Training courses

Kernel and Embedded Linux

Bootlin training courses

Embedded Linux, kernel,
Yocto Project, Buildroot, real-time,
graphics, boot time, debugging...

Bootlin logo

Elixir Cross Referencer

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
/*	$NetBSD: ntp_leapsec.c,v 1.7 2020/05/25 20:47:25 christos Exp $	*/

/*
 * ntp_leapsec.c - leap second processing for NTPD
 *
 * Written by Juergen Perlinger (perlinger@ntp.org) for the NTP project.
 * The contents of 'html/copyright.html' apply.
 * ----------------------------------------------------------------------
 * This is an attempt to get the leap second handling into a dedicated
 * module to make the somewhat convoluted logic testable.
 */

#include <config.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <ctype.h>

#include "ntp_types.h"
#include "ntp_fp.h"
#include "ntp_stdlib.h"
#include "ntp_calendar.h"
#include "ntp_leapsec.h"
#include "ntp.h"
#include "vint64ops.h"
#include "lib_strbuf.h"

#include "isc/sha1.h"

static const char * const logPrefix = "leapsecond file";

/* ---------------------------------------------------------------------
 * GCC is rather sticky with its 'const' attribute. We have to do it more
 * explicit than with a cast if we want to get rid of a CONST qualifier.
 * Greetings from the PASCAL world, where casting was only possible via
 * untagged unions...
 */
static inline void*
noconst(
	const void* ptr
	)
{
	union {
		const void * cp;
		void *       vp;
	} tmp;
	tmp.cp = ptr;
	return tmp.vp;
}

/* ---------------------------------------------------------------------
 * Our internal data structure
 */
#define MAX_HIST 10	/* history of leap seconds */

struct leap_info {
	vint64   ttime;	/* transition time (after the step, ntp scale) */
	uint32_t stime;	/* schedule limit (a month before transition)  */
	int16_t  taiof;	/* TAI offset on and after the transition      */
	uint8_t  dynls; /* dynamic: inserted on peer/clock request     */
};
typedef struct leap_info leap_info_t;

struct leap_head {
	vint64   update; /* time of information update                 */
	vint64   expire; /* table expiration time                      */
	uint16_t size;	 /* number of infos in table	               */
	int16_t  base_tai;	/* total leaps before first entry      */
	int16_t  this_tai;	/* current TAI offset	               */
	int16_t  next_tai;	/* TAI offset after 'when'             */
	vint64   dtime;	 /* due time (current era end)                 */
	vint64   ttime;	 /* nominal transition time (next era start)   */
	vint64   stime;	 /* schedule time (when we take notice)        */
	vint64   ebase;	 /* base time of this leap era                 */
	uint8_t  dynls;	 /* next leap is dynamic (by peer request)     */
};
typedef struct leap_head leap_head_t;

struct leap_table {
	leap_signature_t lsig;
	leap_head_t	 head;
	leap_info_t  	 info[MAX_HIST];
};

/* Where we store our tables */
static leap_table_t _ltab[2], *_lptr;
static int/*BOOL*/  _electric;

/* Forward decls of local helpers */
static int    add_range(leap_table_t*, const leap_info_t*);
static char * get_line(leapsec_reader, void*, char*, size_t);
static char * skipws(const char*);
static int    parsefail(const char * cp, const char * ep);
static void   reload_limits(leap_table_t*, const vint64*);
static void   fetch_leap_era(leap_era_t*, const leap_table_t*,
			     const vint64*);
static int    betweenu32(uint32_t, uint32_t, uint32_t);
static void   reset_times(leap_table_t*);
static int    leapsec_add(leap_table_t*, const vint64*, int);
static int    leapsec_raw(leap_table_t*, const vint64 *, int, int);
static const char * lstostr(const vint64 * ts);

/* =====================================================================
 * Get & Set the current leap table
 */

/* ------------------------------------------------------------------ */
leap_table_t *
leapsec_get_table(
	int alternate)
{
	leap_table_t *p1, *p2;

	p1 = _lptr;
	if (p1 == &_ltab[0]) {
		p2 = &_ltab[1];
	} else if (p1 == &_ltab[1]) {
		p2 = &_ltab[0];
	} else {
		p1 = &_ltab[0];
		p2 = &_ltab[1];
		reset_times(p1);
		reset_times(p2);
		_lptr = p1;
	}
	if (alternate) {
		memcpy(p2, p1, sizeof(leap_table_t));
		p1 = p2;
	}

	return p1;
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_set_table(
	leap_table_t * pt)
{
	if (pt == &_ltab[0] || pt == &_ltab[1])
		_lptr = pt;
	return _lptr == pt;
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_electric(
	int/*BOOL*/ on)
{
	int res = _electric;
	if (on < 0)
		return res;

	_electric = (on != 0);
	if (_electric == res)
		return res;

	if (_lptr == &_ltab[0] || _lptr == &_ltab[1])
		reset_times(_lptr);

	return res;
}

/* =====================================================================
 * API functions that operate on tables
 */

/* ---------------------------------------------------------------------
 * Clear all leap second data. Use it for init & cleanup
 */
void
leapsec_clear(
	leap_table_t * pt)
{
	memset(&pt->lsig, 0, sizeof(pt->lsig));
	memset(&pt->head, 0, sizeof(pt->head));
	reset_times(pt);
}

/* ---------------------------------------------------------------------
 * Load a leap second file and check expiration on the go
 */
int/*BOOL*/
leapsec_load(
	leap_table_t * pt  ,
	leapsec_reader func,
	void *         farg,
	int            use_build_limit)
{
	char   *cp, *ep, linebuf[50];
	vint64 ttime, limit;
	long   taiof;
	struct calendar build;

	leapsec_clear(pt);
	if (use_build_limit && ntpcal_get_build_date(&build)) {
		/* don't prune everything -- permit the last 10yrs
		 * before build.
		 */
		build.year -= 10;
		limit = ntpcal_date_to_ntp64(&build);
	} else {
		memset(&limit, 0, sizeof(limit));
	}

	while (get_line(func, farg, linebuf, sizeof(linebuf))) {
		cp = linebuf;
		if (*cp == '#') {
			cp++;
			if (*cp == '@') {
				cp = skipws(cp+1);
				pt->head.expire = strtouv64(cp, &ep, 10);
				if (parsefail(cp, ep))
					goto fail_read;
				pt->lsig.etime = pt->head.expire.D_s.lo;
			} else if (*cp == '$') {
				cp = skipws(cp+1);
				pt->head.update = strtouv64(cp, &ep, 10);
				if (parsefail(cp, ep))
					goto fail_read;
			}
		} else if (isdigit((u_char)*cp)) {
			ttime = strtouv64(cp, &ep, 10);
			if (parsefail(cp, ep))
				goto fail_read;
			cp = skipws(ep);
			taiof = strtol(cp, &ep, 10);
			if (   parsefail(cp, ep)
			    || taiof > SHRT_MAX || taiof < SHRT_MIN)
				goto fail_read;
			if (ucmpv64(&ttime, &limit) >= 0) {
				if (!leapsec_raw(pt, &ttime,
						 taiof, FALSE))
					goto fail_insn;
			} else {
				pt->head.base_tai = (int16_t)taiof;
			}
			pt->lsig.ttime = ttime.D_s.lo;
			pt->lsig.taiof = (int16_t)taiof;
		}
	}
	return TRUE;

fail_read:
	errno = EILSEQ;
fail_insn:
	leapsec_clear(pt);
	return FALSE;
}

/* ---------------------------------------------------------------------
 * Dump a table in human-readable format. Use 'fprintf' and a FILE
 * pointer if you want to get it printed into a stream.
 */
void
leapsec_dump(
	const leap_table_t * pt  ,
	leapsec_dumper       func,
	void *               farg)
{
	int             idx;
	vint64          ts;
	struct calendar atb, ttb;

	ntpcal_ntp64_to_date(&ttb, &pt->head.expire);
	(*func)(farg, "leap table (%u entries) expires at %04u-%02u-%02u:\n",
		pt->head.size,
		ttb.year, ttb.month, ttb.monthday);
	idx = pt->head.size;
	while (idx-- != 0) {
		ts = pt->info[idx].ttime;
		ntpcal_ntp64_to_date(&ttb, &ts);
		ts = subv64u32(&ts, pt->info[idx].stime);
		ntpcal_ntp64_to_date(&atb, &ts);

		(*func)(farg, "%04u-%02u-%02u [%c] (%04u-%02u-%02u) - %d\n",
			ttb.year, ttb.month, ttb.monthday,
			"-*"[pt->info[idx].dynls != 0],
			atb.year, atb.month, atb.monthday,
			pt->info[idx].taiof);
	}
}

/* =====================================================================
 * usecase driven API functions
 */

int/*BOOL*/
leapsec_query(
	leap_result_t * qr   ,
	uint32_t        ts32 ,
	const time_t *  pivot)
{
	leap_table_t *   pt;
	vint64           ts64, last, next;
	uint32_t         due32;
	int              fired;

	/* preset things we use later on... */
	fired = FALSE;
	ts64  = ntpcal_ntp_to_ntp(ts32, pivot);
	pt    = leapsec_get_table(FALSE);
	memset(qr, 0, sizeof(leap_result_t));

	if (ucmpv64(&ts64, &pt->head.ebase) < 0) {
		/* Most likely after leap frame reset. Could also be a
		 * backstep of the system clock. Anyway, get the new
		 * leap era frame.
		 */
		reload_limits(pt, &ts64);
	} else if (ucmpv64(&ts64, &pt->head.dtime) >= 0) {
		/* Boundary crossed in forward direction. This might
		 * indicate a leap transition, so we prepare for that
		 * case.
		 *
		 * Some operations below are actually NOPs in electric
		 * mode, but having only one code path that works for
		 * both modes is easier to maintain.
		 *
		 * There's another quirk we must keep looking out for:
		 * If we just stepped the clock, the step might have
		 * crossed a leap boundary. As with backward steps, we
		 * do not want to raise the 'fired' event in that case.
		 * So we raise the 'fired' event only if we're close to
		 * the transition and just reload the limits otherwise.
		 */
		last = addv64i32(&pt->head.dtime, 3); /* get boundary */
		if (ucmpv64(&ts64, &last) >= 0) {
			/* that was likely a query after a step */
			reload_limits(pt, &ts64);
		} else {
			/* close enough for deeper examination */
			last = pt->head.ttime;
			qr->warped = (int16_t)(last.D_s.lo -
					       pt->head.dtime.D_s.lo);
			next = addv64i32(&ts64, qr->warped);
			reload_limits(pt, &next);
			fired = ucmpv64(&pt->head.ebase, &last) == 0;
			if (fired) {
				ts64 = next;
				ts32 = next.D_s.lo;
			} else {
				qr->warped = 0;
			}
		}
	}

	qr->tai_offs = pt->head.this_tai;
	qr->ebase    = pt->head.ebase;
	qr->ttime    = pt->head.ttime;

	/* If before the next scheduling alert, we're done. */
	if (ucmpv64(&ts64, &pt->head.stime) < 0)
		return fired;

	/* now start to collect the remaining data */
	due32 = pt->head.dtime.D_s.lo;

	qr->tai_diff  = pt->head.next_tai - pt->head.this_tai;
	qr->ddist     = due32 - ts32;
	qr->dynamic   = pt->head.dynls;
	qr->proximity = LSPROX_SCHEDULE;

	/* if not in the last day before transition, we're done. */
	if (!betweenu32(due32 - SECSPERDAY, ts32, due32))
		return fired;

	qr->proximity = LSPROX_ANNOUNCE;
	if (!betweenu32(due32 - 10, ts32, due32))
		return fired;

	/* The last 10s before the transition. Prepare for action! */
	qr->proximity = LSPROX_ALERT;
	return fired;
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_query_era(
	leap_era_t *   qr   ,
	uint32_t       ntpts,
	const time_t * pivot)
{
	const leap_table_t * pt;
	vint64               ts64;

	pt   = leapsec_get_table(FALSE);
	ts64 = ntpcal_ntp_to_ntp(ntpts, pivot);
	fetch_leap_era(qr, pt, &ts64);
	return TRUE;
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_frame(
        leap_result_t *qr)
{
	const leap_table_t * pt;

        memset(qr, 0, sizeof(leap_result_t));
	pt = leapsec_get_table(FALSE);

	qr->tai_offs = pt->head.this_tai;
	qr->tai_diff = pt->head.next_tai - pt->head.this_tai;
	qr->ebase    = pt->head.ebase;
	qr->ttime    = pt->head.ttime;
	qr->dynamic  = pt->head.dynls;

	return ucmpv64(&pt->head.ttime, &pt->head.stime) >= 0;
}

/* ------------------------------------------------------------------ */
/* Reset the current leap frame */
void
leapsec_reset_frame(void)
{
	reset_times(leapsec_get_table(FALSE));
}

/* ------------------------------------------------------------------ */
/* load a file from a FILE pointer. Note: If hcheck is true, load
 * only after successful signature check. The stream must be seekable
 * or this will fail.
 */
int/*BOOL*/
leapsec_load_stream(
	FILE       * ifp  ,
	const char * fname,
	int/*BOOL*/  logall,
	int/*BOOL*/  vhash)
{
	leap_table_t *pt;
	int           rcheck;

	if (NULL == fname)
		fname = "<unknown>";

	if (vhash) {
		rcheck = leapsec_validate((leapsec_reader)getc, ifp);
		if (logall)
			switch (rcheck)
			{
			case LSVALID_GOODHASH:
				msyslog(LOG_NOTICE, "%s ('%s'): good hash signature",
					logPrefix, fname);
				break;
				
			case LSVALID_NOHASH:
				msyslog(LOG_ERR, "%s ('%s'): no hash signature",
					logPrefix, fname);
				break;
			case LSVALID_BADHASH:
				msyslog(LOG_ERR, "%s ('%s'): signature mismatch",
					logPrefix, fname);
				break;
			case LSVALID_BADFORMAT:
				msyslog(LOG_ERR, "%s ('%s'): malformed hash signature",
					logPrefix, fname);
				break;
			default:
				msyslog(LOG_ERR, "%s ('%s'): unknown error code %d",
					logPrefix, fname, rcheck);
				break;
			}
		if (rcheck < 0)
			return FALSE;
		rewind(ifp);
	}
	pt = leapsec_get_table(TRUE);
	if (!leapsec_load(pt, (leapsec_reader)getc, ifp, TRUE)) {
		switch (errno) {
		case EINVAL:
			msyslog(LOG_ERR, "%s ('%s'): bad transition time",
				logPrefix, fname);
			break;
		case ERANGE:
			msyslog(LOG_ERR, "%s ('%s'): times not ascending",
				logPrefix, fname);
			break;
		default:
			msyslog(LOG_ERR, "%s ('%s'): parsing error",
				logPrefix, fname);
			break;
		}
		return FALSE;
	}

	if (pt->head.size)
		msyslog(LOG_NOTICE, "%s ('%s'): loaded, expire=%s last=%s ofs=%d",
			logPrefix, fname, lstostr(&pt->head.expire),
			lstostr(&pt->info[0].ttime), pt->info[0].taiof);
	else
		msyslog(LOG_NOTICE,
			"%s ('%s'): loaded, expire=%s ofs=%d (no entries after build date)",
			logPrefix, fname, lstostr(&pt->head.expire),
			pt->head.base_tai);

	return leapsec_set_table(pt);
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_load_file(
	const char  * fname,
	struct stat * sb_old,
	int/*BOOL*/   force,
	int/*BOOL*/   logall,
	int/*BOOL*/   vhash)
{
	FILE       * fp;
	struct stat  sb_new;
	int          rc;

	/* just do nothing if there is no leap file */
	if ( !(fname && *fname) )
		return FALSE;

	/* try to stat the leapfile */
	if (0 != stat(fname, &sb_new)) {
		if (logall)
			msyslog(LOG_ERR, "%s ('%s'): stat failed: %m",
				logPrefix, fname);
		return FALSE;
	}

	/* silently skip to postcheck if no new file found */
	if (NULL != sb_old) {
		if (!force
		 && sb_old->st_mtime == sb_new.st_mtime
		 && sb_old->st_ctime == sb_new.st_ctime
		   )
			return FALSE;
		*sb_old = sb_new;
	}

	/* try to open the leap file, complain if that fails
	 *
	 * [perlinger@ntp.org]
	 * coverity raises a TOCTOU (time-of-check/time-of-use) issue
	 * here, which is not entirely helpful: While there is indeed a
	 * possible race condition between the 'stat()' call above and
	 * the 'fopen)' call below, I intentionally want to omit the
	 * overhead of opening the file and calling 'fstat()', because
	 * in most cases the file would have be to closed anyway without
	 * reading the contents.  I chose to disable the coverity
	 * warning instead.
	 *
	 * So unless someone comes up with a reasonable argument why
	 * this could be a real issue, I'll just try to silence coverity
	 * on that topic.
	 */
	/* coverity[toctou] */
	if ((fp = fopen(fname, "r")) == NULL) {
		if (logall)
			msyslog(LOG_ERR,
				"%s ('%s'): open failed: %m",
				logPrefix, fname);
		return FALSE;
	}

	rc = leapsec_load_stream(fp, fname, logall, vhash);
	fclose(fp);
	return rc;
}

/* ------------------------------------------------------------------ */
void
leapsec_getsig(
	leap_signature_t * psig)
{
	const leap_table_t * pt;

	pt = leapsec_get_table(FALSE);
	memcpy(psig, &pt->lsig, sizeof(leap_signature_t));
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_expired(
	uint32_t       when,
	const time_t * tpiv)
{
	const leap_table_t * pt;
	vint64 limit;

	pt = leapsec_get_table(FALSE);
	limit = ntpcal_ntp_to_ntp(when, tpiv);
	return ucmpv64(&limit, &pt->head.expire) >= 0;
}

/* ------------------------------------------------------------------ */
int32_t
leapsec_daystolive(
	uint32_t       when,
	const time_t * tpiv)
{
	const leap_table_t * pt;
	vint64 limit;

	pt = leapsec_get_table(FALSE);
	limit = ntpcal_ntp_to_ntp(when, tpiv);
	limit = subv64(&pt->head.expire, &limit);
	return ntpcal_daysplit(&limit).hi;
}

/* ------------------------------------------------------------------ */
#if 0 /* currently unused -- possibly revived later */
int/*BOOL*/
leapsec_add_fix(
	int            total,
	uint32_t       ttime,
	uint32_t       etime,
	const time_t * pivot)
{
	time_t         tpiv;
	leap_table_t * pt;
	vint64         tt64, et64;

	if (pivot == NULL) {
		time(&tpiv);
		pivot = &tpiv;
	}

	et64 = ntpcal_ntp_to_ntp(etime, pivot);
	tt64 = ntpcal_ntp_to_ntp(ttime, pivot);
	pt   = leapsec_get_table(TRUE);

	if (   ucmpv64(&et64, &pt->head.expire) <= 0
	   || !leapsec_raw(pt, &tt64, total, FALSE) )
		return FALSE;

	pt->lsig.etime = etime;
	pt->lsig.ttime = ttime;
	pt->lsig.taiof = (int16_t)total;

	pt->head.expire = et64;

	return leapsec_set_table(pt);
}
#endif

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_add_dyn(
	int            insert,
	uint32_t       ntpnow,
	const time_t * pivot )
{
	leap_table_t * pt;
	vint64         now64;

	pt = leapsec_get_table(TRUE);
	now64 = ntpcal_ntp_to_ntp(ntpnow, pivot);
	return (   leapsec_add(pt, &now64, (insert != 0))
		&& leapsec_set_table(pt));
}

/* ------------------------------------------------------------------ */
int/*BOOL*/
leapsec_autokey_tai(
	int            tai_offset,
	uint32_t       ntpnow    ,
	const time_t * pivot     )
{
	leap_table_t * pt;
	leap_era_t     era;
	vint64         now64;
	int            idx;

	(void)tai_offset;
	pt = leapsec_get_table(FALSE);

	/* Bail out if the basic offset is not zero and the putative
	 * offset is bigger than 10s. That was in 1972 -- we don't want
	 * to go back that far!
	 */
	if (pt->head.base_tai != 0 || tai_offset < 10)
		return FALSE;

	/* If there's already data in the table, check if an update is
	 * possible. Update is impossible if there are static entries
	 * (since this indicates a valid leapsecond file) or if we're
	 * too close to a leapsecond transition: We do not know on what
	 * side the transition the sender might have been, so we use a
	 * dead zone around the transition.
	 */

	/* Check for static entries */
	for (idx = 0; idx != pt->head.size; idx++)
		if ( ! pt->info[idx].dynls)
			return FALSE;

	/* get the fulll time stamp and leap era for it */
	now64 = ntpcal_ntp_to_ntp(ntpnow, pivot);
	fetch_leap_era(&era, pt, &now64);

	/* check the limits with 20s dead band */
	era.ebase = addv64i32(&era.ebase,  20);
	if (ucmpv64(&now64, &era.ebase) < 0)
		return FALSE;

	era.ttime = addv64i32(&era.ttime, -20);
	if (ucmpv64(&now64, &era.ttime) > 0)
		return FALSE;

	/* Here we can proceed. Calculate the delta update. */
	tai_offset -= era.taiof;

	/* Shift the header info offsets. */
	pt->head.base_tai += tai_offset;
	pt->head.this_tai += tai_offset;
	pt->head.next_tai += tai_offset;

	/* Shift table entry offsets (if any) */
	for (idx = 0; idx != pt->head.size; idx++)
		pt->info[idx].taiof += tai_offset;

	/* claim success... */
	return TRUE;
}


/* =====================================================================
 * internal helpers
 */

/* [internal] Reset / init the time window in the leap processor to
 * force reload on next query. Since a leap transition cannot take place
 * at an odd second, the value chosen avoids spurious leap transition
 * triggers. Making all three times equal forces a reload. Using the
 * maximum value for unsigned 64 bits makes finding the next leap frame
 * a bit easier.
 */
static void
reset_times(
	leap_table_t * pt)
{
	memset(&pt->head.ebase, 0xFF, sizeof(vint64));
	pt->head.stime = pt->head.ebase;
	pt->head.ttime = pt->head.ebase;
	pt->head.dtime = pt->head.ebase;
}

/* [internal] Add raw data to the table, removing old entries on the
 * fly. This cannot fail currently.
 */
static int/*BOOL*/
add_range(
	leap_table_t *      pt,
	const leap_info_t * pi)
{
	/* If the table is full, make room by throwing out the oldest
	 * entry. But remember the accumulated leap seconds!
	 *
	 * Setting the first entry is a bit tricky, too: Simply assuming
	 * it is an insertion is wrong if the first entry is a dynamic
	 * leap second removal. So we decide on the sign -- if the first
	 * entry has a negative offset, we assume that it is a leap
	 * second removal. In both cases the table base offset is set
	 * accordingly to reflect the decision.
	 *
	 * In practice starting with a removal can only happen if the
	 * first entry is a dynamic request without having a leap file
	 * for the history proper.
	 */
	if (pt->head.size == 0) {
		if (pi->taiof >= 0)
			pt->head.base_tai = pi->taiof - 1;
		else
			pt->head.base_tai = pi->taiof + 1;
	} else if (pt->head.size >= MAX_HIST) {
		pt->head.size     = MAX_HIST - 1;
		pt->head.base_tai = pt->info[pt->head.size].taiof;
	}

	/* make room in lower end and insert item */
	memmove(pt->info+1, pt->info, pt->head.size*sizeof(*pt->info));
	pt->info[0] = *pi;
	pt->head.size++;

	/* invalidate the cached limit data -- we might have news ;-)
	 *
	 * This blocks a spurious transition detection. OTOH, if you add
	 * a value after the last query before a leap transition was
	 * expected to occur, this transition trigger is lost. But we
	 * can probably live with that.
	 */
	reset_times(pt);
	return TRUE;
}

/* [internal] given a reader function, read characters into a buffer
 * until either EOL or EOF is reached. Makes sure that the buffer is
 * always NUL terminated, but silently truncates excessive data. The
 * EOL-marker ('\n') is *not* stored in the buffer.
 *
 * Returns the pointer to the buffer, unless EOF was reached when trying
 * to read the first character of a line.
 */
static char *
get_line(
	leapsec_reader func,
	void *         farg,
	char *         buff,
	size_t         size)
{
	int   ch;
	char *ptr;

	/* if we cannot even store the delimiter, declare failure */
	if (buff == NULL || size == 0)
		return NULL;

	ptr = buff;
	while (EOF != (ch = (*func)(farg)) && '\n' != ch)
		if (size > 1) {
			size--;
			*ptr++ = (char)ch;
		}
	/* discard trailing whitespace */
	while (ptr != buff && isspace((u_char)ptr[-1]))
		ptr--;
	*ptr = '\0';
	return (ptr == buff && ch == EOF) ? NULL : buff;
}

/* [internal] skips whitespace characters from a character buffer. */
static char *
skipws(
	const char *ptr)
{
	while (isspace((u_char)*ptr))
		ptr++;
	return (char*)noconst(ptr);
}

/* [internal] check if a strtoXYZ ended at EOL or whitespace and
 * converted something at all. Return TRUE if something went wrong.
 */
static int/*BOOL*/
parsefail(
	const char * cp,
	const char * ep)
{
	return (cp == ep)
	    || (*ep && *ep != '#' && !isspace((u_char)*ep));
}

/* [internal] reload the table limits around the given time stamp. This
 * is where the real work is done when it comes to table lookup and
 * evaluation. Some care has been taken to have correct code for dealing
 * with boundary conditions and empty tables.
 *
 * In electric mode, transition and trip time are the same. In dumb
 * mode, the difference of the TAI offsets must be taken into account
 * and trip time and transition time become different. The difference
 * becomes the warping distance when the trip time is reached.
 */
static void
reload_limits(
	leap_table_t * pt,
	const vint64 * ts)
{
	int idx;

	/* Get full time and search the true lower bound. Use a
	 * simple loop here, since the number of entries does
	 * not warrant a binary search. This also works for an empty
	 * table, so there is no shortcut for that case.
	 */
	for (idx = 0; idx != pt->head.size; idx++)
		if (ucmpv64(ts, &pt->info[idx].ttime) >= 0)
			break;

	/* get time limits with proper bound conditions. Note that the
	 * bounds of the table will be observed even if the table is
	 * empty -- no undefined condition must arise from this code.
	 */
	if (idx >= pt->head.size) {
		memset(&pt->head.ebase, 0x00, sizeof(vint64));
		pt->head.this_tai = pt->head.base_tai;
	} else {
		pt->head.ebase    = pt->info[idx].ttime;
		pt->head.this_tai = pt->info[idx].taiof;
	}
	if (--idx >= 0) {
		pt->head.next_tai = pt->info[idx].taiof;
		pt->head.dynls    = pt->info[idx].dynls;
		pt->head.ttime    = pt->info[idx].ttime;

		if (_electric)
			pt->head.dtime = pt->head.ttime;
                else
			pt->head.dtime = addv64i32(
				&pt->head.ttime,
				pt->head.next_tai - pt->head.this_tai);

		pt->head.stime = subv64u32(
			&pt->head.ttime, pt->info[idx].stime);

	} else {
		memset(&pt->head.ttime, 0xFF, sizeof(vint64));
		pt->head.stime    = pt->head.ttime;
		pt->head.dtime    = pt->head.ttime;
		pt->head.next_tai = pt->head.this_tai;
		pt->head.dynls    = 0;
	}
}

/* [internal] fetch the leap era for a given time stamp.
 * This is a cut-down version the algorithm used to reload the table
 * limits, but it does not update any global state and provides just the
 * era information for a given time stamp.
 */
static void
fetch_leap_era(
	leap_era_t         * into,
	const leap_table_t * pt  ,
	const vint64       * ts  )
{
	int idx;

	/* Simple search loop, also works with empty table. */
	for (idx = 0; idx != pt->head.size; idx++)
		if (ucmpv64(ts, &pt->info[idx].ttime) >= 0)
			break;
	/* fetch era data, keeping an eye on boundary conditions */
	if (idx >= pt->head.size) {
		memset(&into->ebase, 0x00, sizeof(vint64));
		into->taiof = pt->head.base_tai;
	} else {
		into->ebase = pt->info[idx].ttime;
		into->taiof = pt->info[idx].taiof;
	}
	if (--idx >= 0)
		into->ttime = pt->info[idx].ttime;
	else
		memset(&into->ttime, 0xFF, sizeof(vint64));
}

/* [internal] Take a time stamp and create a leap second frame for
 * it. This will schedule a leap second for the beginning of the next
 * month, midnight UTC. The 'insert' argument tells if a leap second is
 * added (!=0) or removed (==0). We do not handle multiple inserts
 * (yet?)
 *
 * Returns 1 if the insert worked, 0 otherwise. (It's not possible to
 * insert a leap second into the current history -- only appending
 * towards the future is allowed!)
 */
static int/*BOOL*/
leapsec_add(
	leap_table_t*  pt    ,
	const vint64 * now64 ,
	int            insert)
{
	vint64		ttime, starttime;
	struct calendar	fts;
	leap_info_t	li;

	/* Check against the table expiration and the latest available
	 * leap entry. Do not permit inserts, only appends, and only if
	 * the extend the table beyond the expiration!
	 */
	if (   ucmpv64(now64, &pt->head.expire) < 0
	    || (pt->head.size && ucmpv64(now64, &pt->info[0].ttime) <= 0)) {
		errno = ERANGE;
		return FALSE;
	}

	ntpcal_ntp64_to_date(&fts, now64);
	/* To guard against dangling leap flags: do not accept leap
	 * second request on the 1st hour of the 1st day of the month.
	 */
	if (fts.monthday == 1 && fts.hour == 0) {
		errno = EINVAL;
		return FALSE;
	}

	/* Ok, do the remaining calculations */
	fts.monthday = 1;
	fts.hour     = 0;
	fts.minute   = 0;
	fts.second   = 0;
	starttime = ntpcal_date_to_ntp64(&fts);
	fts.month++;
	ttime = ntpcal_date_to_ntp64(&fts);

	li.ttime = ttime;
	li.stime = ttime.D_s.lo - starttime.D_s.lo;
	li.taiof = (pt->head.size ? pt->info[0].taiof : pt->head.base_tai)
	         + (insert ? 1 : -1);
	li.dynls = 1;
	return add_range(pt, &li);
}

/* [internal] Given a time stamp for a leap insertion (the exact begin
 * of the new leap era), create new leap frame and put it into the
 * table. This is the work horse for reading a leap file and getting a
 * leap second update via authenticated network packet.
 */
int/*BOOL*/
leapsec_raw(
	leap_table_t * pt,
	const vint64 * ttime,
	int            taiof,
	int            dynls)
{
	vint64		starttime;
	struct calendar	fts;
	leap_info_t	li;

	/* Check that we either extend the table or get a duplicate of
	 * the latest entry. The latter is a benevolent overwrite with
	 * identical data and could happen if we get an autokey message
	 * that extends the lifetime of the current leapsecond table.
	 * Otherwise paranoia rulez!
	 */
	if (pt->head.size) {
		int cmp = ucmpv64(ttime, &pt->info[0].ttime);
		if (cmp == 0)
			cmp -= (taiof != pt->info[0].taiof);
		if (cmp < 0) {
			errno = ERANGE;
			return FALSE;
		}
		if (cmp == 0)
			return TRUE;
	}

	ntpcal_ntp64_to_date(&fts, ttime);
	/* If this does not match the exact month start, bail out. */
	if (fts.monthday != 1 || fts.hour || fts.minute || fts.second) {
		errno = EINVAL;
		return FALSE;
	}
	fts.month--; /* was in range 1..12, no overflow here! */
	starttime = ntpcal_date_to_ntp64(&fts);
	li.ttime = *ttime;
	li.stime = ttime->D_s.lo - starttime.D_s.lo;
	li.taiof = (int16_t)taiof;
	li.dynls = (dynls != 0);
	return add_range(pt, &li);
}

/* [internal] Do a wrap-around save range inclusion check.
 * Returns TRUE if x in [lo,hi[ (intervall open on right side) with full
 * handling of an overflow / wrap-around.
 */
static int/*BOOL*/
betweenu32(
	uint32_t lo,
	uint32_t x,
	uint32_t hi)
{
	int rc;

	if (lo <= hi)
		rc = (lo <= x) && (x < hi);
	else
		rc = (lo <= x) || (x < hi);
	return rc;
}

/* =====================================================================
 * validation stuff
 */

typedef struct {
	unsigned char hv[ISC_SHA1_DIGESTLENGTH];
} sha1_digest;

/* [internal] parse a digest line to get the hash signature
 * The NIST code creating the hash writes them out as 5 hex integers
 * without leading zeros. This makes reading them back as hex-encoded
 * BLOB impossible, because there might be less than 40 hex digits.
 *
 * The solution is to read the values back as integers, and then do the
 * byte twiddle necessary to get it into an array of 20 chars. The
 * drawback is that it permits any acceptable number syntax provided by
 * 'scanf()' and 'strtoul()', including optional signs and '0x'
 * prefixes.
 */
static int/*BOOL*/
do_leap_hash(
	sha1_digest * mac,
	char const  * cp )
{
	int wi, di, num, len;
	unsigned long tmp[5];

	memset(mac, 0, sizeof(*mac));
	num = sscanf(cp, " %lx %lx %lx %lx %lx%n",
		     &tmp[0], &tmp[1], &tmp[2], &tmp[3], &tmp[4],
		     &len);
	if (num != 5 || cp[len] > ' ')
		return FALSE;

	/* now do the byte twiddle */
	for (wi=0; wi < 5; ++wi)
		for (di=3; di >= 0; --di) {
			mac->hv[wi*4 + di] =
				(unsigned char)(tmp[wi] & 0x0FF);
			tmp[wi] >>= 8;
		}
	return TRUE;
}

/* [internal] add the digits of a data line to the hash, stopping at the
 * next hash ('#') character.
 */
static void
do_hash_data(
	isc_sha1_t * mdctx,
	char const * cp   )
{
	unsigned char  text[32]; // must be power of two!
	unsigned int   tlen =  0;
	unsigned char  ch;

	while ('\0' != (ch = *cp++) && '#' != ch)
		if (isdigit(ch)) {
			text[tlen++] = ch;
			tlen &= (sizeof(text)-1);
			if (0 == tlen)
				isc_sha1_update(
					mdctx, text, sizeof(text));
		}

	if (0 < tlen)
		isc_sha1_update(mdctx, text, tlen);
}

/* given a reader and a reader arg, calculate and validate the the hash
 * signature of a NIST leap second file.
 */
int
leapsec_validate(
	leapsec_reader func,
	void *         farg)
{
	isc_sha1_t     mdctx;
	sha1_digest    rdig, ldig; /* remote / local digests */
	char           line[50];
	int            hlseen = -1;

	isc_sha1_init(&mdctx);
	while (get_line(func, farg, line, sizeof(line))) {
		if (!strncmp(line, "#h", 2))
			hlseen = do_leap_hash(&rdig, line+2);
		else if (!strncmp(line, "#@", 2))
			do_hash_data(&mdctx, line+2);
		else if (!strncmp(line, "#$", 2))
			do_hash_data(&mdctx, line+2);
		else if (isdigit((unsigned char)line[0]))
			do_hash_data(&mdctx, line);
	}
	isc_sha1_final(&mdctx, ldig.hv);
	isc_sha1_invalidate(&mdctx);

	if (0 > hlseen)
		return LSVALID_NOHASH;
	if (0 == hlseen)
		return LSVALID_BADFORMAT;
	if (0 != memcmp(&rdig, &ldig, sizeof(sha1_digest)))
		return LSVALID_BADHASH;
	return LSVALID_GOODHASH;
}

/*
 * lstostr - prettyprint NTP seconds
 */
static const char *
lstostr(
	const vint64 * ts)
{
	char *		buf;
	struct calendar tm;

	LIB_GETBUF(buf);

	if ( ! (ts->d_s.hi >= 0 && ntpcal_ntp64_to_date(&tm, ts) >= 0))
		snprintf(buf, LIB_BUFLENGTH, "%s", "9999-12-31T23:59:59Z");
	else
		snprintf(buf, LIB_BUFLENGTH, "%04d-%02d-%02dT%02d:%02d:%02dZ",
			tm.year, tm.month, tm.monthday,
			tm.hour, tm.minute, tm.second);

	return buf;
}

/* reset the global state for unit tests */
void
leapsec_ut_pristine(void)
{
	memset(_ltab, 0, sizeof(_ltab));
	_lptr     = NULL;
	_electric = 0;
}



/* -*- that's all folks! -*- */