xref: /openbmc/qemu/include/fpu/softfloat.h (revision 9c4888c9)
1 /*
2  * QEMU float support
3  *
4  * The code in this source file is derived from release 2a of the SoftFloat
5  * IEC/IEEE Floating-point Arithmetic Package. Those parts of the code (and
6  * some later contributions) are provided under that license, as detailed below.
7  * It has subsequently been modified by contributors to the QEMU Project,
8  * so some portions are provided under:
9  *  the SoftFloat-2a license
10  *  the BSD license
11  *  GPL-v2-or-later
12  *
13  * Any future contributions to this file after December 1st 2014 will be
14  * taken to be licensed under the Softfloat-2a license unless specifically
15  * indicated otherwise.
16  */
17 
18 /*
19 ===============================================================================
20 This C header file is part of the SoftFloat IEC/IEEE Floating-point
21 Arithmetic Package, Release 2a.
22 
23 Written by John R. Hauser.  This work was made possible in part by the
24 International Computer Science Institute, located at Suite 600, 1947 Center
25 Street, Berkeley, California 94704.  Funding was partially provided by the
26 National Science Foundation under grant MIP-9311980.  The original version
27 of this code was written as part of a project to build a fixed-point vector
28 processor in collaboration with the University of California at Berkeley,
29 overseen by Profs. Nelson Morgan and John Wawrzynek.  More information
30 is available through the Web page `http://HTTP.CS.Berkeley.EDU/~jhauser/
31 arithmetic/SoftFloat.html'.
32 
33 THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
34 has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
35 TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
36 PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
37 AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
38 
39 Derivative works are acceptable, even for commercial purposes, so long as
40 (1) they include prominent notice that the work is derivative, and (2) they
41 include prominent notice akin to these four paragraphs for those parts of
42 this code that are retained.
43 
44 ===============================================================================
45 */
46 
47 /* BSD licensing:
48  * Copyright (c) 2006, Fabrice Bellard
49  * All rights reserved.
50  *
51  * Redistribution and use in source and binary forms, with or without
52  * modification, are permitted provided that the following conditions are met:
53  *
54  * 1. Redistributions of source code must retain the above copyright notice,
55  * this list of conditions and the following disclaimer.
56  *
57  * 2. Redistributions in binary form must reproduce the above copyright notice,
58  * this list of conditions and the following disclaimer in the documentation
59  * and/or other materials provided with the distribution.
60  *
61  * 3. Neither the name of the copyright holder nor the names of its contributors
62  * may be used to endorse or promote products derived from this software without
63  * specific prior written permission.
64  *
65  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
66  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
67  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
68  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
69  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
70  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
71  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
72  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
73  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
74  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
75  * THE POSSIBILITY OF SUCH DAMAGE.
76  */
77 
78 /* Portions of this work are licensed under the terms of the GNU GPL,
79  * version 2 or later. See the COPYING file in the top-level directory.
80  */
81 
82 #ifndef SOFTFLOAT_H
83 #define SOFTFLOAT_H
84 
85 /*----------------------------------------------------------------------------
86 | Software IEC/IEEE floating-point ordering relations
87 *----------------------------------------------------------------------------*/
88 
89 typedef enum {
90     float_relation_less      = -1,
91     float_relation_equal     =  0,
92     float_relation_greater   =  1,
93     float_relation_unordered =  2
94 } FloatRelation;
95 
96 #include "fpu/softfloat-types.h"
97 #include "fpu/softfloat-helpers.h"
98 #include "qemu/int128.h"
99 
100 /*----------------------------------------------------------------------------
101 | Routine to raise any or all of the software IEC/IEEE floating-point
102 | exception flags.
103 *----------------------------------------------------------------------------*/
104 static inline void float_raise(uint16_t flags, float_status *status)
105 {
106     status->float_exception_flags |= flags;
107 }
108 
109 /*----------------------------------------------------------------------------
110 | If `a' is denormal and we are in flush-to-zero mode then set the
111 | input-denormal exception and return zero. Otherwise just return the value.
112 *----------------------------------------------------------------------------*/
113 float16 float16_squash_input_denormal(float16 a, float_status *status);
114 float32 float32_squash_input_denormal(float32 a, float_status *status);
115 float64 float64_squash_input_denormal(float64 a, float_status *status);
116 bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status);
117 
118 /*----------------------------------------------------------------------------
119 | Options to indicate which negations to perform in float*_muladd()
120 | Using these differs from negating an input or output before calling
121 | the muladd function in that this means that a NaN doesn't have its
122 | sign bit inverted before it is propagated.
123 | We also support halving the result before rounding, as a special
124 | case to support the ARM fused-sqrt-step instruction FRSQRTS.
125 *----------------------------------------------------------------------------*/
126 enum {
127     float_muladd_negate_c = 1,
128     float_muladd_negate_product = 2,
129     float_muladd_negate_result = 4,
130     float_muladd_halve_result = 8,
131 };
132 
133 /*----------------------------------------------------------------------------
134 | Software IEC/IEEE integer-to-floating-point conversion routines.
135 *----------------------------------------------------------------------------*/
136 
137 float16 int16_to_float16_scalbn(int16_t a, int, float_status *status);
138 float16 int32_to_float16_scalbn(int32_t a, int, float_status *status);
139 float16 int64_to_float16_scalbn(int64_t a, int, float_status *status);
140 float16 uint16_to_float16_scalbn(uint16_t a, int, float_status *status);
141 float16 uint32_to_float16_scalbn(uint32_t a, int, float_status *status);
142 float16 uint64_to_float16_scalbn(uint64_t a, int, float_status *status);
143 
144 float16 int8_to_float16(int8_t a, float_status *status);
145 float16 int16_to_float16(int16_t a, float_status *status);
146 float16 int32_to_float16(int32_t a, float_status *status);
147 float16 int64_to_float16(int64_t a, float_status *status);
148 float16 uint8_to_float16(uint8_t a, float_status *status);
149 float16 uint16_to_float16(uint16_t a, float_status *status);
150 float16 uint32_to_float16(uint32_t a, float_status *status);
151 float16 uint64_to_float16(uint64_t a, float_status *status);
152 
153 float32 int16_to_float32_scalbn(int16_t, int, float_status *status);
154 float32 int32_to_float32_scalbn(int32_t, int, float_status *status);
155 float32 int64_to_float32_scalbn(int64_t, int, float_status *status);
156 float32 uint16_to_float32_scalbn(uint16_t, int, float_status *status);
157 float32 uint32_to_float32_scalbn(uint32_t, int, float_status *status);
158 float32 uint64_to_float32_scalbn(uint64_t, int, float_status *status);
159 
160 float32 int16_to_float32(int16_t, float_status *status);
161 float32 int32_to_float32(int32_t, float_status *status);
162 float32 int64_to_float32(int64_t, float_status *status);
163 float32 uint16_to_float32(uint16_t, float_status *status);
164 float32 uint32_to_float32(uint32_t, float_status *status);
165 float32 uint64_to_float32(uint64_t, float_status *status);
166 
167 float64 int16_to_float64_scalbn(int16_t, int, float_status *status);
168 float64 int32_to_float64_scalbn(int32_t, int, float_status *status);
169 float64 int64_to_float64_scalbn(int64_t, int, float_status *status);
170 float64 uint16_to_float64_scalbn(uint16_t, int, float_status *status);
171 float64 uint32_to_float64_scalbn(uint32_t, int, float_status *status);
172 float64 uint64_to_float64_scalbn(uint64_t, int, float_status *status);
173 
174 float64 int16_to_float64(int16_t, float_status *status);
175 float64 int32_to_float64(int32_t, float_status *status);
176 float64 int64_to_float64(int64_t, float_status *status);
177 float64 uint16_to_float64(uint16_t, float_status *status);
178 float64 uint32_to_float64(uint32_t, float_status *status);
179 float64 uint64_to_float64(uint64_t, float_status *status);
180 
181 floatx80 int32_to_floatx80(int32_t, float_status *status);
182 floatx80 int64_to_floatx80(int64_t, float_status *status);
183 
184 float128 int32_to_float128(int32_t, float_status *status);
185 float128 int64_to_float128(int64_t, float_status *status);
186 float128 int128_to_float128(Int128, float_status *status);
187 float128 uint64_to_float128(uint64_t, float_status *status);
188 float128 uint128_to_float128(Int128, float_status *status);
189 
190 /*----------------------------------------------------------------------------
191 | Software half-precision conversion routines.
192 *----------------------------------------------------------------------------*/
193 
194 float16 float32_to_float16(float32, bool ieee, float_status *status);
195 float32 float16_to_float32(float16, bool ieee, float_status *status);
196 float16 float64_to_float16(float64 a, bool ieee, float_status *status);
197 float64 float16_to_float64(float16 a, bool ieee, float_status *status);
198 
199 int8_t  float16_to_int8_scalbn(float16, FloatRoundMode, int,
200                                float_status *status);
201 int16_t float16_to_int16_scalbn(float16, FloatRoundMode, int, float_status *);
202 int32_t float16_to_int32_scalbn(float16, FloatRoundMode, int, float_status *);
203 int64_t float16_to_int64_scalbn(float16, FloatRoundMode, int, float_status *);
204 
205 int8_t  float16_to_int8(float16, float_status *status);
206 int16_t float16_to_int16(float16, float_status *status);
207 int32_t float16_to_int32(float16, float_status *status);
208 int64_t float16_to_int64(float16, float_status *status);
209 
210 int16_t float16_to_int16_round_to_zero(float16, float_status *status);
211 int32_t float16_to_int32_round_to_zero(float16, float_status *status);
212 int64_t float16_to_int64_round_to_zero(float16, float_status *status);
213 
214 uint8_t float16_to_uint8_scalbn(float16 a, FloatRoundMode,
215                                 int, float_status *status);
216 uint16_t float16_to_uint16_scalbn(float16 a, FloatRoundMode,
217                                   int, float_status *status);
218 uint32_t float16_to_uint32_scalbn(float16 a, FloatRoundMode,
219                                   int, float_status *status);
220 uint64_t float16_to_uint64_scalbn(float16 a, FloatRoundMode,
221                                   int, float_status *status);
222 
223 uint8_t  float16_to_uint8(float16 a, float_status *status);
224 uint16_t float16_to_uint16(float16 a, float_status *status);
225 uint32_t float16_to_uint32(float16 a, float_status *status);
226 uint64_t float16_to_uint64(float16 a, float_status *status);
227 
228 uint16_t float16_to_uint16_round_to_zero(float16 a, float_status *status);
229 uint32_t float16_to_uint32_round_to_zero(float16 a, float_status *status);
230 uint64_t float16_to_uint64_round_to_zero(float16 a, float_status *status);
231 
232 /*----------------------------------------------------------------------------
233 | Software half-precision operations.
234 *----------------------------------------------------------------------------*/
235 
236 float16 float16_round_to_int(float16, float_status *status);
237 float16 float16_add(float16, float16, float_status *status);
238 float16 float16_sub(float16, float16, float_status *status);
239 float16 float16_mul(float16, float16, float_status *status);
240 float16 float16_muladd(float16, float16, float16, int, float_status *status);
241 float16 float16_div(float16, float16, float_status *status);
242 float16 float16_scalbn(float16, int, float_status *status);
243 float16 float16_min(float16, float16, float_status *status);
244 float16 float16_max(float16, float16, float_status *status);
245 float16 float16_minnum(float16, float16, float_status *status);
246 float16 float16_maxnum(float16, float16, float_status *status);
247 float16 float16_minnummag(float16, float16, float_status *status);
248 float16 float16_maxnummag(float16, float16, float_status *status);
249 float16 float16_minimum_number(float16, float16, float_status *status);
250 float16 float16_maximum_number(float16, float16, float_status *status);
251 float16 float16_sqrt(float16, float_status *status);
252 FloatRelation float16_compare(float16, float16, float_status *status);
253 FloatRelation float16_compare_quiet(float16, float16, float_status *status);
254 
255 bool float16_is_quiet_nan(float16, float_status *status);
256 bool float16_is_signaling_nan(float16, float_status *status);
257 float16 float16_silence_nan(float16, float_status *status);
258 
259 static inline bool float16_is_any_nan(float16 a)
260 {
261     return ((float16_val(a) & ~0x8000) > 0x7c00);
262 }
263 
264 static inline bool float16_is_neg(float16 a)
265 {
266     return float16_val(a) >> 15;
267 }
268 
269 static inline bool float16_is_infinity(float16 a)
270 {
271     return (float16_val(a) & 0x7fff) == 0x7c00;
272 }
273 
274 static inline bool float16_is_zero(float16 a)
275 {
276     return (float16_val(a) & 0x7fff) == 0;
277 }
278 
279 static inline bool float16_is_zero_or_denormal(float16 a)
280 {
281     return (float16_val(a) & 0x7c00) == 0;
282 }
283 
284 static inline bool float16_is_normal(float16 a)
285 {
286     return (((float16_val(a) >> 10) + 1) & 0x1f) >= 2;
287 }
288 
289 static inline float16 float16_abs(float16 a)
290 {
291     /* Note that abs does *not* handle NaN specially, nor does
292      * it flush denormal inputs to zero.
293      */
294     return make_float16(float16_val(a) & 0x7fff);
295 }
296 
297 static inline float16 float16_chs(float16 a)
298 {
299     /* Note that chs does *not* handle NaN specially, nor does
300      * it flush denormal inputs to zero.
301      */
302     return make_float16(float16_val(a) ^ 0x8000);
303 }
304 
305 static inline float16 float16_set_sign(float16 a, int sign)
306 {
307     return make_float16((float16_val(a) & 0x7fff) | (sign << 15));
308 }
309 
310 static inline bool float16_eq(float16 a, float16 b, float_status *s)
311 {
312     return float16_compare(a, b, s) == float_relation_equal;
313 }
314 
315 static inline bool float16_le(float16 a, float16 b, float_status *s)
316 {
317     return float16_compare(a, b, s) <= float_relation_equal;
318 }
319 
320 static inline bool float16_lt(float16 a, float16 b, float_status *s)
321 {
322     return float16_compare(a, b, s) < float_relation_equal;
323 }
324 
325 static inline bool float16_unordered(float16 a, float16 b, float_status *s)
326 {
327     return float16_compare(a, b, s) == float_relation_unordered;
328 }
329 
330 static inline bool float16_eq_quiet(float16 a, float16 b, float_status *s)
331 {
332     return float16_compare_quiet(a, b, s) == float_relation_equal;
333 }
334 
335 static inline bool float16_le_quiet(float16 a, float16 b, float_status *s)
336 {
337     return float16_compare_quiet(a, b, s) <= float_relation_equal;
338 }
339 
340 static inline bool float16_lt_quiet(float16 a, float16 b, float_status *s)
341 {
342     return float16_compare_quiet(a, b, s) < float_relation_equal;
343 }
344 
345 static inline bool float16_unordered_quiet(float16 a, float16 b,
346                                            float_status *s)
347 {
348     return float16_compare_quiet(a, b, s) == float_relation_unordered;
349 }
350 
351 #define float16_zero make_float16(0)
352 #define float16_half make_float16(0x3800)
353 #define float16_one make_float16(0x3c00)
354 #define float16_one_point_five make_float16(0x3e00)
355 #define float16_two make_float16(0x4000)
356 #define float16_three make_float16(0x4200)
357 #define float16_infinity make_float16(0x7c00)
358 
359 /*----------------------------------------------------------------------------
360 | Software bfloat16 conversion routines.
361 *----------------------------------------------------------------------------*/
362 
363 bfloat16 bfloat16_round_to_int(bfloat16, float_status *status);
364 bfloat16 float32_to_bfloat16(float32, float_status *status);
365 float32 bfloat16_to_float32(bfloat16, float_status *status);
366 bfloat16 float64_to_bfloat16(float64 a, float_status *status);
367 float64 bfloat16_to_float64(bfloat16 a, float_status *status);
368 
369 int16_t bfloat16_to_int16_scalbn(bfloat16, FloatRoundMode,
370                                  int, float_status *status);
371 int32_t bfloat16_to_int32_scalbn(bfloat16, FloatRoundMode,
372                                  int, float_status *status);
373 int64_t bfloat16_to_int64_scalbn(bfloat16, FloatRoundMode,
374                                  int, float_status *status);
375 
376 int16_t bfloat16_to_int16(bfloat16, float_status *status);
377 int32_t bfloat16_to_int32(bfloat16, float_status *status);
378 int64_t bfloat16_to_int64(bfloat16, float_status *status);
379 
380 int16_t bfloat16_to_int16_round_to_zero(bfloat16, float_status *status);
381 int32_t bfloat16_to_int32_round_to_zero(bfloat16, float_status *status);
382 int64_t bfloat16_to_int64_round_to_zero(bfloat16, float_status *status);
383 
384 uint16_t bfloat16_to_uint16_scalbn(bfloat16 a, FloatRoundMode,
385                                    int, float_status *status);
386 uint32_t bfloat16_to_uint32_scalbn(bfloat16 a, FloatRoundMode,
387                                    int, float_status *status);
388 uint64_t bfloat16_to_uint64_scalbn(bfloat16 a, FloatRoundMode,
389                                    int, float_status *status);
390 
391 uint16_t bfloat16_to_uint16(bfloat16 a, float_status *status);
392 uint32_t bfloat16_to_uint32(bfloat16 a, float_status *status);
393 uint64_t bfloat16_to_uint64(bfloat16 a, float_status *status);
394 
395 uint16_t bfloat16_to_uint16_round_to_zero(bfloat16 a, float_status *status);
396 uint32_t bfloat16_to_uint32_round_to_zero(bfloat16 a, float_status *status);
397 uint64_t bfloat16_to_uint64_round_to_zero(bfloat16 a, float_status *status);
398 
399 bfloat16 int16_to_bfloat16_scalbn(int16_t a, int, float_status *status);
400 bfloat16 int32_to_bfloat16_scalbn(int32_t a, int, float_status *status);
401 bfloat16 int64_to_bfloat16_scalbn(int64_t a, int, float_status *status);
402 bfloat16 uint16_to_bfloat16_scalbn(uint16_t a, int, float_status *status);
403 bfloat16 uint32_to_bfloat16_scalbn(uint32_t a, int, float_status *status);
404 bfloat16 uint64_to_bfloat16_scalbn(uint64_t a, int, float_status *status);
405 
406 bfloat16 int16_to_bfloat16(int16_t a, float_status *status);
407 bfloat16 int32_to_bfloat16(int32_t a, float_status *status);
408 bfloat16 int64_to_bfloat16(int64_t a, float_status *status);
409 bfloat16 uint16_to_bfloat16(uint16_t a, float_status *status);
410 bfloat16 uint32_to_bfloat16(uint32_t a, float_status *status);
411 bfloat16 uint64_to_bfloat16(uint64_t a, float_status *status);
412 
413 /*----------------------------------------------------------------------------
414 | Software bfloat16 operations.
415 *----------------------------------------------------------------------------*/
416 
417 bfloat16 bfloat16_add(bfloat16, bfloat16, float_status *status);
418 bfloat16 bfloat16_sub(bfloat16, bfloat16, float_status *status);
419 bfloat16 bfloat16_mul(bfloat16, bfloat16, float_status *status);
420 bfloat16 bfloat16_div(bfloat16, bfloat16, float_status *status);
421 bfloat16 bfloat16_muladd(bfloat16, bfloat16, bfloat16, int,
422                          float_status *status);
423 float16 bfloat16_scalbn(bfloat16, int, float_status *status);
424 bfloat16 bfloat16_min(bfloat16, bfloat16, float_status *status);
425 bfloat16 bfloat16_max(bfloat16, bfloat16, float_status *status);
426 bfloat16 bfloat16_minnum(bfloat16, bfloat16, float_status *status);
427 bfloat16 bfloat16_maxnum(bfloat16, bfloat16, float_status *status);
428 bfloat16 bfloat16_minnummag(bfloat16, bfloat16, float_status *status);
429 bfloat16 bfloat16_maxnummag(bfloat16, bfloat16, float_status *status);
430 bfloat16 bfloat16_minimum_number(bfloat16, bfloat16, float_status *status);
431 bfloat16 bfloat16_maximum_number(bfloat16, bfloat16, float_status *status);
432 bfloat16 bfloat16_sqrt(bfloat16, float_status *status);
433 FloatRelation bfloat16_compare(bfloat16, bfloat16, float_status *status);
434 FloatRelation bfloat16_compare_quiet(bfloat16, bfloat16, float_status *status);
435 
436 bool bfloat16_is_quiet_nan(bfloat16, float_status *status);
437 bool bfloat16_is_signaling_nan(bfloat16, float_status *status);
438 bfloat16 bfloat16_silence_nan(bfloat16, float_status *status);
439 bfloat16 bfloat16_default_nan(float_status *status);
440 
441 static inline bool bfloat16_is_any_nan(bfloat16 a)
442 {
443     return ((a & ~0x8000) > 0x7F80);
444 }
445 
446 static inline bool bfloat16_is_neg(bfloat16 a)
447 {
448     return a >> 15;
449 }
450 
451 static inline bool bfloat16_is_infinity(bfloat16 a)
452 {
453     return (a & 0x7fff) == 0x7F80;
454 }
455 
456 static inline bool bfloat16_is_zero(bfloat16 a)
457 {
458     return (a & 0x7fff) == 0;
459 }
460 
461 static inline bool bfloat16_is_zero_or_denormal(bfloat16 a)
462 {
463     return (a & 0x7F80) == 0;
464 }
465 
466 static inline bool bfloat16_is_normal(bfloat16 a)
467 {
468     return (((a >> 7) + 1) & 0xff) >= 2;
469 }
470 
471 static inline bfloat16 bfloat16_abs(bfloat16 a)
472 {
473     /* Note that abs does *not* handle NaN specially, nor does
474      * it flush denormal inputs to zero.
475      */
476     return a & 0x7fff;
477 }
478 
479 static inline bfloat16 bfloat16_chs(bfloat16 a)
480 {
481     /* Note that chs does *not* handle NaN specially, nor does
482      * it flush denormal inputs to zero.
483      */
484     return a ^ 0x8000;
485 }
486 
487 static inline bfloat16 bfloat16_set_sign(bfloat16 a, int sign)
488 {
489     return (a & 0x7fff) | (sign << 15);
490 }
491 
492 static inline bool bfloat16_eq(bfloat16 a, bfloat16 b, float_status *s)
493 {
494     return bfloat16_compare(a, b, s) == float_relation_equal;
495 }
496 
497 static inline bool bfloat16_le(bfloat16 a, bfloat16 b, float_status *s)
498 {
499     return bfloat16_compare(a, b, s) <= float_relation_equal;
500 }
501 
502 static inline bool bfloat16_lt(bfloat16 a, bfloat16 b, float_status *s)
503 {
504     return bfloat16_compare(a, b, s) < float_relation_equal;
505 }
506 
507 static inline bool bfloat16_unordered(bfloat16 a, bfloat16 b, float_status *s)
508 {
509     return bfloat16_compare(a, b, s) == float_relation_unordered;
510 }
511 
512 static inline bool bfloat16_eq_quiet(bfloat16 a, bfloat16 b, float_status *s)
513 {
514     return bfloat16_compare_quiet(a, b, s) == float_relation_equal;
515 }
516 
517 static inline bool bfloat16_le_quiet(bfloat16 a, bfloat16 b, float_status *s)
518 {
519     return bfloat16_compare_quiet(a, b, s) <= float_relation_equal;
520 }
521 
522 static inline bool bfloat16_lt_quiet(bfloat16 a, bfloat16 b, float_status *s)
523 {
524     return bfloat16_compare_quiet(a, b, s) < float_relation_equal;
525 }
526 
527 static inline bool bfloat16_unordered_quiet(bfloat16 a, bfloat16 b,
528                                            float_status *s)
529 {
530     return bfloat16_compare_quiet(a, b, s) == float_relation_unordered;
531 }
532 
533 #define bfloat16_zero 0
534 #define bfloat16_half 0x3f00
535 #define bfloat16_one 0x3f80
536 #define bfloat16_one_point_five 0x3fc0
537 #define bfloat16_two 0x4000
538 #define bfloat16_three 0x4040
539 #define bfloat16_infinity 0x7f80
540 
541 /*----------------------------------------------------------------------------
542 | The pattern for a default generated half-precision NaN.
543 *----------------------------------------------------------------------------*/
544 float16 float16_default_nan(float_status *status);
545 
546 /*----------------------------------------------------------------------------
547 | Software IEC/IEEE single-precision conversion routines.
548 *----------------------------------------------------------------------------*/
549 
550 int16_t float32_to_int16_scalbn(float32, FloatRoundMode, int, float_status *);
551 int32_t float32_to_int32_scalbn(float32, FloatRoundMode, int, float_status *);
552 int64_t float32_to_int64_scalbn(float32, FloatRoundMode, int, float_status *);
553 
554 int16_t float32_to_int16(float32, float_status *status);
555 int32_t float32_to_int32(float32, float_status *status);
556 int64_t float32_to_int64(float32, float_status *status);
557 
558 int16_t float32_to_int16_round_to_zero(float32, float_status *status);
559 int32_t float32_to_int32_round_to_zero(float32, float_status *status);
560 int64_t float32_to_int64_round_to_zero(float32, float_status *status);
561 
562 uint16_t float32_to_uint16_scalbn(float32, FloatRoundMode, int, float_status *);
563 uint32_t float32_to_uint32_scalbn(float32, FloatRoundMode, int, float_status *);
564 uint64_t float32_to_uint64_scalbn(float32, FloatRoundMode, int, float_status *);
565 
566 uint16_t float32_to_uint16(float32, float_status *status);
567 uint32_t float32_to_uint32(float32, float_status *status);
568 uint64_t float32_to_uint64(float32, float_status *status);
569 
570 uint16_t float32_to_uint16_round_to_zero(float32, float_status *status);
571 uint32_t float32_to_uint32_round_to_zero(float32, float_status *status);
572 uint64_t float32_to_uint64_round_to_zero(float32, float_status *status);
573 
574 float64 float32_to_float64(float32, float_status *status);
575 floatx80 float32_to_floatx80(float32, float_status *status);
576 float128 float32_to_float128(float32, float_status *status);
577 
578 /*----------------------------------------------------------------------------
579 | Software IEC/IEEE single-precision operations.
580 *----------------------------------------------------------------------------*/
581 float32 float32_round_to_int(float32, float_status *status);
582 float32 float32_add(float32, float32, float_status *status);
583 float32 float32_sub(float32, float32, float_status *status);
584 float32 float32_mul(float32, float32, float_status *status);
585 float32 float32_div(float32, float32, float_status *status);
586 float32 float32_rem(float32, float32, float_status *status);
587 float32 float32_muladd(float32, float32, float32, int, float_status *status);
588 float32 float32_sqrt(float32, float_status *status);
589 float32 float32_exp2(float32, float_status *status);
590 float32 float32_log2(float32, float_status *status);
591 FloatRelation float32_compare(float32, float32, float_status *status);
592 FloatRelation float32_compare_quiet(float32, float32, float_status *status);
593 float32 float32_min(float32, float32, float_status *status);
594 float32 float32_max(float32, float32, float_status *status);
595 float32 float32_minnum(float32, float32, float_status *status);
596 float32 float32_maxnum(float32, float32, float_status *status);
597 float32 float32_minnummag(float32, float32, float_status *status);
598 float32 float32_maxnummag(float32, float32, float_status *status);
599 float32 float32_minimum_number(float32, float32, float_status *status);
600 float32 float32_maximum_number(float32, float32, float_status *status);
601 bool float32_is_quiet_nan(float32, float_status *status);
602 bool float32_is_signaling_nan(float32, float_status *status);
603 float32 float32_silence_nan(float32, float_status *status);
604 float32 float32_scalbn(float32, int, float_status *status);
605 
606 static inline float32 float32_abs(float32 a)
607 {
608     /* Note that abs does *not* handle NaN specially, nor does
609      * it flush denormal inputs to zero.
610      */
611     return make_float32(float32_val(a) & 0x7fffffff);
612 }
613 
614 static inline float32 float32_chs(float32 a)
615 {
616     /* Note that chs does *not* handle NaN specially, nor does
617      * it flush denormal inputs to zero.
618      */
619     return make_float32(float32_val(a) ^ 0x80000000);
620 }
621 
622 static inline bool float32_is_infinity(float32 a)
623 {
624     return (float32_val(a) & 0x7fffffff) == 0x7f800000;
625 }
626 
627 static inline bool float32_is_neg(float32 a)
628 {
629     return float32_val(a) >> 31;
630 }
631 
632 static inline bool float32_is_zero(float32 a)
633 {
634     return (float32_val(a) & 0x7fffffff) == 0;
635 }
636 
637 static inline bool float32_is_any_nan(float32 a)
638 {
639     return ((float32_val(a) & ~(1 << 31)) > 0x7f800000UL);
640 }
641 
642 static inline bool float32_is_zero_or_denormal(float32 a)
643 {
644     return (float32_val(a) & 0x7f800000) == 0;
645 }
646 
647 static inline bool float32_is_normal(float32 a)
648 {
649     return (((float32_val(a) >> 23) + 1) & 0xff) >= 2;
650 }
651 
652 static inline bool float32_is_denormal(float32 a)
653 {
654     return float32_is_zero_or_denormal(a) && !float32_is_zero(a);
655 }
656 
657 static inline bool float32_is_zero_or_normal(float32 a)
658 {
659     return float32_is_normal(a) || float32_is_zero(a);
660 }
661 
662 static inline float32 float32_set_sign(float32 a, int sign)
663 {
664     return make_float32((float32_val(a) & 0x7fffffff) | (sign << 31));
665 }
666 
667 static inline bool float32_eq(float32 a, float32 b, float_status *s)
668 {
669     return float32_compare(a, b, s) == float_relation_equal;
670 }
671 
672 static inline bool float32_le(float32 a, float32 b, float_status *s)
673 {
674     return float32_compare(a, b, s) <= float_relation_equal;
675 }
676 
677 static inline bool float32_lt(float32 a, float32 b, float_status *s)
678 {
679     return float32_compare(a, b, s) < float_relation_equal;
680 }
681 
682 static inline bool float32_unordered(float32 a, float32 b, float_status *s)
683 {
684     return float32_compare(a, b, s) == float_relation_unordered;
685 }
686 
687 static inline bool float32_eq_quiet(float32 a, float32 b, float_status *s)
688 {
689     return float32_compare_quiet(a, b, s) == float_relation_equal;
690 }
691 
692 static inline bool float32_le_quiet(float32 a, float32 b, float_status *s)
693 {
694     return float32_compare_quiet(a, b, s) <= float_relation_equal;
695 }
696 
697 static inline bool float32_lt_quiet(float32 a, float32 b, float_status *s)
698 {
699     return float32_compare_quiet(a, b, s) < float_relation_equal;
700 }
701 
702 static inline bool float32_unordered_quiet(float32 a, float32 b,
703                                            float_status *s)
704 {
705     return float32_compare_quiet(a, b, s) == float_relation_unordered;
706 }
707 
708 #define float32_zero make_float32(0)
709 #define float32_half make_float32(0x3f000000)
710 #define float32_one make_float32(0x3f800000)
711 #define float32_one_point_five make_float32(0x3fc00000)
712 #define float32_two make_float32(0x40000000)
713 #define float32_three make_float32(0x40400000)
714 #define float32_infinity make_float32(0x7f800000)
715 
716 /*----------------------------------------------------------------------------
717 | Packs the sign `zSign', exponent `zExp', and significand `zSig' into a
718 | single-precision floating-point value, returning the result.  After being
719 | shifted into the proper positions, the three fields are simply added
720 | together to form the result.  This means that any integer portion of `zSig'
721 | will be added into the exponent.  Since a properly normalized significand
722 | will have an integer portion equal to 1, the `zExp' input should be 1 less
723 | than the desired result exponent whenever `zSig' is a complete, normalized
724 | significand.
725 *----------------------------------------------------------------------------*/
726 
727 static inline float32 packFloat32(bool zSign, int zExp, uint32_t zSig)
728 {
729     return make_float32(
730           (((uint32_t)zSign) << 31) + (((uint32_t)zExp) << 23) + zSig);
731 }
732 
733 /*----------------------------------------------------------------------------
734 | The pattern for a default generated single-precision NaN.
735 *----------------------------------------------------------------------------*/
736 float32 float32_default_nan(float_status *status);
737 
738 /*----------------------------------------------------------------------------
739 | Software IEC/IEEE double-precision conversion routines.
740 *----------------------------------------------------------------------------*/
741 
742 int16_t float64_to_int16_scalbn(float64, FloatRoundMode, int, float_status *);
743 int32_t float64_to_int32_scalbn(float64, FloatRoundMode, int, float_status *);
744 int64_t float64_to_int64_scalbn(float64, FloatRoundMode, int, float_status *);
745 
746 int16_t float64_to_int16(float64, float_status *status);
747 int32_t float64_to_int32(float64, float_status *status);
748 int64_t float64_to_int64(float64, float_status *status);
749 
750 int16_t float64_to_int16_round_to_zero(float64, float_status *status);
751 int32_t float64_to_int32_round_to_zero(float64, float_status *status);
752 int64_t float64_to_int64_round_to_zero(float64, float_status *status);
753 
754 uint16_t float64_to_uint16_scalbn(float64, FloatRoundMode, int, float_status *);
755 uint32_t float64_to_uint32_scalbn(float64, FloatRoundMode, int, float_status *);
756 uint64_t float64_to_uint64_scalbn(float64, FloatRoundMode, int, float_status *);
757 
758 uint16_t float64_to_uint16(float64, float_status *status);
759 uint32_t float64_to_uint32(float64, float_status *status);
760 uint64_t float64_to_uint64(float64, float_status *status);
761 
762 uint16_t float64_to_uint16_round_to_zero(float64, float_status *status);
763 uint32_t float64_to_uint32_round_to_zero(float64, float_status *status);
764 uint64_t float64_to_uint64_round_to_zero(float64, float_status *status);
765 
766 float32 float64_to_float32(float64, float_status *status);
767 floatx80 float64_to_floatx80(float64, float_status *status);
768 float128 float64_to_float128(float64, float_status *status);
769 
770 /*----------------------------------------------------------------------------
771 | Software IEC/IEEE double-precision operations.
772 *----------------------------------------------------------------------------*/
773 float64 float64_round_to_int(float64, float_status *status);
774 float64 float64_add(float64, float64, float_status *status);
775 float64 float64_sub(float64, float64, float_status *status);
776 float64 float64_mul(float64, float64, float_status *status);
777 float64 float64_div(float64, float64, float_status *status);
778 float64 float64_rem(float64, float64, float_status *status);
779 float64 float64_muladd(float64, float64, float64, int, float_status *status);
780 float64 float64_sqrt(float64, float_status *status);
781 float64 float64_log2(float64, float_status *status);
782 FloatRelation float64_compare(float64, float64, float_status *status);
783 FloatRelation float64_compare_quiet(float64, float64, float_status *status);
784 float64 float64_min(float64, float64, float_status *status);
785 float64 float64_max(float64, float64, float_status *status);
786 float64 float64_minnum(float64, float64, float_status *status);
787 float64 float64_maxnum(float64, float64, float_status *status);
788 float64 float64_minnummag(float64, float64, float_status *status);
789 float64 float64_maxnummag(float64, float64, float_status *status);
790 float64 float64_minimum_number(float64, float64, float_status *status);
791 float64 float64_maximum_number(float64, float64, float_status *status);
792 bool float64_is_quiet_nan(float64 a, float_status *status);
793 bool float64_is_signaling_nan(float64, float_status *status);
794 float64 float64_silence_nan(float64, float_status *status);
795 float64 float64_scalbn(float64, int, float_status *status);
796 
797 static inline float64 float64_abs(float64 a)
798 {
799     /* Note that abs does *not* handle NaN specially, nor does
800      * it flush denormal inputs to zero.
801      */
802     return make_float64(float64_val(a) & 0x7fffffffffffffffLL);
803 }
804 
805 static inline float64 float64_chs(float64 a)
806 {
807     /* Note that chs does *not* handle NaN specially, nor does
808      * it flush denormal inputs to zero.
809      */
810     return make_float64(float64_val(a) ^ 0x8000000000000000LL);
811 }
812 
813 static inline bool float64_is_infinity(float64 a)
814 {
815     return (float64_val(a) & 0x7fffffffffffffffLL ) == 0x7ff0000000000000LL;
816 }
817 
818 static inline bool float64_is_neg(float64 a)
819 {
820     return float64_val(a) >> 63;
821 }
822 
823 static inline bool float64_is_zero(float64 a)
824 {
825     return (float64_val(a) & 0x7fffffffffffffffLL) == 0;
826 }
827 
828 static inline bool float64_is_any_nan(float64 a)
829 {
830     return ((float64_val(a) & ~(1ULL << 63)) > 0x7ff0000000000000ULL);
831 }
832 
833 static inline bool float64_is_zero_or_denormal(float64 a)
834 {
835     return (float64_val(a) & 0x7ff0000000000000LL) == 0;
836 }
837 
838 static inline bool float64_is_normal(float64 a)
839 {
840     return (((float64_val(a) >> 52) + 1) & 0x7ff) >= 2;
841 }
842 
843 static inline bool float64_is_denormal(float64 a)
844 {
845     return float64_is_zero_or_denormal(a) && !float64_is_zero(a);
846 }
847 
848 static inline bool float64_is_zero_or_normal(float64 a)
849 {
850     return float64_is_normal(a) || float64_is_zero(a);
851 }
852 
853 static inline float64 float64_set_sign(float64 a, int sign)
854 {
855     return make_float64((float64_val(a) & 0x7fffffffffffffffULL)
856                         | ((int64_t)sign << 63));
857 }
858 
859 static inline bool float64_eq(float64 a, float64 b, float_status *s)
860 {
861     return float64_compare(a, b, s) == float_relation_equal;
862 }
863 
864 static inline bool float64_le(float64 a, float64 b, float_status *s)
865 {
866     return float64_compare(a, b, s) <= float_relation_equal;
867 }
868 
869 static inline bool float64_lt(float64 a, float64 b, float_status *s)
870 {
871     return float64_compare(a, b, s) < float_relation_equal;
872 }
873 
874 static inline bool float64_unordered(float64 a, float64 b, float_status *s)
875 {
876     return float64_compare(a, b, s) == float_relation_unordered;
877 }
878 
879 static inline bool float64_eq_quiet(float64 a, float64 b, float_status *s)
880 {
881     return float64_compare_quiet(a, b, s) == float_relation_equal;
882 }
883 
884 static inline bool float64_le_quiet(float64 a, float64 b, float_status *s)
885 {
886     return float64_compare_quiet(a, b, s) <= float_relation_equal;
887 }
888 
889 static inline bool float64_lt_quiet(float64 a, float64 b, float_status *s)
890 {
891     return float64_compare_quiet(a, b, s) < float_relation_equal;
892 }
893 
894 static inline bool float64_unordered_quiet(float64 a, float64 b,
895                                            float_status *s)
896 {
897     return float64_compare_quiet(a, b, s) == float_relation_unordered;
898 }
899 
900 #define float64_zero make_float64(0)
901 #define float64_half make_float64(0x3fe0000000000000LL)
902 #define float64_one make_float64(0x3ff0000000000000LL)
903 #define float64_one_point_five make_float64(0x3FF8000000000000ULL)
904 #define float64_two make_float64(0x4000000000000000ULL)
905 #define float64_three make_float64(0x4008000000000000ULL)
906 #define float64_ln2 make_float64(0x3fe62e42fefa39efLL)
907 #define float64_infinity make_float64(0x7ff0000000000000LL)
908 
909 /*----------------------------------------------------------------------------
910 | The pattern for a default generated double-precision NaN.
911 *----------------------------------------------------------------------------*/
912 float64 float64_default_nan(float_status *status);
913 
914 /*----------------------------------------------------------------------------
915 | Software IEC/IEEE double-precision operations, rounding to single precision,
916 | returning a result in double precision, with only one rounding step.
917 *----------------------------------------------------------------------------*/
918 
919 float64 float64r32_add(float64, float64, float_status *status);
920 float64 float64r32_sub(float64, float64, float_status *status);
921 float64 float64r32_mul(float64, float64, float_status *status);
922 float64 float64r32_div(float64, float64, float_status *status);
923 float64 float64r32_muladd(float64, float64, float64, int, float_status *status);
924 float64 float64r32_sqrt(float64, float_status *status);
925 
926 /*----------------------------------------------------------------------------
927 | Software IEC/IEEE extended double-precision conversion routines.
928 *----------------------------------------------------------------------------*/
929 int32_t floatx80_to_int32(floatx80, float_status *status);
930 int32_t floatx80_to_int32_round_to_zero(floatx80, float_status *status);
931 int64_t floatx80_to_int64(floatx80, float_status *status);
932 int64_t floatx80_to_int64_round_to_zero(floatx80, float_status *status);
933 float32 floatx80_to_float32(floatx80, float_status *status);
934 float64 floatx80_to_float64(floatx80, float_status *status);
935 float128 floatx80_to_float128(floatx80, float_status *status);
936 
937 /*----------------------------------------------------------------------------
938 | The pattern for an extended double-precision inf.
939 *----------------------------------------------------------------------------*/
940 extern const floatx80 floatx80_infinity;
941 
942 /*----------------------------------------------------------------------------
943 | Software IEC/IEEE extended double-precision operations.
944 *----------------------------------------------------------------------------*/
945 floatx80 floatx80_round(floatx80 a, float_status *status);
946 floatx80 floatx80_round_to_int(floatx80, float_status *status);
947 floatx80 floatx80_add(floatx80, floatx80, float_status *status);
948 floatx80 floatx80_sub(floatx80, floatx80, float_status *status);
949 floatx80 floatx80_mul(floatx80, floatx80, float_status *status);
950 floatx80 floatx80_div(floatx80, floatx80, float_status *status);
951 floatx80 floatx80_modrem(floatx80, floatx80, bool, uint64_t *,
952                          float_status *status);
953 floatx80 floatx80_mod(floatx80, floatx80, float_status *status);
954 floatx80 floatx80_rem(floatx80, floatx80, float_status *status);
955 floatx80 floatx80_sqrt(floatx80, float_status *status);
956 FloatRelation floatx80_compare(floatx80, floatx80, float_status *status);
957 FloatRelation floatx80_compare_quiet(floatx80, floatx80, float_status *status);
958 int floatx80_is_quiet_nan(floatx80, float_status *status);
959 int floatx80_is_signaling_nan(floatx80, float_status *status);
960 floatx80 floatx80_silence_nan(floatx80, float_status *status);
961 floatx80 floatx80_scalbn(floatx80, int, float_status *status);
962 
963 static inline floatx80 floatx80_abs(floatx80 a)
964 {
965     a.high &= 0x7fff;
966     return a;
967 }
968 
969 static inline floatx80 floatx80_chs(floatx80 a)
970 {
971     a.high ^= 0x8000;
972     return a;
973 }
974 
975 static inline bool floatx80_is_infinity(floatx80 a)
976 {
977 #if defined(TARGET_M68K)
978     return (a.high & 0x7fff) == floatx80_infinity.high && !(a.low << 1);
979 #else
980     return (a.high & 0x7fff) == floatx80_infinity.high &&
981                        a.low == floatx80_infinity.low;
982 #endif
983 }
984 
985 static inline bool floatx80_is_neg(floatx80 a)
986 {
987     return a.high >> 15;
988 }
989 
990 static inline bool floatx80_is_zero(floatx80 a)
991 {
992     return (a.high & 0x7fff) == 0 && a.low == 0;
993 }
994 
995 static inline bool floatx80_is_zero_or_denormal(floatx80 a)
996 {
997     return (a.high & 0x7fff) == 0;
998 }
999 
1000 static inline bool floatx80_is_any_nan(floatx80 a)
1001 {
1002     return ((a.high & 0x7fff) == 0x7fff) && (a.low<<1);
1003 }
1004 
1005 static inline bool floatx80_eq(floatx80 a, floatx80 b, float_status *s)
1006 {
1007     return floatx80_compare(a, b, s) == float_relation_equal;
1008 }
1009 
1010 static inline bool floatx80_le(floatx80 a, floatx80 b, float_status *s)
1011 {
1012     return floatx80_compare(a, b, s) <= float_relation_equal;
1013 }
1014 
1015 static inline bool floatx80_lt(floatx80 a, floatx80 b, float_status *s)
1016 {
1017     return floatx80_compare(a, b, s) < float_relation_equal;
1018 }
1019 
1020 static inline bool floatx80_unordered(floatx80 a, floatx80 b, float_status *s)
1021 {
1022     return floatx80_compare(a, b, s) == float_relation_unordered;
1023 }
1024 
1025 static inline bool floatx80_eq_quiet(floatx80 a, floatx80 b, float_status *s)
1026 {
1027     return floatx80_compare_quiet(a, b, s) == float_relation_equal;
1028 }
1029 
1030 static inline bool floatx80_le_quiet(floatx80 a, floatx80 b, float_status *s)
1031 {
1032     return floatx80_compare_quiet(a, b, s) <= float_relation_equal;
1033 }
1034 
1035 static inline bool floatx80_lt_quiet(floatx80 a, floatx80 b, float_status *s)
1036 {
1037     return floatx80_compare_quiet(a, b, s) < float_relation_equal;
1038 }
1039 
1040 static inline bool floatx80_unordered_quiet(floatx80 a, floatx80 b,
1041                                            float_status *s)
1042 {
1043     return floatx80_compare_quiet(a, b, s) == float_relation_unordered;
1044 }
1045 
1046 /*----------------------------------------------------------------------------
1047 | Return whether the given value is an invalid floatx80 encoding.
1048 | Invalid floatx80 encodings arise when the integer bit is not set, but
1049 | the exponent is not zero. The only times the integer bit is permitted to
1050 | be zero is in subnormal numbers and the value zero.
1051 | This includes what the Intel software developer's manual calls pseudo-NaNs,
1052 | pseudo-infinities and un-normal numbers. It does not include
1053 | pseudo-denormals, which must still be correctly handled as inputs even
1054 | if they are never generated as outputs.
1055 *----------------------------------------------------------------------------*/
1056 static inline bool floatx80_invalid_encoding(floatx80 a)
1057 {
1058 #if defined(TARGET_M68K)
1059     /*-------------------------------------------------------------------------
1060     | With m68k, the explicit integer bit can be zero in the case of:
1061     | - zeros                (exp == 0, mantissa == 0)
1062     | - denormalized numbers (exp == 0, mantissa != 0)
1063     | - unnormalized numbers (exp != 0, exp < 0x7FFF)
1064     | - infinities           (exp == 0x7FFF, mantissa == 0)
1065     | - not-a-numbers        (exp == 0x7FFF, mantissa != 0)
1066     |
1067     | For infinities and NaNs, the explicit integer bit can be either one or
1068     | zero.
1069     |
1070     | The IEEE 754 standard does not define a zero integer bit. Such a number
1071     | is an unnormalized number. Hardware does not directly support
1072     | denormalized and unnormalized numbers, but implicitly supports them by
1073     | trapping them as unimplemented data types, allowing efficient conversion
1074     | in software.
1075     |
1076     | See "M68000 FAMILY PROGRAMMER’S REFERENCE MANUAL",
1077     |     "1.6 FLOATING-POINT DATA TYPES"
1078     *------------------------------------------------------------------------*/
1079     return false;
1080 #else
1081     return (a.low & (1ULL << 63)) == 0 && (a.high & 0x7FFF) != 0;
1082 #endif
1083 }
1084 
1085 #define floatx80_zero make_floatx80(0x0000, 0x0000000000000000LL)
1086 #define floatx80_zero_init make_floatx80_init(0x0000, 0x0000000000000000LL)
1087 #define floatx80_one make_floatx80(0x3fff, 0x8000000000000000LL)
1088 #define floatx80_ln2 make_floatx80(0x3ffe, 0xb17217f7d1cf79acLL)
1089 #define floatx80_pi make_floatx80(0x4000, 0xc90fdaa22168c235LL)
1090 #define floatx80_half make_floatx80(0x3ffe, 0x8000000000000000LL)
1091 
1092 /*----------------------------------------------------------------------------
1093 | Returns the fraction bits of the extended double-precision floating-point
1094 | value `a'.
1095 *----------------------------------------------------------------------------*/
1096 
1097 static inline uint64_t extractFloatx80Frac(floatx80 a)
1098 {
1099     return a.low;
1100 }
1101 
1102 /*----------------------------------------------------------------------------
1103 | Returns the exponent bits of the extended double-precision floating-point
1104 | value `a'.
1105 *----------------------------------------------------------------------------*/
1106 
1107 static inline int32_t extractFloatx80Exp(floatx80 a)
1108 {
1109     return a.high & 0x7FFF;
1110 }
1111 
1112 /*----------------------------------------------------------------------------
1113 | Returns the sign bit of the extended double-precision floating-point value
1114 | `a'.
1115 *----------------------------------------------------------------------------*/
1116 
1117 static inline bool extractFloatx80Sign(floatx80 a)
1118 {
1119     return a.high >> 15;
1120 }
1121 
1122 /*----------------------------------------------------------------------------
1123 | Packs the sign `zSign', exponent `zExp', and significand `zSig' into an
1124 | extended double-precision floating-point value, returning the result.
1125 *----------------------------------------------------------------------------*/
1126 
1127 static inline floatx80 packFloatx80(bool zSign, int32_t zExp, uint64_t zSig)
1128 {
1129     floatx80 z;
1130 
1131     z.low = zSig;
1132     z.high = (((uint16_t)zSign) << 15) + zExp;
1133     return z;
1134 }
1135 
1136 /*----------------------------------------------------------------------------
1137 | Normalizes the subnormal extended double-precision floating-point value
1138 | represented by the denormalized significand `aSig'.  The normalized exponent
1139 | and significand are stored at the locations pointed to by `zExpPtr' and
1140 | `zSigPtr', respectively.
1141 *----------------------------------------------------------------------------*/
1142 
1143 void normalizeFloatx80Subnormal(uint64_t aSig, int32_t *zExpPtr,
1144                                 uint64_t *zSigPtr);
1145 
1146 /*----------------------------------------------------------------------------
1147 | Takes two extended double-precision floating-point values `a' and `b', one
1148 | of which is a NaN, and returns the appropriate NaN result.  If either `a' or
1149 | `b' is a signaling NaN, the invalid exception is raised.
1150 *----------------------------------------------------------------------------*/
1151 
1152 floatx80 propagateFloatx80NaN(floatx80 a, floatx80 b, float_status *status);
1153 
1154 /*----------------------------------------------------------------------------
1155 | Takes an abstract floating-point value having sign `zSign', exponent `zExp',
1156 | and extended significand formed by the concatenation of `zSig0' and `zSig1',
1157 | and returns the proper extended double-precision floating-point value
1158 | corresponding to the abstract input.  Ordinarily, the abstract value is
1159 | rounded and packed into the extended double-precision format, with the
1160 | inexact exception raised if the abstract input cannot be represented
1161 | exactly.  However, if the abstract value is too large, the overflow and
1162 | inexact exceptions are raised and an infinity or maximal finite value is
1163 | returned.  If the abstract value is too small, the input value is rounded to
1164 | a subnormal number, and the underflow and inexact exceptions are raised if
1165 | the abstract input cannot be represented exactly as a subnormal extended
1166 | double-precision floating-point number.
1167 |     If `roundingPrecision' is 32 or 64, the result is rounded to the same
1168 | number of bits as single or double precision, respectively.  Otherwise, the
1169 | result is rounded to the full precision of the extended double-precision
1170 | format.
1171 |     The input significand must be normalized or smaller.  If the input
1172 | significand is not normalized, `zExp' must be 0; in that case, the result
1173 | returned is a subnormal number, and it must not require rounding.  The
1174 | handling of underflow and overflow follows the IEC/IEEE Standard for Binary
1175 | Floating-Point Arithmetic.
1176 *----------------------------------------------------------------------------*/
1177 
1178 floatx80 roundAndPackFloatx80(FloatX80RoundPrec roundingPrecision, bool zSign,
1179                               int32_t zExp, uint64_t zSig0, uint64_t zSig1,
1180                               float_status *status);
1181 
1182 /*----------------------------------------------------------------------------
1183 | Takes an abstract floating-point value having sign `zSign', exponent
1184 | `zExp', and significand formed by the concatenation of `zSig0' and `zSig1',
1185 | and returns the proper extended double-precision floating-point value
1186 | corresponding to the abstract input.  This routine is just like
1187 | `roundAndPackFloatx80' except that the input significand does not have to be
1188 | normalized.
1189 *----------------------------------------------------------------------------*/
1190 
1191 floatx80 normalizeRoundAndPackFloatx80(FloatX80RoundPrec roundingPrecision,
1192                                        bool zSign, int32_t zExp,
1193                                        uint64_t zSig0, uint64_t zSig1,
1194                                        float_status *status);
1195 
1196 /*----------------------------------------------------------------------------
1197 | The pattern for a default generated extended double-precision NaN.
1198 *----------------------------------------------------------------------------*/
1199 floatx80 floatx80_default_nan(float_status *status);
1200 
1201 /*----------------------------------------------------------------------------
1202 | Software IEC/IEEE quadruple-precision conversion routines.
1203 *----------------------------------------------------------------------------*/
1204 int32_t float128_to_int32(float128, float_status *status);
1205 int32_t float128_to_int32_round_to_zero(float128, float_status *status);
1206 int64_t float128_to_int64(float128, float_status *status);
1207 Int128 float128_to_int128(float128, float_status *status);
1208 int64_t float128_to_int64_round_to_zero(float128, float_status *status);
1209 Int128 float128_to_int128_round_to_zero(float128, float_status *status);
1210 uint64_t float128_to_uint64(float128, float_status *status);
1211 Int128 float128_to_uint128(float128, float_status *status);
1212 uint64_t float128_to_uint64_round_to_zero(float128, float_status *status);
1213 Int128 float128_to_uint128_round_to_zero(float128, float_status *status);
1214 uint32_t float128_to_uint32(float128, float_status *status);
1215 uint32_t float128_to_uint32_round_to_zero(float128, float_status *status);
1216 float32 float128_to_float32(float128, float_status *status);
1217 float64 float128_to_float64(float128, float_status *status);
1218 floatx80 float128_to_floatx80(float128, float_status *status);
1219 
1220 /*----------------------------------------------------------------------------
1221 | Software IEC/IEEE quadruple-precision operations.
1222 *----------------------------------------------------------------------------*/
1223 float128 float128_round_to_int(float128, float_status *status);
1224 float128 float128_add(float128, float128, float_status *status);
1225 float128 float128_sub(float128, float128, float_status *status);
1226 float128 float128_mul(float128, float128, float_status *status);
1227 float128 float128_muladd(float128, float128, float128, int,
1228                          float_status *status);
1229 float128 float128_div(float128, float128, float_status *status);
1230 float128 float128_rem(float128, float128, float_status *status);
1231 float128 float128_sqrt(float128, float_status *status);
1232 FloatRelation float128_compare(float128, float128, float_status *status);
1233 FloatRelation float128_compare_quiet(float128, float128, float_status *status);
1234 float128 float128_min(float128, float128, float_status *status);
1235 float128 float128_max(float128, float128, float_status *status);
1236 float128 float128_minnum(float128, float128, float_status *status);
1237 float128 float128_maxnum(float128, float128, float_status *status);
1238 float128 float128_minnummag(float128, float128, float_status *status);
1239 float128 float128_maxnummag(float128, float128, float_status *status);
1240 float128 float128_minimum_number(float128, float128, float_status *status);
1241 float128 float128_maximum_number(float128, float128, float_status *status);
1242 bool float128_is_quiet_nan(float128, float_status *status);
1243 bool float128_is_signaling_nan(float128, float_status *status);
1244 float128 float128_silence_nan(float128, float_status *status);
1245 float128 float128_scalbn(float128, int, float_status *status);
1246 
1247 static inline float128 float128_abs(float128 a)
1248 {
1249     a.high &= 0x7fffffffffffffffLL;
1250     return a;
1251 }
1252 
1253 static inline float128 float128_chs(float128 a)
1254 {
1255     a.high ^= 0x8000000000000000LL;
1256     return a;
1257 }
1258 
1259 static inline bool float128_is_infinity(float128 a)
1260 {
1261     return (a.high & 0x7fffffffffffffffLL) == 0x7fff000000000000LL && a.low == 0;
1262 }
1263 
1264 static inline bool float128_is_neg(float128 a)
1265 {
1266     return a.high >> 63;
1267 }
1268 
1269 static inline bool float128_is_zero(float128 a)
1270 {
1271     return (a.high & 0x7fffffffffffffffLL) == 0 && a.low == 0;
1272 }
1273 
1274 static inline bool float128_is_zero_or_denormal(float128 a)
1275 {
1276     return (a.high & 0x7fff000000000000LL) == 0;
1277 }
1278 
1279 static inline bool float128_is_normal(float128 a)
1280 {
1281     return (((a.high >> 48) + 1) & 0x7fff) >= 2;
1282 }
1283 
1284 static inline bool float128_is_denormal(float128 a)
1285 {
1286     return float128_is_zero_or_denormal(a) && !float128_is_zero(a);
1287 }
1288 
1289 static inline bool float128_is_any_nan(float128 a)
1290 {
1291     return ((a.high >> 48) & 0x7fff) == 0x7fff &&
1292         ((a.low != 0) || ((a.high & 0xffffffffffffLL) != 0));
1293 }
1294 
1295 static inline bool float128_eq(float128 a, float128 b, float_status *s)
1296 {
1297     return float128_compare(a, b, s) == float_relation_equal;
1298 }
1299 
1300 static inline bool float128_le(float128 a, float128 b, float_status *s)
1301 {
1302     return float128_compare(a, b, s) <= float_relation_equal;
1303 }
1304 
1305 static inline bool float128_lt(float128 a, float128 b, float_status *s)
1306 {
1307     return float128_compare(a, b, s) < float_relation_equal;
1308 }
1309 
1310 static inline bool float128_unordered(float128 a, float128 b, float_status *s)
1311 {
1312     return float128_compare(a, b, s) == float_relation_unordered;
1313 }
1314 
1315 static inline bool float128_eq_quiet(float128 a, float128 b, float_status *s)
1316 {
1317     return float128_compare_quiet(a, b, s) == float_relation_equal;
1318 }
1319 
1320 static inline bool float128_le_quiet(float128 a, float128 b, float_status *s)
1321 {
1322     return float128_compare_quiet(a, b, s) <= float_relation_equal;
1323 }
1324 
1325 static inline bool float128_lt_quiet(float128 a, float128 b, float_status *s)
1326 {
1327     return float128_compare_quiet(a, b, s) < float_relation_equal;
1328 }
1329 
1330 static inline bool float128_unordered_quiet(float128 a, float128 b,
1331                                            float_status *s)
1332 {
1333     return float128_compare_quiet(a, b, s) == float_relation_unordered;
1334 }
1335 
1336 #define float128_zero make_float128(0, 0)
1337 
1338 /*----------------------------------------------------------------------------
1339 | The pattern for a default generated quadruple-precision NaN.
1340 *----------------------------------------------------------------------------*/
1341 float128 float128_default_nan(float_status *status);
1342 
1343 #endif /* SOFTFLOAT_H */
1344