff --git a/fpu/softfloat.h b/fpu/softfloat.h
index 58c9b7b..999b95c 100644
--- a/fpu/softfloat.h
+++ b/fpu/softfloat.h
@@ -81,16 +81,9 @@ typedef int64_t int64;
| input or output the `floatx80' type will be defined. The same applies to
| the `FLOAT128' macro and the quadruple-precision format `float128'.
*----------------------------------------------------------------------------*/
-#ifdef CONFIG_SOFTFLOAT
/* bit exact soft float support */
#define FLOATX80
#define FLOAT128
-#else
-/* native float support */
-#if (defined(__i386__) || defined(__x86_64__)) && !defined(CONFIG_BSD)
-#define FLOATX80
-#endif
-#endif /* !CONFIG_SOFTFLOAT */
#define STATUS_PARAM , float_status *status
#define STATUS(field) status->field
@@ -106,7 +99,6 @@ enum {
float_relation_unordered = 2
};
-#ifdef CONFIG_SOFTFLOAT
/*----------------------------------------------------------------------------
| Software IEC/IEEE floating-point types.
*----------------------------------------------------------------------------*/
@@ -699,10 +691,4 @@ INLINE int float128_is_any_nan(float128 a)
#endif
-#else /* CONFIG_SOFTFLOAT */
-
-#include "softfloat-native.h"
-
-#endif /* !CONFIG_SOFTFLOAT */
-
#endif /* !SOFTFLOAT_H */
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 5/9] softfloat: always enable floatx80 and float128 support
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (3 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 4/9] softfloat-native: remove Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0() Aurelien Jarno
` (3 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
Now that softfloat-native is gone, there is no real point on not always
enabling floatx80 and float128 support.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
cpu-all.h | 2 -
fpu/softfloat-specialize.h | 7 -----
fpu/softfloat.c | 62 --------------------------------------------
fpu/softfloat.h | 47 ---------------------------------
4 files changed, 0 insertions(+), 118 deletions(-)
diff --git a/cpu-all.h b/cpu-all.h
index fc252ba..880f570 100644
--- a/cpu-all.h
+++ b/cpu-all.h
@@ -137,7 +137,6 @@ typedef union {
uint64_t ll;
} CPU_DoubleU;
-#if defined(FLOATX80)
typedef union {
floatx80 d;
struct {
@@ -145,7 +144,6 @@ typedef union {
uint16_t upper;
} l;
} CPU_LDoubleU;
-#endif
typedef union {
float128 q;
diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
index 9d68aae..c7d35a1 100644
--- a/fpu/softfloat-specialize.h
+++ b/fpu/softfloat-specialize.h
@@ -523,8 +523,6 @@ static float64 propagateFloat64NaN( float64 a, float64 b STATUS_PARAM)
}
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns 1 if the extended double-precision floating-point value `a' is a
| quiet NaN; otherwise returns 0. This slightly differs from the same
@@ -681,10 +679,6 @@ static floatx80 propagateFloatx80NaN( floatx80 a, floatx80 b STATUS_PARAM)
}
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns 1 if the quadruple-precision floating-point value `a' is a quiet
| NaN; otherwise returns 0.
@@ -820,4 +814,3 @@ static float128 propagateFloat128NaN( float128 a, float128 b STATUS_PARAM)
}
}
-#endif
diff --git a/fpu/softfloat.c b/fpu/softfloat.c
index e3cd8a7..7951a0e 100644
--- a/fpu/softfloat.c
+++ b/fpu/softfloat.c
@@ -64,12 +64,10 @@ void set_float_exception_flags(int val STATUS_PARAM)
STATUS(float_exception_flags) = val;
}
-#ifdef FLOATX80
void set_floatx80_rounding_precision(int val STATUS_PARAM)
{
STATUS(floatx80_rounding_precision) = val;
}
-#endif
/*----------------------------------------------------------------------------
| Returns the fraction bits of the half-precision floating-point value `a'.
@@ -564,8 +562,6 @@ static float64
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the fraction bits of the extended double-precision floating-point
| value `a'.
@@ -851,10 +847,6 @@ static floatx80
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the least-significant 64 fraction bits of the quadruple-precision
| floating-point value `a'.
@@ -1118,8 +1110,6 @@ static float128
}
-#endif
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 32-bit two's complement integer `a'
| to the single-precision floating-point format. The conversion is performed
@@ -1159,8 +1149,6 @@ float64 int32_to_float64( int32 a STATUS_PARAM )
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 32-bit two's complement integer `a'
| to the extended double-precision floating-point format. The conversion
@@ -1184,10 +1172,6 @@ floatx80 int32_to_floatx80( int32 a STATUS_PARAM )
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 32-bit two's complement integer `a' to
| the quadruple-precision floating-point format. The conversion is performed
@@ -1210,8 +1194,6 @@ float128 int32_to_float128( int32 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 64-bit two's complement integer `a'
| to the single-precision floating-point format. The conversion is performed
@@ -1291,8 +1273,6 @@ float64 uint64_to_float64( uint64 a STATUS_PARAM )
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 64-bit two's complement integer `a'
| to the extended double-precision floating-point format. The conversion
@@ -1314,10 +1294,6 @@ floatx80 int64_to_floatx80( int64 a STATUS_PARAM )
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the 64-bit two's complement integer `a' to
| the quadruple-precision floating-point format. The conversion is performed
@@ -1351,8 +1327,6 @@ float128 int64_to_float128( int64 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Returns the result of converting the single-precision floating-point value
| `a' to the 32-bit two's complement integer format. The conversion is
@@ -1590,8 +1564,6 @@ float64 float32_to_float64( float32 a STATUS_PARAM )
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the single-precision floating-point value
| `a' to the extended double-precision floating-point format. The conversion
@@ -1622,10 +1594,6 @@ floatx80 float32_to_floatx80( float32 a STATUS_PARAM )
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the single-precision floating-point value
| `a' to the double-precision floating-point format. The conversion is
@@ -1656,8 +1624,6 @@ float128 float32_to_float128( float32 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Rounds the single-precision floating-point value `a' to an integer, and
| returns the result as a single-precision floating-point value. The
@@ -2939,8 +2905,6 @@ float16 float32_to_float16(float32 a, flag ieee STATUS_PARAM)
return packFloat16(aSign, aExp + 14, aSig >> 13);
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the double-precision floating-point value
| `a' to the extended double-precision floating-point format. The conversion
@@ -2972,10 +2936,6 @@ floatx80 float64_to_floatx80( float64 a STATUS_PARAM )
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the double-precision floating-point value
| `a' to the quadruple-precision floating-point format. The conversion is
@@ -3007,8 +2967,6 @@ float128 float64_to_float128( float64 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Rounds the double-precision floating-point value `a' to an integer, and
| returns the result as a double-precision floating-point value. The
@@ -3816,8 +3774,6 @@ int float64_unordered_quiet( float64 a, float64 b STATUS_PARAM )
return 0;
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the extended double-precision floating-
| point value `a' to the 32-bit two's complement integer format. The
@@ -4030,8 +3986,6 @@ float64 floatx80_to_float64( floatx80 a STATUS_PARAM )
}
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the extended double-precision floating-
| point value `a' to the quadruple-precision floating-point format. The
@@ -4056,8 +4010,6 @@ float128 floatx80_to_float128( floatx80 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Rounds the extended double-precision floating-point value `a' to an integer,
| and returns the result as an extended quadruple-precision floating-point
@@ -4849,10 +4801,6 @@ int floatx80_unordered_quiet( floatx80 a, floatx80 b STATUS_PARAM )
return 0;
}
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Returns the result of converting the quadruple-precision floating-point
| value `a' to the 32-bit two's complement integer format. The conversion
@@ -5102,8 +5050,6 @@ float64 float128_to_float64( float128 a STATUS_PARAM )
}
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Returns the result of converting the quadruple-precision floating-point
| value `a' to the extended double-precision floating-point format. The
@@ -5139,8 +5085,6 @@ floatx80 float128_to_floatx80( float128 a STATUS_PARAM )
}
-#endif
-
/*----------------------------------------------------------------------------
| Rounds the quadruple-precision floating-point value `a' to an integer, and
| returns the result as a quadruple-precision floating-point value. The
@@ -6020,8 +5964,6 @@ int float128_unordered_quiet( float128 a, float128 b STATUS_PARAM )
return 0;
}
-#endif
-
/* misc functions */
float32 uint32_to_float32( unsigned int a STATUS_PARAM )
{
@@ -6423,7 +6365,6 @@ float64 float64_scalbn( float64 a, int n STATUS_PARAM )
return normalizeRoundAndPackFloat64( aSign, aExp, aSig STATUS_VAR );
}
-#ifdef FLOATX80
floatx80 floatx80_scalbn( floatx80 a, int n STATUS_PARAM )
{
flag aSign;
@@ -6454,9 +6395,7 @@ floatx80 floatx80_scalbn( floatx80 a, int n STATUS_PARAM )
return normalizeRoundAndPackFloatx80( STATUS(floatx80_rounding_precision),
aSign, aExp, aSig, 0 STATUS_VAR );
}
-#endif
-#ifdef FLOAT128
float128 float128_scalbn( float128 a, int n STATUS_PARAM )
{
flag aSign;
@@ -6489,4 +6428,3 @@ float128 float128_scalbn( float128 a, int n STATUS_PARAM )
STATUS_VAR );
}
-#endif
diff --git a/fpu/softfloat.h b/fpu/softfloat.h
index 999b95c..8931446 100644
--- a/fpu/softfloat.h
+++ b/fpu/softfloat.h
@@ -74,17 +74,6 @@ typedef int64_t int64;
#define SNAN_BIT_IS_ONE 0
#endif
-/*----------------------------------------------------------------------------
-| The macro `FLOATX80' must be defined to enable the extended double-precision
-| floating-point format `floatx80'. If this macro is not defined, the
-| `floatx80' type will not be defined, and none of the functions that either
-| input or output the `floatx80' type will be defined. The same applies to
-| the `FLOAT128' macro and the quadruple-precision format `float128'.
-*----------------------------------------------------------------------------*/
-/* bit exact soft float support */
-#define FLOATX80
-#define FLOAT128
-
#define STATUS_PARAM , float_status *status
#define STATUS(field) status->field
#define STATUS_VAR , status
@@ -141,14 +130,11 @@ typedef uint64_t float64;
#define const_float32(x) (x)
#define const_float64(x) (x)
#endif
-#ifdef FLOATX80
typedef struct {
uint64_t low;
uint16_t high;
} floatx80;
#define make_floatx80(exp, mant) ((floatx80) { mant, exp })
-#endif
-#ifdef FLOAT128
typedef struct {
#ifdef HOST_WORDS_BIGENDIAN
uint64_t high, low;
@@ -156,7 +142,6 @@ typedef struct {
uint64_t low, high;
#endif
} float128;
-#endif
/*----------------------------------------------------------------------------
| Software IEC/IEEE floating-point underflow tininess-detection mode.
@@ -193,9 +178,7 @@ typedef struct float_status {
signed char float_detect_tininess;
signed char float_rounding_mode;
signed char float_exception_flags;
-#ifdef FLOATX80
signed char floatx80_rounding_precision;
-#endif
/* should denormalised results go to zero and set the inexact flag? */
flag flush_to_zero;
/* should denormalised inputs go to zero and set the input_denormal flag? */
@@ -225,9 +208,7 @@ INLINE int get_float_exception_flags(float_status *status)
{
return STATUS(float_exception_flags);
}
-#ifdef FLOATX80
void set_floatx80_rounding_precision(int val STATUS_PARAM);
-#endif
/*----------------------------------------------------------------------------
| Routine to raise any or all of the software IEC/IEEE floating-point
@@ -242,22 +223,14 @@ float32 int32_to_float32( int32 STATUS_PARAM );
float64 int32_to_float64( int32 STATUS_PARAM );
float32 uint32_to_float32( unsigned int STATUS_PARAM );
float64 uint32_to_float64( unsigned int STATUS_PARAM );
-#ifdef FLOATX80
floatx80 int32_to_floatx80( int32 STATUS_PARAM );
-#endif
-#ifdef FLOAT128
float128 int32_to_float128( int32 STATUS_PARAM );
-#endif
float32 int64_to_float32( int64 STATUS_PARAM );
float32 uint64_to_float32( uint64 STATUS_PARAM );
float64 int64_to_float64( int64 STATUS_PARAM );
float64 uint64_to_float64( uint64 STATUS_PARAM );
-#ifdef FLOATX80
floatx80 int64_to_floatx80( int64 STATUS_PARAM );
-#endif
-#ifdef FLOAT128
float128 int64_to_float128( int64 STATUS_PARAM );
-#endif
/*----------------------------------------------------------------------------
| Software half-precision conversion routines.
@@ -295,12 +268,8 @@ uint32 float32_to_uint32_round_to_zero( float32 STATUS_PARAM );
int64 float32_to_int64( float32 STATUS_PARAM );
int64 float32_to_int64_round_to_zero( float32 STATUS_PARAM );
float64 float32_to_float64( float32 STATUS_PARAM );
-#ifdef FLOATX80
floatx80 float32_to_floatx80( float32 STATUS_PARAM );
-#endif
-#ifdef FLOAT128
float128 float32_to_float128( float32 STATUS_PARAM );
-#endif
/*----------------------------------------------------------------------------
| Software IEC/IEEE single-precision operations.
@@ -412,12 +381,8 @@ int64 float64_to_int64_round_to_zero( float64 STATUS_PARAM );
uint64 float64_to_uint64 (float64 a STATUS_PARAM);
uint64 float64_to_uint64_round_to_zero (float64 a STATUS_PARAM);
float32 float64_to_float32( float64 STATUS_PARAM );
-#ifdef FLOATX80
floatx80 float64_to_floatx80( float64 STATUS_PARAM );
-#endif
-#ifdef FLOAT128
float128 float64_to_float128( float64 STATUS_PARAM );
-#endif
/*----------------------------------------------------------------------------
| Software IEC/IEEE double-precision operations.
@@ -510,8 +475,6 @@ INLINE float64 float64_set_sign(float64 a, int sign)
#define float64_default_nan make_float64(LIT64( 0xFFF8000000000000 ))
#endif
-#ifdef FLOATX80
-
/*----------------------------------------------------------------------------
| Software IEC/IEEE extended double-precision conversion routines.
*----------------------------------------------------------------------------*/
@@ -521,9 +484,7 @@ int64 floatx80_to_int64( floatx80 STATUS_PARAM );
int64 floatx80_to_int64_round_to_zero( floatx80 STATUS_PARAM );
float32 floatx80_to_float32( floatx80 STATUS_PARAM );
float64 floatx80_to_float64( floatx80 STATUS_PARAM );
-#ifdef FLOAT128
float128 floatx80_to_float128( floatx80 STATUS_PARAM );
-#endif
/*----------------------------------------------------------------------------
| Software IEC/IEEE extended double-precision operations.
@@ -602,10 +563,6 @@ INLINE int floatx80_is_any_nan(floatx80 a)
#define floatx80_default_nan_low LIT64( 0xC000000000000000 )
#endif
-#endif
-
-#ifdef FLOAT128
-
/*----------------------------------------------------------------------------
| Software IEC/IEEE quadruple-precision conversion routines.
*----------------------------------------------------------------------------*/
@@ -615,9 +572,7 @@ int64 float128_to_int64( float128 STATUS_PARAM );
int64 float128_to_int64_round_to_zero( float128 STATUS_PARAM );
float32 float128_to_float32( float128 STATUS_PARAM );
float64 float128_to_float64( float128 STATUS_PARAM );
-#ifdef FLOATX80
floatx80 float128_to_floatx80( float128 STATUS_PARAM );
-#endif
/*----------------------------------------------------------------------------
| Software IEC/IEEE quadruple-precision operations.
@@ -689,6 +644,4 @@ INLINE int float128_is_any_nan(float128 a)
#define float128_default_nan_low LIT64( 0x0000000000000000 )
#endif
-#endif
-
#endif /* !SOFTFLOAT_H */
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0()
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (4 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 5/9] softfloat: always enable floatx80 and float128 support Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 22:11 ` Peter Maydell
2011-05-29 10:38 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 7/9] softfloat: add float*_is_zero_or_denormal() Aurelien Jarno
` (2 subsequent siblings)
8 siblings, 2 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Andreas Färber, Aurelien Jarno
Instead of using a table which doesn't correspond to anything from
physical in the CPU, use directly the constants in helper_fld*_ST0().
Cc: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
target-i386/op_helper.c | 27 ++++++++-------------------
1 files changed, 8 insertions(+), 19 deletions(-)
diff --git a/target-i386/op_helper.c b/target-i386/op_helper.c
index 4d309ab..cec0c76 100644
--- a/target-i386/op_helper.c
+++ b/target-i386/op_helper.c
@@ -99,17 +99,6 @@ static const uint8_t rclb_table[32] = {
#define floatx80_l2e make_floatx80( 0x3fff, 0xb8aa3b295c17f0bcLL )
#define floatx80_l2t make_floatx80( 0x4000, 0xd49a784bcd1b8afeLL )
-static const floatx80 f15rk[7] =
-{
- floatx80_zero,
- floatx80_one,
- floatx80_pi,
- floatx80_lg2,
- floatx80_ln2,
- floatx80_l2e,
- floatx80_l2t,
-};
-
/* broken thread support */
static spinlock_t global_cpu_lock = SPIN_LOCK_UNLOCKED;
@@ -3816,42 +3805,42 @@ void helper_fabs_ST0(void)
void helper_fld1_ST0(void)
{
- ST0 = f15rk[1];
+ ST0 = floatx80_one;
}
void helper_fldl2t_ST0(void)
{
- ST0 = f15rk[6];
+ ST0 = floatx80_l2t;
}
void helper_fldl2e_ST0(void)
{
- ST0 = f15rk[5];
+ ST0 = floatx80_l2e;
}
void helper_fldpi_ST0(void)
{
- ST0 = f15rk[2];
+ ST0 = floatx80_pi;
}
void helper_fldlg2_ST0(void)
{
- ST0 = f15rk[3];
+ ST0 = floatx80_lg2;
}
void helper_fldln2_ST0(void)
{
- ST0 = f15rk[4];
+ ST0 = floatx80_ln2;
}
void helper_fldz_ST0(void)
{
- ST0 = f15rk[0];
+ ST0 = floatx80_zero;
}
void helper_fldz_FT0(void)
{
- FT0 = f15rk[0];
+ FT0 = floatx80_zero;
}
uint32_t helper_fnstsw(void)
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 7/9] softfloat: add float*_is_zero_or_denormal()
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (5 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0() Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0() Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions Aurelien Jarno
8 siblings, 0 replies; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
float*_is_zero_or_denormal() is available for float32, but not for
float64, floatx80 and float128. Fix that.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
fpu/softfloat.h | 15 +++++++++++++++
1 files changed, 15 insertions(+), 0 deletions(-)
diff --git a/fpu/softfloat.h b/fpu/softfloat.h
index 8931446..bde2500 100644
--- a/fpu/softfloat.h
+++ b/fpu/softfloat.h
@@ -449,6 +449,11 @@ INLINE int float64_is_any_nan(float64 a)
return ((float64_val(a) & ~(1ULL << 63)) > 0x7ff0000000000000ULL);
}
+INLINE int float64_is_zero_or_denormal(float64 a)
+{
+ return (float64_val(a) & 0x7ff0000000000000LL) == 0;
+}
+
INLINE float64 float64_set_sign(float64 a, int sign)
{
return make_float64((float64_val(a) & 0x7fffffffffffffffULL)
@@ -538,6 +543,11 @@ INLINE int floatx80_is_zero(floatx80 a)
return (a.high & 0x7fff) == 0 && a.low == 0;
}
+INLINE int floatx80_is_zero_or_denormal(floatx80 a)
+{
+ return (a.high & 0x7fff) == 0;
+}
+
INLINE int floatx80_is_any_nan(floatx80 a)
{
return ((a.high & 0x7fff) == 0x7fff) && (a.low<<1);
@@ -626,6 +636,11 @@ INLINE int float128_is_zero(float128 a)
return (a.high & 0x7fffffffffffffffLL) == 0 && a.low == 0;
}
+INLINE int float128_is_zero_or_denormal(float128 a)
+{
+ return (a.high & 0x7fff000000000000LL) == 0;
+}
+
INLINE int float128_is_any_nan(float128 a)
{
return ((a.high >> 48) & 0x7fff) == 0x7fff &&
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0()
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (6 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 7/9] softfloat: add float*_is_zero_or_denormal() Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-23 22:02 ` Peter Maydell
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions Aurelien Jarno
8 siblings, 1 reply; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
Rewrite helper_fxam_ST0() using only softfloat functions.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
target-i386/op_helper.c | 30 ++++++++++++------------------
1 files changed, 12 insertions(+), 18 deletions(-)
diff --git a/target-i386/op_helper.c b/target-i386/op_helper.c
index cec0c76..8ba2b5f 100644
--- a/target-i386/op_helper.c
+++ b/target-i386/op_helper.c
@@ -4241,29 +4241,23 @@ void helper_fcos(void)
void helper_fxam_ST0(void)
{
- CPU_LDoubleU temp;
- int expdif;
-
- temp.d = ST0;
-
env->fpus &= (~0x4700); /* (C3,C2,C1,C0) <-- 0000 */
- if (SIGND(temp))
+
+ if (floatx80_is_neg(ST0)) {
env->fpus |= 0x200; /* C1 <-- 1 */
+ }
/* XXX: test fptags too */
- expdif = EXPD(temp);
- if (expdif == MAXEXPD) {
- if (MANTD(temp) == 0x8000000000000000ULL)
- env->fpus |= 0x500 /*Infinity*/;
- else
- env->fpus |= 0x100 /*NaN*/;
- } else if (expdif == 0) {
- if (MANTD(temp) == 0)
- env->fpus |= 0x4000 /*Zero*/;
- else
- env->fpus |= 0x4400 /*Denormal*/;
+ if (floatx80_is_infinity(ST0)) {
+ env->fpus |= 0x500; /* (C3,C2,C0) <-- 011 */
+ } else if (floatx80_is_any_nan(ST0)) {
+ env->fpus |= 0x100; /* (C3,C2,C0) <-- 001 */
+ } else if (floatx80_is_zero(ST0)) {
+ env->fpus |= 0x4000; /* (C3,C2,C0) <-- 100 */
+ } else if (floatx80_is_zero_or_denormal(ST0)) {
+ env->fpus |= 0x4400; /* (C3,C2,C0) <-- 110 */
} else {
- env->fpus |= 0x400;
+ env->fpus |= 0x400; /* (C3,C2,C0) <-- 010 */
}
}
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
` (7 preceding siblings ...)
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0() Aurelien Jarno
@ 2011-05-23 21:42 ` Aurelien Jarno
2011-05-24 15:55 ` Peter Maydell
8 siblings, 1 reply; 18+ messages in thread
From: Aurelien Jarno @ 2011-05-23 21:42 UTC (permalink / raw)
To: qemu-devel; +Cc: Aurelien Jarno
This patch adds support for FPU exceptions. It keeps the exception in
the softfloat status, and copy them back to env->fpus when needed by
oring them. When loading a new value to env->fpus, it starts with a
clean softfloat status.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
target-i386/cpu.h | 15 +++++++++++++++
target-i386/exec.h | 12 ------------
target-i386/helper.c | 36 ++++++++++++++++++++++++++++++++++++
target-i386/machine.c | 3 ++-
target-i386/op_helper.c | 32 ++++++++++++++++----------------
5 files changed, 69 insertions(+), 29 deletions(-)
diff --git a/target-i386/cpu.h b/target-i386/cpu.h
index fe65886..d216356 100644
--- a/target-i386/cpu.h
+++ b/target-i386/cpu.h
@@ -121,6 +121,19 @@
#define VIP_MASK 0x00100000
#define ID_MASK 0x00200000
+/* FPU flags */
+#define FPUS_IE (1 << 0)
+#define FPUS_DE (1 << 1)
+#define FPUS_ZE (1 << 2)
+#define FPUS_OE (1 << 3)
+#define FPUS_UE (1 << 4)
+#define FPUS_PE (1 << 5)
+#define FPUS_SF (1 << 6)
+#define FPUS_SE (1 << 7)
+#define FPUS_B (1 << 15)
+
+#define FPUC_EM 0x3f
+
/* hidden flags - used internally by qemu to represent additional cpu
states. Only the CPL, INHIBIT_IRQ, SMM and SVMI are not
redundant. We avoid using the IOPL_MASK, TF_MASK and VM_MASK bit
@@ -877,6 +890,8 @@ void host_cpuid(uint32_t function, uint32_t count,
uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx);
/* helper.c */
+void cpu_x86_set_fpus(CPUX86State *s, uint16_t val);
+void cpu_x86_update_fpus(CPUX86State *s);
int cpu_x86_handle_mmu_fault(CPUX86State *env, target_ulong addr,
int is_write, int mmu_idx, int is_softmmu);
#define cpu_handle_mmu_fault cpu_x86_handle_mmu_fault
diff --git a/target-i386/exec.h b/target-i386/exec.h
index 9bd080e..e14d0d9 100644
--- a/target-i386/exec.h
+++ b/target-i386/exec.h
@@ -144,18 +144,6 @@ static inline void helper_fstt(floatx80 f, target_ulong ptr)
stw(ptr + 8, temp.l.upper);
}
-#define FPUS_IE (1 << 0)
-#define FPUS_DE (1 << 1)
-#define FPUS_ZE (1 << 2)
-#define FPUS_OE (1 << 3)
-#define FPUS_UE (1 << 4)
-#define FPUS_PE (1 << 5)
-#define FPUS_SF (1 << 6)
-#define FPUS_SE (1 << 7)
-#define FPUS_B (1 << 15)
-
-#define FPUC_EM 0x3f
-
static inline uint32_t compute_eflags(void)
{
return env->eflags | helper_cc_compute_all(CC_OP) | (DF & DF_MASK);
diff --git a/target-i386/helper.c b/target-i386/helper.c
index 5c4b288..23646d3 100644
--- a/target-i386/helper.c
+++ b/target-i386/helper.c
@@ -396,6 +396,7 @@ void cpu_dump_state(CPUState *env, FILE *f, fprintf_function cpu_fprintf,
for(i = 0; i < 8; i++) {
fptag |= ((!env->fptags[i]) << i);
}
+ cpu_x86_update_fpus(env);
cpu_fprintf(f, "FCW=%04x FSW=%04x [ST=%d] FTW=%02x MXCSR=%08x\n",
env->fpuc,
(env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11,
@@ -1238,6 +1239,41 @@ int cpu_x86_get_descr_debug(CPUX86State *env, unsigned int selector,
return 1;
}
+void cpu_x86_set_fpus(CPUX86State *s, uint16_t val)
+{
+ set_float_exception_flags(0, &s->fp_status);
+ s->fpus = val;
+}
+
+void cpu_x86_update_fpus(CPUX86State *s)
+{
+ int xcpt = get_float_exception_flags(&s->fp_status);
+
+ if (xcpt) {
+ if (xcpt & float_flag_invalid) {
+ s->fpus |= FPUS_IE;
+ }
+ if (xcpt & float_flag_input_denormal) {
+ s->fpus |= FPUS_DE;
+ }
+ if (xcpt & float_flag_divbyzero) {
+ s->fpus |= FPUS_ZE;
+ }
+ if (xcpt & float_flag_overflow) {
+ s->fpus |= FPUS_OE;
+ }
+ if (xcpt & float_flag_underflow) {
+ s->fpus |= FPUS_UE;
+ }
+ if (xcpt & float_flag_inexact) {
+ s->fpus |= FPUS_PE;
+ }
+ if (s->fpus & (~s->fpuc & FPUC_EM)) {
+ s->fpus |= FPUS_SE | FPUS_B;
+ }
+ }
+}
+
CPUX86State *cpu_x86_init(const char *cpu_model)
{
CPUX86State *env;
diff --git a/target-i386/machine.c b/target-i386/machine.c
index bbeae88..82c5c01 100644
--- a/target-i386/machine.c
+++ b/target-i386/machine.c
@@ -241,6 +241,7 @@ static void cpu_pre_save(void *opaque)
int i;
/* FPU */
+ cpu_x86_update_fpus(env);
env->fpus_vmstate = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
env->fptag_vmstate = 0;
for(i = 0; i < 8; i++) {
@@ -257,7 +258,7 @@ static int cpu_post_load(void *opaque, int version_id)
/* XXX: restore FPU round state */
env->fpstt = (env->fpus_vmstate >> 11) & 7;
- env->fpus = env->fpus_vmstate & ~0x3800;
+ cpu_x86_set_fpus(env, env->fpus_vmstate & ~0x3800);
env->fptag_vmstate ^= 0xff;
for(i = 0; i < 8; i++) {
env->fptags[i] = (env->fptag_vmstate >> i) & 1;
diff --git a/target-i386/op_helper.c b/target-i386/op_helper.c
index 8ba2b5f..eccb957 100644
--- a/target-i386/op_helper.c
+++ b/target-i386/op_helper.c
@@ -3447,18 +3447,8 @@ static inline floatx80 double_to_floatx80(double a)
return float64_to_floatx80(u.f64, &env->fp_status);
}
-static void fpu_set_exception(int mask)
-{
- env->fpus |= mask;
- if (env->fpus & (~env->fpuc & FPUC_EM))
- env->fpus |= FPUS_SE | FPUS_B;
-}
-
static inline floatx80 helper_fdiv(floatx80 a, floatx80 b)
{
- if (floatx80_is_zero(b)) {
- fpu_set_exception(FPUS_ZE);
- }
return floatx80_div(a, b, &env->fp_status);
}
@@ -3845,6 +3835,7 @@ void helper_fldz_FT0(void)
uint32_t helper_fnstsw(void)
{
+ cpu_x86_update_fpus(env);
return (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
}
@@ -3897,18 +3888,19 @@ void helper_fldcw(uint32_t val)
void helper_fclex(void)
{
- env->fpus &= 0x7f00;
+ cpu_x86_set_fpus(env, env->fpus & 0x7f00);
}
void helper_fwait(void)
{
+ cpu_x86_update_fpus(env);
if (env->fpus & FPUS_SE)
fpu_raise_exception();
}
void helper_fninit(void)
{
- env->fpus = 0;
+ cpu_x86_set_fpus(env, 0);
env->fpstt = 0;
env->fpuc = 0x37f;
env->fptags[0] = 1;
@@ -4208,7 +4200,13 @@ void helper_fscale(void)
if (floatx80_is_any_nan(ST1)) {
ST0 = ST1;
} else {
- int n = floatx80_to_int32_round_to_zero(ST1, &env->fp_status);
+ int n, x;
+
+ /* The float to int conversion should not generate any exception. */
+ x = get_float_exception_flags(&env->fp_status);
+ n = floatx80_to_int32_round_to_zero(ST1, &env->fp_status);
+ set_float_exception_flags(x, &env->fp_status);
+
ST0 = floatx80_scalbn(ST0, n, &env->fp_status);
}
}
@@ -4267,6 +4265,7 @@ void helper_fstenv(target_ulong ptr, int data32)
uint64_t mant;
CPU_LDoubleU tmp;
+ cpu_x86_update_fpus(env);
fpus = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
fptag = 0;
for (i=7; i>=0; i--) {
@@ -4324,7 +4323,7 @@ void helper_fldenv(target_ulong ptr, int data32)
fptag = lduw(ptr + 4);
}
env->fpstt = (fpus >> 11) & 7;
- env->fpus = fpus & ~0x3800;
+ cpu_x86_set_fpus(env, fpus & ~0x3800);
for(i = 0;i < 8; i++) {
env->fptags[i] = ((fptag & 3) == 3);
fptag >>= 2;
@@ -4346,7 +4345,7 @@ void helper_fsave(target_ulong ptr, int data32)
}
/* fninit */
- env->fpus = 0;
+ cpu_x86_set_fpus(env, 0);
env->fpstt = 0;
env->fpuc = 0x37f;
env->fptags[0] = 1;
@@ -4385,6 +4384,7 @@ void helper_fxsave(target_ulong ptr, int data64)
raise_exception(EXCP0D_GPF);
}
+ cpu_x86_update_fpus(env);
fpus = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
fptag = 0;
for(i = 0; i < 8; i++) {
@@ -4450,7 +4450,7 @@ void helper_fxrstor(target_ulong ptr, int data64)
fpus = lduw(ptr + 2);
fptag = lduw(ptr + 4);
env->fpstt = (fpus >> 11) & 7;
- env->fpus = fpus & ~0x3800;
+ cpu_x86_set_fpus(env, fpus & ~0x3800);
fptag ^= 0xff;
for(i = 0;i < 8; i++) {
env->fptags[i] = ((fptag >> i) & 1);
--
1.7.5.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0()
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0() Aurelien Jarno
@ 2011-05-23 22:02 ` Peter Maydell
0 siblings, 0 replies; 18+ messages in thread
From: Peter Maydell @ 2011-05-23 22:02 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
On 23 May 2011 22:42, Aurelien Jarno <aurelien@aurel32.net> wrote:
> Rewrite helper_fxam_ST0() using only softfloat functions.
>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
(Sorry I didn't get round to this one earlier; I think I did
the easy patches and stalled on the ones I'd have to pull out
the x86 manuals for :-))
> /* XXX: test fptags too */
You could fix this XXX while you were here; it's just adding
if (env->fptags[env->fpstt]) {
env->fpus |= 0x4100; /* (C3,C2,C0) <-- 101 */
} else ...
to the top of the ladder, right?
(Catching the "Unsupported" case would be slightly more tedious,
we'd need to identify all the floatx80 weirdo bitpatterns
as per the table in the architecture manual. Could use a
FIXME/XXX comment on that account, maybe.)
> - expdif = EXPD(temp);
> - if (expdif == MAXEXPD) {
> - if (MANTD(temp) == 0x8000000000000000ULL)
> - env->fpus |= 0x500 /*Infinity*/;
> - else
> - env->fpus |= 0x100 /*NaN*/;
> - } else if (expdif == 0) {
> - if (MANTD(temp) == 0)
> - env->fpus |= 0x4000 /*Zero*/;
> - else
> - env->fpus |= 0x4400 /*Denormal*/;
> + if (floatx80_is_infinity(ST0)) {
> + env->fpus |= 0x500; /* (C3,C2,C0) <-- 011 */
> + } else if (floatx80_is_any_nan(ST0)) {
> + env->fpus |= 0x100; /* (C3,C2,C0) <-- 001 */
> + } else if (floatx80_is_zero(ST0)) {
> + env->fpus |= 0x4000; /* (C3,C2,C0) <-- 100 */
> + } else if (floatx80_is_zero_or_denormal(ST0)) {
> + env->fpus |= 0x4400; /* (C3,C2,C0) <-- 110 */
> } else {
> - env->fpus |= 0x400;
> + env->fpus |= 0x400; /* (C3,C2,C0) <-- 010 */
> }
> }
This all looks right. The C0/C1/C2/C3 bits seem to get enough
use in this and other functions to merit FPUS_* constants,
although that should probably be a different patch.
There's also code in helper_fstenv() which is currently
doing manual identification of zero/NaN/inf/denorm.
-- PMM
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef Aurelien Jarno
@ 2011-05-23 22:09 ` Peter Maydell
2011-05-29 11:01 ` Andreas Färber
1 sibling, 0 replies; 18+ messages in thread
From: Peter Maydell @ 2011-05-23 22:09 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
On 23 May 2011 22:42, Aurelien Jarno <aurelien@aurel32.net> wrote:
> target-mips has been switched to softfloat only long ago, but
> a #ifndef CONFIG_SOFTFLOAT has been forgotten. Remove it.
>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0()
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0() Aurelien Jarno
@ 2011-05-23 22:11 ` Peter Maydell
2011-05-29 10:38 ` Andreas Färber
1 sibling, 0 replies; 18+ messages in thread
From: Peter Maydell @ 2011-05-23 22:11 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: Andreas Färber, qemu-devel
On 23 May 2011 22:42, Aurelien Jarno <aurelien@aurel32.net> wrote:
> Instead of using a table which doesn't correspond to anything from
> physical in the CPU, use directly the constants in helper_fld*_ST0().
>
> Cc: Andreas Färber <andreas.faerber@web.de>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64 Aurelien Jarno
@ 2011-05-24 15:12 ` Peter Maydell
0 siblings, 0 replies; 18+ messages in thread
From: Peter Maydell @ 2011-05-24 15:12 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
On 23 May 2011 22:42, Aurelien Jarno <aurelien@aurel32.net> wrote:
> Now that target-i386 uses softfloat, floatx80 is always available and
> there is no need anymore to have code handling both float64 and floax80.
>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions Aurelien Jarno
@ 2011-05-24 15:55 ` Peter Maydell
0 siblings, 0 replies; 18+ messages in thread
From: Peter Maydell @ 2011-05-24 15:55 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
On 23 May 2011 22:42, Aurelien Jarno <aurelien@aurel32.net> wrote:
> This patch adds support for FPU exceptions. It keeps the exception in
> the softfloat status, and copy them back to env->fpus when needed by
> oring them. When loading a new value to env->fpus, it starts with a
> clean softfloat status.
>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
gdbstub.c still looks at env->fpus directly without calling
cpu_x86_update_fpus() first; it should probably go through
helper_fnstsw() now (or if you like a utility function that does what
helper_fnstsw() does now).
Similarly the code in helper_fstenv() and cpu_pre_save() that does:
cpu_x86_update_fpus(env);
SOMETHING = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
could be usefully abstracted away into a call to this utility fn.
For the rest, it looks OK, but the cpu_x86_update_fpus() and
cpu_x86_set_fpus() functions feel a bit like they're only half
an abstraction layer -- so sometimes it's OK to tweak env->fpus
directly and sometimes you need to use the functions. That means
it's tricky to tell from eyeballing or grepping the code whether an
update_fpus() call got missed out. Maybe it would be better to have
a set of functions such that you could mandate that all env->fpus
accesses went via the functions?
A random nitpick:
> @@ -4208,7 +4200,13 @@ void helper_fscale(void)
> if (floatx80_is_any_nan(ST1)) {
> ST0 = ST1;
> } else {
> - int n = floatx80_to_int32_round_to_zero(ST1, &env->fp_status);
> + int n, x;
> +
> + /* The float to int conversion should not generate any exception. */
> + x = get_float_exception_flags(&env->fp_status);
> + n = floatx80_to_int32_round_to_zero(ST1, &env->fp_status);
> + set_float_exception_flags(x, &env->fp_status);
> +
> ST0 = floatx80_scalbn(ST0, n, &env->fp_status);
> }
> }
...doesn't this mean you won't set the #D flag if ST1 is denormal?
I'm unconvinced about scalbn as a softfloat primitive. It gets used:
* by ARM for the VCVT fixed-point conversions, as a combination of
scalbn + float-to-int or int-to-float + scalbn
* by PPC for the same reason (vctuxs, vctsxs, vcfux, vcfsx)
* by x86 here as a float-to-int + scalbn, for FSCALE
and I think that in all three cases the attempt to implement a floating
point op by composing two softfloat primitives causes problems with
getting the floating point exception flags right. (In the fixedpoint
conversion case the issue is that for instance if the input is a
negative number we should only set InvalidOp but the scalbn is likely
to result in our also setting Inexact. But if you don't let the
scalbn set any flags at all then you have to special case when the
input is a NaN or denormal, which is equally awkward.)
-- PMM
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0()
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0() Aurelien Jarno
2011-05-23 22:11 ` Peter Maydell
@ 2011-05-29 10:38 ` Andreas Färber
1 sibling, 0 replies; 18+ messages in thread
From: Andreas Färber @ 2011-05-29 10:38 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
Am 23.05.2011 um 23:42 schrieb Aurelien Jarno:
> Instead of using a table which doesn't correspond to anything from
> physical in the CPU, use directly the constants in helper_fld*_ST0().
>
> Cc: Andreas Färber <andreas.faerber@web.de>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Tested-by: Andreas Färber <andreas.faerber@web.de>
I'm still wondering *why* the table doesn't work though... cf. reply
to Peter. I'll try some simplified sample code.
Andreas
> ---
> target-i386/op_helper.c | 27 ++++++++-------------------
> 1 files changed, 8 insertions(+), 19 deletions(-)
>
> diff --git a/target-i386/op_helper.c b/target-i386/op_helper.c
> index 4d309ab..cec0c76 100644
> --- a/target-i386/op_helper.c
> +++ b/target-i386/op_helper.c
> @@ -99,17 +99,6 @@ static const uint8_t rclb_table[32] = {
> #define floatx80_l2e make_floatx80( 0x3fff, 0xb8aa3b295c17f0bcLL )
> #define floatx80_l2t make_floatx80( 0x4000, 0xd49a784bcd1b8afeLL )
>
> -static const floatx80 f15rk[7] =
> -{
> - floatx80_zero,
> - floatx80_one,
> - floatx80_pi,
> - floatx80_lg2,
> - floatx80_ln2,
> - floatx80_l2e,
> - floatx80_l2t,
> -};
> -
> /* broken thread support */
>
> static spinlock_t global_cpu_lock = SPIN_LOCK_UNLOCKED;
> @@ -3816,42 +3805,42 @@ void helper_fabs_ST0(void)
>
> void helper_fld1_ST0(void)
> {
> - ST0 = f15rk[1];
> + ST0 = floatx80_one;
> }
>
> void helper_fldl2t_ST0(void)
> {
> - ST0 = f15rk[6];
> + ST0 = floatx80_l2t;
> }
>
> void helper_fldl2e_ST0(void)
> {
> - ST0 = f15rk[5];
> + ST0 = floatx80_l2e;
> }
>
> void helper_fldpi_ST0(void)
> {
> - ST0 = f15rk[2];
> + ST0 = floatx80_pi;
> }
>
> void helper_fldlg2_ST0(void)
> {
> - ST0 = f15rk[3];
> + ST0 = floatx80_lg2;
> }
>
> void helper_fldln2_ST0(void)
> {
> - ST0 = f15rk[4];
> + ST0 = floatx80_ln2;
> }
>
> void helper_fldz_ST0(void)
> {
> - ST0 = f15rk[0];
> + ST0 = floatx80_zero;
> }
>
> void helper_fldz_FT0(void)
> {
> - FT0 = f15rk[0];
> + FT0 = floatx80_zero;
> }
>
> uint32_t helper_fnstsw(void)
> --
> 1.7.5.1
>
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef Aurelien Jarno
@ 2011-05-29 10:58 ` Andreas Färber
0 siblings, 0 replies; 18+ messages in thread
From: Andreas Färber @ 2011-05-29 10:58 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel, Alexander Graf
Am 23.05.2011 um 23:42 schrieb Aurelien Jarno:
> target-ppc has been switched to softfloat only long ago, but a
> few #ifdef CONFIG_SOFTFLOAT have been forgotten. Remove them.
>
> Cc: Alexander Graf <agraf@suse.de>
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Only removes now-useless #ifdefs and dead code. The dead code in
helper_float_check_status() differs from the softfloat code path in
that the softfloat path also evaluations the softfloat status and
calls some exception helpers. So we're not losing anything.
Acked-by: Andreas Färber <andreas.faerber@web.de>
> ---
> target-ppc/helper.h | 2 --
> target-ppc/op_helper.c | 11 -----------
> target-ppc/translate.c | 2 --
> 3 files changed, 0 insertions(+), 15 deletions(-)
>
> diff --git a/target-ppc/helper.h b/target-ppc/helper.h
> index 51c99c8..470e42f 100644
> --- a/target-ppc/helper.h
> +++ b/target-ppc/helper.h
> @@ -51,9 +51,7 @@ DEF_HELPER_FLAGS_1(cntlzw32, TCG_CALL_CONST |
> TCG_CALL_PURE, i32, i32)
> DEF_HELPER_FLAGS_2(brinc, TCG_CALL_CONST | TCG_CALL_PURE, tl, tl, tl)
>
> DEF_HELPER_0(float_check_status, void)
> -#ifdef CONFIG_SOFTFLOAT
> DEF_HELPER_0(reset_fpstatus, void)
> -#endif
> DEF_HELPER_2(compute_fprf, i32, i64, i32)
> DEF_HELPER_2(store_fpscr, void, i64, i32)
> DEF_HELPER_1(fpscr_clrbit, void, i32)
> diff --git a/target-ppc/op_helper.c b/target-ppc/op_helper.c
> index c6cb448..fe66f14 100644
> --- a/target-ppc/op_helper.c
> +++ b/target-ppc/op_helper.c
> @@ -971,7 +971,6 @@ void helper_store_fpscr (uint64_t arg, uint32_t
> mask)
>
> void helper_float_check_status (void)
> {
> -#ifdef CONFIG_SOFTFLOAT
> if (env->exception_index == POWERPC_EXCP_PROGRAM &&
> (env->error_code & POWERPC_EXCP_FP)) {
> /* Differred floating-point exception after target FPR
> update */
> @@ -989,22 +988,12 @@ void helper_float_check_status (void)
> float_inexact_excp();
> }
> }
> -#else
> - if (env->exception_index == POWERPC_EXCP_PROGRAM &&
> - (env->error_code & POWERPC_EXCP_FP)) {
> - /* Differred floating-point exception after target FPR
> update */
> - if (msr_fe0 != 0 || msr_fe1 != 0)
> - helper_raise_exception_err(env->exception_index, env-
> >error_code);
> - }
> -#endif
> }
>
> -#ifdef CONFIG_SOFTFLOAT
> void helper_reset_fpstatus (void)
> {
> set_float_exception_flags(0, &env->fp_status);
> }
> -#endif
>
> /* fadd - fadd. */
> uint64_t helper_fadd (uint64_t arg1, uint64_t arg2)
> diff --git a/target-ppc/translate.c b/target-ppc/translate.c
> index 9b3f90c..59aef85 100644
> --- a/target-ppc/translate.c
> +++ b/target-ppc/translate.c
> @@ -215,9 +215,7 @@ struct opc_handler_t {
>
> static inline void gen_reset_fpstatus(void)
> {
> -#ifdef CONFIG_SOFTFLOAT
> gen_helper_reset_fpstatus();
> -#endif
> }
>
> static inline void gen_compute_fprf(TCGv_i64 arg, int set_fprf, int
> set_rc)
> --
> 1.7.5.1
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef Aurelien Jarno
2011-05-23 22:09 ` Peter Maydell
@ 2011-05-29 11:01 ` Andreas Färber
1 sibling, 0 replies; 18+ messages in thread
From: Andreas Färber @ 2011-05-29 11:01 UTC (permalink / raw)
To: Aurelien Jarno; +Cc: qemu-devel
Am 23.05.2011 um 23:42 schrieb Aurelien Jarno:
> target-mips has been switched to softfloat only long ago, but
> a #ifndef CONFIG_SOFTFLOAT has been forgotten. Remove it.
>
> Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Andreas Färber <andreas.faerber@web.de>
> ---
> gdbstub.c | 4 ----
> 1 files changed, 0 insertions(+), 4 deletions(-)
>
> diff --git a/gdbstub.c b/gdbstub.c
> index ae856f9..b9ae30d 100644
> --- a/gdbstub.c
> +++ b/gdbstub.c
> @@ -1105,10 +1105,6 @@ static int cpu_gdb_write_register(CPUState
> *env, uint8_t *mem_buf, int n)
> env->active_fpu.fcr31 = tmp & 0xFF83FFFF;
> /* set rounding mode */
> RESTORE_ROUNDING_MODE;
> -#ifndef CONFIG_SOFTFLOAT
> - /* no floating point exception for native float */
> - SET_FP_ENABLE(env->active_fpu.fcr31, 0);
> -#endif
> break;
> case 71: env->active_fpu.fcr0 = tmp; break;
> }
> --
> 1.7.5.1
^ permalink raw reply [flat|nested] 18+ messages in thread
end of thread, other threads:[~2011-05-29 11:01 UTC | newest]
Thread overview: 18+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-05-23 21:42 [Qemu-devel] [PATCH v2 0/9] softfloat-native removal and i386 improvement Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 1/9] target-ppc: remove old CONFIG_SOFTFLOAT #ifdef Aurelien Jarno
2011-05-29 10:58 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 2/9] target-mips/gdbstub: remove old CONFIG_SOFTFLOAT #ifndef Aurelien Jarno
2011-05-23 22:09 ` Peter Maydell
2011-05-29 11:01 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 3/9] target-i386: remove old code handling float64 Aurelien Jarno
2011-05-24 15:12 ` Peter Maydell
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 4/9] softfloat-native: remove Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 5/9] softfloat: always enable floatx80 and float128 support Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 6/9] target-i386: use floatx80 constants in helper_fld*_ST0() Aurelien Jarno
2011-05-23 22:11 ` Peter Maydell
2011-05-29 10:38 ` Andreas Färber
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 7/9] softfloat: add float*_is_zero_or_denormal() Aurelien Jarno
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 8/9] target-i386: cleanup helper_fxam_ST0() Aurelien Jarno
2011-05-23 22:02 ` Peter Maydell
2011-05-23 21:42 ` [Qemu-devel] [PATCH v2 9/9] target-i386: add support for FPU exceptions Aurelien Jarno
2011-05-24 15:55 ` Peter Maydell
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.