-
Notifications
You must be signed in to change notification settings - Fork 432
/
main.cpp
3246 lines (2755 loc) · 90.6 KB
/
main.cpp
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
#ifdef MTRACE
#include <mcheck.h>
#endif
#ifdef __APPLE__
#define _DARWIN_UNLIMITED_STREAMS
#endif
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <string.h>
#include <unistd.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/mman.h>
#include <string.h>
#include <fcntl.h>
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <sqlite3.h>
#include <stdarg.h>
#include <sys/resource.h>
#include <pthread.h>
#include <getopt.h>
#include <signal.h>
#include <sys/time.h>
#include <zlib.h>
#include <algorithm>
#include <vector>
#include <string>
#include <set>
#include <map>
#include <cmath>
#ifdef __APPLE__
#include <sys/types.h>
#include <sys/sysctl.h>
#include <sys/param.h>
#include <sys/mount.h>
#else
#include <sys/statfs.h>
#endif
#include "jsonpull/jsonpull.h"
#include "mbtiles.hpp"
#include "tile.hpp"
#include "pool.hpp"
#include "projection.hpp"
#include "version.hpp"
#include "memfile.hpp"
#include "main.hpp"
#include "geojson.hpp"
#include "geobuf.hpp"
#include "geocsv.hpp"
#include "geometry.hpp"
#include "serial.hpp"
#include "options.hpp"
#include "mvt.hpp"
#include "dirtiles.hpp"
#include "evaluator.hpp"
#include "text.hpp"
static int low_detail = 12;
static int full_detail = -1;
static int min_detail = 7;
int quiet = 0;
int quiet_progress = 0;
double progress_interval = 0;
std::atomic<double> last_progress(0);
int geometry_scale = 0;
double simplification = 1;
size_t max_tile_size = 500000;
size_t max_tile_features = 200000;
int cluster_distance = 0;
long justx = -1, justy = -1;
std::string attribute_for_id = "";
int prevent[256];
int additional[256];
struct source {
std::string layer = "";
std::string file = "";
std::string description = "";
std::string format = "";
};
size_t CPUS;
size_t TEMP_FILES;
long long MAX_FILES;
static long long diskfree;
char **av;
std::vector<clipbbox> clipbboxes;
void checkdisk(std::vector<struct reader> *r) {
long long used = 0;
for (size_t i = 0; i < r->size(); i++) {
// Meta, pool, and tree are used once.
// Geometry and index will be duplicated during sorting and tiling.
used += (*r)[i].metapos + 2 * (*r)[i].geompos + 2 * (*r)[i].indexpos + (*r)[i].poolfile->len + (*r)[i].treefile->len;
}
static int warned = 0;
if (used > diskfree * .9 && !warned) {
fprintf(stderr, "You will probably run out of disk space.\n%lld bytes used or committed, of %lld originally available\n", used, diskfree);
warned = 1;
}
};
int atoi_require(const char *s, const char *what) {
char *err = NULL;
if (*s == '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
int ret = strtol(s, &err, 10);
if (*err != '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
return ret;
}
double atof_require(const char *s, const char *what) {
char *err = NULL;
if (*s == '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
double ret = strtod(s, &err);
if (*err != '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
return ret;
}
long long atoll_require(const char *s, const char *what) {
char *err = NULL;
if (*s == '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
long long ret = strtoll(s, &err, 10);
if (*err != '\0') {
fprintf(stderr, "%s: %s must be a number (got %s)\n", *av, what, s);
exit(EXIT_FAILURE);
}
return ret;
}
void init_cpus() {
const char *TIPPECANOE_MAX_THREADS = getenv("TIPPECANOE_MAX_THREADS");
if (TIPPECANOE_MAX_THREADS != NULL) {
CPUS = atoi_require(TIPPECANOE_MAX_THREADS, "TIPPECANOE_MAX_THREADS");
} else {
CPUS = sysconf(_SC_NPROCESSORS_ONLN);
}
if (CPUS < 1) {
CPUS = 1;
}
// Guard against short struct index.segment
if (CPUS > 32767) {
CPUS = 32767;
}
// Round down to a power of 2
CPUS = 1 << (int) (log(CPUS) / log(2));
struct rlimit rl;
if (getrlimit(RLIMIT_NOFILE, &rl) != 0) {
perror("getrlimit");
exit(EXIT_FAILURE);
} else {
MAX_FILES = rl.rlim_cur;
}
// Don't really want too many temporary files, because the file system
// will start to bog down eventually
if (MAX_FILES > 2000) {
MAX_FILES = 2000;
}
// MacOS can run out of system file descriptors
// even if we stay under the rlimit, so try to
// find out the real limit.
long long fds[MAX_FILES];
long long i;
for (i = 0; i < MAX_FILES; i++) {
fds[i] = open("/dev/null", O_RDONLY | O_CLOEXEC);
if (fds[i] < 0) {
break;
}
}
long long j;
for (j = 0; j < i; j++) {
if (close(fds[j]) < 0) {
perror("close");
exit(EXIT_FAILURE);
}
}
// Scale down because we really don't want to run the system out of files
MAX_FILES = i * 3 / 4;
if (MAX_FILES < 32) {
fprintf(stderr, "Can't open a useful number of files: %lld\n", MAX_FILES);
exit(EXIT_FAILURE);
}
TEMP_FILES = (MAX_FILES - 10) / 2;
if (TEMP_FILES > CPUS * 4) {
TEMP_FILES = CPUS * 4;
}
}
int indexcmp(const void *v1, const void *v2) {
const struct index *i1 = (const struct index *) v1;
const struct index *i2 = (const struct index *) v2;
if (i1->ix < i2->ix) {
return -1;
} else if (i1->ix > i2->ix) {
return 1;
}
if (i1->seq < i2->seq) {
return -1;
} else if (i1->seq > i2->seq) {
return 1;
}
return 0;
}
struct mergelist {
long long start;
long long end;
struct mergelist *next;
};
static void insert(struct mergelist *m, struct mergelist **head, unsigned char *map) {
while (*head != NULL && indexcmp(map + m->start, map + (*head)->start) > 0) {
head = &((*head)->next);
}
m->next = *head;
*head = m;
}
struct drop_state {
double gap;
unsigned long long previndex;
double interval;
double scale;
double seq;
long long included;
unsigned x;
unsigned y;
};
int calc_feature_minzoom(struct index *ix, struct drop_state *ds, int maxzoom, double gamma) {
int feature_minzoom = 0;
unsigned xx, yy;
decode_index(ix->ix, &xx, &yy);
if (gamma >= 0 && (ix->t == VT_POINT ||
(additional[A_LINE_DROP] && ix->t == VT_LINE) ||
(additional[A_POLYGON_DROP] && ix->t == VT_POLYGON))) {
for (ssize_t i = maxzoom; i >= 0; i--) {
ds[i].seq++;
}
for (ssize_t i = maxzoom; i >= 0; i--) {
if (ds[i].seq >= 0) {
ds[i].seq -= ds[i].interval;
ds[i].included++;
} else {
feature_minzoom = i + 1;
break;
}
}
// XXX manage_gap
}
return feature_minzoom;
}
static void merge(struct mergelist *merges, size_t nmerges, unsigned char *map, FILE *indexfile, int bytes, char *geom_map, FILE *geom_out, std::atomic<long long> *geompos, long long *progress, long long *progress_max, long long *progress_reported, int maxzoom, double gamma, struct drop_state *ds) {
struct mergelist *head = NULL;
for (size_t i = 0; i < nmerges; i++) {
if (merges[i].start < merges[i].end) {
insert(&(merges[i]), &head, map);
}
}
last_progress = 0;
while (head != NULL) {
struct index ix = *((struct index *) (map + head->start));
long long pos = *geompos;
fwrite_check(geom_map + ix.start, 1, ix.end - ix.start, geom_out, "merge geometry");
*geompos += ix.end - ix.start;
int feature_minzoom = calc_feature_minzoom(&ix, ds, maxzoom, gamma);
serialize_byte(geom_out, feature_minzoom, geompos, "merge geometry");
// Count this as an 75%-accomplishment, since we already 25%-counted it
*progress += (ix.end - ix.start) * 3 / 4;
if (!quiet && !quiet_progress && progress_time() && 100 * *progress / *progress_max != *progress_reported) {
fprintf(stderr, "Reordering geometry: %lld%% \r", 100 * *progress / *progress_max);
*progress_reported = 100 * *progress / *progress_max;
}
ix.start = pos;
ix.end = *geompos;
fwrite_check(&ix, bytes, 1, indexfile, "merge temporary");
head->start += bytes;
struct mergelist *m = head;
head = m->next;
m->next = NULL;
if (m->start < m->end) {
insert(m, &head, map);
}
}
}
struct sort_arg {
int task;
int cpus;
long long indexpos;
struct mergelist *merges;
int indexfd;
size_t nmerges;
long long unit;
int bytes;
sort_arg(int task1, int cpus1, long long indexpos1, struct mergelist *merges1, int indexfd1, size_t nmerges1, long long unit1, int bytes1)
: task(task1), cpus(cpus1), indexpos(indexpos1), merges(merges1), indexfd(indexfd1), nmerges(nmerges1), unit(unit1), bytes(bytes1) {
}
};
void *run_sort(void *v) {
struct sort_arg *a = (struct sort_arg *) v;
long long start;
for (start = a->task * a->unit; start < a->indexpos; start += a->unit * a->cpus) {
long long end = start + a->unit;
if (end > a->indexpos) {
end = a->indexpos;
}
a->merges[start / a->unit].start = start;
a->merges[start / a->unit].end = end;
a->merges[start / a->unit].next = NULL;
// MAP_PRIVATE to avoid disk writes if it fits in memory
void *map = mmap(NULL, end - start, PROT_READ | PROT_WRITE, MAP_PRIVATE, a->indexfd, start);
if (map == MAP_FAILED) {
perror("mmap in run_sort");
exit(EXIT_FAILURE);
}
madvise(map, end - start, MADV_RANDOM);
madvise(map, end - start, MADV_WILLNEED);
qsort(map, (end - start) / a->bytes, a->bytes, indexcmp);
// Sorting and then copying avoids disk access to
// write out intermediate stages of the sort.
void *map2 = mmap(NULL, end - start, PROT_READ | PROT_WRITE, MAP_SHARED, a->indexfd, start);
if (map2 == MAP_FAILED) {
perror("mmap (write)");
exit(EXIT_FAILURE);
}
madvise(map2, end - start, MADV_SEQUENTIAL);
memcpy(map2, map, end - start);
// No madvise, since caller will want the sorted data
munmap(map, end - start);
munmap(map2, end - start);
}
return NULL;
}
void do_read_parallel(char *map, long long len, long long initial_offset, const char *reading, std::vector<struct reader> *readers, std::atomic<long long> *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, int basezoom, int source, std::vector<std::map<std::string, layermap_entry> > *layermaps, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator, double *dist_sum, size_t *dist_count, bool want_dist, bool filters) {
long long segs[CPUS + 1];
segs[0] = 0;
segs[CPUS] = len;
for (size_t i = 1; i < CPUS; i++) {
segs[i] = len * i / CPUS;
while (segs[i] < len && map[segs[i]] != separator) {
segs[i]++;
}
}
double dist_sums[CPUS];
size_t dist_counts[CPUS];
std::atomic<long long> layer_seq[CPUS];
for (size_t i = 0; i < CPUS; i++) {
// To preserve feature ordering, unique id for each segment
// begins with that segment's offset into the input
layer_seq[i] = segs[i] + initial_offset;
dist_sums[i] = dist_counts[i] = 0;
}
std::vector<parse_json_args> pja;
std::vector<serialization_state> sst;
sst.resize(CPUS);
pthread_t pthreads[CPUS];
std::vector<std::set<type_and_string> > file_subkeys;
for (size_t i = 0; i < CPUS; i++) {
file_subkeys.push_back(std::set<type_and_string>());
}
for (size_t i = 0; i < CPUS; i++) {
sst[i].fname = reading;
sst[i].line = 0;
sst[i].layer_seq = &layer_seq[i];
sst[i].progress_seq = progress_seq;
sst[i].readers = readers;
sst[i].segment = i;
sst[i].initialized = &initialized[i];
sst[i].initial_x = &initial_x[i];
sst[i].initial_y = &initial_y[i];
sst[i].dist_sum = &(dist_sums[i]);
sst[i].dist_count = &(dist_counts[i]);
sst[i].want_dist = want_dist;
sst[i].maxzoom = maxzoom;
sst[i].uses_gamma = uses_gamma;
sst[i].filters = filters;
sst[i].layermap = &(*layermaps)[i];
sst[i].exclude = exclude;
sst[i].include = include;
sst[i].exclude_all = exclude_all;
sst[i].basezoom = basezoom;
sst[i].attribute_types = attribute_types;
pja.push_back(parse_json_args(
json_begin_map(map + segs[i], segs[i + 1] - segs[i]),
source,
&layername,
&sst[i]));
}
for (size_t i = 0; i < CPUS; i++) {
if (pthread_create(&pthreads[i], NULL, run_parse_json, &pja[i]) != 0) {
perror("pthread_create");
exit(EXIT_FAILURE);
}
}
for (size_t i = 0; i < CPUS; i++) {
void *retval;
if (pthread_join(pthreads[i], &retval) != 0) {
perror("pthread_join 370");
}
*dist_sum += dist_sums[i];
*dist_count += dist_counts[i];
json_end_map(pja[i].jp);
}
}
static ssize_t read_stream(json_pull *j, char *buffer, size_t n);
struct STREAM {
FILE *fp = NULL;
gzFile gz = NULL;
int fclose() {
int ret;
if (gz != NULL) {
ret = gzclose(gz);
} else {
ret = ::fclose(fp);
}
delete this;
return ret;
}
int peekc() {
if (gz != NULL) {
int c = gzgetc(gz);
if (c != EOF) {
gzungetc(c, gz);
}
return c;
} else {
int c = getc(fp);
if (c != EOF) {
ungetc(c, fp);
}
return c;
}
}
size_t read(char *out, size_t count) {
if (gz != NULL) {
int ret = gzread(gz, out, count);
if (ret < 0) {
fprintf(stderr, "%s: Error reading compressed data\n", *av);
exit(EXIT_FAILURE);
}
return ret;
} else {
return ::fread(out, 1, count, fp);
}
}
json_pull *json_begin() {
return ::json_begin(read_stream, this);
}
};
static ssize_t read_stream(json_pull *j, char *buffer, size_t n) {
return ((STREAM *) j->source)->read(buffer, n);
}
STREAM *streamfdopen(int fd, const char *mode, std::string const &fname) {
STREAM *s = new STREAM;
s->fp = NULL;
s->gz = NULL;
if (fname.size() > 3 && fname.substr(fname.size() - 3) == std::string(".gz")) {
s->gz = gzdopen(fd, mode);
if (s->gz == NULL) {
fprintf(stderr, "%s: %s: Decompression error\n", *av, fname.c_str());
exit(EXIT_FAILURE);
}
} else {
s->fp = fdopen(fd, mode);
if (s->fp == NULL) {
perror(fname.c_str());
exit(EXIT_FAILURE);
}
}
return s;
}
STREAM *streamfpopen(FILE *fp) {
STREAM *s = new STREAM;
s->fp = fp;
s->gz = NULL;
return s;
}
struct read_parallel_arg {
int fd = 0;
STREAM *fp = NULL;
long long offset = 0;
long long len = 0;
std::atomic<int> *is_parsing = NULL;
int separator = 0;
const char *reading = NULL;
std::vector<struct reader> *readers = NULL;
std::atomic<long long> *progress_seq = NULL;
std::set<std::string> *exclude = NULL;
std::set<std::string> *include = NULL;
int exclude_all = 0;
int maxzoom = 0;
int basezoom = 0;
int source = 0;
std::vector<std::map<std::string, layermap_entry> > *layermaps = NULL;
int *initialized = NULL;
unsigned *initial_x = NULL;
unsigned *initial_y = NULL;
std::string layername = "";
bool uses_gamma = false;
std::map<std::string, int> const *attribute_types = NULL;
double *dist_sum = NULL;
size_t *dist_count = NULL;
bool want_dist = false;
bool filters = false;
};
void *run_read_parallel(void *v) {
struct read_parallel_arg *rpa = (struct read_parallel_arg *) v;
struct stat st;
if (fstat(rpa->fd, &st) != 0) {
perror("stat read temp");
}
if (rpa->len != st.st_size) {
fprintf(stderr, "wrong number of bytes in temporary: %lld vs %lld\n", rpa->len, (long long) st.st_size);
}
rpa->len = st.st_size;
char *map = (char *) mmap(NULL, rpa->len, PROT_READ, MAP_PRIVATE, rpa->fd, 0);
if (map == NULL || map == MAP_FAILED) {
perror("map intermediate input");
exit(EXIT_FAILURE);
}
madvise(map, rpa->len, MADV_RANDOM); // sequential, but from several pointers at once
do_read_parallel(map, rpa->len, rpa->offset, rpa->reading, rpa->readers, rpa->progress_seq, rpa->exclude, rpa->include, rpa->exclude_all, rpa->basezoom, rpa->source, rpa->layermaps, rpa->initialized, rpa->initial_x, rpa->initial_y, rpa->maxzoom, rpa->layername, rpa->uses_gamma, rpa->attribute_types, rpa->separator, rpa->dist_sum, rpa->dist_count, rpa->want_dist, rpa->filters);
madvise(map, rpa->len, MADV_DONTNEED);
if (munmap(map, rpa->len) != 0) {
perror("munmap source file");
}
if (rpa->fp->fclose() != 0) {
perror("close source file");
exit(EXIT_FAILURE);
}
*(rpa->is_parsing) = 0;
delete rpa;
return NULL;
}
void start_parsing(int fd, STREAM *fp, long long offset, long long len, std::atomic<int> *is_parsing, pthread_t *parallel_parser, bool &parser_created, const char *reading, std::vector<struct reader> *readers, std::atomic<long long> *progress_seq, std::set<std::string> *exclude, std::set<std::string> *include, int exclude_all, int basezoom, int source, std::vector<std::map<std::string, layermap_entry> > &layermaps, int *initialized, unsigned *initial_x, unsigned *initial_y, int maxzoom, std::string layername, bool uses_gamma, std::map<std::string, int> const *attribute_types, int separator, double *dist_sum, size_t *dist_count, bool want_dist, bool filters) {
// This has to kick off an intermediate thread to start the parser threads,
// so the main thread can get back to reading the next input stage while
// the intermediate thread waits for the completion of the parser threads.
*is_parsing = 1;
struct read_parallel_arg *rpa = new struct read_parallel_arg;
if (rpa == NULL) {
perror("Out of memory");
exit(EXIT_FAILURE);
}
rpa->fd = fd;
rpa->fp = fp;
rpa->offset = offset;
rpa->len = len;
rpa->is_parsing = is_parsing;
rpa->separator = separator;
rpa->reading = reading;
rpa->readers = readers;
rpa->progress_seq = progress_seq;
rpa->exclude = exclude;
rpa->include = include;
rpa->exclude_all = exclude_all;
rpa->basezoom = basezoom;
rpa->source = source;
rpa->layermaps = &layermaps;
rpa->initialized = initialized;
rpa->initial_x = initial_x;
rpa->initial_y = initial_y;
rpa->maxzoom = maxzoom;
rpa->layername = layername;
rpa->uses_gamma = uses_gamma;
rpa->attribute_types = attribute_types;
rpa->dist_sum = dist_sum;
rpa->dist_count = dist_count;
rpa->want_dist = want_dist;
rpa->filters = filters;
if (pthread_create(parallel_parser, NULL, run_read_parallel, rpa) != 0) {
perror("pthread_create");
exit(EXIT_FAILURE);
}
parser_created = true;
}
void radix1(int *geomfds_in, int *indexfds_in, int inputs, int prefix, int splits, long long mem, const char *tmpdir, long long *availfiles, FILE *geomfile, FILE *indexfile, std::atomic<long long> *geompos_out, long long *progress, long long *progress_max, long long *progress_reported, int maxzoom, int basezoom, double droprate, double gamma, struct drop_state *ds) {
// Arranged as bits to facilitate subdividing again if a subdivided file is still huge
int splitbits = log(splits) / log(2);
splits = 1 << splitbits;
FILE *geomfiles[splits];
FILE *indexfiles[splits];
int geomfds[splits];
int indexfds[splits];
std::atomic<long long> sub_geompos[splits];
int i;
for (i = 0; i < splits; i++) {
sub_geompos[i] = 0;
char geomname[strlen(tmpdir) + strlen("/geom.XXXXXXXX") + 1];
sprintf(geomname, "%s%s", tmpdir, "/geom.XXXXXXXX");
char indexname[strlen(tmpdir) + strlen("/index.XXXXXXXX") + 1];
sprintf(indexname, "%s%s", tmpdir, "/index.XXXXXXXX");
geomfds[i] = mkstemp_cloexec(geomname);
if (geomfds[i] < 0) {
perror(geomname);
exit(EXIT_FAILURE);
}
indexfds[i] = mkstemp_cloexec(indexname);
if (indexfds[i] < 0) {
perror(indexname);
exit(EXIT_FAILURE);
}
geomfiles[i] = fopen_oflag(geomname, "wb", O_WRONLY | O_CLOEXEC);
if (geomfiles[i] == NULL) {
perror(geomname);
exit(EXIT_FAILURE);
}
indexfiles[i] = fopen_oflag(indexname, "wb", O_WRONLY | O_CLOEXEC);
if (indexfiles[i] == NULL) {
perror(indexname);
exit(EXIT_FAILURE);
}
*availfiles -= 4;
unlink(geomname);
unlink(indexname);
}
for (i = 0; i < inputs; i++) {
struct stat geomst, indexst;
if (fstat(geomfds_in[i], &geomst) < 0) {
perror("stat geom");
exit(EXIT_FAILURE);
}
if (fstat(indexfds_in[i], &indexst) < 0) {
perror("stat index");
exit(EXIT_FAILURE);
}
if (indexst.st_size != 0) {
struct index *indexmap = (struct index *) mmap(NULL, indexst.st_size, PROT_READ, MAP_PRIVATE, indexfds_in[i], 0);
if (indexmap == MAP_FAILED) {
fprintf(stderr, "fd %lld, len %lld\n", (long long) indexfds_in[i], (long long) indexst.st_size);
perror("map index");
exit(EXIT_FAILURE);
}
madvise(indexmap, indexst.st_size, MADV_SEQUENTIAL);
madvise(indexmap, indexst.st_size, MADV_WILLNEED);
char *geommap = (char *) mmap(NULL, geomst.st_size, PROT_READ, MAP_PRIVATE, geomfds_in[i], 0);
if (geommap == MAP_FAILED) {
perror("map geom");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_SEQUENTIAL);
madvise(geommap, geomst.st_size, MADV_WILLNEED);
for (size_t a = 0; a < indexst.st_size / sizeof(struct index); a++) {
struct index ix = indexmap[a];
unsigned long long which = (ix.ix << prefix) >> (64 - splitbits);
long long pos = sub_geompos[which];
fwrite_check(geommap + ix.start, ix.end - ix.start, 1, geomfiles[which], "geom");
sub_geompos[which] += ix.end - ix.start;
// Count this as a 25%-accomplishment, since we will copy again
*progress += (ix.end - ix.start) / 4;
if (!quiet && !quiet_progress && progress_time() && 100 * *progress / *progress_max != *progress_reported) {
fprintf(stderr, "Reordering geometry: %lld%% \r", 100 * *progress / *progress_max);
*progress_reported = 100 * *progress / *progress_max;
}
ix.start = pos;
ix.end = sub_geompos[which];
fwrite_check(&ix, sizeof(struct index), 1, indexfiles[which], "index");
}
madvise(indexmap, indexst.st_size, MADV_DONTNEED);
if (munmap(indexmap, indexst.st_size) < 0) {
perror("unmap index");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_DONTNEED);
if (munmap(geommap, geomst.st_size) < 0) {
perror("unmap geom");
exit(EXIT_FAILURE);
}
}
if (close(geomfds_in[i]) < 0) {
perror("close geom");
exit(EXIT_FAILURE);
}
if (close(indexfds_in[i]) < 0) {
perror("close index");
exit(EXIT_FAILURE);
}
*availfiles += 2;
}
for (i = 0; i < splits; i++) {
if (fclose(geomfiles[i]) != 0) {
perror("fclose geom");
exit(EXIT_FAILURE);
}
if (fclose(indexfiles[i]) != 0) {
perror("fclose index");
exit(EXIT_FAILURE);
}
*availfiles += 2;
}
for (i = 0; i < splits; i++) {
int already_closed = 0;
struct stat geomst, indexst;
if (fstat(geomfds[i], &geomst) < 0) {
perror("stat geom");
exit(EXIT_FAILURE);
}
if (fstat(indexfds[i], &indexst) < 0) {
perror("stat index");
exit(EXIT_FAILURE);
}
if (indexst.st_size > 0) {
if (indexst.st_size + geomst.st_size < mem) {
std::atomic<long long> indexpos(indexst.st_size);
int bytes = sizeof(struct index);
int page = sysconf(_SC_PAGESIZE);
// Don't try to sort more than 2GB at once,
// which used to crash Macs and may still
long long max_unit = 2LL * 1024 * 1024 * 1024;
long long unit = ((indexpos / CPUS + bytes - 1) / bytes) * bytes;
if (unit > max_unit) {
unit = max_unit;
}
unit = ((unit + page - 1) / page) * page;
if (unit < page) {
unit = page;
}
size_t nmerges = (indexpos + unit - 1) / unit;
struct mergelist merges[nmerges];
for (size_t a = 0; a < nmerges; a++) {
merges[a].start = merges[a].end = 0;
}
pthread_t pthreads[CPUS];
std::vector<sort_arg> args;
for (size_t a = 0; a < CPUS; a++) {
args.push_back(sort_arg(
a,
CPUS,
indexpos,
merges,
indexfds[i],
nmerges,
unit,
bytes));
}
for (size_t a = 0; a < CPUS; a++) {
if (pthread_create(&pthreads[a], NULL, run_sort, &args[a]) != 0) {
perror("pthread_create");
exit(EXIT_FAILURE);
}
}
for (size_t a = 0; a < CPUS; a++) {
void *retval;
if (pthread_join(pthreads[a], &retval) != 0) {
perror("pthread_join 679");
}
}
struct indexmap *indexmap = (struct indexmap *) mmap(NULL, indexst.st_size, PROT_READ, MAP_PRIVATE, indexfds[i], 0);
if (indexmap == MAP_FAILED) {
fprintf(stderr, "fd %lld, len %lld\n", (long long) indexfds[i], (long long) indexst.st_size);
perror("map index");
exit(EXIT_FAILURE);
}
madvise(indexmap, indexst.st_size, MADV_RANDOM); // sequential, but from several pointers at once
madvise(indexmap, indexst.st_size, MADV_WILLNEED);
char *geommap = (char *) mmap(NULL, geomst.st_size, PROT_READ, MAP_PRIVATE, geomfds[i], 0);
if (geommap == MAP_FAILED) {
perror("map geom");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_RANDOM);
madvise(geommap, geomst.st_size, MADV_WILLNEED);
merge(merges, nmerges, (unsigned char *) indexmap, indexfile, bytes, geommap, geomfile, geompos_out, progress, progress_max, progress_reported, maxzoom, gamma, ds);
madvise(indexmap, indexst.st_size, MADV_DONTNEED);
if (munmap(indexmap, indexst.st_size) < 0) {
perror("unmap index");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_DONTNEED);
if (munmap(geommap, geomst.st_size) < 0) {
perror("unmap geom");
exit(EXIT_FAILURE);
}
} else if (indexst.st_size == sizeof(struct index) || prefix + splitbits >= 64) {
struct index *indexmap = (struct index *) mmap(NULL, indexst.st_size, PROT_READ, MAP_PRIVATE, indexfds[i], 0);
if (indexmap == MAP_FAILED) {
fprintf(stderr, "fd %lld, len %lld\n", (long long) indexfds[i], (long long) indexst.st_size);
perror("map index");
exit(EXIT_FAILURE);
}
madvise(indexmap, indexst.st_size, MADV_SEQUENTIAL);
madvise(indexmap, indexst.st_size, MADV_WILLNEED);
char *geommap = (char *) mmap(NULL, geomst.st_size, PROT_READ, MAP_PRIVATE, geomfds[i], 0);
if (geommap == MAP_FAILED) {
perror("map geom");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_RANDOM);
madvise(geommap, geomst.st_size, MADV_WILLNEED);
for (size_t a = 0; a < indexst.st_size / sizeof(struct index); a++) {
struct index ix = indexmap[a];
long long pos = *geompos_out;
fwrite_check(geommap + ix.start, ix.end - ix.start, 1, geomfile, "geom");
*geompos_out += ix.end - ix.start;
int feature_minzoom = calc_feature_minzoom(&ix, ds, maxzoom, gamma);
serialize_byte(geomfile, feature_minzoom, geompos_out, "merge geometry");
// Count this as an 75%-accomplishment, since we already 25%-counted it
*progress += (ix.end - ix.start) * 3 / 4;
if (!quiet && !quiet_progress && progress_time() && 100 * *progress / *progress_max != *progress_reported) {
fprintf(stderr, "Reordering geometry: %lld%% \r", 100 * *progress / *progress_max);
*progress_reported = 100 * *progress / *progress_max;
}
ix.start = pos;
ix.end = *geompos_out;
fwrite_check(&ix, sizeof(struct index), 1, indexfile, "index");
}
madvise(indexmap, indexst.st_size, MADV_DONTNEED);
if (munmap(indexmap, indexst.st_size) < 0) {
perror("unmap index");
exit(EXIT_FAILURE);
}
madvise(geommap, geomst.st_size, MADV_DONTNEED);
if (munmap(geommap, geomst.st_size) < 0) {
perror("unmap geom");
exit(EXIT_FAILURE);
}
} else {
// We already reported the progress from splitting this radix out
// but we need to split it again, which will be credited with more
// progress. So increase the total amount of progress to report by
// the additional progress that will happpen, which may move the
// counter backward but will be an honest estimate of the work remaining.
*progress_max += geomst.st_size / 4;
radix1(&geomfds[i], &indexfds[i], 1, prefix + splitbits, *availfiles / 4, mem, tmpdir, availfiles, geomfile, indexfile, geompos_out, progress, progress_max, progress_reported, maxzoom, basezoom, droprate, gamma, ds);
already_closed = 1;
}
}
if (!already_closed) {
if (close(geomfds[i]) < 0) {
perror("close geom");
exit(EXIT_FAILURE);
}
if (close(indexfds[i]) < 0) {
perror("close index");
exit(EXIT_FAILURE);
}
*availfiles += 2;
}
}
}
void prep_drop_states(struct drop_state *ds, int maxzoom, int basezoom, double droprate) {
// Needs to be signed for interval calculation
for (ssize_t i = 0; i <= maxzoom; i++) {
ds[i].gap = 0;
ds[i].previndex = 0;
ds[i].interval = 0;
if (i < basezoom) {
ds[i].interval = std::exp(std::log(droprate) * (basezoom - i));
}