diff options
author | Laurent Vivier <laurent@vivier.eu> | 2018-02-24 21:17:59 +0100 |
---|---|---|
committer | Laurent Vivier <laurent@vivier.eu> | 2018-03-04 17:22:55 +0100 |
commit | 88857aca93f6ec8f372fb9c8201394b0e5582034 (patch) | |
tree | c0573c016f5973f0a020e4f6a48fb052cc85cf26 /include | |
parent | 24989f0e21cc9cd90237d4836a456c3ffb824b3e (diff) |
softfloat: export some functions
Move fpu/softfloat-macros.h to include/fpu/
Export floatx80 functions to be used by target floatx80
specific implementations.
Exports:
propagateFloatx80NaN(), extractFloatx80Frac(),
extractFloatx80Exp(), extractFloatx80Sign(),
normalizeFloatx80Subnormal(), packFloatx80(),
roundAndPackFloatx80(), normalizeRoundAndPackFloatx80()
Also exports packFloat32() that will be used to implement
m68k fsinh, fcos, fsin, ftan operations.
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20180224201802.911-2-laurent@vivier.eu>
Diffstat (limited to 'include')
-rw-r--r-- | include/fpu/softfloat-macros.h | 841 | ||||
-rw-r--r-- | include/fpu/softfloat.h | 121 |
2 files changed, 962 insertions, 0 deletions
diff --git a/include/fpu/softfloat-macros.h b/include/fpu/softfloat-macros.h new file mode 100644 index 0000000000..35e1603a5e --- /dev/null +++ b/include/fpu/softfloat-macros.h @@ -0,0 +1,841 @@ +/* + * QEMU float support macros + * + * The code in this source file is derived from release 2a of the SoftFloat + * IEC/IEEE Floating-point Arithmetic Package. Those parts of the code (and + * some later contributions) are provided under that license, as detailed below. + * It has subsequently been modified by contributors to the QEMU Project, + * so some portions are provided under: + * the SoftFloat-2a license + * the BSD license + * GPL-v2-or-later + * + * Any future contributions to this file after December 1st 2014 will be + * taken to be licensed under the Softfloat-2a license unless specifically + * indicated otherwise. + */ + +/* +=============================================================================== +This C source fragment is part of the SoftFloat IEC/IEEE Floating-point +Arithmetic Package, Release 2a. + +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 four paragraphs for those parts of +this code that are retained. + +=============================================================================== +*/ + +/* BSD licensing: + * Copyright (c) 2006, Fabrice Bellard + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are met: + * + * 1. Redistributions of source code must retain the above copyright notice, + * this list of conditions and the following disclaimer. + * + * 2. Redistributions in binary form must reproduce the above copyright notice, + * this list of conditions and the following disclaimer in the documentation + * and/or other materials provided with the distribution. + * + * 3. Neither the name of the copyright holder nor the names of its contributors + * may be used to endorse or promote products derived from this software without + * specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF + * THE POSSIBILITY OF SUCH DAMAGE. + */ + +/* Portions of this work are licensed under the terms of the GNU GPL, + * version 2 or later. See the COPYING file in the top-level directory. + */ + +/*---------------------------------------------------------------------------- +| This macro tests for minimum version of the GNU C compiler. +*----------------------------------------------------------------------------*/ +#if defined(__GNUC__) && defined(__GNUC_MINOR__) +# define SOFTFLOAT_GNUC_PREREQ(maj, min) \ + ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min)) +#else +# define SOFTFLOAT_GNUC_PREREQ(maj, min) 0 +#endif + + +/*---------------------------------------------------------------------------- +| Shifts `a' right by the number of bits given in `count'. If any nonzero +| bits are shifted off, they are ``jammed'' into the least significant bit of +| the result by setting the least significant bit to 1. The value of `count' +| can be arbitrarily large; in particular, if `count' is greater than 32, the +| result will be either 0 or 1, depending on whether `a' is zero or nonzero. +| The result is stored in the location pointed to by `zPtr'. +*----------------------------------------------------------------------------*/ + +static inline void shift32RightJamming(uint32_t a, int count, uint32_t *zPtr) +{ + uint32_t z; + + if ( count == 0 ) { + z = a; + } + else if ( count < 32 ) { + z = ( a>>count ) | ( ( a<<( ( - count ) & 31 ) ) != 0 ); + } + else { + z = ( a != 0 ); + } + *zPtr = z; + +} + +/*---------------------------------------------------------------------------- +| Shifts `a' right by the number of bits given in `count'. If any nonzero +| bits are shifted off, they are ``jammed'' into the least significant bit of +| the result by setting the least significant bit to 1. The value of `count' +| can be arbitrarily large; in particular, if `count' is greater than 64, the +| result will be either 0 or 1, depending on whether `a' is zero or nonzero. +| The result is stored in the location pointed to by `zPtr'. +*----------------------------------------------------------------------------*/ + +static inline void shift64RightJamming(uint64_t a, int count, uint64_t *zPtr) +{ + uint64_t z; + + if ( count == 0 ) { + z = a; + } + else if ( count < 64 ) { + z = ( a>>count ) | ( ( a<<( ( - count ) & 63 ) ) != 0 ); + } + else { + z = ( a != 0 ); + } + *zPtr = z; + +} + +/*---------------------------------------------------------------------------- +| Shifts the 128-bit value formed by concatenating `a0' and `a1' right by 64 +| _plus_ the number of bits given in `count'. The shifted result is at most +| 64 nonzero bits; this is stored at the location pointed to by `z0Ptr'. The +| bits shifted off form a second 64-bit result as follows: The _last_ bit +| shifted off is the most-significant bit of the extra result, and the other +| 63 bits of the extra result are all zero if and only if _all_but_the_last_ +| bits shifted off were all zero. This extra result is stored in the location +| pointed to by `z1Ptr'. The value of `count' can be arbitrarily large. +| (This routine makes more sense if `a0' and `a1' are considered to form a +| fixed-point value with binary point between `a0' and `a1'. This fixed-point +| value is shifted right by the number of bits given in `count', and the +| integer part of the result is returned at the location pointed to by +| `z0Ptr'. The fractional part of the result may be slightly corrupted as +| described above, and is returned at the location pointed to by `z1Ptr'.) +*----------------------------------------------------------------------------*/ + +static inline void + shift64ExtraRightJamming( + uint64_t a0, uint64_t a1, int count, uint64_t *z0Ptr, uint64_t *z1Ptr) +{ + uint64_t z0, z1; + int8_t negCount = ( - count ) & 63; + + if ( count == 0 ) { + z1 = a1; + z0 = a0; + } + else if ( count < 64 ) { + z1 = ( a0<<negCount ) | ( a1 != 0 ); + z0 = a0>>count; + } + else { + if ( count == 64 ) { + z1 = a0 | ( a1 != 0 ); + } + else { + z1 = ( ( a0 | a1 ) != 0 ); + } + z0 = 0; + } + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Shifts the 128-bit value formed by concatenating `a0' and `a1' right by the +| number of bits given in `count'. Any bits shifted off are lost. The value +| of `count' can be arbitrarily large; in particular, if `count' is greater +| than 128, the result will be 0. The result is broken into two 64-bit pieces +| which are stored at the locations pointed to by `z0Ptr' and `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + shift128Right( + uint64_t a0, uint64_t a1, int count, uint64_t *z0Ptr, uint64_t *z1Ptr) +{ + uint64_t z0, z1; + int8_t negCount = ( - count ) & 63; + + if ( count == 0 ) { + z1 = a1; + z0 = a0; + } + else if ( count < 64 ) { + z1 = ( a0<<negCount ) | ( a1>>count ); + z0 = a0>>count; + } + else { + z1 = (count < 128) ? (a0 >> (count & 63)) : 0; + z0 = 0; + } + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Shifts the 128-bit value formed by concatenating `a0' and `a1' right by the +| number of bits given in `count'. If any nonzero bits are shifted off, they +| are ``jammed'' into the least significant bit of the result by setting the +| least significant bit to 1. The value of `count' can be arbitrarily large; +| in particular, if `count' is greater than 128, the result will be either +| 0 or 1, depending on whether the concatenation of `a0' and `a1' is zero or +| nonzero. The result is broken into two 64-bit pieces which are stored at +| the locations pointed to by `z0Ptr' and `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + shift128RightJamming( + uint64_t a0, uint64_t a1, int count, uint64_t *z0Ptr, uint64_t *z1Ptr) +{ + uint64_t z0, z1; + int8_t negCount = ( - count ) & 63; + + if ( count == 0 ) { + z1 = a1; + z0 = a0; + } + else if ( count < 64 ) { + z1 = ( a0<<negCount ) | ( a1>>count ) | ( ( a1<<negCount ) != 0 ); + z0 = a0>>count; + } + else { + if ( count == 64 ) { + z1 = a0 | ( a1 != 0 ); + } + else if ( count < 128 ) { + z1 = ( a0>>( count & 63 ) ) | ( ( ( a0<<negCount ) | a1 ) != 0 ); + } + else { + z1 = ( ( a0 | a1 ) != 0 ); + } + z0 = 0; + } + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Shifts the 192-bit value formed by concatenating `a0', `a1', and `a2' right +| by 64 _plus_ the number of bits given in `count'. The shifted result is +| at most 128 nonzero bits; these are broken into two 64-bit pieces which are +| stored at the locations pointed to by `z0Ptr' and `z1Ptr'. The bits shifted +| off form a third 64-bit result as follows: The _last_ bit shifted off is +| the most-significant bit of the extra result, and the other 63 bits of the +| extra result are all zero if and only if _all_but_the_last_ bits shifted off +| were all zero. This extra result is stored in the location pointed to by +| `z2Ptr'. The value of `count' can be arbitrarily large. +| (This routine makes more sense if `a0', `a1', and `a2' are considered +| to form a fixed-point value with binary point between `a1' and `a2'. This +| fixed-point value is shifted right by the number of bits given in `count', +| and the integer part of the result is returned at the locations pointed to +| by `z0Ptr' and `z1Ptr'. The fractional part of the result may be slightly +| corrupted as described above, and is returned at the location pointed to by +| `z2Ptr'.) +*----------------------------------------------------------------------------*/ + +static inline void + shift128ExtraRightJamming( + uint64_t a0, + uint64_t a1, + uint64_t a2, + int count, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr + ) +{ + uint64_t z0, z1, z2; + int8_t negCount = ( - count ) & 63; + + if ( count == 0 ) { + z2 = a2; + z1 = a1; + z0 = a0; + } + else { + if ( count < 64 ) { + z2 = a1<<negCount; + z1 = ( a0<<negCount ) | ( a1>>count ); + z0 = a0>>count; + } + else { + if ( count == 64 ) { + z2 = a1; + z1 = a0; + } + else { + a2 |= a1; + if ( count < 128 ) { + z2 = a0<<negCount; + z1 = a0>>( count & 63 ); + } + else { + z2 = ( count == 128 ) ? a0 : ( a0 != 0 ); + z1 = 0; + } + } + z0 = 0; + } + z2 |= ( a2 != 0 ); + } + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Shifts the 128-bit value formed by concatenating `a0' and `a1' left by the +| number of bits given in `count'. Any bits shifted off are lost. The value +| of `count' must be less than 64. The result is broken into two 64-bit +| pieces which are stored at the locations pointed to by `z0Ptr' and `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + shortShift128Left( + uint64_t a0, uint64_t a1, int count, uint64_t *z0Ptr, uint64_t *z1Ptr) +{ + + *z1Ptr = a1<<count; + *z0Ptr = + ( count == 0 ) ? a0 : ( a0<<count ) | ( a1>>( ( - count ) & 63 ) ); + +} + +/*---------------------------------------------------------------------------- +| Shifts the 192-bit value formed by concatenating `a0', `a1', and `a2' left +| by the number of bits given in `count'. Any bits shifted off are lost. +| The value of `count' must be less than 64. The result is broken into three +| 64-bit pieces which are stored at the locations pointed to by `z0Ptr', +| `z1Ptr', and `z2Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + shortShift192Left( + uint64_t a0, + uint64_t a1, + uint64_t a2, + int count, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr + ) +{ + uint64_t z0, z1, z2; + int8_t negCount; + + z2 = a2<<count; + z1 = a1<<count; + z0 = a0<<count; + if ( 0 < count ) { + negCount = ( ( - count ) & 63 ); + z1 |= a2>>negCount; + z0 |= a1>>negCount; + } + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Adds the 128-bit value formed by concatenating `a0' and `a1' to the 128-bit +| value formed by concatenating `b0' and `b1'. Addition is modulo 2^128, so +| any carry out is lost. The result is broken into two 64-bit pieces which +| are stored at the locations pointed to by `z0Ptr' and `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + add128( + uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1, uint64_t *z0Ptr, uint64_t *z1Ptr ) +{ + uint64_t z1; + + z1 = a1 + b1; + *z1Ptr = z1; + *z0Ptr = a0 + b0 + ( z1 < a1 ); + +} + +/*---------------------------------------------------------------------------- +| Adds the 192-bit value formed by concatenating `a0', `a1', and `a2' to the +| 192-bit value formed by concatenating `b0', `b1', and `b2'. Addition is +| modulo 2^192, so any carry out is lost. The result is broken into three +| 64-bit pieces which are stored at the locations pointed to by `z0Ptr', +| `z1Ptr', and `z2Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + add192( + uint64_t a0, + uint64_t a1, + uint64_t a2, + uint64_t b0, + uint64_t b1, + uint64_t b2, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr + ) +{ + uint64_t z0, z1, z2; + int8_t carry0, carry1; + + z2 = a2 + b2; + carry1 = ( z2 < a2 ); + z1 = a1 + b1; + carry0 = ( z1 < a1 ); + z0 = a0 + b0; + z1 += carry1; + z0 += ( z1 < carry1 ); + z0 += carry0; + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Subtracts the 128-bit value formed by concatenating `b0' and `b1' from the +| 128-bit value formed by concatenating `a0' and `a1'. Subtraction is modulo +| 2^128, so any borrow out (carry out) is lost. The result is broken into two +| 64-bit pieces which are stored at the locations pointed to by `z0Ptr' and +| `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + sub128( + uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1, uint64_t *z0Ptr, uint64_t *z1Ptr ) +{ + + *z1Ptr = a1 - b1; + *z0Ptr = a0 - b0 - ( a1 < b1 ); + +} + +/*---------------------------------------------------------------------------- +| Subtracts the 192-bit value formed by concatenating `b0', `b1', and `b2' +| from the 192-bit value formed by concatenating `a0', `a1', and `a2'. +| Subtraction is modulo 2^192, so any borrow out (carry out) is lost. The +| result is broken into three 64-bit pieces which are stored at the locations +| pointed to by `z0Ptr', `z1Ptr', and `z2Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + sub192( + uint64_t a0, + uint64_t a1, + uint64_t a2, + uint64_t b0, + uint64_t b1, + uint64_t b2, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr + ) +{ + uint64_t z0, z1, z2; + int8_t borrow0, borrow1; + + z2 = a2 - b2; + borrow1 = ( a2 < b2 ); + z1 = a1 - b1; + borrow0 = ( a1 < b1 ); + z0 = a0 - b0; + z0 -= ( z1 < borrow1 ); + z1 -= borrow1; + z0 -= borrow0; + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Multiplies `a' by `b' to obtain a 128-bit product. The product is broken +| into two 64-bit pieces which are stored at the locations pointed to by +| `z0Ptr' and `z1Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void mul64To128( uint64_t a, uint64_t b, uint64_t *z0Ptr, uint64_t *z1Ptr ) +{ + uint32_t aHigh, aLow, bHigh, bLow; + uint64_t z0, zMiddleA, zMiddleB, z1; + + aLow = a; + aHigh = a>>32; + bLow = b; + bHigh = b>>32; + z1 = ( (uint64_t) aLow ) * bLow; + zMiddleA = ( (uint64_t) aLow ) * bHigh; + zMiddleB = ( (uint64_t) aHigh ) * bLow; + z0 = ( (uint64_t) aHigh ) * bHigh; + zMiddleA += zMiddleB; + z0 += ( ( (uint64_t) ( zMiddleA < zMiddleB ) )<<32 ) + ( zMiddleA>>32 ); + zMiddleA <<= 32; + z1 += zMiddleA; + z0 += ( z1 < zMiddleA ); + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Multiplies the 128-bit value formed by concatenating `a0' and `a1' by +| `b' to obtain a 192-bit product. The product is broken into three 64-bit +| pieces which are stored at the locations pointed to by `z0Ptr', `z1Ptr', and +| `z2Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + mul128By64To192( + uint64_t a0, + uint64_t a1, + uint64_t b, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr + ) +{ + uint64_t z0, z1, z2, more1; + + mul64To128( a1, b, &z1, &z2 ); + mul64To128( a0, b, &z0, &more1 ); + add128( z0, more1, 0, z1, &z0, &z1 ); + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Multiplies the 128-bit value formed by concatenating `a0' and `a1' to the +| 128-bit value formed by concatenating `b0' and `b1' to obtain a 256-bit +| product. The product is broken into four 64-bit pieces which are stored at +| the locations pointed to by `z0Ptr', `z1Ptr', `z2Ptr', and `z3Ptr'. +*----------------------------------------------------------------------------*/ + +static inline void + mul128To256( + uint64_t a0, + uint64_t a1, + uint64_t b0, + uint64_t b1, + uint64_t *z0Ptr, + uint64_t *z1Ptr, + uint64_t *z2Ptr, + uint64_t *z3Ptr + ) +{ + uint64_t z0, z1, z2, z3; + uint64_t more1, more2; + + mul64To128( a1, b1, &z2, &z3 ); + mul64To128( a1, b0, &z1, &more2 ); + add128( z1, more2, 0, z2, &z1, &z2 ); + mul64To128( a0, b0, &z0, &more1 ); + add128( z0, more1, 0, z1, &z0, &z1 ); + mul64To128( a0, b1, &more1, &more2 ); + add128( more1, more2, 0, z2, &more1, &z2 ); + add128( z0, z1, 0, more1, &z0, &z1 ); + *z3Ptr = z3; + *z2Ptr = z2; + *z1Ptr = z1; + *z0Ptr = z0; + +} + +/*---------------------------------------------------------------------------- +| Returns an approximation to the 64-bit integer quotient obtained by dividing +| `b' into the 128-bit value formed by concatenating `a0' and `a1'. The +| divisor `b' must be at least 2^63. If q is the exact quotient truncated +| toward zero, the approximation returned lies between q and q + 2 inclusive. +| If the exact quotient q is larger than 64 bits, the maximum positive 64-bit +| unsigned integer is returned. +*----------------------------------------------------------------------------*/ + +static inline uint64_t estimateDiv128To64(uint64_t a0, uint64_t a1, uint64_t b) +{ + uint64_t b0, b1; + uint64_t rem0, rem1, term0, term1; + uint64_t z; + + if ( b <= a0 ) return LIT64( 0xFFFFFFFFFFFFFFFF ); + b0 = b>>32; + z = ( b0<<32 <= a0 ) ? LIT64( 0xFFFFFFFF00000000 ) : ( a0 / b0 )<<32; + mul64To128( b, z, &term0, &term1 ); + sub128( a0, a1, term0, term1, &rem0, &rem1 ); + while ( ( (int64_t) rem0 ) < 0 ) { + z -= LIT64( 0x100000000 ); + b1 = b<<32; + add128( rem0, rem1, b0, b1, &rem0, &rem1 ); + } + rem0 = ( rem0<<32 ) | ( rem1>>32 ); + z |= ( b0<<32 <= rem0 ) ? 0xFFFFFFFF : rem0 / b0; + return z; + +} + +/* From the GNU Multi Precision Library - longlong.h __udiv_qrnnd + * (https://gmplib.org/repo/gmp/file/tip/longlong.h) + * + * Licensed under the GPLv2/LGPLv3 + */ +static inline uint64_t div128To64(uint64_t n0, uint64_t n1, uint64_t d) +{ + uint64_t d0, d1, q0, q1, r1, r0, m; + + d0 = (uint32_t)d; + d1 = d >> 32; + + r1 = n1 % d1; + q1 = n1 / d1; + m = q1 * d0; + r1 = (r1 << 32) | (n0 >> 32); + if (r1 < m) { + q1 -= 1; + r1 += d; + if (r1 >= d) { + if (r1 < m) { + q1 -= 1; + r1 += d; + } + } + } + r1 -= m; + + r0 = r1 % d1; + q0 = r1 / d1; + m = q0 * d0; + r0 = (r0 << 32) | (uint32_t)n0; + if (r0 < m) { + q0 -= 1; + r0 += d; + if (r0 >= d) { + if (r0 < m) { + q0 -= 1; + r0 += d; + } + } + } + r0 -= m; + + /* Return remainder in LSB */ + return (q1 << 32) | q0 | (r0 != 0); +} + +/*---------------------------------------------------------------------------- +| Returns an approximation to the square root of the 32-bit significand given +| by `a'. Considered as an integer, `a' must be at least 2^31. If bit 0 of +| `aExp' (the least significant bit) is 1, the integer returned approximates +| 2^31*sqrt(`a'/2^31), where `a' is considered an integer. If bit 0 of `aExp' +| is 0, the integer returned approximates 2^31*sqrt(`a'/2^30). In either +| case, the approximation returned lies strictly within +/-2 of the exact +| value. +*----------------------------------------------------------------------------*/ + +static inline uint32_t estimateSqrt32(int aExp, uint32_t a) +{ + static const uint16_t sqrtOddAdjustments[] = { + 0x0004, 0x0022, 0x005D, 0x00B1, 0x011D, 0x019F, 0x0236, 0x02E0, + 0x039C, 0x0468, 0x0545, 0x0631, 0x072B, 0x0832, 0x0946, 0x0A67 + }; + static const uint16_t sqrtEvenAdjustments[] = { + 0x0A2D, 0x08AF, 0x075A, 0x0629, 0x051A, 0x0429, 0x0356, 0x029E, + 0x0200, 0x0179, 0x0109, 0x00AF, 0x0068, 0x0034, 0x0012, 0x0002 + }; + int8_t index; + uint32_t z; + + index = ( a>>27 ) & 15; + if ( aExp & 1 ) { + z = 0x4000 + ( a>>17 ) - sqrtOddAdjustments[ (int)index ]; + z = ( ( a / z )<<14 ) + ( z<<15 ); + a >>= 1; + } + else { + z = 0x8000 + ( a>>17 ) - sqrtEvenAdjustments[ (int)index ]; + z = a / z + z; + z = ( 0x20000 <= z ) ? 0xFFFF8000 : ( z<<15 ); + if ( z <= a ) return (uint32_t) ( ( (int32_t) a )>>1 ); + } + return ( (uint32_t) ( ( ( (uint64_t) a )<<31 ) / z ) ) + ( z>>1 ); + +} + +/*---------------------------------------------------------------------------- +| Returns the number of leading 0 bits before the most-significant 1 bit of +| `a'. If `a' is zero, 32 is returned. +*----------------------------------------------------------------------------*/ + +static inline int8_t countLeadingZeros32(uint32_t a) +{ +#if SOFTFLOAT_GNUC_PREREQ(3, 4) + if (a) { + return __builtin_clz(a); + } else { + return 32; + } +#else + static const int8_t countLeadingZerosHigh[] = { + 8, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 + }; + int8_t shiftCount; + + shiftCount = 0; + if ( a < 0x10000 ) { + shiftCount += 16; + a <<= 16; + } + if ( a < 0x1000000 ) { + shiftCount += 8; + a <<= 8; + } + shiftCount += countLeadingZerosHigh[ a>>24 ]; + return shiftCount; +#endif +} + +/*---------------------------------------------------------------------------- +| Returns the number of leading 0 bits before the most-significant 1 bit of +| `a'. If `a' is zero, 64 is returned. +*----------------------------------------------------------------------------*/ + +static inline int8_t countLeadingZeros64(uint64_t a) +{ +#if SOFTFLOAT_GNUC_PREREQ(3, 4) + if (a) { + return __builtin_clzll(a); + } else { + return 64; + } +#else + int8_t shiftCount; + + shiftCount = 0; + if ( a < ( (uint64_t) 1 )<<32 ) { + shiftCount += 32; + } + else { + a >>= 32; + } + shiftCount += countLeadingZeros32( a ); + return shiftCount; +#endif +} + +/*---------------------------------------------------------------------------- +| Returns 1 if the 128-bit value formed by concatenating `a0' and `a1' +| is equal to the 128-bit value formed by concatenating `b0' and `b1'. +| Otherwise, returns 0. +*----------------------------------------------------------------------------*/ + +static inline flag eq128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 ) +{ + + return ( a0 == b0 ) && ( a1 == b1 ); + +} + +/*---------------------------------------------------------------------------- +| Returns 1 if the 128-bit value formed by concatenating `a0' and `a1' is less +| than or equal to the 128-bit value formed by concatenating `b0' and `b1'. +| Otherwise, returns 0. +*----------------------------------------------------------------------------*/ + +static inline flag le128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 ) +{ + + return ( a0 < b0 ) || ( ( a0 == b0 ) && ( a1 <= b1 ) ); + +} + +/*---------------------------------------------------------------------------- +| Returns 1 if the 128-bit value formed by concatenating `a0' and `a1' is less +| than the 128-bit value formed by concatenating `b0' and `b1'. Otherwise, +| returns 0. +*----------------------------------------------------------------------------*/ + +static inline flag lt128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 ) +{ + + return ( a0 < b0 ) || ( ( a0 == b0 ) && ( a1 < b1 ) ); + +} + +/*---------------------------------------------------------------------------- +| Returns 1 if the 128-bit value formed by concatenating `a0' and `a1' is +| not equal to the 128-bit value formed by concatenating `b0' and `b1'. +| Otherwise, returns 0. +*----------------------------------------------------------------------------*/ + +static inline flag ne128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 ) +{ + + return ( a0 != b0 ) || ( a1 != b1 ); + +} diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index 27876e711c..fa4fae224f 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -426,6 +426,23 @@ static inline float32 float32_set_sign(float32 a, int sign) #define float32_infinity make_float32(0x7f800000) /*---------------------------------------------------------------------------- +| 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. +*----------------------------------------------------------------------------*/ + +static inline float32 packFloat32(flag zSign, int zExp, uint32_t zSig) +{ + return make_float32( + (((uint32_t)zSign) << 31) + (((uint32_t)zExp) << 23) + zSig); +} + +/*---------------------------------------------------------------------------- | The pattern for a default generated single-precision NaN. *----------------------------------------------------------------------------*/ float32 float32_default_nan(float_status *status); @@ -641,6 +658,110 @@ static inline bool floatx80_invalid_encoding(floatx80 a) #define floatx80_infinity make_floatx80(0x7fff, 0x8000000000000000LL) /*---------------------------------------------------------------------------- +| Returns the fraction bits of the extended double-precision floating-point +| value `a'. +*----------------------------------------------------------------------------*/ + +static inline uint64_t extractFloatx80Frac(floatx80 a) +{ + return a.low; +} + +/*---------------------------------------------------------------------------- +| Returns the exponent bits of the extended double-precision floating-point +| value `a'. +*----------------------------------------------------------------------------*/ + +static inline int32_t extractFloatx80Exp(floatx80 a) +{ + return a.high & 0x7FFF; +} + +/*---------------------------------------------------------------------------- +| Returns the sign bit of the extended double-precision floating-point value +| `a'. +*----------------------------------------------------------------------------*/ + +static inline flag extractFloatx80Sign(floatx80 a) +{ + return a.high >> 15; +} + +/*---------------------------------------------------------------------------- +| Packs the sign `zSign', exponent `zExp', and significand `zSig' into an +| extended double-precision floating-point value, returning the result. +*----------------------------------------------------------------------------*/ + +static inline floatx80 packFloatx80(flag zSign, int32_t zExp, uint64_t zSig) +{ + floatx80 z; + + z.low = zSig; + z.high = (((uint16_t)zSign) << 15) + zExp; + return z; +} + +/*---------------------------------------------------------------------------- +| 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. +*----------------------------------------------------------------------------*/ + +void normalizeFloatx80Subnormal(uint64_t aSig, int32_t *zExpPtr, + uint64_t *zSigPtr); + +/*---------------------------------------------------------------------------- +| Takes two extended double-precision floating-point values `a' and `b', one +| of which is a NaN, and returns the appropriate NaN result. If either `a' or +| `b' is a signaling NaN, the invalid exception is raised. +*----------------------------------------------------------------------------*/ + +floatx80 propagateFloatx80NaN(floatx80 a, floatx80 b, float_status *status); + +/*---------------------------------------------------------------------------- +| 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. However, if the abstract value is too large, 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. +*----------------------------------------------------------------------------*/ + +floatx80 roundAndPackFloatx80(int8_t roundingPrecision, flag zSign, + int32_t zExp, uint64_t zSig0, uint64_t zSig1, + float_status *status); + +/*---------------------------------------------------------------------------- +| 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. +*----------------------------------------------------------------------------*/ + +floatx80 normalizeRoundAndPackFloatx80(int8_t roundingPrecision, + flag zSign, int32_t zExp, + uint64_t zSig0, uint64_t zSig1, + float_status *status); + +/*---------------------------------------------------------------------------- | The pattern for a default generated extended double-precision NaN. *----------------------------------------------------------------------------*/ floatx80 floatx80_default_nan(float_status *status); |