diff options
Diffstat (limited to 'fpu/softfloat.c')
| -rw-r--r-- | fpu/softfloat.c | 107 |
1 files changed, 55 insertions, 52 deletions
diff --git a/fpu/softfloat.c b/fpu/softfloat.c index 59ca356d0e..e1eef954e6 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -726,8 +726,7 @@ static FloatParts addsub_floats(FloatParts a, FloatParts b, bool subtract, * IEC/IEEE Standard for Binary Floating-Point Arithmetic. */ -float16 __attribute__((flatten)) float16_add(float16 a, float16 b, - float_status *status) +float16 QEMU_FLATTEN float16_add(float16 a, float16 b, float_status *status) { FloatParts pa = float16_unpack_canonical(a, status); FloatParts pb = float16_unpack_canonical(b, status); @@ -736,8 +735,7 @@ float16 __attribute__((flatten)) float16_add(float16 a, float16 b, return float16_round_pack_canonical(pr, status); } -float32 __attribute__((flatten)) float32_add(float32 a, float32 b, - float_status *status) +float32 QEMU_FLATTEN float32_add(float32 a, float32 b, float_status *status) { FloatParts pa = float32_unpack_canonical(a, status); FloatParts pb = float32_unpack_canonical(b, status); @@ -746,8 +744,7 @@ float32 __attribute__((flatten)) float32_add(float32 a, float32 b, return float32_round_pack_canonical(pr, status); } -float64 __attribute__((flatten)) float64_add(float64 a, float64 b, - float_status *status) +float64 QEMU_FLATTEN float64_add(float64 a, float64 b, float_status *status) { FloatParts pa = float64_unpack_canonical(a, status); FloatParts pb = float64_unpack_canonical(b, status); @@ -756,8 +753,7 @@ float64 __attribute__((flatten)) float64_add(float64 a, float64 b, return float64_round_pack_canonical(pr, status); } -float16 __attribute__((flatten)) float16_sub(float16 a, float16 b, - float_status *status) +float16 QEMU_FLATTEN float16_sub(float16 a, float16 b, float_status *status) { FloatParts pa = float16_unpack_canonical(a, status); FloatParts pb = float16_unpack_canonical(b, status); @@ -766,8 +762,7 @@ float16 __attribute__((flatten)) float16_sub(float16 a, float16 b, return float16_round_pack_canonical(pr, status); } -float32 __attribute__((flatten)) float32_sub(float32 a, float32 b, - float_status *status) +float32 QEMU_FLATTEN float32_sub(float32 a, float32 b, float_status *status) { FloatParts pa = float32_unpack_canonical(a, status); FloatParts pb = float32_unpack_canonical(b, status); @@ -776,8 +771,7 @@ float32 __attribute__((flatten)) float32_sub(float32 a, float32 b, return float32_round_pack_canonical(pr, status); } -float64 __attribute__((flatten)) float64_sub(float64 a, float64 b, - float_status *status) +float64 QEMU_FLATTEN float64_sub(float64 a, float64 b, float_status *status) { FloatParts pa = float64_unpack_canonical(a, status); FloatParts pb = float64_unpack_canonical(b, status); @@ -835,8 +829,7 @@ static FloatParts mul_floats(FloatParts a, FloatParts b, float_status *s) g_assert_not_reached(); } -float16 __attribute__((flatten)) float16_mul(float16 a, float16 b, - float_status *status) +float16 QEMU_FLATTEN float16_mul(float16 a, float16 b, float_status *status) { FloatParts pa = float16_unpack_canonical(a, status); FloatParts pb = float16_unpack_canonical(b, status); @@ -845,8 +838,7 @@ float16 __attribute__((flatten)) float16_mul(float16 a, float16 b, return float16_round_pack_canonical(pr, status); } -float32 __attribute__((flatten)) float32_mul(float32 a, float32 b, - float_status *status) +float32 QEMU_FLATTEN float32_mul(float32 a, float32 b, float_status *status) { FloatParts pa = float32_unpack_canonical(a, status); FloatParts pb = float32_unpack_canonical(b, status); @@ -855,8 +847,7 @@ float32 __attribute__((flatten)) float32_mul(float32 a, float32 b, return float32_round_pack_canonical(pr, status); } -float64 __attribute__((flatten)) float64_mul(float64 a, float64 b, - float_status *status) +float64 QEMU_FLATTEN float64_mul(float64 a, float64 b, float_status *status) { FloatParts pa = float64_unpack_canonical(a, status); FloatParts pb = float64_unpack_canonical(b, status); @@ -1068,7 +1059,7 @@ static FloatParts muladd_floats(FloatParts a, FloatParts b, FloatParts c, return a; } -float16 __attribute__((flatten)) float16_muladd(float16 a, float16 b, float16 c, +float16 QEMU_FLATTEN float16_muladd(float16 a, float16 b, float16 c, int flags, float_status *status) { FloatParts pa = float16_unpack_canonical(a, status); @@ -1079,7 +1070,7 @@ float16 __attribute__((flatten)) float16_muladd(float16 a, float16 b, float16 c, return float16_round_pack_canonical(pr, status); } -float32 __attribute__((flatten)) float32_muladd(float32 a, float32 b, float32 c, +float32 QEMU_FLATTEN float32_muladd(float32 a, float32 b, float32 c, int flags, float_status *status) { FloatParts pa = float32_unpack_canonical(a, status); @@ -1090,7 +1081,7 @@ float32 __attribute__((flatten)) float32_muladd(float32 a, float32 b, float32 c, return float32_round_pack_canonical(pr, status); } -float64 __attribute__((flatten)) float64_muladd(float64 a, float64 b, float64 c, +float64 QEMU_FLATTEN float64_muladd(float64 a, float64 b, float64 c, int flags, float_status *status) { FloatParts pa = float64_unpack_canonical(a, status); @@ -1112,19 +1103,38 @@ static FloatParts div_floats(FloatParts a, FloatParts b, float_status *s) bool sign = a.sign ^ b.sign; if (a.cls == float_class_normal && b.cls == float_class_normal) { - uint64_t temp_lo, temp_hi; + uint64_t n0, n1, q, r; int exp = a.exp - b.exp; + + /* + * We want a 2*N / N-bit division to produce exactly an N-bit + * result, so that we do not lose any precision and so that we + * do not have to renormalize afterward. If A.frac < B.frac, + * then division would produce an (N-1)-bit result; shift A left + * by one to produce the an N-bit result, and decrement the + * exponent to match. + * + * The udiv_qrnnd algorithm that we're using requires normalization, + * i.e. the msb of the denominator must be set. Since we know that + * DECOMPOSED_BINARY_POINT is msb-1, the inputs must be shifted left + * by one (more), and the remainder must be shifted right by one. + */ if (a.frac < b.frac) { exp -= 1; - shortShift128Left(0, a.frac, DECOMPOSED_BINARY_POINT + 1, - &temp_hi, &temp_lo); + shift128Left(0, a.frac, DECOMPOSED_BINARY_POINT + 2, &n1, &n0); } else { - shortShift128Left(0, a.frac, DECOMPOSED_BINARY_POINT, - &temp_hi, &temp_lo); + shift128Left(0, a.frac, DECOMPOSED_BINARY_POINT + 1, &n1, &n0); } - /* LSB of quot is set if inexact which roundandpack will use - * to set flags. Yet again we re-use a for the result */ - a.frac = div128To64(temp_lo, temp_hi, b.frac); + q = udiv_qrnnd(&r, n1, n0, b.frac << 1); + + /* + * Set lsb if there is a remainder, to set inexact. + * As mentioned above, to find the actual value of the remainder we + * would need to shift right, but (1) we are only concerned about + * non-zero-ness, and (2) the remainder will always be even because + * both inputs to the division primitive are even. + */ + a.frac = q | (r != 0); a.sign = sign; a.exp = exp; return a; @@ -1409,13 +1419,6 @@ float64 float64_round_to_int(float64 a, float_status *s) return float64_round_pack_canonical(pr, s); } -float64 float64_trunc_to_int(float64 a, float_status *s) -{ - FloatParts pa = float64_unpack_canonical(a, s); - FloatParts pr = round_to_int(pa, float_round_to_zero, 0, s); - return float64_round_pack_canonical(pr, s); -} - /* * Returns the result of converting the floating-point value `a' to * the two's complement integer format. The conversion is performed @@ -2402,21 +2405,21 @@ static FloatParts sqrt_float(FloatParts a, float_status *s, const FloatFmt *p) return a; } -float16 __attribute__((flatten)) float16_sqrt(float16 a, float_status *status) +float16 QEMU_FLATTEN float16_sqrt(float16 a, float_status *status) { FloatParts pa = float16_unpack_canonical(a, status); FloatParts pr = sqrt_float(pa, status, &float16_params); return float16_round_pack_canonical(pr, status); } -float32 __attribute__((flatten)) float32_sqrt(float32 a, float_status *status) +float32 QEMU_FLATTEN float32_sqrt(float32 a, float_status *status) { FloatParts pa = float32_unpack_canonical(a, status); FloatParts pr = sqrt_float(pa, status, &float32_params); return float32_round_pack_canonical(pr, status); } -float64 __attribute__((flatten)) float64_sqrt(float64 a, float_status *status) +float64 QEMU_FLATTEN float64_sqrt(float64 a, float_status *status) { FloatParts pa = float64_unpack_canonical(a, status); FloatParts pr = sqrt_float(pa, status, &float64_params); @@ -2690,7 +2693,7 @@ static void { int8_t shiftCount; - shiftCount = countLeadingZeros32( aSig ) - 8; + shiftCount = clz32(aSig) - 8; *zSigPtr = aSig<<shiftCount; *zExpPtr = 1 - shiftCount; @@ -2798,7 +2801,7 @@ static float32 { int8_t shiftCount; - shiftCount = countLeadingZeros32( zSig ) - 1; + shiftCount = clz32(zSig) - 1; return roundAndPackFloat32(zSign, zExp - shiftCount, zSig<<shiftCount, status); @@ -2831,7 +2834,7 @@ static void { int8_t shiftCount; - shiftCount = countLeadingZeros64( aSig ) - 11; + shiftCount = clz64(aSig) - 11; *zSigPtr = aSig<<shiftCount; *zExpPtr = 1 - shiftCount; @@ -2969,7 +2972,7 @@ static float64 { int8_t shiftCount; - shiftCount = countLeadingZeros64( zSig ) - 1; + shiftCount = clz64(zSig) - 1; return roundAndPackFloat64(zSign, zExp - shiftCount, zSig<<shiftCount, status); @@ -2987,7 +2990,7 @@ void normalizeFloatx80Subnormal(uint64_t aSig, int32_t *zExpPtr, { int8_t shiftCount; - shiftCount = countLeadingZeros64( aSig ); + shiftCount = clz64(aSig); *zSigPtr = aSig<<shiftCount; *zExpPtr = 1 - shiftCount; } @@ -3226,7 +3229,7 @@ floatx80 normalizeRoundAndPackFloatx80(int8_t roundingPrecision, zSig1 = 0; zExp -= 64; } - shiftCount = countLeadingZeros64( zSig0 ); + shiftCount = clz64(zSig0); shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); zExp -= shiftCount; return roundAndPackFloatx80(roundingPrecision, zSign, zExp, @@ -3303,7 +3306,7 @@ static void int8_t shiftCount; if ( aSig0 == 0 ) { - shiftCount = countLeadingZeros64( aSig1 ) - 15; + shiftCount = clz64(aSig1) - 15; if ( shiftCount < 0 ) { *zSig0Ptr = aSig1>>( - shiftCount ); *zSig1Ptr = aSig1<<( shiftCount & 63 ); @@ -3315,7 +3318,7 @@ static void *zExpPtr = - shiftCount - 63; } else { - shiftCount = countLeadingZeros64( aSig0 ) - 15; + shiftCount = clz64(aSig0) - 15; shortShift128Left( aSig0, aSig1, shiftCount, zSig0Ptr, zSig1Ptr ); *zExpPtr = 1 - shiftCount; } @@ -3504,7 +3507,7 @@ static float128 normalizeRoundAndPackFloat128(flag zSign, int32_t zExp, zSig1 = 0; zExp -= 64; } - shiftCount = countLeadingZeros64( zSig0 ) - 15; + shiftCount = clz64(zSig0) - 15; if ( 0 <= shiftCount ) { zSig2 = 0; shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); @@ -3536,7 +3539,7 @@ floatx80 int32_to_floatx80(int32_t a, float_status *status) if ( a == 0 ) return packFloatx80( 0, 0, 0 ); zSign = ( a < 0 ); absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 32; + shiftCount = clz32(absA) + 32; zSig = absA; return packFloatx80( zSign, 0x403E - shiftCount, zSig<<shiftCount ); @@ -3558,7 +3561,7 @@ float128 int32_to_float128(int32_t a, float_status *status) if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); zSign = ( a < 0 ); absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 17; + shiftCount = clz32(absA) + 17; zSig0 = absA; return packFloat128( zSign, 0x402E - shiftCount, zSig0<<shiftCount, 0 ); @@ -3580,7 +3583,7 @@ floatx80 int64_to_floatx80(int64_t a, float_status *status) if ( a == 0 ) return packFloatx80( 0, 0, 0 ); zSign = ( a < 0 ); absA = zSign ? - a : a; - shiftCount = countLeadingZeros64( absA ); + shiftCount = clz64(absA); return packFloatx80( zSign, 0x403E - shiftCount, absA<<shiftCount ); } @@ -3602,7 +3605,7 @@ float128 int64_to_float128(int64_t a, float_status *status) if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); zSign = ( a < 0 ); absA = zSign ? - a : a; - shiftCount = countLeadingZeros64( absA ) + 49; + shiftCount = clz64(absA) + 49; zExp = 0x406E - shiftCount; if ( 64 <= shiftCount ) { zSig1 = 0; |