Hi!
The following patch fixes
FAIL: gcc.dg/dfp/bitint-1.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-2.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-3.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-4.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-5.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-6.c (test for excess errors)
FAIL: gcc.dg/dfp/bitint-8.c (test for excess errors)
FAIL: gcc.dg/dfp/int128-1.c (test for excess errors)
FAIL: gcc.dg/dfp/int128-2.c (test for excess errors)
FAIL: gcc.dg/dfp/int128-4.c (test for excess errors)
on s390x-linux (with the 3 not yet posted patches).
The patch does multiple things:
1) the routines were written for the DFP BID (binary integer decimal)
format which is used on all arches but powerpc*/s390* (those use
DPD - densely packed decimal format); as most of the code is actually
the same for both BID and DPD formats, I haven't copied the sources
+ slightly modified them, but added the DPD support directly, + renaming
of the exported symbols from __bid_* prefixed to __dpd_* prefixed that
GCC expects on the DPD targets
2) while testing that I've found some big-endian issues in the existing
support
3) testing also revealed that in some cases __builtin_clzll (~msb) was
called with msb set to all ones, so invoking UB; apparently on aarch64
and x86 we were lucky and got some value that happened to work well,
but that wasn't the case on s390x
For 1), the patch uses two ~ 2KB tables to speed up the decoding/encoding.
I haven't found such tables in what is added into libgcc.a, though they
are in libdecnumber/bid/bid2dpd_dpd2bid.h, but there they are just huge
and next to other huge tables - there is d2b which is like __dpd_d2bbitint
in the patch but it uses 64-bit entries rather than 16-bit, then there is
d2b2 with 64-bit entries like in d2b all multiplied by 1000, then d2b3
similarly multiplied by 1000000, then d2b4 similarly multiplied by
1000000000, then d2b5 similarly multiplied by 1000000000000ULL and
d2b6 similarly multipled by 1000000000000000ULL. Arguably it can
save some of the multiplications, but on the other side accesses memory
which is unlikely in the caches, and the 2048 bytes in the patch vs.
24 times more for d2b is IMHO significant.
For b2d, libdecnumber/bid/bid2dpd_dpd2bid.h has again b2d table like
__dpd_b2dbitint in the patch, except that it has 64-bit entries rather
than 16-bit (this time 1000 entries), but then has b2d2 which has the
same entries shifted left by 10, then b2d3 shifted left by 20, b2d4 shifted
left by 30 and b2d5 shifted left by 40. I can understand for d2b paying
memory cost to speed up multiplications, but don't understand paying
extra 4 * 8 * 1000 bytes (+ 6 * 1000 bytes for b2d not using ushort)
just to avoid shifts.
Tested on x86_64-linux, i686-linux and s390x-linux with
make check-gcc dfp.exp
ok for trunk?
2025-05-20 Jakub Jelinek <ja...@redhat.com>
* config/t-softfp (softfp_bid_list): Don't guard with
$(enable_decimal_float) == bid.
* soft-fp/bitint.h (__bid_pow10bitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_pow10bitint.
(__dpd_d2bbitint, __dpd_b2dbitint): Declare.
* soft-fp/bitintpow10.c (__dpd_d2bbitint, __dpd_b2dbitint): New
variables.
* soft-fp/fixsdbitint.c (__bid_fixsdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixsdbitint.
Add DPD support. Fix big-endian support.
* soft-fp/fixddbitint.c (__bid_fixddbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixddbitint.
Add DPD support. Fix big-endian support.
* soft-fp/fixtdbitint.c (__bid_fixtdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixtdbitint.
Add DPD support. Fix big-endian support.
* soft-fp/fixsdti.c (__bid_fixsdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixsdbitint.
(__bid_fixsdti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_fixsdti.
* soft-fp/fixddti.c (__bid_fixddbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixddbitint.
(__bid_fixddti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_fixddti.
* soft-fp/fixtdti.c (__bid_fixtdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixtdbitint.
(__bid_fixtdti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_fixtdti.
* soft-fp/fixunssdti.c (__bid_fixsdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixsdbitint.
(__bid_fixunssdti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_fixunssdti.
* soft-fp/fixunsddti.c (__bid_fixddbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixddbitint.
(__bid_fixunsddti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_fixunsddti.
* soft-fp/fixunstdti.c (__bid_fixtdbitint): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_fixtdbitint.
(__bid_fixunstdti): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_fixunstdti.
* soft-fp/floatbitintsd.c (__bid_floatbitintsd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintsd.
Add DPD support. Avoid calling __builtin_clzll with 0 argument. Fix
big-endian support.
* soft-fp/floatbitintdd.c (__bid_floatbitintdd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintdd.
Add DPD support. Avoid calling __builtin_clzll with 0 argument. Fix
big-endian support.
* soft-fp/floatbitinttd.c (__bid_floatbitinttd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitinttd.
Add DPD support. Avoid calling __builtin_clzll with 0 argument. Fix
big-endian support.
* soft-fp/floattisd.c (__bid_floatbitintsd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintsd.
(__bid_floattisd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_floattisd.
* soft-fp/floattidd.c (__bid_floatbitintdd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintdd.
(__bid_floattidd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_floattidd.
* soft-fp/floattitd.c (__bid_floatbitinttd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitinttd.
(__bid_floattitd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine to
__dpd_floattitd.
* soft-fp/floatuntisd.c (__bid_floatbitintsd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintsd.
(__bid_floatuntisd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_floatuntisd.
* soft-fp/floatuntidd.c (__bid_floatbitintdd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitintdd.
(__bid_floatuntidd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_floatuntidd.
* soft-fp/floatuntitd.c (__bid_floatbitinttd): For
!defined(ENABLE_DECIMAL_BID_FORMAT) redefine to __dpd_floatbitinttd.
(__bid_floatuntitd): For !defined(ENABLE_DECIMAL_BID_FORMAT) redefine
to __dpd_floatuntitd.
--- libgcc/config/t-softfp.jj 2025-04-08 14:09:53.626413530 +0200
+++ libgcc/config/t-softfp 2025-05-20 12:28:59.377848126 +0200
@@ -67,13 +67,11 @@ softfp_floatint_funcs = fix$(m)$(i) fixu
softfp_floatbitint_funcs = fix$(m)bitint floatbitint$(m)
softfp_bid_list :=
ifeq ($(decimal_float),yes)
-ifeq ($(enable_decimal_float),bid)
softfp_bid_list += bitintpow10 \
$(foreach m,sd dd td,fix$(m)bitint floatbitint$(m) \
fix$(m)ti fixuns$(m)ti \
floatti$(m) floatunti$(m))
endif
-endif
softfp_func_list := \
$(foreach m,$(softfp_float_modes), \
--- libgcc/soft-fp/bitint.h.jj 2025-05-20 08:20:13.267366111 +0200
+++ libgcc/soft-fp/bitint.h 2025-05-20 14:22:40.023732555 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Definitions for _BitInt implementation details.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -352,6 +352,10 @@ extern void __divmodbitint4 (UBILtype *,
const UBILtype *, SItype,
const UBILtype *, SItype);
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_pow10bitint __dpd_pow10bitint
+extern const unsigned short __dpd_d2bbitint[1024], __dpd_b2dbitint[1000];
+#endif
extern USItype __bid_pow10bitint (UBILtype *, SItype, USItype);
#endif /* __BITINT_MAXWIDTH__ */
--- libgcc/soft-fp/bitintpow10.c.jj 2025-05-20 08:13:28.240930549 +0200
+++ libgcc/soft-fp/bitintpow10.c 2025-05-20 09:44:54.171713272 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Compute powers of 10 into _BitInt.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -130,3 +130,341 @@ __bid_pow10bitint (UBILtype *r, SItype r
}
}
#endif
+
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+/* DPD encoded 10-bit number to decimal table. */
+const unsigned short __dpd_d2bbitint[1024] = {
+ 0, 1, 2, 3, 4, 5, 6, 7,
+ 8, 9, 80, 81, 800, 801, 880, 881,
+ 10, 11, 12, 13, 14, 15, 16, 17,
+ 18, 19, 90, 91, 810, 811, 890, 891,
+ 20, 21, 22, 23, 24, 25, 26, 27,
+ 28, 29, 82, 83, 820, 821, 808, 809,
+ 30, 31, 32, 33, 34, 35, 36, 37,
+ 38, 39, 92, 93, 830, 831, 818, 819,
+ 40, 41, 42, 43, 44, 45, 46, 47,
+ 48, 49, 84, 85, 840, 841, 88, 89,
+ 50, 51, 52, 53, 54, 55, 56, 57,
+ 58, 59, 94, 95, 850, 851, 98, 99,
+ 60, 61, 62, 63, 64, 65, 66, 67,
+ 68, 69, 86, 87, 860, 861, 888, 889,
+ 70, 71, 72, 73, 74, 75, 76, 77,
+ 78, 79, 96, 97, 870, 871, 898, 899,
+ 100, 101, 102, 103, 104, 105, 106, 107,
+ 108, 109, 180, 181, 900, 901, 980, 981,
+ 110, 111, 112, 113, 114, 115, 116, 117,
+ 118, 119, 190, 191, 910, 911, 990, 991,
+ 120, 121, 122, 123, 124, 125, 126, 127,
+ 128, 129, 182, 183, 920, 921, 908, 909,
+ 130, 131, 132, 133, 134, 135, 136, 137,
+ 138, 139, 192, 193, 930, 931, 918, 919,
+ 140, 141, 142, 143, 144, 145, 146, 147,
+ 148, 149, 184, 185, 940, 941, 188, 189,
+ 150, 151, 152, 153, 154, 155, 156, 157,
+ 158, 159, 194, 195, 950, 951, 198, 199,
+ 160, 161, 162, 163, 164, 165, 166, 167,
+ 168, 169, 186, 187, 960, 961, 988, 989,
+ 170, 171, 172, 173, 174, 175, 176, 177,
+ 178, 179, 196, 197, 970, 971, 998, 999,
+ 200, 201, 202, 203, 204, 205, 206, 207,
+ 208, 209, 280, 281, 802, 803, 882, 883,
+ 210, 211, 212, 213, 214, 215, 216, 217,
+ 218, 219, 290, 291, 812, 813, 892, 893,
+ 220, 221, 222, 223, 224, 225, 226, 227,
+ 228, 229, 282, 283, 822, 823, 828, 829,
+ 230, 231, 232, 233, 234, 235, 236, 237,
+ 238, 239, 292, 293, 832, 833, 838, 839,
+ 240, 241, 242, 243, 244, 245, 246, 247,
+ 248, 249, 284, 285, 842, 843, 288, 289,
+ 250, 251, 252, 253, 254, 255, 256, 257,
+ 258, 259, 294, 295, 852, 853, 298, 299,
+ 260, 261, 262, 263, 264, 265, 266, 267,
+ 268, 269, 286, 287, 862, 863, 888, 889,
+ 270, 271, 272, 273, 274, 275, 276, 277,
+ 278, 279, 296, 297, 872, 873, 898, 899,
+ 300, 301, 302, 303, 304, 305, 306, 307,
+ 308, 309, 380, 381, 902, 903, 982, 983,
+ 310, 311, 312, 313, 314, 315, 316, 317,
+ 318, 319, 390, 391, 912, 913, 992, 993,
+ 320, 321, 322, 323, 324, 325, 326, 327,
+ 328, 329, 382, 383, 922, 923, 928, 929,
+ 330, 331, 332, 333, 334, 335, 336, 337,
+ 338, 339, 392, 393, 932, 933, 938, 939,
+ 340, 341, 342, 343, 344, 345, 346, 347,
+ 348, 349, 384, 385, 942, 943, 388, 389,
+ 350, 351, 352, 353, 354, 355, 356, 357,
+ 358, 359, 394, 395, 952, 953, 398, 399,
+ 360, 361, 362, 363, 364, 365, 366, 367,
+ 368, 369, 386, 387, 962, 963, 988, 989,
+ 370, 371, 372, 373, 374, 375, 376, 377,
+ 378, 379, 396, 397, 972, 973, 998, 999,
+ 400, 401, 402, 403, 404, 405, 406, 407,
+ 408, 409, 480, 481, 804, 805, 884, 885,
+ 410, 411, 412, 413, 414, 415, 416, 417,
+ 418, 419, 490, 491, 814, 815, 894, 895,
+ 420, 421, 422, 423, 424, 425, 426, 427,
+ 428, 429, 482, 483, 824, 825, 848, 849,
+ 430, 431, 432, 433, 434, 435, 436, 437,
+ 438, 439, 492, 493, 834, 835, 858, 859,
+ 440, 441, 442, 443, 444, 445, 446, 447,
+ 448, 449, 484, 485, 844, 845, 488, 489,
+ 450, 451, 452, 453, 454, 455, 456, 457,
+ 458, 459, 494, 495, 854, 855, 498, 499,
+ 460, 461, 462, 463, 464, 465, 466, 467,
+ 468, 469, 486, 487, 864, 865, 888, 889,
+ 470, 471, 472, 473, 474, 475, 476, 477,
+ 478, 479, 496, 497, 874, 875, 898, 899,
+ 500, 501, 502, 503, 504, 505, 506, 507,
+ 508, 509, 580, 581, 904, 905, 984, 985,
+ 510, 511, 512, 513, 514, 515, 516, 517,
+ 518, 519, 590, 591, 914, 915, 994, 995,
+ 520, 521, 522, 523, 524, 525, 526, 527,
+ 528, 529, 582, 583, 924, 925, 948, 949,
+ 530, 531, 532, 533, 534, 535, 536, 537,
+ 538, 539, 592, 593, 934, 935, 958, 959,
+ 540, 541, 542, 543, 544, 545, 546, 547,
+ 548, 549, 584, 585, 944, 945, 588, 589,
+ 550, 551, 552, 553, 554, 555, 556, 557,
+ 558, 559, 594, 595, 954, 955, 598, 599,
+ 560, 561, 562, 563, 564, 565, 566, 567,
+ 568, 569, 586, 587, 964, 965, 988, 989,
+ 570, 571, 572, 573, 574, 575, 576, 577,
+ 578, 579, 596, 597, 974, 975, 998, 999,
+ 600, 601, 602, 603, 604, 605, 606, 607,
+ 608, 609, 680, 681, 806, 807, 886, 887,
+ 610, 611, 612, 613, 614, 615, 616, 617,
+ 618, 619, 690, 691, 816, 817, 896, 897,
+ 620, 621, 622, 623, 624, 625, 626, 627,
+ 628, 629, 682, 683, 826, 827, 868, 869,
+ 630, 631, 632, 633, 634, 635, 636, 637,
+ 638, 639, 692, 693, 836, 837, 878, 879,
+ 640, 641, 642, 643, 644, 645, 646, 647,
+ 648, 649, 684, 685, 846, 847, 688, 689,
+ 650, 651, 652, 653, 654, 655, 656, 657,
+ 658, 659, 694, 695, 856, 857, 698, 699,
+ 660, 661, 662, 663, 664, 665, 666, 667,
+ 668, 669, 686, 687, 866, 867, 888, 889,
+ 670, 671, 672, 673, 674, 675, 676, 677,
+ 678, 679, 696, 697, 876, 877, 898, 899,
+ 700, 701, 702, 703, 704, 705, 706, 707,
+ 708, 709, 780, 781, 906, 907, 986, 987,
+ 710, 711, 712, 713, 714, 715, 716, 717,
+ 718, 719, 790, 791, 916, 917, 996, 997,
+ 720, 721, 722, 723, 724, 725, 726, 727,
+ 728, 729, 782, 783, 926, 927, 968, 969,
+ 730, 731, 732, 733, 734, 735, 736, 737,
+ 738, 739, 792, 793, 936, 937, 978, 979,
+ 740, 741, 742, 743, 744, 745, 746, 747,
+ 748, 749, 784, 785, 946, 947, 788, 789,
+ 750, 751, 752, 753, 754, 755, 756, 757,
+ 758, 759, 794, 795, 956, 957, 798, 799,
+ 760, 761, 762, 763, 764, 765, 766, 767,
+ 768, 769, 786, 787, 966, 967, 988, 989,
+ 770, 771, 772, 773, 774, 775, 776, 777,
+ 778, 779, 796, 797, 976, 977, 998, 999
+};
+
+/* Decimal to DPD encoded 10-bit number table. */
+const unsigned short __dpd_b2dbitint[1000] = {
+ 0x000, 0x001, 0x002, 0x003, 0x004,
+ 0x005, 0x006, 0x007, 0x008, 0x009,
+ 0x010, 0x011, 0x012, 0x013, 0x014,
+ 0x015, 0x016, 0x017, 0x018, 0x019,
+ 0x020, 0x021, 0x022, 0x023, 0x024,
+ 0x025, 0x026, 0x027, 0x028, 0x029,
+ 0x030, 0x031, 0x032, 0x033, 0x034,
+ 0x035, 0x036, 0x037, 0x038, 0x039,
+ 0x040, 0x041, 0x042, 0x043, 0x044,
+ 0x045, 0x046, 0x047, 0x048, 0x049,
+ 0x050, 0x051, 0x052, 0x053, 0x054,
+ 0x055, 0x056, 0x057, 0x058, 0x059,
+ 0x060, 0x061, 0x062, 0x063, 0x064,
+ 0x065, 0x066, 0x067, 0x068, 0x069,
+ 0x070, 0x071, 0x072, 0x073, 0x074,
+ 0x075, 0x076, 0x077, 0x078, 0x079,
+ 0x00a, 0x00b, 0x02a, 0x02b, 0x04a,
+ 0x04b, 0x06a, 0x06b, 0x04e, 0x04f,
+ 0x01a, 0x01b, 0x03a, 0x03b, 0x05a,
+ 0x05b, 0x07a, 0x07b, 0x05e, 0x05f,
+ 0x080, 0x081, 0x082, 0x083, 0x084,
+ 0x085, 0x086, 0x087, 0x088, 0x089,
+ 0x090, 0x091, 0x092, 0x093, 0x094,
+ 0x095, 0x096, 0x097, 0x098, 0x099,
+ 0x0a0, 0x0a1, 0x0a2, 0x0a3, 0x0a4,
+ 0x0a5, 0x0a6, 0x0a7, 0x0a8, 0x0a9,
+ 0x0b0, 0x0b1, 0x0b2, 0x0b3, 0x0b4,
+ 0x0b5, 0x0b6, 0x0b7, 0x0b8, 0x0b9,
+ 0x0c0, 0x0c1, 0x0c2, 0x0c3, 0x0c4,
+ 0x0c5, 0x0c6, 0x0c7, 0x0c8, 0x0c9,
+ 0x0d0, 0x0d1, 0x0d2, 0x0d3, 0x0d4,
+ 0x0d5, 0x0d6, 0x0d7, 0x0d8, 0x0d9,
+ 0x0e0, 0x0e1, 0x0e2, 0x0e3, 0x0e4,
+ 0x0e5, 0x0e6, 0x0e7, 0x0e8, 0x0e9,
+ 0x0f0, 0x0f1, 0x0f2, 0x0f3, 0x0f4,
+ 0x0f5, 0x0f6, 0x0f7, 0x0f8, 0x0f9,
+ 0x08a, 0x08b, 0x0aa, 0x0ab, 0x0ca,
+ 0x0cb, 0x0ea, 0x0eb, 0x0ce, 0x0cf,
+ 0x09a, 0x09b, 0x0ba, 0x0bb, 0x0da,
+ 0x0db, 0x0fa, 0x0fb, 0x0de, 0x0df,
+ 0x100, 0x101, 0x102, 0x103, 0x104,
+ 0x105, 0x106, 0x107, 0x108, 0x109,
+ 0x110, 0x111, 0x112, 0x113, 0x114,
+ 0x115, 0x116, 0x117, 0x118, 0x119,
+ 0x120, 0x121, 0x122, 0x123, 0x124,
+ 0x125, 0x126, 0x127, 0x128, 0x129,
+ 0x130, 0x131, 0x132, 0x133, 0x134,
+ 0x135, 0x136, 0x137, 0x138, 0x139,
+ 0x140, 0x141, 0x142, 0x143, 0x144,
+ 0x145, 0x146, 0x147, 0x148, 0x149,
+ 0x150, 0x151, 0x152, 0x153, 0x154,
+ 0x155, 0x156, 0x157, 0x158, 0x159,
+ 0x160, 0x161, 0x162, 0x163, 0x164,
+ 0x165, 0x166, 0x167, 0x168, 0x169,
+ 0x170, 0x171, 0x172, 0x173, 0x174,
+ 0x175, 0x176, 0x177, 0x178, 0x179,
+ 0x10a, 0x10b, 0x12a, 0x12b, 0x14a,
+ 0x14b, 0x16a, 0x16b, 0x14e, 0x14f,
+ 0x11a, 0x11b, 0x13a, 0x13b, 0x15a,
+ 0x15b, 0x17a, 0x17b, 0x15e, 0x15f,
+ 0x180, 0x181, 0x182, 0x183, 0x184,
+ 0x185, 0x186, 0x187, 0x188, 0x189,
+ 0x190, 0x191, 0x192, 0x193, 0x194,
+ 0x195, 0x196, 0x197, 0x198, 0x199,
+ 0x1a0, 0x1a1, 0x1a2, 0x1a3, 0x1a4,
+ 0x1a5, 0x1a6, 0x1a7, 0x1a8, 0x1a9,
+ 0x1b0, 0x1b1, 0x1b2, 0x1b3, 0x1b4,
+ 0x1b5, 0x1b6, 0x1b7, 0x1b8, 0x1b9,
+ 0x1c0, 0x1c1, 0x1c2, 0x1c3, 0x1c4,
+ 0x1c5, 0x1c6, 0x1c7, 0x1c8, 0x1c9,
+ 0x1d0, 0x1d1, 0x1d2, 0x1d3, 0x1d4,
+ 0x1d5, 0x1d6, 0x1d7, 0x1d8, 0x1d9,
+ 0x1e0, 0x1e1, 0x1e2, 0x1e3, 0x1e4,
+ 0x1e5, 0x1e6, 0x1e7, 0x1e8, 0x1e9,
+ 0x1f0, 0x1f1, 0x1f2, 0x1f3, 0x1f4,
+ 0x1f5, 0x1f6, 0x1f7, 0x1f8, 0x1f9,
+ 0x18a, 0x18b, 0x1aa, 0x1ab, 0x1ca,
+ 0x1cb, 0x1ea, 0x1eb, 0x1ce, 0x1cf,
+ 0x19a, 0x19b, 0x1ba, 0x1bb, 0x1da,
+ 0x1db, 0x1fa, 0x1fb, 0x1de, 0x1df,
+ 0x200, 0x201, 0x202, 0x203, 0x204,
+ 0x205, 0x206, 0x207, 0x208, 0x209,
+ 0x210, 0x211, 0x212, 0x213, 0x214,
+ 0x215, 0x216, 0x217, 0x218, 0x219,
+ 0x220, 0x221, 0x222, 0x223, 0x224,
+ 0x225, 0x226, 0x227, 0x228, 0x229,
+ 0x230, 0x231, 0x232, 0x233, 0x234,
+ 0x235, 0x236, 0x237, 0x238, 0x239,
+ 0x240, 0x241, 0x242, 0x243, 0x244,
+ 0x245, 0x246, 0x247, 0x248, 0x249,
+ 0x250, 0x251, 0x252, 0x253, 0x254,
+ 0x255, 0x256, 0x257, 0x258, 0x259,
+ 0x260, 0x261, 0x262, 0x263, 0x264,
+ 0x265, 0x266, 0x267, 0x268, 0x269,
+ 0x270, 0x271, 0x272, 0x273, 0x274,
+ 0x275, 0x276, 0x277, 0x278, 0x279,
+ 0x20a, 0x20b, 0x22a, 0x22b, 0x24a,
+ 0x24b, 0x26a, 0x26b, 0x24e, 0x24f,
+ 0x21a, 0x21b, 0x23a, 0x23b, 0x25a,
+ 0x25b, 0x27a, 0x27b, 0x25e, 0x25f,
+ 0x280, 0x281, 0x282, 0x283, 0x284,
+ 0x285, 0x286, 0x287, 0x288, 0x289,
+ 0x290, 0x291, 0x292, 0x293, 0x294,
+ 0x295, 0x296, 0x297, 0x298, 0x299,
+ 0x2a0, 0x2a1, 0x2a2, 0x2a3, 0x2a4,
+ 0x2a5, 0x2a6, 0x2a7, 0x2a8, 0x2a9,
+ 0x2b0, 0x2b1, 0x2b2, 0x2b3, 0x2b4,
+ 0x2b5, 0x2b6, 0x2b7, 0x2b8, 0x2b9,
+ 0x2c0, 0x2c1, 0x2c2, 0x2c3, 0x2c4,
+ 0x2c5, 0x2c6, 0x2c7, 0x2c8, 0x2c9,
+ 0x2d0, 0x2d1, 0x2d2, 0x2d3, 0x2d4,
+ 0x2d5, 0x2d6, 0x2d7, 0x2d8, 0x2d9,
+ 0x2e0, 0x2e1, 0x2e2, 0x2e3, 0x2e4,
+ 0x2e5, 0x2e6, 0x2e7, 0x2e8, 0x2e9,
+ 0x2f0, 0x2f1, 0x2f2, 0x2f3, 0x2f4,
+ 0x2f5, 0x2f6, 0x2f7, 0x2f8, 0x2f9,
+ 0x28a, 0x28b, 0x2aa, 0x2ab, 0x2ca,
+ 0x2cb, 0x2ea, 0x2eb, 0x2ce, 0x2cf,
+ 0x29a, 0x29b, 0x2ba, 0x2bb, 0x2da,
+ 0x2db, 0x2fa, 0x2fb, 0x2de, 0x2df,
+ 0x300, 0x301, 0x302, 0x303, 0x304,
+ 0x305, 0x306, 0x307, 0x308, 0x309,
+ 0x310, 0x311, 0x312, 0x313, 0x314,
+ 0x315, 0x316, 0x317, 0x318, 0x319,
+ 0x320, 0x321, 0x322, 0x323, 0x324,
+ 0x325, 0x326, 0x327, 0x328, 0x329,
+ 0x330, 0x331, 0x332, 0x333, 0x334,
+ 0x335, 0x336, 0x337, 0x338, 0x339,
+ 0x340, 0x341, 0x342, 0x343, 0x344,
+ 0x345, 0x346, 0x347, 0x348, 0x349,
+ 0x350, 0x351, 0x352, 0x353, 0x354,
+ 0x355, 0x356, 0x357, 0x358, 0x359,
+ 0x360, 0x361, 0x362, 0x363, 0x364,
+ 0x365, 0x366, 0x367, 0x368, 0x369,
+ 0x370, 0x371, 0x372, 0x373, 0x374,
+ 0x375, 0x376, 0x377, 0x378, 0x379,
+ 0x30a, 0x30b, 0x32a, 0x32b, 0x34a,
+ 0x34b, 0x36a, 0x36b, 0x34e, 0x34f,
+ 0x31a, 0x31b, 0x33a, 0x33b, 0x35a,
+ 0x35b, 0x37a, 0x37b, 0x35e, 0x35f,
+ 0x380, 0x381, 0x382, 0x383, 0x384,
+ 0x385, 0x386, 0x387, 0x388, 0x389,
+ 0x390, 0x391, 0x392, 0x393, 0x394,
+ 0x395, 0x396, 0x397, 0x398, 0x399,
+ 0x3a0, 0x3a1, 0x3a2, 0x3a3, 0x3a4,
+ 0x3a5, 0x3a6, 0x3a7, 0x3a8, 0x3a9,
+ 0x3b0, 0x3b1, 0x3b2, 0x3b3, 0x3b4,
+ 0x3b5, 0x3b6, 0x3b7, 0x3b8, 0x3b9,
+ 0x3c0, 0x3c1, 0x3c2, 0x3c3, 0x3c4,
+ 0x3c5, 0x3c6, 0x3c7, 0x3c8, 0x3c9,
+ 0x3d0, 0x3d1, 0x3d2, 0x3d3, 0x3d4,
+ 0x3d5, 0x3d6, 0x3d7, 0x3d8, 0x3d9,
+ 0x3e0, 0x3e1, 0x3e2, 0x3e3, 0x3e4,
+ 0x3e5, 0x3e6, 0x3e7, 0x3e8, 0x3e9,
+ 0x3f0, 0x3f1, 0x3f2, 0x3f3, 0x3f4,
+ 0x3f5, 0x3f6, 0x3f7, 0x3f8, 0x3f9,
+ 0x38a, 0x38b, 0x3aa, 0x3ab, 0x3ca,
+ 0x3cb, 0x3ea, 0x3eb, 0x3ce, 0x3cf,
+ 0x39a, 0x39b, 0x3ba, 0x3bb, 0x3da,
+ 0x3db, 0x3fa, 0x3fb, 0x3de, 0x3df,
+ 0x00c, 0x00d, 0x10c, 0x10d, 0x20c,
+ 0x20d, 0x30c, 0x30d, 0x02e, 0x02f,
+ 0x01c, 0x01d, 0x11c, 0x11d, 0x21c,
+ 0x21d, 0x31c, 0x31d, 0x03e, 0x03f,
+ 0x02c, 0x02d, 0x12c, 0x12d, 0x22c,
+ 0x22d, 0x32c, 0x32d, 0x12e, 0x12f,
+ 0x03c, 0x03d, 0x13c, 0x13d, 0x23c,
+ 0x23d, 0x33c, 0x33d, 0x13e, 0x13f,
+ 0x04c, 0x04d, 0x14c, 0x14d, 0x24c,
+ 0x24d, 0x34c, 0x34d, 0x22e, 0x22f,
+ 0x05c, 0x05d, 0x15c, 0x15d, 0x25c,
+ 0x25d, 0x35c, 0x35d, 0x23e, 0x23f,
+ 0x06c, 0x06d, 0x16c, 0x16d, 0x26c,
+ 0x26d, 0x36c, 0x36d, 0x32e, 0x32f,
+ 0x07c, 0x07d, 0x17c, 0x17d, 0x27c,
+ 0x27d, 0x37c, 0x37d, 0x33e, 0x33f,
+ 0x00e, 0x00f, 0x10e, 0x10f, 0x20e,
+ 0x20f, 0x30e, 0x30f, 0x06e, 0x06f,
+ 0x01e, 0x01f, 0x11e, 0x11f, 0x21e,
+ 0x21f, 0x31e, 0x31f, 0x07e, 0x07f,
+ 0x08c, 0x08d, 0x18c, 0x18d, 0x28c,
+ 0x28d, 0x38c, 0x38d, 0x0ae, 0x0af,
+ 0x09c, 0x09d, 0x19c, 0x19d, 0x29c,
+ 0x29d, 0x39c, 0x39d, 0x0be, 0x0bf,
+ 0x0ac, 0x0ad, 0x1ac, 0x1ad, 0x2ac,
+ 0x2ad, 0x3ac, 0x3ad, 0x1ae, 0x1af,
+ 0x0bc, 0x0bd, 0x1bc, 0x1bd, 0x2bc,
+ 0x2bd, 0x3bc, 0x3bd, 0x1be, 0x1bf,
+ 0x0cc, 0x0cd, 0x1cc, 0x1cd, 0x2cc,
+ 0x2cd, 0x3cc, 0x3cd, 0x2ae, 0x2af,
+ 0x0dc, 0x0dd, 0x1dc, 0x1dd, 0x2dc,
+ 0x2dd, 0x3dc, 0x3dd, 0x2be, 0x2bf,
+ 0x0ec, 0x0ed, 0x1ec, 0x1ed, 0x2ec,
+ 0x2ed, 0x3ec, 0x3ed, 0x3ae, 0x3af,
+ 0x0fc, 0x0fd, 0x1fc, 0x1fd, 0x2fc,
+ 0x2fd, 0x3fc, 0x3fd, 0x3be, 0x3bf,
+ 0x08e, 0x08f, 0x18e, 0x18f, 0x28e,
+ 0x28f, 0x38e, 0x38f, 0x0ee, 0x0ef,
+ 0x09e, 0x09f, 0x19e, 0x19f, 0x29e,
+ 0x29f, 0x39e, 0x39f, 0x0fe, 0x0ff
+};
+#endif
--- libgcc/soft-fp/fixsdbitint.c.jj 2025-05-20 08:13:28.281929985 +0200
+++ libgcc/soft-fp/fixsdbitint.c 2025-05-20 14:23:02.637427170 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal32 to signed or unsigned _BitInt.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixsdbitint __dpd_fixsdbitint
+#endif
extern void __bid_fixsdbitint (UBILtype *, SItype, _Decimal32);
void
@@ -48,6 +51,7 @@ __bid_fixsdbitint (UBILtype *r, SItype r
u.d = a;
t = u.u >> 21;
sgn = (SItype) u.u < 0;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
if ((t & (3 << 8)) != (3 << 8))
{
mantissa = u.u & ((((USItype) 1) << 23) - 1);
@@ -61,6 +65,25 @@ __bid_fixsdbitint (UBILtype *r, SItype r
if (mantissa > (USItype) 9999999)
mantissa = 0;
}
+#else
+ if ((t & (15 << 6)) != (15 << 6))
+ {
+ exponent = (u.u >> 20) & 63;
+ if ((t & (3 << 8)) != (3 << 8))
+ {
+ mantissa = ((t >> 5) & 7) * 1000;
+ exponent += (t >> 2) & (3 << 6);
+ }
+ else
+ {
+ mantissa = ((t >> 5) & 1) ? 9000 : 8000;
+ exponent += t & (3 << 6);
+ }
+ mantissa += __dpd_d2bbitint[(u.u >> 10) & 1023];
+ mantissa *= 1000;
+ mantissa += __dpd_d2bbitint[u.u & 1023];
+ }
+#endif
else
{
FP_SET_EXCEPTION (FP_EX_INVALID
@@ -153,7 +176,7 @@ __bid_fixsdbitint (UBILtype *r, SItype r
if (res_limbs + low_zeros > rn && resv[BITINT_END (0, res_limbs - 1)])
goto ovf_ex;
if ((arprec % BIL_TYPE_SIZE) != 0
- && (resv[BITINT_END (rn - res_limbs, rn - 1) - low_zeros]
+ && (resv[BITINT_END (res_limbs + low_zeros - rn, rn - 1 - low_zeros)]
& ((UBILtype) -1 << (arprec % BIL_TYPE_SIZE))) != 0)
goto ovf_ex;
min_limbs = rn - low_zeros;
--- libgcc/soft-fp/fixddbitint.c.jj 2025-05-20 08:13:28.265930205 +0200
+++ libgcc/soft-fp/fixddbitint.c 2025-05-20 13:21:06.227096189 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal64 to signed or unsigned _BitInt.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixddbitint __dpd_fixddbitint
+#endif
extern void __bid_fixddbitint (UBILtype *, SItype, _Decimal64);
void
@@ -48,6 +51,7 @@ __bid_fixddbitint (UBILtype *r, SItype r
u.d = a;
t = u.u >> 51;
sgn = (DItype) u.u < 0;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
if ((t & (3 << 10)) != (3 << 10))
{
mantissa = u.u & ((((UDItype) 1) << 53) - 1);
@@ -61,6 +65,31 @@ __bid_fixddbitint (UBILtype *r, SItype r
if (mantissa > (UDItype) 9999999999999999)
mantissa = 0;
}
+#else
+ if ((t & (15 << 8)) != (15 << 8))
+ {
+ exponent = (u.u >> 50) & 255;
+ if ((t & (3 << 10)) != (3 << 10))
+ {
+ mantissa = ((t >> 7) & 7) * 1000;
+ exponent += (t >> 2) & (3 << 8);
+ }
+ else
+ {
+ mantissa = ((t >> 7) & 1) ? 9000 : 8000;
+ exponent += t & (3 << 8);
+ }
+ mantissa += __dpd_d2bbitint[(u.u >> 40) & 1023];
+ mantissa *= 1000;
+ mantissa += __dpd_d2bbitint[(u.u >> 30) & 1023];
+ mantissa *= 1000;
+ mantissa += __dpd_d2bbitint[(u.u >> 20) & 1023];
+ mantissa *= 1000;
+ mantissa += __dpd_d2bbitint[(u.u >> 10) & 1023];
+ mantissa *= 1000;
+ mantissa += __dpd_d2bbitint[u.u & 1023];
+ }
+#endif
else
{
FP_SET_EXCEPTION (FP_EX_INVALID
@@ -162,7 +191,7 @@ __bid_fixddbitint (UBILtype *r, SItype r
if (res_limbs + low_zeros > rn && resv[BITINT_END (0, res_limbs - 1)])
goto ovf_ex;
if ((arprec % BIL_TYPE_SIZE) != 0
- && (resv[BITINT_END (rn - res_limbs, rn - 1) - low_zeros]
+ && (resv[BITINT_END (res_limbs + low_zeros - rn, rn - 1 - low_zeros)]
& ((UBILtype) -1 << (arprec % BIL_TYPE_SIZE))) != 0)
goto ovf_ex;
min_limbs = rn - low_zeros;
--- libgcc/soft-fp/fixtdbitint.c.jj 2025-05-20 08:13:28.309929601 +0200
+++ libgcc/soft-fp/fixtdbitint.c 2025-05-20 13:20:38.461475817 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal128 to signed or unsigned _BitInt.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixtdbitint __dpd_fixtdbitint
+#endif
extern void __bid_fixtdbitint (UBILtype *, SItype, _Decimal128);
void
@@ -50,6 +53,7 @@ __bid_fixtdbitint (UBILtype *r, SItype r
mantissalo = u.u[__FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__];
t = mantissahi >> 47;
sgn = (DItype) mantissahi < 0;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
if ((t & (3 << 14)) != (3 << 14))
{
mantissahi &= ((((UDItype) 1) << 49) - 1);
@@ -68,6 +72,52 @@ __bid_fixtdbitint (UBILtype *r, SItype r
mantissalo = 0;
exponent = t & 0x3fff;
}
+#else
+ if ((t & (15 << 12)) != (15 << 12))
+ {
+ exponent = (mantissahi >> 46) & 4095;
+ if ((t & (3 << 14)) != (3 << 14))
+ {
+ exponent += (t >> 2) & (3 << 12);
+ t = ((t >> 11) & 7) * 1000;
+ }
+ else
+ {
+ exponent += t & (3 << 12);
+ t = ((t >> 11) & 1) ? 9000 : 8000;
+ }
+ t += __dpd_d2bbitint[(mantissahi >> (100 - 64)) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissahi >> (90 - 64)) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissahi >> (80 - 64)) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissahi >> (70 - 64)) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[((mantissahi & 63) << 4)
+ + ((mantissalo >> 60) & 15)];
+ t *= 1000;
+ t += __dpd_d2bbitint[mantissalo >> 50 & 1023];
+#ifdef __SIZEOF_INT128__
+ unsigned __int128 m = t;
+#else
+ unsigned _BitInt(128) m = t;
+#endif
+ m *= 1000000000000000ULL;
+ t = __dpd_d2bbitint[(mantissalo >> 40) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissalo >> 30) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissalo >> 20) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[(mantissalo >> 10) & 1023];
+ t *= 1000;
+ t += __dpd_d2bbitint[mantissalo & 1023];
+ m += t;
+ mantissahi = m >> 64;
+ mantissalo = m;
+ }
+#endif
else
{
FP_SET_EXCEPTION (FP_EX_INVALID
@@ -199,7 +249,7 @@ __bid_fixtdbitint (UBILtype *r, SItype r
if (res_limbs + low_zeros > rn && resv[BITINT_END (0, res_limbs - 1)])
goto ovf_ex;
if ((arprec % BIL_TYPE_SIZE) != 0
- && (resv[BITINT_END (rn - res_limbs, rn - 1) - low_zeros]
+ && (resv[BITINT_END (res_limbs + low_zeros - rn, rn - 1 - low_zeros)]
& ((UBILtype) -1 << (arprec % BIL_TYPE_SIZE))) != 0)
goto ovf_ex;
min_limbs = rn - low_zeros;
--- libgcc/soft-fp/fixsdti.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/fixsdti.c 2025-05-20 12:37:41.202712881 +0200
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixsdbitint __dpd_fixsdbitint
+#define __bid_fixsdti __dpd_fixsdti
+#endif
extern void __bid_fixsdbitint (UBILtype *, SItype, _Decimal32);
extern TItype __bid_fixsdti (_Decimal32);
--- libgcc/soft-fp/fixddti.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/fixddti.c 2025-05-20 12:37:24.716938294 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal64 to 128bit signed integer.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixddbitint __dpd_fixddbitint
+#define __bid_fixddti __dpd_fixddti
+#endif
extern void __bid_fixddbitint (UBILtype *, SItype, _Decimal64);
extern TItype __bid_fixddti (_Decimal64);
--- libgcc/soft-fp/fixtdti.c.jj 2023-11-09 09:13:13.578963846 +0100
+++ libgcc/soft-fp/fixtdti.c 2025-05-20 12:38:05.598379297 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal128 to 128bit signed integer.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixtdbitint __dpd_fixtdbitint
+#define __bid_fixtdti __dpd_fixtdti
+#endif
extern void __bid_fixtdbitint (UBILtype *, SItype, _Decimal128);
extern TItype __bid_fixtdti (_Decimal128);
--- libgcc/soft-fp/fixunssdti.c.jj 2023-11-09 09:13:13.578963846 +0100
+++ libgcc/soft-fp/fixunssdti.c 2025-05-20 12:38:47.574805325 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal32 to 128bit unsigned integer.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixsdbitint __dpd_fixsdbitint
+#define __bid_fixunssdti __dpd_fixunssdti
+#endif
extern void __bid_fixsdbitint (UBILtype *, SItype, _Decimal32);
extern UTItype __bid_fixunssdti (_Decimal32);
--- libgcc/soft-fp/fixunsddti.c.jj 2023-11-09 09:13:13.578963846 +0100
+++ libgcc/soft-fp/fixunsddti.c 2025-05-20 12:38:29.601051095 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal64 to 128bit unsigned integer.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixddbitint __dpd_fixddbitint
+#define __bid_fixunsddti __dpd_fixunsddti
+#endif
extern void __bid_fixddbitint (UBILtype *, SItype, _Decimal64);
extern UTItype __bid_fixunsddti (_Decimal64);
--- libgcc/soft-fp/fixunstdti.c.jj 2023-11-09 09:13:13.578963846 +0100
+++ libgcc/soft-fp/fixunstdti.c 2025-05-20 12:39:05.223564004 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert _Decimal128 to 128bit unsigned integer.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_fixtdbitint __dpd_fixtdbitint
+#define __bid_fixunstdti __dpd_fixunstdti
+#endif
extern void __bid_fixtdbitint (UBILtype *, SItype, _Decimal128);
extern UTItype __bid_fixunstdti (_Decimal128);
--- libgcc/soft-fp/floatbitintsd.c.jj 2024-02-10 12:51:25.913534391 +0100
+++ libgcc/soft-fp/floatbitintsd.c 2025-05-20 14:24:13.801466134 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a _BitInt to _Decimal32.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintsd __dpd_floatbitintsd
+#endif
extern _Decimal32 __bid_floatbitintsd (const UBILtype *, SItype);
_Decimal32
@@ -51,7 +54,11 @@ __bid_floatbitintsd (const UBILtype *i,
}
if (iprec < 0)
{
- SItype n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
+ SItype n;
+ if (msb == ~(UBILtype) 0)
+ n = 1;
+ else
+ n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
aiprec = (in - 1) * BIL_TYPE_SIZE + n;
}
else if (msb == 0)
@@ -117,7 +124,8 @@ __bid_floatbitintsd (const UBILtype *i,
}
else
{
- __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0), i,
+ __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0),
+ i + BITINT_END (1, 0),
(in - 1) * sizeof (UBILtype));
buf[BITINT_END (q_limbs - in, in - 1)] = msb;
if (iprec < 0)
@@ -210,6 +218,7 @@ __bid_floatbitintsd (const UBILtype *i,
}
exponent += 101;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
if (mantissa >= (USItype) 0x800000)
u.u = (((((iprec < 0) << 2) | (USItype) 3) << 29)
| (((USItype) exponent) << 21)
@@ -218,10 +227,34 @@ __bid_floatbitintsd (const UBILtype *i,
u.u = ((((USItype) (iprec < 0)) << 31)
| (((USItype) exponent) << 23)
| mantissa);
+#else
+ u.u = mantissa;
+ mantissa = __dpd_b2dbitint[u.u % 1000];
+ u.u /= 1000;
+ mantissa |= ((USItype) __dpd_b2dbitint[u.u % 1000]) << 10;
+ u.u /= 1000;
+ if (u.u >= 8)
+ u.u = (((((iprec < 0) << 2) | (USItype) 3) << 29)
+ | (((USItype) exponent & (3 << 6)) << 21)
+ | ((u.u & 1) << 26)
+ | (((USItype) exponent & 63) << 20)
+ | mantissa);
+ else
+ u.u = ((((USItype) (iprec < 0)) << 31)
+ | (((USItype) exponent & (3 << 6)) << 23)
+ | (u.u << 26)
+ | (((USItype) exponent & 63) << 20)
+ | mantissa);
+#endif
if (inexact)
{
ui.u = ((((USItype) (iprec < 0)) << 31)
+#ifdef ENABLE_DECIMAL_BID_FORMAT
| (((USItype) (exponent - 1)) << 23)
+#else
+ | (((USItype) (exponent - 1) & (3 << 6)) << 23)
+ | (((USItype) (exponent - 1) & 63) << 20)
+#endif
| (inexact + 3));
__asm ("" : "+g" (u.d));
__asm ("" : "+g" (ui.d));
--- libgcc/soft-fp/floatbitintdd.c.jj 2024-02-10 12:51:25.894534653 +0100
+++ libgcc/soft-fp/floatbitintdd.c 2025-05-20 14:23:57.903680829 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a _BitInt to _Decimal64.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintdd __dpd_floatbitintdd
+#endif
extern _Decimal64 __bid_floatbitintdd (const UBILtype *, SItype);
_Decimal64
@@ -51,7 +54,11 @@ __bid_floatbitintdd (const UBILtype *i,
}
if (iprec < 0)
{
- SItype n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
+ SItype n;
+ if (msb == ~(UBILtype) 0)
+ n = 1;
+ else
+ n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
aiprec = (in - 1) * BIL_TYPE_SIZE + n;
}
else if (msb == 0)
@@ -118,7 +125,8 @@ __bid_floatbitintdd (const UBILtype *i,
}
else
{
- __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0), i,
+ __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0),
+ i + BITINT_END (1, 0),
(in - 1) * sizeof (UBILtype));
buf[BITINT_END (q_limbs - in, in - 1)] = msb;
if (iprec < 0)
@@ -239,6 +247,7 @@ __bid_floatbitintdd (const UBILtype *i,
}
exponent += 398;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
if (mantissa >= (UDItype) 0x20000000000000)
u.u = (((((iprec < 0) << 2) | (UDItype) 3) << 61)
| (((UDItype) exponent) << 51)
@@ -247,10 +256,40 @@ __bid_floatbitintdd (const UBILtype *i,
u.u = ((((UDItype) (iprec < 0)) << 63)
| (((UDItype) exponent) << 53)
| mantissa);
+#else
+ u.u = mantissa;
+ mantissa = __dpd_b2dbitint[u.u % 1000];
+ u.u /= 1000;
+ mantissa |= ((UDItype) __dpd_b2dbitint[u.u % 1000]) << 10;
+ u.u /= 1000;
+ mantissa |= ((UDItype) __dpd_b2dbitint[u.u % 1000]) << 20;
+ u.u /= 1000;
+ mantissa |= ((UDItype) __dpd_b2dbitint[u.u % 1000]) << 30;
+ u.u /= 1000;
+ mantissa |= ((UDItype) __dpd_b2dbitint[u.u % 1000]) << 40;
+ u.u /= 1000;
+ if (u.u >= 8)
+ u.u = (((((iprec < 0) << 2) | (UDItype) 3) << 61)
+ | (((UDItype) exponent & (3 << 8)) << 51)
+ | ((u.u & 1) << 58)
+ | (((UDItype) exponent & 255) << 50)
+ | mantissa);
+ else
+ u.u = ((((UDItype) (iprec < 0)) << 63)
+ | (((UDItype) exponent & (3 << 8)) << 53)
+ | (u.u << 58)
+ | (((UDItype) exponent & 255) << 50)
+ | mantissa);
+#endif
if (inexact)
{
ui.u = ((((UDItype) (iprec < 0)) << 63)
+#ifdef ENABLE_DECIMAL_BID_FORMAT
| (((UDItype) (exponent - 1)) << 53)
+#else
+ | (((UDItype) (exponent - 1) & (3 << 8)) << 53)
+ | (((UDItype) (exponent - 1) & 255) << 50)
+#endif
| (inexact + 3));
__asm ("" : "+g" (u.d));
__asm ("" : "+g" (ui.d));
--- libgcc/soft-fp/floatbitinttd.c.jj 2024-02-10 12:51:25.913534391 +0100
+++ libgcc/soft-fp/floatbitinttd.c 2025-05-20 14:24:26.080300315 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a _BitInt to _Decimal128.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,9 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#ifdef __BITINT_MAXWIDTH__
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitinttd __dpd_floatbitinttd
+#endif
extern _Decimal128 __bid_floatbitinttd (const UBILtype *, SItype);
_Decimal128
@@ -51,7 +54,11 @@ __bid_floatbitinttd (const UBILtype *i,
}
if (iprec < 0)
{
- SItype n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
+ SItype n;
+ if (msb == ~(UBILtype) 0)
+ n = 1;
+ else
+ n = sizeof (0ULL) * __CHAR_BIT__ + 1 - __builtin_clzll (~msb);
aiprec = (in - 1) * BIL_TYPE_SIZE + n;
}
else if (msb == 0)
@@ -117,7 +124,8 @@ __bid_floatbitinttd (const UBILtype *i,
}
else
{
- __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0), i,
+ __builtin_memcpy (buf + BITINT_END (q_limbs - in + 1, 0),
+ i + BITINT_END (1, 0),
(in - 1) * sizeof (UBILtype));
buf[BITINT_END (q_limbs - in, in - 1)] = msb;
if (iprec < 0)
@@ -253,15 +261,69 @@ __bid_floatbitinttd (const UBILtype *i,
}
exponent += 6176;
+#ifdef ENABLE_DECIMAL_BID_FORMAT
u.u[__FLOAT_WORD_ORDER__ != __ORDER_BIG_ENDIAN__]
= ((((UDItype) (iprec < 0)) << 63)
| (((UDItype) exponent) << 49)
| mantissahi);
+#else
+#ifdef __SIZEOF_INT128__
+ unsigned __int128 m = mantissahi;
+#else
+ unsigned _BitInt(128) m = mantissahi;
+#endif
+ m = (m << 64) | mantissalo;
+ u.u[0] = m / 1000000000000000ULL;
+ u.u[1] = m % 1000000000000000ULL;
+ mantissalo = __dpd_b2dbitint[u.u[1] % 1000];
+ u.u[1] /= 1000;
+ mantissalo |= ((UDItype) __dpd_b2dbitint[u.u[1] % 1000]) << 10;
+ u.u[1] /= 1000;
+ mantissalo |= ((UDItype) __dpd_b2dbitint[u.u[1] % 1000]) << 20;
+ u.u[1] /= 1000;
+ mantissalo |= ((UDItype) __dpd_b2dbitint[u.u[1] % 1000]) << 30;
+ u.u[1] /= 1000;
+ mantissalo |= ((UDItype) __dpd_b2dbitint[u.u[1] % 1000]) << 40;
+ mantissalo |= ((UDItype) __dpd_b2dbitint[u.u[0] % 1000]) << 50;
+ u.u[0] /= 1000;
+ mantissahi = __dpd_b2dbitint[u.u[0] % 1000];
+ u.u[0] /= 1000;
+ mantissalo |= mantissahi << 60;
+ mantissahi >>= 4;
+ mantissahi |= ((UDItype) __dpd_b2dbitint[u.u[0] % 1000]) << 6;
+ u.u[0] /= 1000;
+ mantissahi |= ((UDItype) __dpd_b2dbitint[u.u[0] % 1000]) << 16;
+ u.u[0] /= 1000;
+ mantissahi |= ((UDItype) __dpd_b2dbitint[u.u[0] % 1000]) << 26;
+ u.u[0] /= 1000;
+ mantissahi |= ((UDItype) __dpd_b2dbitint[u.u[0] % 1000]) << 36;
+ u.u[0] /= 1000;
+ if (u.u[0] >= 8)
+ u.u[__FLOAT_WORD_ORDER__ != __ORDER_BIG_ENDIAN__]
+ = (((((iprec < 0) << 2) | (UDItype) 3) << 61)
+ | (((UDItype) exponent & (3 << 12)) << 47)
+ | ((u.u[0] & 1) << 58)
+ | (((UDItype) exponent & 4095) << 46)
+ | mantissahi);
+ else
+ u.u[__FLOAT_WORD_ORDER__ != __ORDER_BIG_ENDIAN__]
+ = ((((UDItype) (iprec < 0)) << 63)
+ | (((UDItype) exponent & (3 << 12)) << 49)
+ | (u.u[0] << 58)
+ | (((UDItype) exponent & 4095) << 46)
+ | mantissahi);
+#endif
u.u[__FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__] = mantissalo;
if (inexact)
{
ui.u[__FLOAT_WORD_ORDER__ != __ORDER_BIG_ENDIAN__]
+#ifdef ENABLE_DECIMAL_BID_FORMAT
= (((UDItype) (iprec < 0)) << 63) | (((UDItype) exponent - 1) << 49);
+#else
+ = ((((UDItype) (iprec < 0)) << 63)
+ | ((((UDItype) exponent - 1) & (3 << 12)) << 49)
+ | ((((UDItype) exponent - 1) & 4095) << 46));
+#endif
ui.u[__FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__] = inexact + 3;
__asm ("" : "+g" (u.d));
__asm ("" : "+g" (ui.d));
--- libgcc/soft-fp/floattisd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floattisd.c 2025-05-20 12:32:47.993722112 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit signed integer to _Decimal32.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintsd __dpd_floatbitintsd
+#define __bid_floattisd __dpd_floattisd
+#endif
extern _Decimal32 __bid_floatbitintsd (const UBILtype *, SItype);
extern _Decimal32 __bid_floattisd (TItype);
--- libgcc/soft-fp/floattidd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floattidd.c 2025-05-20 12:32:18.550124716 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit signed integer to _Decimal64.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintdd __dpd_floatbitintdd
+#define __bid_floattidd __dpd_floattidd
+#endif
extern _Decimal64 __bid_floatbitintdd (const UBILtype *, SItype);
extern _Decimal64 __bid_floattidd (TItype);
--- libgcc/soft-fp/floattitd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floattitd.c 2025-05-20 12:33:13.165377928 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit signed integer to _Decimal128.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitinttd __dpd_floatbitinttd
+#define __bid_floattitd __dpd_floattitd
+#endif
extern _Decimal128 __bid_floatbitinttd (const UBILtype *, SItype);
extern _Decimal128 __bid_floattitd (TItype);
--- libgcc/soft-fp/floatuntisd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floatuntisd.c 2025-05-20 12:35:09.116792447 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit unsigned integer to _Decimal32.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintsd __dpd_floatbitintsd
+#define __bid_floatunstisd __dpd_floatunstisd
+#endif
extern _Decimal32 __bid_floatbitintsd (const UBILtype *, SItype);
extern _Decimal32 __bid_floatunstisd (UTItype);
--- libgcc/soft-fp/floatuntidd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floatuntidd.c 2025-05-20 12:34:48.160078998 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit unsigned integer to _Decimal64.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitintdd __dpd_floatbitintdd
+#define __bid_floatunstidd __dpd_floatunstidd
+#endif
extern _Decimal64 __bid_floatbitintdd (const UBILtype *, SItype);
extern _Decimal64 __bid_floatunstidd (UTItype);
--- libgcc/soft-fp/floatuntitd.c.jj 2023-11-09 09:13:13.577963860 +0100
+++ libgcc/soft-fp/floatuntitd.c 2025-05-20 12:35:28.230531090 +0200
@@ -1,7 +1,7 @@
/* Software floating-point emulation.
Convert a 128bit unsigned integer to _Decimal128.
- Copyright (C) 2023 Free Software Foundation, Inc.
+ Copyright (C) 2023-2025 Free Software Foundation, Inc.
This file is part of GCC.
@@ -28,6 +28,10 @@ see the files COPYING3 and COPYING.RUNTI
#include "bitint.h"
#if defined(__BITINT_MAXWIDTH__) && defined(__SIZEOF_INT128__)
+#ifndef ENABLE_DECIMAL_BID_FORMAT
+#define __bid_floatbitinttd __dpd_floatbitinttd
+#define __bid_floatunstitd __dpd_floatunstitd
+#endif
extern _Decimal128 __bid_floatbitinttd (const UBILtype *, SItype);
extern _Decimal128 __bid_floatunstitd (UTItype);
Jakub
