File: | build/gcc/double-int.cc |
Warning: | line 624, column 7 Value stored to 'words' during its initialization is never read |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
1 | /* Operations with long integers. |
2 | Copyright (C) 2006-2023 Free Software Foundation, Inc. |
3 | |
4 | This file is part of GCC. |
5 | |
6 | GCC is free software; you can redistribute it and/or modify it |
7 | under the terms of the GNU General Public License as published by the |
8 | Free Software Foundation; either version 3, or (at your option) any |
9 | later version. |
10 | |
11 | GCC is distributed in the hope that it will be useful, but WITHOUT |
12 | ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
13 | FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
14 | for more details. |
15 | |
16 | You should have received a copy of the GNU General Public License |
17 | along with GCC; see the file COPYING3. If not see |
18 | <http://www.gnu.org/licenses/>. */ |
19 | |
20 | #include "config.h" |
21 | #include "system.h" |
22 | #include "coretypes.h" |
23 | #include "tm.h" /* For BITS_PER_UNIT and *_BIG_ENDIAN. */ |
24 | #include "tree.h" |
25 | |
26 | static int add_double_with_sign (unsigned HOST_WIDE_INTlong, HOST_WIDE_INTlong, |
27 | unsigned HOST_WIDE_INTlong, HOST_WIDE_INTlong, |
28 | unsigned HOST_WIDE_INTlong *, HOST_WIDE_INTlong *, |
29 | bool); |
30 | |
31 | #define add_double(l1,h1,l2,h2,lv,hv)add_double_with_sign (l1, h1, l2, h2, lv, hv, false) \ |
32 | add_double_with_sign (l1, h1, l2, h2, lv, hv, false) |
33 | |
34 | static int neg_double (unsigned HOST_WIDE_INTlong, HOST_WIDE_INTlong, |
35 | unsigned HOST_WIDE_INTlong *, HOST_WIDE_INTlong *); |
36 | |
37 | static int mul_double_wide_with_sign (unsigned HOST_WIDE_INTlong, HOST_WIDE_INTlong, |
38 | unsigned HOST_WIDE_INTlong, HOST_WIDE_INTlong, |
39 | unsigned HOST_WIDE_INTlong *, HOST_WIDE_INTlong *, |
40 | unsigned HOST_WIDE_INTlong *, HOST_WIDE_INTlong *, |
41 | bool); |
42 | |
43 | #define mul_double(l1,h1,l2,h2,lv,hv)mul_double_wide_with_sign (l1, h1, l2, h2, lv, hv, nullptr, nullptr , false) \ |
44 | mul_double_wide_with_sign (l1, h1, l2, h2, lv, hv, NULLnullptr, NULLnullptr, false) |
45 | |
46 | static int div_and_round_double (unsigned, int, unsigned HOST_WIDE_INTlong, |
47 | HOST_WIDE_INTlong, unsigned HOST_WIDE_INTlong, |
48 | HOST_WIDE_INTlong, unsigned HOST_WIDE_INTlong *, |
49 | HOST_WIDE_INTlong *, unsigned HOST_WIDE_INTlong *, |
50 | HOST_WIDE_INTlong *); |
51 | |
52 | /* We know that A1 + B1 = SUM1, using 2's complement arithmetic and ignoring |
53 | overflow. Suppose A, B and SUM have the same respective signs as A1, B1, |
54 | and SUM1. Then this yields nonzero if overflow occurred during the |
55 | addition. |
56 | |
57 | Overflow occurs if A and B have the same sign, but A and SUM differ in |
58 | sign. Use `^' to test whether signs differ, and `< 0' to isolate the |
59 | sign. */ |
60 | #define OVERFLOW_SUM_SIGN(a, b, sum)((~((a) ^ (b)) & ((a) ^ (sum))) < 0) ((~((a) ^ (b)) & ((a) ^ (sum))) < 0) |
61 | |
62 | /* To do constant folding on INTEGER_CST nodes requires two-word arithmetic. |
63 | We do that by representing the two-word integer in 4 words, with only |
64 | HOST_BITS_PER_WIDE_INT / 2 bits stored in each word, as a positive |
65 | number. The value of the word is LOWPART + HIGHPART * BASE. */ |
66 | |
67 | #define LOWPART(x)((x) & ((1UL << (64 / 2)) - 1)) \ |
68 | ((x) & ((HOST_WIDE_INT_1U1UL << (HOST_BITS_PER_WIDE_INT64 / 2)) - 1)) |
69 | #define HIGHPART(x)((unsigned long) (x) >> 64 / 2) \ |
70 | ((unsigned HOST_WIDE_INTlong) (x) >> HOST_BITS_PER_WIDE_INT64 / 2) |
71 | #define BASE(1UL << 64 / 2) (HOST_WIDE_INT_1U1UL << HOST_BITS_PER_WIDE_INT64 / 2) |
72 | |
73 | /* Unpack a two-word integer into 4 words. |
74 | LOW and HI are the integer, as two `HOST_WIDE_INT' pieces. |
75 | WORDS points to the array of HOST_WIDE_INTs. */ |
76 | |
77 | static void |
78 | encode (HOST_WIDE_INTlong *words, unsigned HOST_WIDE_INTlong low, HOST_WIDE_INTlong hi) |
79 | { |
80 | words[0] = LOWPART (low)((low) & ((1UL << (64 / 2)) - 1)); |
81 | words[1] = HIGHPART (low)((unsigned long) (low) >> 64 / 2); |
82 | words[2] = LOWPART (hi)((hi) & ((1UL << (64 / 2)) - 1)); |
83 | words[3] = HIGHPART (hi)((unsigned long) (hi) >> 64 / 2); |
84 | } |
85 | |
86 | /* Pack an array of 4 words into a two-word integer. |
87 | WORDS points to the array of words. |
88 | The integer is stored into *LOW and *HI as two `HOST_WIDE_INT' pieces. */ |
89 | |
90 | static void |
91 | decode (HOST_WIDE_INTlong *words, unsigned HOST_WIDE_INTlong *low, |
92 | HOST_WIDE_INTlong *hi) |
93 | { |
94 | *low = words[0] + words[1] * BASE(1UL << 64 / 2); |
95 | *hi = words[2] + words[3] * BASE(1UL << 64 / 2); |
96 | } |
97 | |
98 | /* Add two doubleword integers with doubleword result. |
99 | Return nonzero if the operation overflows according to UNSIGNED_P. |
100 | Each argument is given as two `HOST_WIDE_INT' pieces. |
101 | One argument is L1 and H1; the other, L2 and H2. |
102 | The value is stored as two `HOST_WIDE_INT' pieces in *LV and *HV. */ |
103 | |
104 | static int |
105 | add_double_with_sign (unsigned HOST_WIDE_INTlong l1, HOST_WIDE_INTlong h1, |
106 | unsigned HOST_WIDE_INTlong l2, HOST_WIDE_INTlong h2, |
107 | unsigned HOST_WIDE_INTlong *lv, HOST_WIDE_INTlong *hv, |
108 | bool unsigned_p) |
109 | { |
110 | unsigned HOST_WIDE_INTlong l; |
111 | HOST_WIDE_INTlong h; |
112 | |
113 | l = l1 + l2; |
114 | h = (HOST_WIDE_INTlong) ((unsigned HOST_WIDE_INTlong) h1 |
115 | + (unsigned HOST_WIDE_INTlong) h2 |
116 | + (l < l1)); |
117 | |
118 | *lv = l; |
119 | *hv = h; |
120 | |
121 | if (unsigned_p) |
122 | return ((unsigned HOST_WIDE_INTlong) h < (unsigned HOST_WIDE_INTlong) h1 |
123 | || (h == h1 |
124 | && l < l1)); |
125 | else |
126 | return OVERFLOW_SUM_SIGN (h1, h2, h)((~((h1) ^ (h2)) & ((h1) ^ (h))) < 0); |
127 | } |
128 | |
129 | /* Negate a doubleword integer with doubleword result. |
130 | Return nonzero if the operation overflows, assuming it's signed. |
131 | The argument is given as two `HOST_WIDE_INT' pieces in L1 and H1. |
132 | The value is stored as two `HOST_WIDE_INT' pieces in *LV and *HV. */ |
133 | |
134 | static int |
135 | neg_double (unsigned HOST_WIDE_INTlong l1, HOST_WIDE_INTlong h1, |
136 | unsigned HOST_WIDE_INTlong *lv, HOST_WIDE_INTlong *hv) |
137 | { |
138 | if (l1 == 0) |
139 | { |
140 | *lv = 0; |
141 | *hv = - (unsigned HOST_WIDE_INTlong) h1; |
142 | return (*hv & h1) < 0; |
143 | } |
144 | else |
145 | { |
146 | *lv = -l1; |
147 | *hv = ~h1; |
148 | return 0; |
149 | } |
150 | } |
151 | |
152 | /* Multiply two doubleword integers with quadword result. |
153 | Return nonzero if the operation overflows according to UNSIGNED_P. |
154 | Each argument is given as two `HOST_WIDE_INT' pieces. |
155 | One argument is L1 and H1; the other, L2 and H2. |
156 | The value is stored as four `HOST_WIDE_INT' pieces in *LV and *HV, |
157 | *LW and *HW. |
158 | If lw is NULL then only the low part and no overflow is computed. */ |
159 | |
160 | static int |
161 | mul_double_wide_with_sign (unsigned HOST_WIDE_INTlong l1, HOST_WIDE_INTlong h1, |
162 | unsigned HOST_WIDE_INTlong l2, HOST_WIDE_INTlong h2, |
163 | unsigned HOST_WIDE_INTlong *lv, HOST_WIDE_INTlong *hv, |
164 | unsigned HOST_WIDE_INTlong *lw, HOST_WIDE_INTlong *hw, |
165 | bool unsigned_p) |
166 | { |
167 | HOST_WIDE_INTlong arg1[4]; |
168 | HOST_WIDE_INTlong arg2[4]; |
169 | HOST_WIDE_INTlong prod[4 * 2]; |
170 | unsigned HOST_WIDE_INTlong carry; |
171 | int i, j, k; |
172 | unsigned HOST_WIDE_INTlong neglow; |
173 | HOST_WIDE_INTlong neghigh; |
174 | |
175 | encode (arg1, l1, h1); |
176 | encode (arg2, l2, h2); |
177 | |
178 | memset (prod, 0, sizeof prod); |
179 | |
180 | for (i = 0; i < 4; i++) |
181 | { |
182 | carry = 0; |
183 | for (j = 0; j < 4; j++) |
184 | { |
185 | k = i + j; |
186 | /* This product is <= 0xFFFE0001, the sum <= 0xFFFF0000. */ |
187 | carry += (unsigned HOST_WIDE_INTlong) arg1[i] * arg2[j]; |
188 | /* Since prod[p] < 0xFFFF, this sum <= 0xFFFFFFFF. */ |
189 | carry += prod[k]; |
190 | prod[k] = LOWPART (carry)((carry) & ((1UL << (64 / 2)) - 1)); |
191 | carry = HIGHPART (carry)((unsigned long) (carry) >> 64 / 2); |
192 | } |
193 | prod[i + 4] = carry; |
194 | } |
195 | |
196 | decode (prod, lv, hv); |
197 | |
198 | /* We are not interested in the wide part nor in overflow. */ |
199 | if (lw == NULLnullptr) |
200 | return 0; |
201 | |
202 | decode (prod + 4, lw, hw); |
203 | |
204 | /* Unsigned overflow is immediate. */ |
205 | if (unsigned_p) |
206 | return (*lw | *hw) != 0; |
207 | |
208 | /* Check for signed overflow by calculating the signed representation of the |
209 | top half of the result; it should agree with the low half's sign bit. */ |
210 | if (h1 < 0) |
211 | { |
212 | neg_double (l2, h2, &neglow, &neghigh); |
213 | add_double (neglow, neghigh, *lw, *hw, lw, hw)add_double_with_sign (neglow, neghigh, *lw, *hw, lw, hw, false ); |
214 | } |
215 | if (h2 < 0) |
216 | { |
217 | neg_double (l1, h1, &neglow, &neghigh); |
218 | add_double (neglow, neghigh, *lw, *hw, lw, hw)add_double_with_sign (neglow, neghigh, *lw, *hw, lw, hw, false ); |
219 | } |
220 | return (*hv < 0 ? ~(*lw & *hw) : *lw | *hw) != 0; |
221 | } |
222 | |
223 | /* Shift the doubleword integer in L1, H1 right by COUNT places |
224 | keeping only PREC bits of result. ARITH nonzero specifies |
225 | arithmetic shifting; otherwise use logical shift. |
226 | Store the value as two `HOST_WIDE_INT' pieces in *LV and *HV. */ |
227 | |
228 | static void |
229 | rshift_double (unsigned HOST_WIDE_INTlong l1, HOST_WIDE_INTlong h1, |
230 | unsigned HOST_WIDE_INTlong count, unsigned int prec, |
231 | unsigned HOST_WIDE_INTlong *lv, HOST_WIDE_INTlong *hv, |
232 | bool arith) |
233 | { |
234 | unsigned HOST_WIDE_INTlong signmask; |
235 | |
236 | signmask = (arith |
237 | ? -((unsigned HOST_WIDE_INTlong) h1 >> (HOST_BITS_PER_WIDE_INT64 - 1)) |
238 | : 0); |
239 | |
240 | if (count >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
241 | { |
242 | /* Shifting by the host word size is undefined according to the |
243 | ANSI standard, so we must handle this as a special case. */ |
244 | *hv = 0; |
245 | *lv = 0; |
246 | } |
247 | else if (count >= HOST_BITS_PER_WIDE_INT64) |
248 | { |
249 | *hv = 0; |
250 | *lv = (unsigned HOST_WIDE_INTlong) h1 >> (count - HOST_BITS_PER_WIDE_INT64); |
251 | } |
252 | else |
253 | { |
254 | *hv = (unsigned HOST_WIDE_INTlong) h1 >> count; |
255 | *lv = ((l1 >> count) |
256 | | ((unsigned HOST_WIDE_INTlong) h1 |
257 | << (HOST_BITS_PER_WIDE_INT64 - count - 1) << 1)); |
258 | } |
259 | |
260 | /* Zero / sign extend all bits that are beyond the precision. */ |
261 | |
262 | if (count >= prec) |
263 | { |
264 | *hv = signmask; |
265 | *lv = signmask; |
266 | } |
267 | else if ((prec - count) >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
268 | ; |
269 | else if ((prec - count) >= HOST_BITS_PER_WIDE_INT64) |
270 | { |
271 | *hv &= ~(HOST_WIDE_INT_M1U-1UL << (prec - count - HOST_BITS_PER_WIDE_INT64)); |
272 | *hv |= signmask << (prec - count - HOST_BITS_PER_WIDE_INT64); |
273 | } |
274 | else |
275 | { |
276 | *hv = signmask; |
277 | *lv &= ~(HOST_WIDE_INT_M1U-1UL << (prec - count)); |
278 | *lv |= signmask << (prec - count); |
279 | } |
280 | } |
281 | |
282 | /* Shift the doubleword integer in L1, H1 left by COUNT places |
283 | keeping only PREC bits of result. |
284 | Shift right if COUNT is negative. |
285 | ARITH nonzero specifies arithmetic shifting; otherwise use logical shift. |
286 | Store the value as two `HOST_WIDE_INT' pieces in *LV and *HV. */ |
287 | |
288 | static void |
289 | lshift_double (unsigned HOST_WIDE_INTlong l1, HOST_WIDE_INTlong h1, |
290 | unsigned HOST_WIDE_INTlong count, unsigned int prec, |
291 | unsigned HOST_WIDE_INTlong *lv, HOST_WIDE_INTlong *hv) |
292 | { |
293 | unsigned HOST_WIDE_INTlong signmask; |
294 | |
295 | if (count >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
296 | { |
297 | /* Shifting by the host word size is undefined according to the |
298 | ANSI standard, so we must handle this as a special case. */ |
299 | *hv = 0; |
300 | *lv = 0; |
301 | } |
302 | else if (count >= HOST_BITS_PER_WIDE_INT64) |
303 | { |
304 | *hv = l1 << (count - HOST_BITS_PER_WIDE_INT64); |
305 | *lv = 0; |
306 | } |
307 | else |
308 | { |
309 | *hv = (((unsigned HOST_WIDE_INTlong) h1 << count) |
310 | | (l1 >> (HOST_BITS_PER_WIDE_INT64 - count - 1) >> 1)); |
311 | *lv = l1 << count; |
312 | } |
313 | |
314 | /* Sign extend all bits that are beyond the precision. */ |
315 | |
316 | signmask = -((prec > HOST_BITS_PER_WIDE_INT64 |
317 | ? ((unsigned HOST_WIDE_INTlong) *hv |
318 | >> (prec - HOST_BITS_PER_WIDE_INT64 - 1)) |
319 | : (*lv >> (prec - 1))) & 1); |
320 | |
321 | if (prec >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
322 | ; |
323 | else if (prec >= HOST_BITS_PER_WIDE_INT64) |
324 | { |
325 | *hv &= ~(HOST_WIDE_INT_M1U-1UL << (prec - HOST_BITS_PER_WIDE_INT64)); |
326 | *hv |= signmask << (prec - HOST_BITS_PER_WIDE_INT64); |
327 | } |
328 | else |
329 | { |
330 | *hv = signmask; |
331 | *lv &= ~(HOST_WIDE_INT_M1U-1UL << prec); |
332 | *lv |= signmask << prec; |
333 | } |
334 | } |
335 | |
336 | /* Divide doubleword integer LNUM, HNUM by doubleword integer LDEN, HDEN |
337 | for a quotient (stored in *LQUO, *HQUO) and remainder (in *LREM, *HREM). |
338 | CODE is a tree code for a kind of division, one of |
339 | TRUNC_DIV_EXPR, FLOOR_DIV_EXPR, CEIL_DIV_EXPR, ROUND_DIV_EXPR |
340 | or EXACT_DIV_EXPR |
341 | It controls how the quotient is rounded to an integer. |
342 | Return nonzero if the operation overflows. |
343 | UNS nonzero says do unsigned division. */ |
344 | |
345 | static int |
346 | div_and_round_double (unsigned code, int uns, |
347 | /* num == numerator == dividend */ |
348 | unsigned HOST_WIDE_INTlong lnum_orig, |
349 | HOST_WIDE_INTlong hnum_orig, |
350 | /* den == denominator == divisor */ |
351 | unsigned HOST_WIDE_INTlong lden_orig, |
352 | HOST_WIDE_INTlong hden_orig, |
353 | unsigned HOST_WIDE_INTlong *lquo, |
354 | HOST_WIDE_INTlong *hquo, unsigned HOST_WIDE_INTlong *lrem, |
355 | HOST_WIDE_INTlong *hrem) |
356 | { |
357 | int quo_neg = 0; |
358 | HOST_WIDE_INTlong num[4 + 1]; /* extra element for scaling. */ |
359 | HOST_WIDE_INTlong den[4], quo[4]; |
360 | int i, j; |
361 | unsigned HOST_WIDE_INTlong work; |
362 | unsigned HOST_WIDE_INTlong carry = 0; |
363 | unsigned HOST_WIDE_INTlong lnum = lnum_orig; |
364 | HOST_WIDE_INTlong hnum = hnum_orig; |
365 | unsigned HOST_WIDE_INTlong lden = lden_orig; |
366 | HOST_WIDE_INTlong hden = hden_orig; |
367 | int overflow = 0; |
368 | |
369 | if (hden == 0 && lden == 0) |
370 | overflow = 1, lden = 1; |
371 | |
372 | /* Calculate quotient sign and convert operands to unsigned. */ |
373 | if (!uns) |
374 | { |
375 | if (hnum < 0) |
376 | { |
377 | quo_neg = ~ quo_neg; |
378 | /* (minimum integer) / (-1) is the only overflow case. */ |
379 | if (neg_double (lnum, hnum, &lnum, &hnum) |
380 | && ((HOST_WIDE_INTlong) lden & hden) == -1) |
381 | overflow = 1; |
382 | } |
383 | if (hden < 0) |
384 | { |
385 | quo_neg = ~ quo_neg; |
386 | neg_double (lden, hden, &lden, &hden); |
387 | } |
388 | } |
389 | |
390 | if (hnum == 0 && hden == 0) |
391 | { /* single precision */ |
392 | *hquo = *hrem = 0; |
393 | /* This unsigned division rounds toward zero. */ |
394 | *lquo = lnum / lden; |
395 | goto finish_up; |
396 | } |
397 | |
398 | if (hnum == 0) |
399 | { /* trivial case: dividend < divisor */ |
400 | /* hden != 0 already checked. */ |
401 | *hquo = *lquo = 0; |
402 | *hrem = hnum; |
403 | *lrem = lnum; |
404 | goto finish_up; |
405 | } |
406 | |
407 | memset (quo, 0, sizeof quo); |
408 | |
409 | memset (num, 0, sizeof num); /* to zero 9th element */ |
410 | memset (den, 0, sizeof den); |
411 | |
412 | encode (num, lnum, hnum); |
413 | encode (den, lden, hden); |
414 | |
415 | /* Special code for when the divisor < BASE. */ |
416 | if (hden == 0 && lden < (unsigned HOST_WIDE_INTlong) BASE(1UL << 64 / 2)) |
417 | { |
418 | /* hnum != 0 already checked. */ |
419 | for (i = 4 - 1; i >= 0; i--) |
420 | { |
421 | work = num[i] + carry * BASE(1UL << 64 / 2); |
422 | quo[i] = work / lden; |
423 | carry = work % lden; |
424 | } |
425 | } |
426 | else |
427 | { |
428 | /* Full double precision division, |
429 | with thanks to Don Knuth's "Seminumerical Algorithms". */ |
430 | int num_hi_sig, den_hi_sig; |
431 | unsigned HOST_WIDE_INTlong quo_est, scale; |
432 | |
433 | /* Find the highest nonzero divisor digit. */ |
434 | for (i = 4 - 1;; i--) |
435 | if (den[i] != 0) |
436 | { |
437 | den_hi_sig = i; |
438 | break; |
439 | } |
440 | |
441 | /* Insure that the first digit of the divisor is at least BASE/2. |
442 | This is required by the quotient digit estimation algorithm. */ |
443 | |
444 | scale = BASE(1UL << 64 / 2) / (den[den_hi_sig] + 1); |
445 | if (scale > 1) |
446 | { /* scale divisor and dividend */ |
447 | carry = 0; |
448 | for (i = 0; i <= 4 - 1; i++) |
449 | { |
450 | work = (num[i] * scale) + carry; |
451 | num[i] = LOWPART (work)((work) & ((1UL << (64 / 2)) - 1)); |
452 | carry = HIGHPART (work)((unsigned long) (work) >> 64 / 2); |
453 | } |
454 | |
455 | num[4] = carry; |
456 | carry = 0; |
457 | for (i = 0; i <= 4 - 1; i++) |
458 | { |
459 | work = (den[i] * scale) + carry; |
460 | den[i] = LOWPART (work)((work) & ((1UL << (64 / 2)) - 1)); |
461 | carry = HIGHPART (work)((unsigned long) (work) >> 64 / 2); |
462 | if (den[i] != 0) den_hi_sig = i; |
463 | } |
464 | } |
465 | |
466 | num_hi_sig = 4; |
467 | |
468 | /* Main loop */ |
469 | for (i = num_hi_sig - den_hi_sig - 1; i >= 0; i--) |
470 | { |
471 | /* Guess the next quotient digit, quo_est, by dividing the first |
472 | two remaining dividend digits by the high order quotient digit. |
473 | quo_est is never low and is at most 2 high. */ |
474 | unsigned HOST_WIDE_INTlong tmp; |
475 | |
476 | num_hi_sig = i + den_hi_sig + 1; |
477 | work = num[num_hi_sig] * BASE(1UL << 64 / 2) + num[num_hi_sig - 1]; |
478 | if (num[num_hi_sig] != den[den_hi_sig]) |
479 | quo_est = work / den[den_hi_sig]; |
480 | else |
481 | quo_est = BASE(1UL << 64 / 2) - 1; |
482 | |
483 | /* Refine quo_est so it's usually correct, and at most one high. */ |
484 | tmp = work - quo_est * den[den_hi_sig]; |
485 | if (tmp < BASE(1UL << 64 / 2) |
486 | && (den[den_hi_sig - 1] * quo_est |
487 | > (tmp * BASE(1UL << 64 / 2) + num[num_hi_sig - 2]))) |
488 | quo_est--; |
489 | |
490 | /* Try QUO_EST as the quotient digit, by multiplying the |
491 | divisor by QUO_EST and subtracting from the remaining dividend. |
492 | Keep in mind that QUO_EST is the I - 1st digit. */ |
493 | |
494 | carry = 0; |
495 | for (j = 0; j <= den_hi_sig; j++) |
496 | { |
497 | work = quo_est * den[j] + carry; |
498 | carry = HIGHPART (work)((unsigned long) (work) >> 64 / 2); |
499 | work = num[i + j] - LOWPART (work)((work) & ((1UL << (64 / 2)) - 1)); |
500 | num[i + j] = LOWPART (work)((work) & ((1UL << (64 / 2)) - 1)); |
501 | carry += HIGHPART (work)((unsigned long) (work) >> 64 / 2) != 0; |
502 | } |
503 | |
504 | /* If quo_est was high by one, then num[i] went negative and |
505 | we need to correct things. */ |
506 | if (num[num_hi_sig] < (HOST_WIDE_INTlong) carry) |
507 | { |
508 | quo_est--; |
509 | carry = 0; /* add divisor back in */ |
510 | for (j = 0; j <= den_hi_sig; j++) |
511 | { |
512 | work = num[i + j] + den[j] + carry; |
513 | carry = HIGHPART (work)((unsigned long) (work) >> 64 / 2); |
514 | num[i + j] = LOWPART (work)((work) & ((1UL << (64 / 2)) - 1)); |
515 | } |
516 | |
517 | num [num_hi_sig] += carry; |
518 | } |
519 | |
520 | /* Store the quotient digit. */ |
521 | quo[i] = quo_est; |
522 | } |
523 | } |
524 | |
525 | decode (quo, lquo, hquo); |
526 | |
527 | finish_up: |
528 | /* If result is negative, make it so. */ |
529 | if (quo_neg) |
530 | neg_double (*lquo, *hquo, lquo, hquo); |
531 | |
532 | /* Compute trial remainder: rem = num - (quo * den) */ |
533 | mul_double (*lquo, *hquo, lden_orig, hden_orig, lrem, hrem)mul_double_wide_with_sign (*lquo, *hquo, lden_orig, hden_orig , lrem, hrem, nullptr, nullptr, false); |
534 | neg_double (*lrem, *hrem, lrem, hrem); |
535 | add_double (lnum_orig, hnum_orig, *lrem, *hrem, lrem, hrem)add_double_with_sign (lnum_orig, hnum_orig, *lrem, *hrem, lrem , hrem, false); |
536 | |
537 | switch (code) |
538 | { |
539 | case TRUNC_DIV_EXPR: |
540 | case TRUNC_MOD_EXPR: /* round toward zero */ |
541 | case EXACT_DIV_EXPR: /* for this one, it shouldn't matter */ |
542 | return overflow; |
543 | |
544 | case FLOOR_DIV_EXPR: |
545 | case FLOOR_MOD_EXPR: /* round toward negative infinity */ |
546 | if (quo_neg && (*lrem != 0 || *hrem != 0)) /* ratio < 0 && rem != 0 */ |
547 | { |
548 | /* quo = quo - 1; */ |
549 | add_double (*lquo, *hquo, HOST_WIDE_INT_M1, HOST_WIDE_INT_M1,add_double_with_sign (*lquo, *hquo, -1L, -1L, lquo, hquo, false ) |
550 | lquo, hquo)add_double_with_sign (*lquo, *hquo, -1L, -1L, lquo, hquo, false ); |
551 | } |
552 | else |
553 | return overflow; |
554 | break; |
555 | |
556 | case CEIL_DIV_EXPR: |
557 | case CEIL_MOD_EXPR: /* round toward positive infinity */ |
558 | if (!quo_neg && (*lrem != 0 || *hrem != 0)) /* ratio > 0 && rem != 0 */ |
559 | { |
560 | add_double (*lquo, *hquo, HOST_WIDE_INT_1, HOST_WIDE_INT_0,add_double_with_sign (*lquo, *hquo, 1L, 0L, lquo, hquo, false ) |
561 | lquo, hquo)add_double_with_sign (*lquo, *hquo, 1L, 0L, lquo, hquo, false ); |
562 | } |
563 | else |
564 | return overflow; |
565 | break; |
566 | |
567 | case ROUND_DIV_EXPR: |
568 | case ROUND_MOD_EXPR: /* round to closest integer */ |
569 | { |
570 | unsigned HOST_WIDE_INTlong labs_rem = *lrem; |
571 | HOST_WIDE_INTlong habs_rem = *hrem; |
572 | unsigned HOST_WIDE_INTlong labs_den = lden, lnegabs_rem, ldiff; |
573 | HOST_WIDE_INTlong habs_den = hden, hnegabs_rem, hdiff; |
574 | |
575 | /* Get absolute values. */ |
576 | if (!uns && *hrem < 0) |
577 | neg_double (*lrem, *hrem, &labs_rem, &habs_rem); |
578 | if (!uns && hden < 0) |
579 | neg_double (lden, hden, &labs_den, &habs_den); |
580 | |
581 | /* If abs(rem) >= abs(den) - abs(rem), adjust the quotient. */ |
582 | neg_double (labs_rem, habs_rem, &lnegabs_rem, &hnegabs_rem); |
583 | add_double (labs_den, habs_den, lnegabs_rem, hnegabs_rem,add_double_with_sign (labs_den, habs_den, lnegabs_rem, hnegabs_rem , &ldiff, &hdiff, false) |
584 | &ldiff, &hdiff)add_double_with_sign (labs_den, habs_den, lnegabs_rem, hnegabs_rem , &ldiff, &hdiff, false); |
585 | |
586 | if (((unsigned HOST_WIDE_INTlong) habs_rem |
587 | > (unsigned HOST_WIDE_INTlong) hdiff) |
588 | || (habs_rem == hdiff && labs_rem >= ldiff)) |
589 | { |
590 | if (quo_neg) |
591 | /* quo = quo - 1; */ |
592 | add_double (*lquo, *hquo,add_double_with_sign (*lquo, *hquo, -1L, -1L, lquo, hquo, false ) |
593 | HOST_WIDE_INT_M1, HOST_WIDE_INT_M1, lquo, hquo)add_double_with_sign (*lquo, *hquo, -1L, -1L, lquo, hquo, false ); |
594 | else |
595 | /* quo = quo + 1; */ |
596 | add_double (*lquo, *hquo, HOST_WIDE_INT_1, HOST_WIDE_INT_0,add_double_with_sign (*lquo, *hquo, 1L, 0L, lquo, hquo, false ) |
597 | lquo, hquo)add_double_with_sign (*lquo, *hquo, 1L, 0L, lquo, hquo, false ); |
598 | } |
599 | else |
600 | return overflow; |
601 | } |
602 | break; |
603 | |
604 | default: |
605 | gcc_unreachable ()(fancy_abort ("/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 605, __FUNCTION__)); |
606 | } |
607 | |
608 | /* Compute true remainder: rem = num - (quo * den) */ |
609 | mul_double (*lquo, *hquo, lden_orig, hden_orig, lrem, hrem)mul_double_wide_with_sign (*lquo, *hquo, lden_orig, hden_orig , lrem, hrem, nullptr, nullptr, false); |
610 | neg_double (*lrem, *hrem, lrem, hrem); |
611 | add_double (lnum_orig, hnum_orig, *lrem, *hrem, lrem, hrem)add_double_with_sign (lnum_orig, hnum_orig, *lrem, *hrem, lrem , hrem, false); |
612 | return overflow; |
613 | } |
614 | |
615 | |
616 | /* Construct from a buffer of length LEN. BUFFER will be read according |
617 | to byte endianness and word endianness. Only the lower LEN bytes |
618 | of the result are set; the remaining high bytes are cleared. */ |
619 | |
620 | double_int |
621 | double_int::from_buffer (const unsigned char *buffer, int len) |
622 | { |
623 | double_int result = double_int_zero(double_int::from_shwi (0)); |
624 | int words = len / UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4); |
Value stored to 'words' during its initialization is never read | |
625 | |
626 | gcc_assert (len * BITS_PER_UNIT <= HOST_BITS_PER_DOUBLE_INT)((void)(!(len * (8) <= (2 * 64)) ? fancy_abort ("/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 626, __FUNCTION__), 0 : 0)); |
627 | |
628 | for (int byte = 0; byte < len; byte++) |
629 | { |
630 | int offset; |
631 | int bitpos = byte * BITS_PER_UNIT(8); |
632 | unsigned HOST_WIDE_INTlong value; |
633 | |
634 | if (len > UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4)) |
635 | { |
636 | int word = byte / UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4); |
637 | |
638 | if (WORDS_BIG_ENDIAN0) |
639 | word = (words - 1) - word; |
640 | |
641 | offset = word * UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4); |
642 | |
643 | if (BYTES_BIG_ENDIAN0) |
644 | offset += (UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4) - 1) - (byte % UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4)); |
645 | else |
646 | offset += byte % UNITS_PER_WORD(((global_options.x_ix86_isa_flags & (1UL << 1)) != 0) ? 8 : 4); |
647 | } |
648 | else |
649 | offset = BYTES_BIG_ENDIAN0 ? (len - 1) - byte : byte; |
650 | |
651 | value = (unsigned HOST_WIDE_INTlong) buffer[offset]; |
652 | |
653 | if (bitpos < HOST_BITS_PER_WIDE_INT64) |
654 | result.low |= value << bitpos; |
655 | else |
656 | result.high |= value << (bitpos - HOST_BITS_PER_WIDE_INT64); |
657 | } |
658 | |
659 | return result; |
660 | } |
661 | |
662 | |
663 | /* Returns mask for PREC bits. */ |
664 | |
665 | double_int |
666 | double_int::mask (unsigned prec) |
667 | { |
668 | unsigned HOST_WIDE_INTlong m; |
669 | double_int mask; |
670 | |
671 | if (prec > HOST_BITS_PER_WIDE_INT64) |
672 | { |
673 | prec -= HOST_BITS_PER_WIDE_INT64; |
674 | m = ((unsigned HOST_WIDE_INTlong) 2 << (prec - 1)) - 1; |
675 | mask.high = (HOST_WIDE_INTlong) m; |
676 | mask.low = ALL_ONES-1UL; |
677 | } |
678 | else |
679 | { |
680 | mask.high = 0; |
681 | mask.low = prec ? ((unsigned HOST_WIDE_INTlong) 2 << (prec - 1)) - 1 : 0; |
682 | } |
683 | |
684 | return mask; |
685 | } |
686 | |
687 | /* Returns a maximum value for signed or unsigned integer |
688 | of precision PREC. */ |
689 | |
690 | double_int |
691 | double_int::max_value (unsigned int prec, bool uns) |
692 | { |
693 | return double_int::mask (prec - (uns ? 0 : 1)); |
694 | } |
695 | |
696 | /* Returns a minimum value for signed or unsigned integer |
697 | of precision PREC. */ |
698 | |
699 | double_int |
700 | double_int::min_value (unsigned int prec, bool uns) |
701 | { |
702 | if (uns) |
703 | return double_int_zero(double_int::from_shwi (0)); |
704 | return double_int_one(double_int::from_shwi (1)).lshift (prec - 1, prec, false); |
705 | } |
706 | |
707 | /* Clears the bits of CST over the precision PREC. If UNS is false, the bits |
708 | outside of the precision are set to the sign bit (i.e., the PREC-th one), |
709 | otherwise they are set to zero. |
710 | |
711 | This corresponds to returning the value represented by PREC lowermost bits |
712 | of CST, with the given signedness. */ |
713 | |
714 | double_int |
715 | double_int::ext (unsigned prec, bool uns) const |
716 | { |
717 | if (uns) |
718 | return this->zext (prec); |
719 | else |
720 | return this->sext (prec); |
721 | } |
722 | |
723 | /* The same as double_int::ext with UNS = true. */ |
724 | |
725 | double_int |
726 | double_int::zext (unsigned prec) const |
727 | { |
728 | const double_int &cst = *this; |
729 | double_int mask = double_int::mask (prec); |
730 | double_int r; |
731 | |
732 | r.low = cst.low & mask.low; |
733 | r.high = cst.high & mask.high; |
734 | |
735 | return r; |
736 | } |
737 | |
738 | /* The same as double_int::ext with UNS = false. */ |
739 | |
740 | double_int |
741 | double_int::sext (unsigned prec) const |
742 | { |
743 | const double_int &cst = *this; |
744 | double_int mask = double_int::mask (prec); |
745 | double_int r; |
746 | unsigned HOST_WIDE_INTlong snum; |
747 | |
748 | if (prec <= HOST_BITS_PER_WIDE_INT64) |
749 | snum = cst.low; |
750 | else |
751 | { |
752 | prec -= HOST_BITS_PER_WIDE_INT64; |
753 | snum = (unsigned HOST_WIDE_INTlong) cst.high; |
754 | } |
755 | if (((snum >> (prec - 1)) & 1) == 1) |
756 | { |
757 | r.low = cst.low | ~mask.low; |
758 | r.high = cst.high | ~mask.high; |
759 | } |
760 | else |
761 | { |
762 | r.low = cst.low & mask.low; |
763 | r.high = cst.high & mask.high; |
764 | } |
765 | |
766 | return r; |
767 | } |
768 | |
769 | /* Returns true if CST fits in signed HOST_WIDE_INT. */ |
770 | |
771 | bool |
772 | double_int::fits_shwi () const |
773 | { |
774 | const double_int &cst = *this; |
775 | if (cst.high == 0) |
776 | return (HOST_WIDE_INTlong) cst.low >= 0; |
777 | else if (cst.high == -1) |
778 | return (HOST_WIDE_INTlong) cst.low < 0; |
779 | else |
780 | return false; |
781 | } |
782 | |
783 | /* Returns true if CST fits in HOST_WIDE_INT if UNS is false, or in |
784 | unsigned HOST_WIDE_INT if UNS is true. */ |
785 | |
786 | bool |
787 | double_int::fits_hwi (bool uns) const |
788 | { |
789 | if (uns) |
790 | return this->fits_uhwi (); |
791 | else |
792 | return this->fits_shwi (); |
793 | } |
794 | |
795 | /* Returns A * B. */ |
796 | |
797 | double_int |
798 | double_int::operator * (double_int b) const |
799 | { |
800 | const double_int &a = *this; |
801 | double_int ret; |
802 | mul_double (a.low, a.high, b.low, b.high, &ret.low, &ret.high)mul_double_wide_with_sign (a.low, a.high, b.low, b.high, & ret.low, &ret.high, nullptr, nullptr, false); |
803 | return ret; |
804 | } |
805 | |
806 | /* Multiplies *this with B and returns a reference to *this. */ |
807 | |
808 | double_int & |
809 | double_int::operator *= (double_int b) |
810 | { |
811 | mul_double (low, high, b.low, b.high, &low, &high)mul_double_wide_with_sign (low, high, b.low, b.high, &low , &high, nullptr, nullptr, false); |
812 | return *this; |
813 | } |
814 | |
815 | /* Returns A * B. If the operation overflows according to UNSIGNED_P, |
816 | *OVERFLOW is set to nonzero. */ |
817 | |
818 | double_int |
819 | double_int::mul_with_sign (double_int b, bool unsigned_p, bool *overflow) const |
820 | { |
821 | const double_int &a = *this; |
822 | double_int ret, tem; |
823 | *overflow = mul_double_wide_with_sign (a.low, a.high, b.low, b.high, |
824 | &ret.low, &ret.high, |
825 | &tem.low, &tem.high, unsigned_p); |
826 | return ret; |
827 | } |
828 | |
829 | double_int |
830 | double_int::wide_mul_with_sign (double_int b, bool unsigned_p, |
831 | double_int *higher, bool *overflow) const |
832 | |
833 | { |
834 | double_int lower; |
835 | *overflow = mul_double_wide_with_sign (low, high, b.low, b.high, |
836 | &lower.low, &lower.high, |
837 | &higher->low, &higher->high, |
838 | unsigned_p); |
839 | return lower; |
840 | } |
841 | |
842 | /* Returns A + B. */ |
843 | |
844 | double_int |
845 | double_int::operator + (double_int b) const |
846 | { |
847 | const double_int &a = *this; |
848 | double_int ret; |
849 | add_double (a.low, a.high, b.low, b.high, &ret.low, &ret.high)add_double_with_sign (a.low, a.high, b.low, b.high, &ret. low, &ret.high, false); |
850 | return ret; |
851 | } |
852 | |
853 | /* Adds B to *this and returns a reference to *this. */ |
854 | |
855 | double_int & |
856 | double_int::operator += (double_int b) |
857 | { |
858 | add_double (low, high, b.low, b.high, &low, &high)add_double_with_sign (low, high, b.low, b.high, &low, & high, false); |
859 | return *this; |
860 | } |
861 | |
862 | |
863 | /* Returns A + B. If the operation overflows according to UNSIGNED_P, |
864 | *OVERFLOW is set to nonzero. */ |
865 | |
866 | double_int |
867 | double_int::add_with_sign (double_int b, bool unsigned_p, bool *overflow) const |
868 | { |
869 | const double_int &a = *this; |
870 | double_int ret; |
871 | *overflow = add_double_with_sign (a.low, a.high, b.low, b.high, |
872 | &ret.low, &ret.high, unsigned_p); |
873 | return ret; |
874 | } |
875 | |
876 | /* Returns A - B. */ |
877 | |
878 | double_int |
879 | double_int::operator - (double_int b) const |
880 | { |
881 | const double_int &a = *this; |
882 | double_int ret; |
883 | neg_double (b.low, b.high, &b.low, &b.high); |
884 | add_double (a.low, a.high, b.low, b.high, &ret.low, &ret.high)add_double_with_sign (a.low, a.high, b.low, b.high, &ret. low, &ret.high, false); |
885 | return ret; |
886 | } |
887 | |
888 | /* Subtracts B from *this and returns a reference to *this. */ |
889 | |
890 | double_int & |
891 | double_int::operator -= (double_int b) |
892 | { |
893 | neg_double (b.low, b.high, &b.low, &b.high); |
894 | add_double (low, high, b.low, b.high, &low, &high)add_double_with_sign (low, high, b.low, b.high, &low, & high, false); |
895 | return *this; |
896 | } |
897 | |
898 | |
899 | /* Returns A - B. If the operation overflows via inconsistent sign bits, |
900 | *OVERFLOW is set to nonzero. */ |
901 | |
902 | double_int |
903 | double_int::sub_with_overflow (double_int b, bool *overflow) const |
904 | { |
905 | double_int ret; |
906 | neg_double (b.low, b.high, &ret.low, &ret.high); |
907 | add_double (low, high, ret.low, ret.high, &ret.low, &ret.high)add_double_with_sign (low, high, ret.low, ret.high, &ret. low, &ret.high, false); |
908 | *overflow = OVERFLOW_SUM_SIGN (ret.high, b.high, high)((~((ret.high) ^ (b.high)) & ((ret.high) ^ (high))) < 0 ); |
909 | return ret; |
910 | } |
911 | |
912 | /* Returns -A. */ |
913 | |
914 | double_int |
915 | double_int::operator - () const |
916 | { |
917 | const double_int &a = *this; |
918 | double_int ret; |
919 | neg_double (a.low, a.high, &ret.low, &ret.high); |
920 | return ret; |
921 | } |
922 | |
923 | double_int |
924 | double_int::neg_with_overflow (bool *overflow) const |
925 | { |
926 | double_int ret; |
927 | *overflow = neg_double (low, high, &ret.low, &ret.high); |
928 | return ret; |
929 | } |
930 | |
931 | /* Returns A / B (computed as unsigned depending on UNS, and rounded as |
932 | specified by CODE). CODE is enum tree_code in fact, but double_int.h |
933 | must be included before tree.h. The remainder after the division is |
934 | stored to MOD. */ |
935 | |
936 | double_int |
937 | double_int::divmod_with_overflow (double_int b, bool uns, unsigned code, |
938 | double_int *mod, bool *overflow) const |
939 | { |
940 | const double_int &a = *this; |
941 | double_int ret; |
942 | |
943 | *overflow = div_and_round_double (code, uns, a.low, a.high, |
944 | b.low, b.high, &ret.low, &ret.high, |
945 | &mod->low, &mod->high); |
946 | return ret; |
947 | } |
948 | |
949 | double_int |
950 | double_int::divmod (double_int b, bool uns, unsigned code, |
951 | double_int *mod) const |
952 | { |
953 | const double_int &a = *this; |
954 | double_int ret; |
955 | |
956 | div_and_round_double (code, uns, a.low, a.high, |
957 | b.low, b.high, &ret.low, &ret.high, |
958 | &mod->low, &mod->high); |
959 | return ret; |
960 | } |
961 | |
962 | /* The same as double_int::divmod with UNS = false. */ |
963 | |
964 | double_int |
965 | double_int::sdivmod (double_int b, unsigned code, double_int *mod) const |
966 | { |
967 | return this->divmod (b, false, code, mod); |
968 | } |
969 | |
970 | /* The same as double_int::divmod with UNS = true. */ |
971 | |
972 | double_int |
973 | double_int::udivmod (double_int b, unsigned code, double_int *mod) const |
974 | { |
975 | return this->divmod (b, true, code, mod); |
976 | } |
977 | |
978 | /* Returns A / B (computed as unsigned depending on UNS, and rounded as |
979 | specified by CODE). CODE is enum tree_code in fact, but double_int.h |
980 | must be included before tree.h. */ |
981 | |
982 | double_int |
983 | double_int::div (double_int b, bool uns, unsigned code) const |
984 | { |
985 | double_int mod; |
986 | |
987 | return this->divmod (b, uns, code, &mod); |
988 | } |
989 | |
990 | /* The same as double_int::div with UNS = false. */ |
991 | |
992 | double_int |
993 | double_int::sdiv (double_int b, unsigned code) const |
994 | { |
995 | return this->div (b, false, code); |
996 | } |
997 | |
998 | /* The same as double_int::div with UNS = true. */ |
999 | |
1000 | double_int |
1001 | double_int::udiv (double_int b, unsigned code) const |
1002 | { |
1003 | return this->div (b, true, code); |
1004 | } |
1005 | |
1006 | /* Returns A % B (computed as unsigned depending on UNS, and rounded as |
1007 | specified by CODE). CODE is enum tree_code in fact, but double_int.h |
1008 | must be included before tree.h. */ |
1009 | |
1010 | double_int |
1011 | double_int::mod (double_int b, bool uns, unsigned code) const |
1012 | { |
1013 | double_int mod; |
1014 | |
1015 | this->divmod (b, uns, code, &mod); |
1016 | return mod; |
1017 | } |
1018 | |
1019 | /* The same as double_int::mod with UNS = false. */ |
1020 | |
1021 | double_int |
1022 | double_int::smod (double_int b, unsigned code) const |
1023 | { |
1024 | return this->mod (b, false, code); |
1025 | } |
1026 | |
1027 | /* The same as double_int::mod with UNS = true. */ |
1028 | |
1029 | double_int |
1030 | double_int::umod (double_int b, unsigned code) const |
1031 | { |
1032 | return this->mod (b, true, code); |
1033 | } |
1034 | |
1035 | /* Return TRUE iff PRODUCT is an integral multiple of FACTOR, and return |
1036 | the multiple in *MULTIPLE. Otherwise return FALSE and leave *MULTIPLE |
1037 | unchanged. */ |
1038 | |
1039 | bool |
1040 | double_int::multiple_of (double_int factor, |
1041 | bool unsigned_p, double_int *multiple) const |
1042 | { |
1043 | double_int remainder; |
1044 | double_int quotient = this->divmod (factor, unsigned_p, |
1045 | TRUNC_DIV_EXPR, &remainder); |
1046 | if (remainder.is_zero ()) |
1047 | { |
1048 | *multiple = quotient; |
1049 | return true; |
1050 | } |
1051 | |
1052 | return false; |
1053 | } |
1054 | |
1055 | /* Set BITPOS bit in A. */ |
1056 | double_int |
1057 | double_int::set_bit (unsigned bitpos) const |
1058 | { |
1059 | double_int a = *this; |
1060 | if (bitpos < HOST_BITS_PER_WIDE_INT64) |
1061 | a.low |= HOST_WIDE_INT_1U1UL << bitpos; |
1062 | else |
1063 | a.high |= HOST_WIDE_INT_11L << (bitpos - HOST_BITS_PER_WIDE_INT64); |
1064 | |
1065 | return a; |
1066 | } |
1067 | |
1068 | /* Count trailing zeros in A. */ |
1069 | int |
1070 | double_int::trailing_zeros () const |
1071 | { |
1072 | const double_int &a = *this; |
1073 | unsigned HOST_WIDE_INTlong w = a.low ? a.low : (unsigned HOST_WIDE_INTlong) a.high; |
1074 | unsigned bits = a.low ? 0 : HOST_BITS_PER_WIDE_INT64; |
1075 | if (!w) |
1076 | return HOST_BITS_PER_DOUBLE_INT(2 * 64); |
1077 | bits += ctz_hwi (w); |
1078 | return bits; |
1079 | } |
1080 | |
1081 | /* Shift A left by COUNT places. */ |
1082 | |
1083 | double_int |
1084 | double_int::lshift (HOST_WIDE_INTlong count) const |
1085 | { |
1086 | double_int ret; |
1087 | |
1088 | gcc_checking_assert (count >= 0)((void)(!(count >= 0) ? fancy_abort ("/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 1088, __FUNCTION__), 0 : 0)); |
1089 | |
1090 | if (count >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
1091 | { |
1092 | /* Shifting by the host word size is undefined according to the |
1093 | ANSI standard, so we must handle this as a special case. */ |
1094 | ret.high = 0; |
1095 | ret.low = 0; |
1096 | } |
1097 | else if (count >= HOST_BITS_PER_WIDE_INT64) |
1098 | { |
1099 | ret.high = low << (count - HOST_BITS_PER_WIDE_INT64); |
1100 | ret.low = 0; |
1101 | } |
1102 | else |
1103 | { |
1104 | ret.high = (((unsigned HOST_WIDE_INTlong) high << count) |
1105 | | (low >> (HOST_BITS_PER_WIDE_INT64 - count - 1) >> 1)); |
1106 | ret.low = low << count; |
1107 | } |
1108 | |
1109 | return ret; |
1110 | } |
1111 | |
1112 | /* Shift A right by COUNT places. */ |
1113 | |
1114 | double_int |
1115 | double_int::rshift (HOST_WIDE_INTlong count) const |
1116 | { |
1117 | double_int ret; |
1118 | |
1119 | gcc_checking_assert (count >= 0)((void)(!(count >= 0) ? fancy_abort ("/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 1119, __FUNCTION__), 0 : 0)); |
1120 | |
1121 | if (count >= HOST_BITS_PER_DOUBLE_INT(2 * 64)) |
1122 | { |
1123 | /* Shifting by the host word size is undefined according to the |
1124 | ANSI standard, so we must handle this as a special case. */ |
1125 | ret.high = 0; |
1126 | ret.low = 0; |
1127 | } |
1128 | else if (count >= HOST_BITS_PER_WIDE_INT64) |
1129 | { |
1130 | ret.high = 0; |
1131 | ret.low |
1132 | = (unsigned HOST_WIDE_INTlong) (high >> (count - HOST_BITS_PER_WIDE_INT64)); |
1133 | } |
1134 | else |
1135 | { |
1136 | ret.high = high >> count; |
1137 | ret.low = ((low >> count) |
1138 | | ((unsigned HOST_WIDE_INTlong) high |
1139 | << (HOST_BITS_PER_WIDE_INT64 - count - 1) << 1)); |
1140 | } |
1141 | |
1142 | return ret; |
1143 | } |
1144 | |
1145 | /* Shift A left by COUNT places keeping only PREC bits of result. Shift |
1146 | right if COUNT is negative. ARITH true specifies arithmetic shifting; |
1147 | otherwise use logical shift. */ |
1148 | |
1149 | double_int |
1150 | double_int::lshift (HOST_WIDE_INTlong count, unsigned int prec, bool arith) const |
1151 | { |
1152 | double_int ret; |
1153 | if (count > 0) |
1154 | lshift_double (low, high, count, prec, &ret.low, &ret.high); |
1155 | else |
1156 | rshift_double (low, high, absu_hwi (count), prec, &ret.low, &ret.high, arith); |
1157 | return ret; |
1158 | } |
1159 | |
1160 | /* Shift A right by COUNT places keeping only PREC bits of result. Shift |
1161 | left if COUNT is negative. ARITH true specifies arithmetic shifting; |
1162 | otherwise use logical shift. */ |
1163 | |
1164 | double_int |
1165 | double_int::rshift (HOST_WIDE_INTlong count, unsigned int prec, bool arith) const |
1166 | { |
1167 | double_int ret; |
1168 | if (count > 0) |
1169 | rshift_double (low, high, count, prec, &ret.low, &ret.high, arith); |
1170 | else |
1171 | lshift_double (low, high, absu_hwi (count), prec, &ret.low, &ret.high); |
1172 | return ret; |
1173 | } |
1174 | |
1175 | /* Arithmetic shift A left by COUNT places keeping only PREC bits of result. |
1176 | Shift right if COUNT is negative. */ |
1177 | |
1178 | double_int |
1179 | double_int::alshift (HOST_WIDE_INTlong count, unsigned int prec) const |
1180 | { |
1181 | double_int r; |
1182 | if (count > 0) |
1183 | lshift_double (low, high, count, prec, &r.low, &r.high); |
1184 | else |
1185 | rshift_double (low, high, absu_hwi (count), prec, &r.low, &r.high, true); |
1186 | return r; |
1187 | } |
1188 | |
1189 | /* Arithmetic shift A right by COUNT places keeping only PREC bits of result. |
1190 | Shift left if COUNT is negative. */ |
1191 | |
1192 | double_int |
1193 | double_int::arshift (HOST_WIDE_INTlong count, unsigned int prec) const |
1194 | { |
1195 | double_int r; |
1196 | if (count > 0) |
1197 | rshift_double (low, high, count, prec, &r.low, &r.high, true); |
1198 | else |
1199 | lshift_double (low, high, absu_hwi (count), prec, &r.low, &r.high); |
1200 | return r; |
1201 | } |
1202 | |
1203 | /* Logical shift A left by COUNT places keeping only PREC bits of result. |
1204 | Shift right if COUNT is negative. */ |
1205 | |
1206 | double_int |
1207 | double_int::llshift (HOST_WIDE_INTlong count, unsigned int prec) const |
1208 | { |
1209 | double_int r; |
1210 | if (count > 0) |
1211 | lshift_double (low, high, count, prec, &r.low, &r.high); |
1212 | else |
1213 | rshift_double (low, high, absu_hwi (count), prec, &r.low, &r.high, false); |
1214 | return r; |
1215 | } |
1216 | |
1217 | /* Logical shift A right by COUNT places keeping only PREC bits of result. |
1218 | Shift left if COUNT is negative. */ |
1219 | |
1220 | double_int |
1221 | double_int::lrshift (HOST_WIDE_INTlong count, unsigned int prec) const |
1222 | { |
1223 | double_int r; |
1224 | if (count > 0) |
1225 | rshift_double (low, high, count, prec, &r.low, &r.high, false); |
1226 | else |
1227 | lshift_double (low, high, absu_hwi (count), prec, &r.low, &r.high); |
1228 | return r; |
1229 | } |
1230 | |
1231 | /* Rotate A left by COUNT places keeping only PREC bits of result. |
1232 | Rotate right if COUNT is negative. */ |
1233 | |
1234 | double_int |
1235 | double_int::lrotate (HOST_WIDE_INTlong count, unsigned int prec) const |
1236 | { |
1237 | double_int t1, t2; |
1238 | |
1239 | count %= prec; |
1240 | if (count < 0) |
1241 | count += prec; |
1242 | |
1243 | t1 = this->llshift (count, prec); |
1244 | t2 = this->lrshift (prec - count, prec); |
1245 | |
1246 | return t1 | t2; |
1247 | } |
1248 | |
1249 | /* Rotate A rigth by COUNT places keeping only PREC bits of result. |
1250 | Rotate right if COUNT is negative. */ |
1251 | |
1252 | double_int |
1253 | double_int::rrotate (HOST_WIDE_INTlong count, unsigned int prec) const |
1254 | { |
1255 | double_int t1, t2; |
1256 | |
1257 | count %= prec; |
1258 | if (count < 0) |
1259 | count += prec; |
1260 | |
1261 | t1 = this->lrshift (count, prec); |
1262 | t2 = this->llshift (prec - count, prec); |
1263 | |
1264 | return t1 | t2; |
1265 | } |
1266 | |
1267 | /* Returns -1 if A < B, 0 if A == B and 1 if A > B. Signedness of the |
1268 | comparison is given by UNS. */ |
1269 | |
1270 | int |
1271 | double_int::cmp (double_int b, bool uns) const |
1272 | { |
1273 | if (uns) |
1274 | return this->ucmp (b); |
1275 | else |
1276 | return this->scmp (b); |
1277 | } |
1278 | |
1279 | /* Compares two unsigned values A and B. Returns -1 if A < B, 0 if A == B, |
1280 | and 1 if A > B. */ |
1281 | |
1282 | int |
1283 | double_int::ucmp (double_int b) const |
1284 | { |
1285 | const double_int &a = *this; |
1286 | if ((unsigned HOST_WIDE_INTlong) a.high < (unsigned HOST_WIDE_INTlong) b.high) |
1287 | return -1; |
1288 | if ((unsigned HOST_WIDE_INTlong) a.high > (unsigned HOST_WIDE_INTlong) b.high) |
1289 | return 1; |
1290 | if (a.low < b.low) |
1291 | return -1; |
1292 | if (a.low > b.low) |
1293 | return 1; |
1294 | |
1295 | return 0; |
1296 | } |
1297 | |
1298 | /* Compares two signed values A and B. Returns -1 if A < B, 0 if A == B, |
1299 | and 1 if A > B. */ |
1300 | |
1301 | int |
1302 | double_int::scmp (double_int b) const |
1303 | { |
1304 | const double_int &a = *this; |
1305 | if (a.high < b.high) |
1306 | return -1; |
1307 | if (a.high > b.high) |
1308 | return 1; |
1309 | if (a.low < b.low) |
1310 | return -1; |
1311 | if (a.low > b.low) |
1312 | return 1; |
1313 | |
1314 | return 0; |
1315 | } |
1316 | |
1317 | /* Compares two unsigned values A and B for less-than. */ |
1318 | |
1319 | bool |
1320 | double_int::ult (double_int b) const |
1321 | { |
1322 | if ((unsigned HOST_WIDE_INTlong) high < (unsigned HOST_WIDE_INTlong) b.high) |
1323 | return true; |
1324 | if ((unsigned HOST_WIDE_INTlong) high > (unsigned HOST_WIDE_INTlong) b.high) |
1325 | return false; |
1326 | if (low < b.low) |
1327 | return true; |
1328 | return false; |
1329 | } |
1330 | |
1331 | /* Compares two unsigned values A and B for less-than or equal-to. */ |
1332 | |
1333 | bool |
1334 | double_int::ule (double_int b) const |
1335 | { |
1336 | if ((unsigned HOST_WIDE_INTlong) high < (unsigned HOST_WIDE_INTlong) b.high) |
1337 | return true; |
1338 | if ((unsigned HOST_WIDE_INTlong) high > (unsigned HOST_WIDE_INTlong) b.high) |
1339 | return false; |
1340 | if (low <= b.low) |
1341 | return true; |
1342 | return false; |
1343 | } |
1344 | |
1345 | /* Compares two unsigned values A and B for greater-than. */ |
1346 | |
1347 | bool |
1348 | double_int::ugt (double_int b) const |
1349 | { |
1350 | if ((unsigned HOST_WIDE_INTlong) high > (unsigned HOST_WIDE_INTlong) b.high) |
1351 | return true; |
1352 | if ((unsigned HOST_WIDE_INTlong) high < (unsigned HOST_WIDE_INTlong) b.high) |
1353 | return false; |
1354 | if (low > b.low) |
1355 | return true; |
1356 | return false; |
1357 | } |
1358 | |
1359 | /* Compares two signed values A and B for less-than. */ |
1360 | |
1361 | bool |
1362 | double_int::slt (double_int b) const |
1363 | { |
1364 | if (high < b.high) |
1365 | return true; |
1366 | if (high > b.high) |
1367 | return false; |
1368 | if (low < b.low) |
1369 | return true; |
1370 | return false; |
1371 | } |
1372 | |
1373 | /* Compares two signed values A and B for less-than or equal-to. */ |
1374 | |
1375 | bool |
1376 | double_int::sle (double_int b) const |
1377 | { |
1378 | if (high < b.high) |
1379 | return true; |
1380 | if (high > b.high) |
1381 | return false; |
1382 | if (low <= b.low) |
1383 | return true; |
1384 | return false; |
1385 | } |
1386 | |
1387 | /* Compares two signed values A and B for greater-than. */ |
1388 | |
1389 | bool |
1390 | double_int::sgt (double_int b) const |
1391 | { |
1392 | if (high > b.high) |
1393 | return true; |
1394 | if (high < b.high) |
1395 | return false; |
1396 | if (low > b.low) |
1397 | return true; |
1398 | return false; |
1399 | } |
1400 | |
1401 | |
1402 | /* Compares two values A and B. Returns max value. Signedness of the |
1403 | comparison is given by UNS. */ |
1404 | |
1405 | double_int |
1406 | double_int::max (double_int b, bool uns) |
1407 | { |
1408 | return (this->cmp (b, uns) == 1) ? *this : b; |
1409 | } |
1410 | |
1411 | /* Compares two signed values A and B. Returns max value. */ |
1412 | |
1413 | double_int |
1414 | double_int::smax (double_int b) |
1415 | { |
1416 | return (this->scmp (b) == 1) ? *this : b; |
1417 | } |
1418 | |
1419 | /* Compares two unsigned values A and B. Returns max value. */ |
1420 | |
1421 | double_int |
1422 | double_int::umax (double_int b) |
1423 | { |
1424 | return (this->ucmp (b) == 1) ? *this : b; |
1425 | } |
1426 | |
1427 | /* Compares two values A and B. Returns mix value. Signedness of the |
1428 | comparison is given by UNS. */ |
1429 | |
1430 | double_int |
1431 | double_int::min (double_int b, bool uns) |
1432 | { |
1433 | return (this->cmp (b, uns) == -1) ? *this : b; |
1434 | } |
1435 | |
1436 | /* Compares two signed values A and B. Returns min value. */ |
1437 | |
1438 | double_int |
1439 | double_int::smin (double_int b) |
1440 | { |
1441 | return (this->scmp (b) == -1) ? *this : b; |
1442 | } |
1443 | |
1444 | /* Compares two unsigned values A and B. Returns min value. */ |
1445 | |
1446 | double_int |
1447 | double_int::umin (double_int b) |
1448 | { |
1449 | return (this->ucmp (b) == -1) ? *this : b; |
1450 | } |
1451 | |
1452 | /* Splits last digit of *CST (taken as unsigned) in BASE and returns it. */ |
1453 | |
1454 | static unsigned |
1455 | double_int_split_digit (double_int *cst, unsigned base) |
1456 | { |
1457 | unsigned HOST_WIDE_INTlong resl, reml; |
1458 | HOST_WIDE_INTlong resh, remh; |
1459 | |
1460 | div_and_round_double (FLOOR_DIV_EXPR, true, cst->low, cst->high, base, 0, |
1461 | &resl, &resh, &reml, &remh); |
1462 | cst->high = resh; |
1463 | cst->low = resl; |
1464 | |
1465 | return reml; |
1466 | } |
1467 | |
1468 | /* Dumps CST to FILE. If UNS is true, CST is considered to be unsigned, |
1469 | otherwise it is signed. */ |
1470 | |
1471 | void |
1472 | dump_double_int (FILE *file, double_int cst, bool uns) |
1473 | { |
1474 | unsigned digits[100], n; |
1475 | int i; |
1476 | |
1477 | if (cst.is_zero ()) |
1478 | { |
1479 | fprintf (file, "0"); |
1480 | return; |
1481 | } |
1482 | |
1483 | if (!uns && cst.is_negative ()) |
1484 | { |
1485 | fprintf (file, "-"); |
1486 | cst = -cst; |
1487 | } |
1488 | |
1489 | for (n = 0; !cst.is_zero (); n++) |
1490 | digits[n] = double_int_split_digit (&cst, 10); |
1491 | for (i = n - 1; i >= 0; i--) |
1492 | fprintf (file, "%u", digits[i]); |
1493 | } |
1494 | |
1495 | |
1496 | /* Sets RESULT to VAL, taken unsigned if UNS is true and as signed |
1497 | otherwise. */ |
1498 | |
1499 | void |
1500 | mpz_set_double_int (mpz_t result, double_int val, bool uns) |
1501 | { |
1502 | bool negate = false; |
1503 | unsigned HOST_WIDE_INTlong vp[2]; |
1504 | |
1505 | if (!uns && val.is_negative ()) |
1506 | { |
1507 | negate = true; |
1508 | val = -val; |
1509 | } |
1510 | |
1511 | vp[0] = val.low; |
1512 | vp[1] = (unsigned HOST_WIDE_INTlong) val.high; |
1513 | mpz_import__gmpz_import (result, 2, -1, sizeof (HOST_WIDE_INTlong), 0, 0, vp); |
1514 | |
1515 | if (negate) |
1516 | mpz_neg__gmpz_neg (result, result); |
1517 | } |
1518 | |
1519 | /* Returns VAL converted to TYPE. If WRAP is true, then out-of-range |
1520 | values of VAL will be wrapped; otherwise, they will be set to the |
1521 | appropriate minimum or maximum TYPE bound. */ |
1522 | |
1523 | double_int |
1524 | mpz_get_double_int (const_tree type, mpz_t val, bool wrap) |
1525 | { |
1526 | unsigned HOST_WIDE_INTlong *vp; |
1527 | size_t count, numb; |
1528 | double_int res; |
1529 | |
1530 | if (!wrap) |
1531 | { |
1532 | mpz_t min, max; |
1533 | |
1534 | mpz_init__gmpz_init (min); |
1535 | mpz_init__gmpz_init (max); |
1536 | get_type_static_bounds (type, min, max); |
1537 | |
1538 | if (mpz_cmp__gmpz_cmp (val, min) < 0) |
1539 | mpz_set__gmpz_set (val, min); |
1540 | else if (mpz_cmp__gmpz_cmp (val, max) > 0) |
1541 | mpz_set__gmpz_set (val, max); |
1542 | |
1543 | mpz_clear__gmpz_clear (min); |
1544 | mpz_clear__gmpz_clear (max); |
1545 | } |
1546 | |
1547 | /* Determine the number of unsigned HOST_WIDE_INT that are required |
1548 | for representing the value. The code to calculate count is |
1549 | extracted from the GMP manual, section "Integer Import and Export": |
1550 | http://gmplib.org/manual/Integer-Import-and-Export.html */ |
1551 | numb = 8 * sizeof (HOST_WIDE_INTlong); |
1552 | count = (mpz_sizeinbase__gmpz_sizeinbase (val, 2) + numb-1) / numb; |
1553 | if (count < 2) |
1554 | count = 2; |
1555 | vp = (unsigned HOST_WIDE_INTlong *) alloca (count * sizeof (HOST_WIDE_INT))__builtin_alloca(count * sizeof (long)); |
1556 | |
1557 | vp[0] = 0; |
1558 | vp[1] = 0; |
1559 | mpz_export__gmpz_export (vp, &count, -1, sizeof (HOST_WIDE_INTlong), 0, 0, val); |
1560 | |
1561 | gcc_assert (wrap || count <= 2)((void)(!(wrap || count <= 2) ? fancy_abort ("/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 1561, __FUNCTION__), 0 : 0)); |
1562 | |
1563 | res.low = vp[0]; |
1564 | res.high = (HOST_WIDE_INTlong) vp[1]; |
1565 | |
1566 | res = res.ext (TYPE_PRECISION (type)((tree_class_check ((type), (tcc_type), "/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 1566, __FUNCTION__))->type_common.precision), TYPE_UNSIGNED (type)((tree_class_check ((type), (tcc_type), "/buildworker/marxinbox-gcc-clang-static-analyzer/build/gcc/double-int.cc" , 1566, __FUNCTION__))->base.u.bits.unsigned_flag)); |
1567 | if (mpz_sgn (val)((val)->_mp_size < 0 ? -1 : (val)->_mp_size > 0) < 0) |
1568 | res = -res; |
1569 | |
1570 | return res; |
1571 | } |