diff options
Diffstat (limited to 'arch/arm26/nwfpe/softfloat.c')
| -rw-r--r-- | arch/arm26/nwfpe/softfloat.c | 3439 |
1 files changed, 0 insertions, 3439 deletions
diff --git a/arch/arm26/nwfpe/softfloat.c b/arch/arm26/nwfpe/softfloat.c deleted file mode 100644 index 26c1b916e52..00000000000 --- a/arch/arm26/nwfpe/softfloat.c +++ /dev/null @@ -1,3439 +0,0 @@ -/* -=============================================================================== - -This C source file is part of the SoftFloat IEC/IEEE Floating-point -Arithmetic Package, Release 2. - -Written by John R. Hauser. This work was made possible in part by the -International Computer Science Institute, located at Suite 600, 1947 Center -Street, Berkeley, California 94704. Funding was partially provided by the -National Science Foundation under grant MIP-9311980. The original version -of this code was written as part of a project to build a fixed-point vector -processor in collaboration with the University of California at Berkeley, -overseen by Profs. Nelson Morgan and John Wawrzynek. More information -is available through the web page `http://HTTP.CS.Berkeley.EDU/~jhauser/ -arithmetic/softfloat.html'. - -THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE. Although reasonable effort -has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT -TIMES RESULT IN INCORRECT BEHAVIOR. USE OF THIS SOFTWARE IS RESTRICTED TO -PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY -AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE. - -Derivative works are acceptable, even for commercial purposes, so long as -(1) they include prominent notice that the work is derivative, and (2) they -include prominent notice akin to these three paragraphs for those parts of -this code that are retained. - -=============================================================================== -*/ - -#include "fpa11.h" -#include "milieu.h" -#include "softfloat.h" - -/* -------------------------------------------------------------------------------- -Floating-point rounding mode, extended double-precision rounding precision, -and exception flags. -------------------------------------------------------------------------------- -*/ -int8 float_rounding_mode = float_round_nearest_even; -int8 floatx80_rounding_precision = 80; -int8 float_exception_flags; - -/* -------------------------------------------------------------------------------- -Primitive arithmetic functions, including multi-word arithmetic, and -division and square root approximations. (Can be specialized to target if -desired.) -------------------------------------------------------------------------------- -*/ -#include "softfloat-macros" - -/* -------------------------------------------------------------------------------- -Functions and definitions to determine: (1) whether tininess for underflow -is detected before or after rounding by default, (2) what (if anything) -happens when exceptions are raised, (3) how signaling NaNs are distinguished -from quiet NaNs, (4) the default generated quiet NaNs, and (5) how NaNs -are propagated from function inputs to output. These details are target- -specific. -------------------------------------------------------------------------------- -*/ -#include "softfloat-specialize" - -/* -------------------------------------------------------------------------------- -Takes a 64-bit fixed-point value `absZ' with binary point between bits 6 -and 7, and returns the properly rounded 32-bit integer corresponding to the -input. If `zSign' is nonzero, the input is negated before being converted -to an integer. Bit 63 of `absZ' must be zero. Ordinarily, the fixed-point -input is simply rounded to an integer, with the inexact exception raised if -the input cannot be represented exactly as an integer. If the fixed-point -input is too large, however, the invalid exception is raised and the largest -positive or negative integer is returned. -------------------------------------------------------------------------------- -*/ -static int32 roundAndPackInt32( flag zSign, bits64 absZ ) -{ - int8 roundingMode; - flag roundNearestEven; - int8 roundIncrement, roundBits; - int32 z; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x40; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x7F; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = absZ & 0x7F; - absZ = ( absZ + roundIncrement )>>7; - absZ &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); - z = absZ; - if ( zSign ) z = - z; - if ( ( absZ>>32 ) || ( z && ( ( z < 0 ) ^ zSign ) ) ) { - float_exception_flags |= float_flag_invalid; - return zSign ? 0x80000000 : 0x7FFFFFFF; - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - return z; - -} - -/* -------------------------------------------------------------------------------- -Returns the fraction bits of the single-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE bits32 extractFloat32Frac( float32 a ) -{ - - return a & 0x007FFFFF; - -} - -/* -------------------------------------------------------------------------------- -Returns the exponent bits of the single-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE int16 extractFloat32Exp( float32 a ) -{ - - return ( a>>23 ) & 0xFF; - -} - -/* -------------------------------------------------------------------------------- -Returns the sign bit of the single-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE flag extractFloat32Sign( float32 a ) -{ - - return a>>31; - -} - -/* -------------------------------------------------------------------------------- -Normalizes the subnormal single-precision floating-point value represented -by the denormalized significand `aSig'. The normalized exponent and -significand are stored at the locations pointed to by `zExpPtr' and -`zSigPtr', respectively. -------------------------------------------------------------------------------- -*/ -static void - normalizeFloat32Subnormal( bits32 aSig, int16 *zExpPtr, bits32 *zSigPtr ) -{ - int8 shiftCount; - - shiftCount = countLeadingZeros32( aSig ) - 8; - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; - -} - -/* -------------------------------------------------------------------------------- -Packs the sign `zSign', exponent `zExp', and significand `zSig' into a -single-precision floating-point value, returning the result. After being -shifted into the proper positions, the three fields are simply added -together to form the result. This means that any integer portion of `zSig' -will be added into the exponent. Since a properly normalized significand -will have an integer portion equal to 1, the `zExp' input should be 1 less -than the desired result exponent whenever `zSig' is a complete, normalized -significand. -------------------------------------------------------------------------------- -*/ -INLINE float32 packFloat32( flag zSign, int16 zExp, bits32 zSig ) -{ -#if 0 - float32 f; - __asm__("@ packFloat32; \n\ - mov %0, %1, asl #31; \n\ - orr %0, %2, asl #23; \n\ - orr %0, %3" - : /* no outputs */ - : "g" (f), "g" (zSign), "g" (zExp), "g" (zSig) - : "cc"); - return f; -#else - return ( ( (bits32) zSign )<<31 ) + ( ( (bits32) zExp )<<23 ) + zSig; -#endif -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent `zExp', -and significand `zSig', and returns the proper single-precision floating- -point value corresponding to the abstract input. Ordinarily, the abstract -value is simply rounded and packed into the single-precision format, with -the inexact exception raised if the abstract input cannot be represented -exactly. If the abstract value is too large, however, the overflow and -inexact exceptions are raised and an infinity or maximal finite value is -returned. If the abstract value is too small, the input value is rounded to -a subnormal number, and the underflow and inexact exceptions are raised if -the abstract input cannot be represented exactly as a subnormal single- -precision floating-point number. - The input significand `zSig' has its binary point between bits 30 -and 29, which is 7 bits to the left of the usual location. This shifted -significand must be normalized or smaller. If `zSig' is not normalized, -`zExp' must be 0; in that case, the result returned is a subnormal number, -and it must not require rounding. In the usual case that `zSig' is -normalized, `zExp' must be 1 less than the ``true'' floating-point exponent. -The handling of underflow and overflow follows the IEC/IEEE Standard for -Binary Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -static float32 roundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) -{ - int8 roundingMode; - flag roundNearestEven; - int8 roundIncrement, roundBits; - flag isTiny; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x40; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x7F; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig & 0x7F; - if ( 0xFD <= (bits16) zExp ) { - if ( ( 0xFD < zExp ) - || ( ( zExp == 0xFD ) - && ( (sbits32) ( zSig + roundIncrement ) < 0 ) ) - ) { - float_raise( float_flag_overflow | float_flag_inexact ); - return packFloat32( zSign, 0xFF, 0 ) - ( roundIncrement == 0 ); - } - if ( zExp < 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < -1 ) - || ( zSig + roundIncrement < 0x80000000 ); - shift32RightJamming( zSig, - zExp, &zSig ); - zExp = 0; - roundBits = zSig & 0x7F; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig = ( zSig + roundIncrement )>>7; - zSig &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); - if ( zSig == 0 ) zExp = 0; - return packFloat32( zSign, zExp, zSig ); - -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent `zExp', -and significand `zSig', and returns the proper single-precision floating- -point value corresponding to the abstract input. This routine is just like -`roundAndPackFloat32' except that `zSig' does not have to be normalized in -any way. In all cases, `zExp' must be 1 less than the ``true'' floating- -point exponent. -------------------------------------------------------------------------------- -*/ -static float32 - normalizeRoundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) -{ - int8 shiftCount; - - shiftCount = countLeadingZeros32( zSig ) - 1; - return roundAndPackFloat32( zSign, zExp - shiftCount, zSig<<shiftCount ); - -} - -/* -------------------------------------------------------------------------------- -Returns the fraction bits of the double-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE bits64 extractFloat64Frac( float64 a ) -{ - - return a & LIT64( 0x000FFFFFFFFFFFFF ); - -} - -/* -------------------------------------------------------------------------------- -Returns the exponent bits of the double-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE int16 extractFloat64Exp( float64 a ) -{ - - return ( a>>52 ) & 0x7FF; - -} - -/* -------------------------------------------------------------------------------- -Returns the sign bit of the double-precision floating-point value `a'. -------------------------------------------------------------------------------- -*/ -INLINE flag extractFloat64Sign( float64 a ) -{ - - return a>>63; - -} - -/* -------------------------------------------------------------------------------- -Normalizes the subnormal double-precision floating-point value represented -by the denormalized significand `aSig'. The normalized exponent and -significand are stored at the locations pointed to by `zExpPtr' and -`zSigPtr', respectively. -------------------------------------------------------------------------------- -*/ -static void - normalizeFloat64Subnormal( bits64 aSig, int16 *zExpPtr, bits64 *zSigPtr ) -{ - int8 shiftCount; - - shiftCount = countLeadingZeros64( aSig ) - 11; - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; - -} - -/* -------------------------------------------------------------------------------- -Packs the sign `zSign', exponent `zExp', and significand `zSig' into a -double-precision floating-point value, returning the result. After being -shifted into the proper positions, the three fields are simply added -together to form the result. This means that any integer portion of `zSig' -will be added into the exponent. Since a properly normalized significand -will have an integer portion equal to 1, the `zExp' input should be 1 less -than the desired result exponent whenever `zSig' is a complete, normalized -significand. -------------------------------------------------------------------------------- -*/ -INLINE float64 packFloat64( flag zSign, int16 zExp, bits64 zSig ) -{ - - return ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<52 ) + zSig; - -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent `zExp', -and significand `zSig', and returns the proper double-precision floating- -point value corresponding to the abstract input. Ordinarily, the abstract -value is simply rounded and packed into the double-precision format, with -the inexact exception raised if the abstract input cannot be represented -exactly. If the abstract value is too large, however, the overflow and -inexact exceptions are raised and an infinity or maximal finite value is -returned. If the abstract value is too small, the input value is rounded to -a subnormal number, and the underflow and inexact exceptions are raised if -the abstract input cannot be represented exactly as a subnormal double- -precision floating-point number. - The input significand `zSig' has its binary point between bits 62 -and 61, which is 10 bits to the left of the usual location. This shifted -significand must be normalized or smaller. If `zSig' is not normalized, -`zExp' must be 0; in that case, the result returned is a subnormal number, -and it must not require rounding. In the usual case that `zSig' is -normalized, `zExp' must be 1 less than the ``true'' floating-point exponent. -The handling of underflow and overflow follows the IEC/IEEE Standard for -Binary Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -static float64 roundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) -{ - int8 roundingMode; - flag roundNearestEven; - int16 roundIncrement, roundBits; - flag isTiny; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x200; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x3FF; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig & 0x3FF; - if ( 0x7FD <= (bits16) zExp ) { - if ( ( 0x7FD < zExp ) - || ( ( zExp == 0x7FD ) - && ( (sbits64) ( zSig + roundIncrement ) < 0 ) ) - ) { - //register int lr = __builtin_return_address(0); - //printk("roundAndPackFloat64 called from 0x%08x\n",lr); - float_raise( float_flag_overflow | float_flag_inexact ); - return packFloat64( zSign, 0x7FF, 0 ) - ( roundIncrement == 0 ); - } - if ( zExp < 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < -1 ) - || ( zSig + roundIncrement < LIT64( 0x8000000000000000 ) ); - shift64RightJamming( zSig, - zExp, &zSig ); - zExp = 0; - roundBits = zSig & 0x3FF; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig = ( zSig + roundIncrement )>>10; - zSig &= ~ ( ( ( roundBits ^ 0x200 ) == 0 ) & roundNearestEven ); - if ( zSig == 0 ) zExp = 0; - return packFloat64( zSign, zExp, zSig ); - -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent `zExp', -and significand `zSig', and returns the proper double-precision floating- -point value corresponding to the abstract input. This routine is just like -`roundAndPackFloat64' except that `zSig' does not have to be normalized in -any way. In all cases, `zExp' must be 1 less than the ``true'' floating- -point exponent. -------------------------------------------------------------------------------- -*/ -static float64 - normalizeRoundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) -{ - int8 shiftCount; - - shiftCount = countLeadingZeros64( zSig ) - 1; - return roundAndPackFloat64( zSign, zExp - shiftCount, zSig<<shiftCount ); - -} - -#ifdef FLOATX80 - -/* -------------------------------------------------------------------------------- -Returns the fraction bits of the extended double-precision floating-point -value `a'. -------------------------------------------------------------------------------- -*/ -INLINE bits64 extractFloatx80Frac( floatx80 a ) -{ - - return a.low; - -} - -/* -------------------------------------------------------------------------------- -Returns the exponent bits of the extended double-precision floating-point -value `a'. -------------------------------------------------------------------------------- -*/ -INLINE int32 extractFloatx80Exp( floatx80 a ) -{ - - return a.high & 0x7FFF; - -} - -/* -------------------------------------------------------------------------------- -Returns the sign bit of the extended double-precision floating-point value -`a'. -------------------------------------------------------------------------------- -*/ -INLINE flag extractFloatx80Sign( floatx80 a ) -{ - - return a.high>>15; - -} - -/* -------------------------------------------------------------------------------- -Normalizes the subnormal extended double-precision floating-point value -represented by the denormalized significand `aSig'. The normalized exponent -and significand are stored at the locations pointed to by `zExpPtr' and -`zSigPtr', respectively. -------------------------------------------------------------------------------- -*/ -static void - normalizeFloatx80Subnormal( bits64 aSig, int32 *zExpPtr, bits64 *zSigPtr ) -{ - int8 shiftCount; - - shiftCount = countLeadingZeros64( aSig ); - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; - -} - -/* -------------------------------------------------------------------------------- -Packs the sign `zSign', exponent `zExp', and significand `zSig' into an -extended double-precision floating-point value, returning the result. -------------------------------------------------------------------------------- -*/ -INLINE floatx80 packFloatx80( flag zSign, int32 zExp, bits64 zSig ) -{ - floatx80 z; - - z.low = zSig; - z.high = ( ( (bits16) zSign )<<15 ) + zExp; - return z; - -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent `zExp', -and extended significand formed by the concatenation of `zSig0' and `zSig1', -and returns the proper extended double-precision floating-point value -corresponding to the abstract input. Ordinarily, the abstract value is -rounded and packed into the extended double-precision format, with the -inexact exception raised if the abstract input cannot be represented -exactly. If the abstract value is too large, however, the overflow and -inexact exceptions are raised and an infinity or maximal finite value is -returned. If the abstract value is too small, the input value is rounded to -a subnormal number, and the underflow and inexact exceptions are raised if -the abstract input cannot be represented exactly as a subnormal extended -double-precision floating-point number. - If `roundingPrecision' is 32 or 64, the result is rounded to the same -number of bits as single or double precision, respectively. Otherwise, the -result is rounded to the full precision of the extended double-precision -format. - The input significand must be normalized or smaller. If the input -significand is not normalized, `zExp' must be 0; in that case, the result -returned is a subnormal number, and it must not require rounding. The -handling of underflow and overflow follows the IEC/IEEE Standard for Binary -Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -static floatx80 - roundAndPackFloatx80( - int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 - ) -{ - int8 roundingMode; - flag roundNearestEven, increment, isTiny; - int64 roundIncrement, roundMask, roundBits; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - if ( roundingPrecision == 80 ) goto precision80; - if ( roundingPrecision == 64 ) { - roundIncrement = LIT64( 0x0000000000000400 ); - roundMask = LIT64( 0x00000000000007FF ); - } - else if ( roundingPrecision == 32 ) { - roundIncrement = LIT64( 0x0000008000000000 ); - roundMask = LIT64( 0x000000FFFFFFFFFF ); - } - else { - goto precision80; - } - zSig0 |= ( zSig1 != 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = roundMask; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig0 & roundMask; - if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { - if ( ( 0x7FFE < zExp ) - || ( ( zExp == 0x7FFE ) && ( zSig0 + roundIncrement < zSig0 ) ) - ) { - goto overflow; - } - if ( zExp <= 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < 0 ) - || ( zSig0 <= zSig0 + roundIncrement ); - shift64RightJamming( zSig0, 1 - zExp, &zSig0 ); - zExp = 0; - roundBits = zSig0 & roundMask; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig0 += roundIncrement; - if ( (sbits64) zSig0 < 0 ) zExp = 1; - roundIncrement = roundMask + 1; - if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { - roundMask |= roundIncrement; - } - zSig0 &= ~ roundMask; - return packFloatx80( zSign, zExp, zSig0 ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig0 += roundIncrement; - if ( zSig0 < roundIncrement ) { - ++zExp; - zSig0 = LIT64( 0x8000000000000000 ); - } - roundIncrement = roundMask + 1; - if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { - roundMask |= roundIncrement; - } - zSig0 &= ~ roundMask; - if ( zSig0 == 0 ) zExp = 0; - return packFloatx80( zSign, zExp, zSig0 ); - precision80: - increment = ( (sbits64) zSig1 < 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - increment = 0; - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig1; - } - else { - increment = ( roundingMode == float_round_up ) && zSig1; - } - } - } - if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { - if ( ( 0x7FFE < zExp ) - || ( ( zExp == 0x7FFE ) - && ( zSig0 == LIT64( 0xFFFFFFFFFFFFFFFF ) ) - && increment - ) - ) { - roundMask = 0; - overflow: - float_raise( float_flag_overflow | float_flag_inexact ); - if ( ( roundingMode == float_round_to_zero ) - || ( zSign && ( roundingMode == float_round_up ) ) - || ( ! zSign && ( roundingMode == float_round_down ) ) - ) { - return packFloatx80( zSign, 0x7FFE, ~ roundMask ); - } - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( zExp <= 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < 0 ) - || ! increment - || ( zSig0 < LIT64( 0xFFFFFFFFFFFFFFFF ) ); - shift64ExtraRightJamming( zSig0, zSig1, 1 - zExp, &zSig0, &zSig1 ); - zExp = 0; - if ( isTiny && zSig1 ) float_raise( float_flag_underflow ); - if ( zSig1 ) float_exception_flags |= float_flag_inexact; - if ( roundNearestEven ) { - increment = ( (sbits64) zSig1 < 0 ); - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig1; - } - else { - increment = ( roundingMode == float_round_up ) && zSig1; - } - } - if ( increment ) { - ++zSig0; - zSig0 &= ~ ( ( zSig1 + zSig1 == 0 ) & roundNearestEven ); - if ( (sbits64) zSig0 < 0 ) zExp = 1; - } - return packFloatx80( zSign, zExp, zSig0 ); - } - } - if ( zSig1 ) float_exception_flags |= float_flag_inexact; - if ( increment ) { - ++zSig0; - if ( zSig0 == 0 ) { - ++zExp; - zSig0 = LIT64( 0x8000000000000000 ); - } - else { - zSig0 &= ~ ( ( zSig1 + zSig1 == 0 ) & roundNearestEven ); - } - } - else { - if ( zSig0 == 0 ) zExp = 0; - } - - return packFloatx80( zSign, zExp, zSig0 ); -} - -/* -------------------------------------------------------------------------------- -Takes an abstract floating-point value having sign `zSign', exponent -`zExp', and significand formed by the concatenation of `zSig0' and `zSig1', -and returns the proper extended double-precision floating-point value -corresponding to the abstract input. This routine is just like -`roundAndPackFloatx80' except that the input significand does not have to be -normalized. -------------------------------------------------------------------------------- -*/ -static floatx80 - normalizeRoundAndPackFloatx80( - int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 - ) -{ - int8 shiftCount; - - if ( zSig0 == 0 ) { - zSig0 = zSig1; - zSig1 = 0; - zExp -= 64; - } - shiftCount = countLeadingZeros64( zSig0 ); - shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); - zExp -= shiftCount; - return - roundAndPackFloatx80( roundingPrecision, zSign, zExp, zSig0, zSig1 ); - -} - -#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 -according to the IEC/IEEE Standard for Binary Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -float32 int32_to_float32( int32 a ) -{ - flag zSign; - - if ( a == 0 ) return 0; - if ( a == 0x80000000 ) return packFloat32( 1, 0x9E, 0 ); - zSign = ( a < 0 ); - return normalizeRoundAndPackFloat32( zSign, 0x9C, zSign ? - a : a ); - -} - -/* -------------------------------------------------------------------------------- -Returns the result of converting the 32-bit two's complement integer `a' to -the double-precision floating-point format. The conversion is performed -according to the IEC/IEEE Standard for Binary Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -float64 int32_to_float64( int32 a ) -{ - flag aSign; - uint32 absA; - int8 shiftCount; - bits64 zSig; - - if ( a == 0 ) return 0; - aSign = ( a < 0 ); - absA = aSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 21; - zSig = absA; - return packFloat64( aSign, 0x432 - shiftCount, zSig<<shiftCount ); - -} - -#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 -is performed according to the IEC/IEEE Standard for Binary Floating-point -Arithmetic. -------------------------------------------------------------------------------- -*/ -floatx80 int32_to_floatx80( int32 a ) -{ - flag zSign; - uint32 absA; - int8 shiftCount; - bits64 zSig; - - if ( a == 0 ) return packFloatx80( 0, 0, 0 ); - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 32; - zSig = absA; - return packFloatx80( zSign, 0x403E - shiftCount, zSig<<shiftCount ); - -} - -#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 -performed according to the IEC/IEEE Standard for Binary Floating-point -Arithmetic---which means in particular that the conversion is rounded -according to the current rounding mode. If `a' is a NaN, the largest -positive integer is returned. Otherwise, if the conversion overflows, the -largest integer with the same sign as `a' is returned. -------------------------------------------------------------------------------- -*/ -int32 float32_to_int32( float32 a ) -{ - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - bits64 zSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; - if ( aExp ) aSig |= 0x00800000; - shiftCount = 0xAF - aExp; - zSig = aSig; - zSig <<= 32; - if ( 0 < shiftCount ) shift64RightJamming( zSig, shiftCount, &zSig ); - return roundAndPackInt32( aSign, zSig ); - -} - -/* -------------------------------------------------------------------------------- -Returns the result of converting the single-precision floating-point value -`a' to the 32-bit two's complement integer format. The conversion is -performed according to the IEC/IEEE Standard for Binary Floating-point -Arithmetic, except that the conversion is always rounded toward zero. If -`a' is a NaN, the largest positive integer is returned. Otherwise, if the -conversion overflows, the largest integer with the same sign as `a' is -returned. -------------------------------------------------------------------------------- -*/ -int32 float32_to_int32_round_to_zero( float32 a ) -{ - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - int32 z; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - shiftCount = aExp - 0x9E; - if ( 0 <= shiftCount ) { - if ( a == 0xCF000000 ) return 0x80000000; - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) return 0x7FFFFFFF; - return 0x80000000; - } - else if ( aExp <= 0x7E ) { - if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - aSig = ( aSig | 0x00800000 )<<8; - z = aSig>>( - shiftCount ); - if ( (bits32) ( aSig<<( shiftCount & 31 ) ) ) { - float_exception_flags |= float_flag_inexact; - } - return aSign ? - z : z; - -} - -/* -------------------------------------------------------------------------------- -Returns the result of converting the single-precision floating-point value -`a' to the double-precision floating-point format. The conversion is -performed according to the IEC/IEEE Standard for Binary Floating-point -Arithmetic. -------------------------------------------------------------------------------- -*/ -float64 float32_to_float64( float32 a ) -{ - flag aSign; - int16 aExp; - bits32 aSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return commonNaNToFloat64( float32ToCommonNaN( a ) ); - return packFloat64( aSign, 0x7FF, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat64( aSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - --aExp; - } - return packFloat64( aSign, aExp + 0x380, ( (bits64) aSig )<<29 ); - -} - -#ifdef FLOATX80 - -/* -------------------------------------------------------------------------------- -Returns the result of converting the single-precision floating-point value -`a' to the extended double-precision floating-point format. The conversion -is performed according to the IEC/IEEE Standard for Binary Floating-point -Arithmetic. -------------------------------------------------------------------------------- -*/ -floatx80 float32_to_floatx80( float32 a ) -{ - flag aSign; - int16 aExp; - bits32 aSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return commonNaNToFloatx80( float32ToCommonNaN( a ) ); - return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - aSig |= 0x00800000; - return packFloatx80( aSign, aExp + 0x3F80, ( (bits64) aSig )<<40 ); - -} - -#endif - -/* -------------------------------------------------------------------------------- -Rounds the single-precision floating-point value `a' to an integer, and -returns the result as a single-precision floating-point value. The -operation is performed according to the IEC/IEEE Standard for Binary -Floating-point Arithmetic. -------------------------------------------------------------------------------- -*/ -float32 float32_round_to_int( float32 a ) -{ - flag aSign; - int16 aExp; - bits32 lastBitMask, roundBitsMask; - int8 roundingMode; - float32 z; - - aExp = extractFloat32Exp( a ); - if ( 0x96 <= aExp ) { - if ( ( aExp == 0xFF ) && extractFloat32Frac( a ) ) { - return propagateFloat32NaN( a, a ); - } - return a; - } - if ( aExp <= 0x7E ) { - if ( (bits32) ( a<<1 ) == 0 ) return a; - float_exception_flags |= float_flag_inexact; - aSign = extractFloat32Sign( a ); - switch ( float_rounding_mode ) { - case float_round_nearest_even: - if ( ( aExp == 0x7E ) && extractFloat32Frac( a ) ) { - return packFloat32( aSign, 0x7F, 0 ); - } - break; - case float_round_down: - return aSign ? 0xBF800000 : 0; |