Line data Source code
1 : /*-------------------------------------------------------------------------
2 : *
3 : * tuplesort.c
4 : * Generalized tuple sorting routines.
5 : *
6 : * This module provides a generalized facility for tuple sorting, which can be
7 : * applied to different kinds of sortable objects. Implementation of
8 : * the particular sorting variants is given in tuplesortvariants.c.
9 : * This module works efficiently for both small and large amounts
10 : * of data. Small amounts are sorted in-memory using qsort(). Large
11 : * amounts are sorted using temporary files and a standard external sort
12 : * algorithm.
13 : *
14 : * See Knuth, volume 3, for more than you want to know about external
15 : * sorting algorithms. The algorithm we use is a balanced k-way merge.
16 : * Before PostgreSQL 15, we used the polyphase merge algorithm (Knuth's
17 : * Algorithm 5.4.2D), but with modern hardware, a straightforward balanced
18 : * merge is better. Knuth is assuming that tape drives are expensive
19 : * beasts, and in particular that there will always be many more runs than
20 : * tape drives. The polyphase merge algorithm was good at keeping all the
21 : * tape drives busy, but in our implementation a "tape drive" doesn't cost
22 : * much more than a few Kb of memory buffers, so we can afford to have
23 : * lots of them. In particular, if we can have as many tape drives as
24 : * sorted runs, we can eliminate any repeated I/O at all.
25 : *
26 : * Historically, we divided the input into sorted runs using replacement
27 : * selection, in the form of a priority tree implemented as a heap
28 : * (essentially Knuth's Algorithm 5.2.3H), but now we always use quicksort
29 : * for run generation.
30 : *
31 : * The approximate amount of memory allowed for any one sort operation
32 : * is specified in kilobytes by the caller (most pass work_mem). Initially,
33 : * we absorb tuples and simply store them in an unsorted array as long as
34 : * we haven't exceeded workMem. If we reach the end of the input without
35 : * exceeding workMem, we sort the array using qsort() and subsequently return
36 : * tuples just by scanning the tuple array sequentially. If we do exceed
37 : * workMem, we begin to emit tuples into sorted runs in temporary tapes.
38 : * When tuples are dumped in batch after quicksorting, we begin a new run
39 : * with a new output tape. If we reach the max number of tapes, we write
40 : * subsequent runs on the existing tapes in a round-robin fashion. We will
41 : * need multiple merge passes to finish the merge in that case. After the
42 : * end of the input is reached, we dump out remaining tuples in memory into
43 : * a final run, then merge the runs.
44 : *
45 : * When merging runs, we use a heap containing just the frontmost tuple from
46 : * each source run; we repeatedly output the smallest tuple and replace it
47 : * with the next tuple from its source tape (if any). When the heap empties,
48 : * the merge is complete. The basic merge algorithm thus needs very little
49 : * memory --- only M tuples for an M-way merge, and M is constrained to a
50 : * small number. However, we can still make good use of our full workMem
51 : * allocation by pre-reading additional blocks from each source tape. Without
52 : * prereading, our access pattern to the temporary file would be very erratic;
53 : * on average we'd read one block from each of M source tapes during the same
54 : * time that we're writing M blocks to the output tape, so there is no
55 : * sequentiality of access at all, defeating the read-ahead methods used by
56 : * most Unix kernels. Worse, the output tape gets written into a very random
57 : * sequence of blocks of the temp file, ensuring that things will be even
58 : * worse when it comes time to read that tape. A straightforward merge pass
59 : * thus ends up doing a lot of waiting for disk seeks. We can improve matters
60 : * by prereading from each source tape sequentially, loading about workMem/M
61 : * bytes from each tape in turn, and making the sequential blocks immediately
62 : * available for reuse. This approach helps to localize both read and write
63 : * accesses. The pre-reading is handled by logtape.c, we just tell it how
64 : * much memory to use for the buffers.
65 : *
66 : * In the current code we determine the number of input tapes M on the basis
67 : * of workMem: we want workMem/M to be large enough that we read a fair
68 : * amount of data each time we read from a tape, so as to maintain the
69 : * locality of access described above. Nonetheless, with large workMem we
70 : * can have many tapes. The logical "tapes" are implemented by logtape.c,
71 : * which avoids space wastage by recycling disk space as soon as each block
72 : * is read from its "tape".
73 : *
74 : * When the caller requests random access to the sort result, we form
75 : * the final sorted run on a logical tape which is then "frozen", so
76 : * that we can access it randomly. When the caller does not need random
77 : * access, we return from tuplesort_performsort() as soon as we are down
78 : * to one run per logical tape. The final merge is then performed
79 : * on-the-fly as the caller repeatedly calls tuplesort_getXXX; this
80 : * saves one cycle of writing all the data out to disk and reading it in.
81 : *
82 : * This module supports parallel sorting. Parallel sorts involve coordination
83 : * among one or more worker processes, and a leader process, each with its own
84 : * tuplesort state. The leader process (or, more accurately, the
85 : * Tuplesortstate associated with a leader process) creates a full tapeset
86 : * consisting of worker tapes with one run to merge; a run for every
87 : * worker process. This is then merged. Worker processes are guaranteed to
88 : * produce exactly one output run from their partial input.
89 : *
90 : *
91 : * Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group
92 : * Portions Copyright (c) 1994, Regents of the University of California
93 : *
94 : * IDENTIFICATION
95 : * src/backend/utils/sort/tuplesort.c
96 : *
97 : *-------------------------------------------------------------------------
98 : */
99 :
100 : #include "postgres.h"
101 :
102 : #include <limits.h>
103 :
104 : #include "commands/tablespace.h"
105 : #include "miscadmin.h"
106 : #include "pg_trace.h"
107 : #include "storage/shmem.h"
108 : #include "utils/guc.h"
109 : #include "utils/memutils.h"
110 : #include "utils/pg_rusage.h"
111 : #include "utils/tuplesort.h"
112 :
113 : /*
114 : * Initial size of memtuples array. We're trying to select this size so that
115 : * array doesn't exceed ALLOCSET_SEPARATE_THRESHOLD and so that the overhead of
116 : * allocation might possibly be lowered. However, we don't consider array sizes
117 : * less than 1024.
118 : *
119 : */
120 : #define INITIAL_MEMTUPSIZE Max(1024, \
121 : ALLOCSET_SEPARATE_THRESHOLD / sizeof(SortTuple) + 1)
122 :
123 : /* GUC variables */
124 : bool trace_sort = false;
125 :
126 : #ifdef DEBUG_BOUNDED_SORT
127 : bool optimize_bounded_sort = true;
128 : #endif
129 :
130 :
131 : /*
132 : * During merge, we use a pre-allocated set of fixed-size slots to hold
133 : * tuples. To avoid palloc/pfree overhead.
134 : *
135 : * Merge doesn't require a lot of memory, so we can afford to waste some,
136 : * by using gratuitously-sized slots. If a tuple is larger than 1 kB, the
137 : * palloc() overhead is not significant anymore.
138 : *
139 : * 'nextfree' is valid when this chunk is in the free list. When in use, the
140 : * slot holds a tuple.
141 : */
142 : #define SLAB_SLOT_SIZE 1024
143 :
144 : typedef union SlabSlot
145 : {
146 : union SlabSlot *nextfree;
147 : char buffer[SLAB_SLOT_SIZE];
148 : } SlabSlot;
149 :
150 : /*
151 : * Possible states of a Tuplesort object. These denote the states that
152 : * persist between calls of Tuplesort routines.
153 : */
154 : typedef enum
155 : {
156 : TSS_INITIAL, /* Loading tuples; still within memory limit */
157 : TSS_BOUNDED, /* Loading tuples into bounded-size heap */
158 : TSS_BUILDRUNS, /* Loading tuples; writing to tape */
159 : TSS_SORTEDINMEM, /* Sort completed entirely in memory */
160 : TSS_SORTEDONTAPE, /* Sort completed, final run is on tape */
161 : TSS_FINALMERGE, /* Performing final merge on-the-fly */
162 : } TupSortStatus;
163 :
164 : /*
165 : * Parameters for calculation of number of tapes to use --- see inittapes()
166 : * and tuplesort_merge_order().
167 : *
168 : * In this calculation we assume that each tape will cost us about 1 blocks
169 : * worth of buffer space. This ignores the overhead of all the other data
170 : * structures needed for each tape, but it's probably close enough.
171 : *
172 : * MERGE_BUFFER_SIZE is how much buffer space we'd like to allocate for each
173 : * input tape, for pre-reading (see discussion at top of file). This is *in
174 : * addition to* the 1 block already included in TAPE_BUFFER_OVERHEAD.
175 : */
176 : #define MINORDER 6 /* minimum merge order */
177 : #define MAXORDER 500 /* maximum merge order */
178 : #define TAPE_BUFFER_OVERHEAD BLCKSZ
179 : #define MERGE_BUFFER_SIZE (BLCKSZ * 32)
180 :
181 :
182 : /*
183 : * Private state of a Tuplesort operation.
184 : */
185 : struct Tuplesortstate
186 : {
187 : TuplesortPublic base;
188 : TupSortStatus status; /* enumerated value as shown above */
189 : bool bounded; /* did caller specify a maximum number of
190 : * tuples to return? */
191 : bool boundUsed; /* true if we made use of a bounded heap */
192 : int bound; /* if bounded, the maximum number of tuples */
193 : int64 tupleMem; /* memory consumed by individual tuples.
194 : * storing this separately from what we track
195 : * in availMem allows us to subtract the
196 : * memory consumed by all tuples when dumping
197 : * tuples to tape */
198 : int64 availMem; /* remaining memory available, in bytes */
199 : int64 allowedMem; /* total memory allowed, in bytes */
200 : int maxTapes; /* max number of input tapes to merge in each
201 : * pass */
202 : int64 maxSpace; /* maximum amount of space occupied among sort
203 : * of groups, either in-memory or on-disk */
204 : bool isMaxSpaceDisk; /* true when maxSpace is value for on-disk
205 : * space, false when its value for in-memory
206 : * space */
207 : TupSortStatus maxSpaceStatus; /* sort status when maxSpace was reached */
208 : LogicalTapeSet *tapeset; /* logtape.c object for tapes in a temp file */
209 :
210 : /*
211 : * This array holds the tuples now in sort memory. If we are in state
212 : * INITIAL, the tuples are in no particular order; if we are in state
213 : * SORTEDINMEM, the tuples are in final sorted order; in states BUILDRUNS
214 : * and FINALMERGE, the tuples are organized in "heap" order per Algorithm
215 : * H. In state SORTEDONTAPE, the array is not used.
216 : */
217 : SortTuple *memtuples; /* array of SortTuple structs */
218 : int memtupcount; /* number of tuples currently present */
219 : int memtupsize; /* allocated length of memtuples array */
220 : bool growmemtuples; /* memtuples' growth still underway? */
221 :
222 : /*
223 : * Memory for tuples is sometimes allocated using a simple slab allocator,
224 : * rather than with palloc(). Currently, we switch to slab allocation
225 : * when we start merging. Merging only needs to keep a small, fixed
226 : * number of tuples in memory at any time, so we can avoid the
227 : * palloc/pfree overhead by recycling a fixed number of fixed-size slots
228 : * to hold the tuples.
229 : *
230 : * For the slab, we use one large allocation, divided into SLAB_SLOT_SIZE
231 : * slots. The allocation is sized to have one slot per tape, plus one
232 : * additional slot. We need that many slots to hold all the tuples kept
233 : * in the heap during merge, plus the one we have last returned from the
234 : * sort, with tuplesort_gettuple.
235 : *
236 : * Initially, all the slots are kept in a linked list of free slots. When
237 : * a tuple is read from a tape, it is put to the next available slot, if
238 : * it fits. If the tuple is larger than SLAB_SLOT_SIZE, it is palloc'd
239 : * instead.
240 : *
241 : * When we're done processing a tuple, we return the slot back to the free
242 : * list, or pfree() if it was palloc'd. We know that a tuple was
243 : * allocated from the slab, if its pointer value is between
244 : * slabMemoryBegin and -End.
245 : *
246 : * When the slab allocator is used, the USEMEM/LACKMEM mechanism of
247 : * tracking memory usage is not used.
248 : */
249 : bool slabAllocatorUsed;
250 :
251 : char *slabMemoryBegin; /* beginning of slab memory arena */
252 : char *slabMemoryEnd; /* end of slab memory arena */
253 : SlabSlot *slabFreeHead; /* head of free list */
254 :
255 : /* Memory used for input and output tape buffers. */
256 : size_t tape_buffer_mem;
257 :
258 : /*
259 : * When we return a tuple to the caller in tuplesort_gettuple_XXX, that
260 : * came from a tape (that is, in TSS_SORTEDONTAPE or TSS_FINALMERGE
261 : * modes), we remember the tuple in 'lastReturnedTuple', so that we can
262 : * recycle the memory on next gettuple call.
263 : */
264 : void *lastReturnedTuple;
265 :
266 : /*
267 : * While building initial runs, this is the current output run number.
268 : * Afterwards, it is the number of initial runs we made.
269 : */
270 : int currentRun;
271 :
272 : /*
273 : * Logical tapes, for merging.
274 : *
275 : * The initial runs are written in the output tapes. In each merge pass,
276 : * the output tapes of the previous pass become the input tapes, and new
277 : * output tapes are created as needed. When nInputTapes equals
278 : * nInputRuns, there is only one merge pass left.
279 : */
280 : LogicalTape **inputTapes;
281 : int nInputTapes;
282 : int nInputRuns;
283 :
284 : LogicalTape **outputTapes;
285 : int nOutputTapes;
286 : int nOutputRuns;
287 :
288 : LogicalTape *destTape; /* current output tape */
289 :
290 : /*
291 : * These variables are used after completion of sorting to keep track of
292 : * the next tuple to return. (In the tape case, the tape's current read
293 : * position is also critical state.)
294 : */
295 : LogicalTape *result_tape; /* actual tape of finished output */
296 : int current; /* array index (only used if SORTEDINMEM) */
297 : bool eof_reached; /* reached EOF (needed for cursors) */
298 :
299 : /* markpos_xxx holds marked position for mark and restore */
300 : int64 markpos_block; /* tape block# (only used if SORTEDONTAPE) */
301 : int markpos_offset; /* saved "current", or offset in tape block */
302 : bool markpos_eof; /* saved "eof_reached" */
303 :
304 : /*
305 : * These variables are used during parallel sorting.
306 : *
307 : * worker is our worker identifier. Follows the general convention that
308 : * -1 value relates to a leader tuplesort, and values >= 0 worker
309 : * tuplesorts. (-1 can also be a serial tuplesort.)
310 : *
311 : * shared is mutable shared memory state, which is used to coordinate
312 : * parallel sorts.
313 : *
314 : * nParticipants is the number of worker Tuplesortstates known by the
315 : * leader to have actually been launched, which implies that they must
316 : * finish a run that the leader needs to merge. Typically includes a
317 : * worker state held by the leader process itself. Set in the leader
318 : * Tuplesortstate only.
319 : */
320 : int worker;
321 : Sharedsort *shared;
322 : int nParticipants;
323 :
324 : /*
325 : * Additional state for managing "abbreviated key" sortsupport routines
326 : * (which currently may be used by all cases except the hash index case).
327 : * Tracks the intervals at which the optimization's effectiveness is
328 : * tested.
329 : */
330 : int64 abbrevNext; /* Tuple # at which to next check
331 : * applicability */
332 :
333 : /*
334 : * Resource snapshot for time of sort start.
335 : */
336 : PGRUsage ru_start;
337 : };
338 :
339 : /*
340 : * Private mutable state of tuplesort-parallel-operation. This is allocated
341 : * in shared memory.
342 : */
343 : struct Sharedsort
344 : {
345 : /* mutex protects all fields prior to tapes */
346 : slock_t mutex;
347 :
348 : /*
349 : * currentWorker generates ordinal identifier numbers for parallel sort
350 : * workers. These start from 0, and are always gapless.
351 : *
352 : * Workers increment workersFinished to indicate having finished. If this
353 : * is equal to state.nParticipants within the leader, leader is ready to
354 : * merge worker runs.
355 : */
356 : int currentWorker;
357 : int workersFinished;
358 :
359 : /* Temporary file space */
360 : SharedFileSet fileset;
361 :
362 : /* Size of tapes flexible array */
363 : int nTapes;
364 :
365 : /*
366 : * Tapes array used by workers to report back information needed by the
367 : * leader to concatenate all worker tapes into one for merging
368 : */
369 : TapeShare tapes[FLEXIBLE_ARRAY_MEMBER];
370 : };
371 :
372 : /*
373 : * Is the given tuple allocated from the slab memory arena?
374 : */
375 : #define IS_SLAB_SLOT(state, tuple) \
376 : ((char *) (tuple) >= (state)->slabMemoryBegin && \
377 : (char *) (tuple) < (state)->slabMemoryEnd)
378 :
379 : /*
380 : * Return the given tuple to the slab memory free list, or free it
381 : * if it was palloc'd.
382 : */
383 : #define RELEASE_SLAB_SLOT(state, tuple) \
384 : do { \
385 : SlabSlot *buf = (SlabSlot *) tuple; \
386 : \
387 : if (IS_SLAB_SLOT((state), buf)) \
388 : { \
389 : buf->nextfree = (state)->slabFreeHead; \
390 : (state)->slabFreeHead = buf; \
391 : } else \
392 : pfree(buf); \
393 : } while(0)
394 :
395 : #define REMOVEABBREV(state,stup,count) ((*(state)->base.removeabbrev) (state, stup, count))
396 : #define COMPARETUP(state,a,b) ((*(state)->base.comparetup) (a, b, state))
397 : #define WRITETUP(state,tape,stup) ((*(state)->base.writetup) (state, tape, stup))
398 : #define READTUP(state,stup,tape,len) ((*(state)->base.readtup) (state, stup, tape, len))
399 : #define FREESTATE(state) ((state)->base.freestate ? (*(state)->base.freestate) (state) : (void) 0)
400 : #define LACKMEM(state) ((state)->availMem < 0 && !(state)->slabAllocatorUsed)
401 : #define USEMEM(state,amt) ((state)->availMem -= (amt))
402 : #define FREEMEM(state,amt) ((state)->availMem += (amt))
403 : #define SERIAL(state) ((state)->shared == NULL)
404 : #define WORKER(state) ((state)->shared && (state)->worker != -1)
405 : #define LEADER(state) ((state)->shared && (state)->worker == -1)
406 :
407 : /*
408 : * NOTES about on-tape representation of tuples:
409 : *
410 : * We require the first "unsigned int" of a stored tuple to be the total size
411 : * on-tape of the tuple, including itself (so it is never zero; an all-zero
412 : * unsigned int is used to delimit runs). The remainder of the stored tuple
413 : * may or may not match the in-memory representation of the tuple ---
414 : * any conversion needed is the job of the writetup and readtup routines.
415 : *
416 : * If state->sortopt contains TUPLESORT_RANDOMACCESS, then the stored
417 : * representation of the tuple must be followed by another "unsigned int" that
418 : * is a copy of the length --- so the total tape space used is actually
419 : * sizeof(unsigned int) more than the stored length value. This allows
420 : * read-backwards. When the random access flag was not specified, the
421 : * write/read routines may omit the extra length word.
422 : *
423 : * writetup is expected to write both length words as well as the tuple
424 : * data. When readtup is called, the tape is positioned just after the
425 : * front length word; readtup must read the tuple data and advance past
426 : * the back length word (if present).
427 : *
428 : * The write/read routines can make use of the tuple description data
429 : * stored in the Tuplesortstate record, if needed. They are also expected
430 : * to adjust state->availMem by the amount of memory space (not tape space!)
431 : * released or consumed. There is no error return from either writetup
432 : * or readtup; they should ereport() on failure.
433 : *
434 : *
435 : * NOTES about memory consumption calculations:
436 : *
437 : * We count space allocated for tuples against the workMem limit, plus
438 : * the space used by the variable-size memtuples array. Fixed-size space
439 : * is not counted; it's small enough to not be interesting.
440 : *
441 : * Note that we count actual space used (as shown by GetMemoryChunkSpace)
442 : * rather than the originally-requested size. This is important since
443 : * palloc can add substantial overhead. It's not a complete answer since
444 : * we won't count any wasted space in palloc allocation blocks, but it's
445 : * a lot better than what we were doing before 7.3. As of 9.6, a
446 : * separate memory context is used for caller passed tuples. Resetting
447 : * it at certain key increments significantly ameliorates fragmentation.
448 : * readtup routines use the slab allocator (they cannot use
449 : * the reset context because it gets deleted at the point that merging
450 : * begins).
451 : */
452 :
453 :
454 : static void tuplesort_begin_batch(Tuplesortstate *state);
455 : static bool consider_abort_common(Tuplesortstate *state);
456 : static void inittapes(Tuplesortstate *state, bool mergeruns);
457 : static void inittapestate(Tuplesortstate *state, int maxTapes);
458 : static void selectnewtape(Tuplesortstate *state);
459 : static void init_slab_allocator(Tuplesortstate *state, int numSlots);
460 : static void mergeruns(Tuplesortstate *state);
461 : static void mergeonerun(Tuplesortstate *state);
462 : static void beginmerge(Tuplesortstate *state);
463 : static bool mergereadnext(Tuplesortstate *state, LogicalTape *srcTape, SortTuple *stup);
464 : static void dumptuples(Tuplesortstate *state, bool alltuples);
465 : static void make_bounded_heap(Tuplesortstate *state);
466 : static void sort_bounded_heap(Tuplesortstate *state);
467 : static void tuplesort_sort_memtuples(Tuplesortstate *state);
468 : static void tuplesort_heap_insert(Tuplesortstate *state, SortTuple *tuple);
469 : static void tuplesort_heap_replace_top(Tuplesortstate *state, SortTuple *tuple);
470 : static void tuplesort_heap_delete_top(Tuplesortstate *state);
471 : static void reversedirection(Tuplesortstate *state);
472 : static unsigned int getlen(LogicalTape *tape, bool eofOK);
473 : static void markrunend(LogicalTape *tape);
474 : static int worker_get_identifier(Tuplesortstate *state);
475 : static void worker_freeze_result_tape(Tuplesortstate *state);
476 : static void worker_nomergeruns(Tuplesortstate *state);
477 : static void leader_takeover_tapes(Tuplesortstate *state);
478 : static void free_sort_tuple(Tuplesortstate *state, SortTuple *stup);
479 : static void tuplesort_free(Tuplesortstate *state);
480 : static void tuplesort_updatemax(Tuplesortstate *state);
481 :
482 : /*
483 : * Specialized comparators that we can inline into specialized sorts. The goal
484 : * is to try to sort two tuples without having to follow the pointers to the
485 : * comparator or the tuple.
486 : *
487 : * XXX: For now, there is no specialization for cases where datum1 is
488 : * authoritative and we don't even need to fall back to a callback at all (that
489 : * would be true for types like int4/int8/timestamp/date, but not true for
490 : * abbreviations of text or multi-key sorts. There could be! Is it worth it?
491 : */
492 :
493 : /* Used if first key's comparator is ssup_datum_unsigned_cmp */
494 : static pg_attribute_always_inline int
495 46521886 : qsort_tuple_unsigned_compare(SortTuple *a, SortTuple *b, Tuplesortstate *state)
496 : {
497 : int compare;
498 :
499 46521886 : compare = ApplyUnsignedSortComparator(a->datum1, a->isnull1,
500 46521886 : b->datum1, b->isnull1,
501 : &state->base.sortKeys[0]);
502 46521886 : if (compare != 0)
503 41970120 : return compare;
504 :
505 : /*
506 : * No need to waste effort calling the tiebreak function when there are no
507 : * other keys to sort on.
508 : */
509 4551766 : if (state->base.onlyKey != NULL)
510 0 : return 0;
511 :
512 4551766 : return state->base.comparetup_tiebreak(a, b, state);
513 : }
514 :
515 : /* Used if first key's comparator is ssup_datum_signed_cmp */
516 : static pg_attribute_always_inline int
517 8995712 : qsort_tuple_signed_compare(SortTuple *a, SortTuple *b, Tuplesortstate *state)
518 : {
519 : int compare;
520 :
521 8995712 : compare = ApplySignedSortComparator(a->datum1, a->isnull1,
522 8995712 : b->datum1, b->isnull1,
523 : &state->base.sortKeys[0]);
524 :
525 8995712 : if (compare != 0)
526 8954576 : return compare;
527 :
528 : /*
529 : * No need to waste effort calling the tiebreak function when there are no
530 : * other keys to sort on.
531 : */
532 41136 : if (state->base.onlyKey != NULL)
533 26674 : return 0;
534 :
535 14462 : return state->base.comparetup_tiebreak(a, b, state);
536 : }
537 :
538 : /* Used if first key's comparator is ssup_datum_int32_cmp */
539 : static pg_attribute_always_inline int
540 55280204 : qsort_tuple_int32_compare(SortTuple *a, SortTuple *b, Tuplesortstate *state)
541 : {
542 : int compare;
543 :
544 55280204 : compare = ApplyInt32SortComparator(a->datum1, a->isnull1,
545 55280204 : b->datum1, b->isnull1,
546 : &state->base.sortKeys[0]);
547 :
548 55280204 : if (compare != 0)
549 39727124 : return compare;
550 :
551 : /*
552 : * No need to waste effort calling the tiebreak function when there are no
553 : * other keys to sort on.
554 : */
555 15553080 : if (state->base.onlyKey != NULL)
556 1828242 : return 0;
557 :
558 13724838 : return state->base.comparetup_tiebreak(a, b, state);
559 : }
560 :
561 : /*
562 : * Special versions of qsort just for SortTuple objects. qsort_tuple() sorts
563 : * any variant of SortTuples, using the appropriate comparetup function.
564 : * qsort_ssup() is specialized for the case where the comparetup function
565 : * reduces to ApplySortComparator(), that is single-key MinimalTuple sorts
566 : * and Datum sorts. qsort_tuple_{unsigned,signed,int32} are specialized for
567 : * common comparison functions on pass-by-value leading datums.
568 : */
569 :
570 : #define ST_SORT qsort_tuple_unsigned
571 : #define ST_ELEMENT_TYPE SortTuple
572 : #define ST_COMPARE(a, b, state) qsort_tuple_unsigned_compare(a, b, state)
573 : #define ST_COMPARE_ARG_TYPE Tuplesortstate
574 : #define ST_CHECK_FOR_INTERRUPTS
575 : #define ST_SCOPE static
576 : #define ST_DEFINE
577 : #include "lib/sort_template.h"
578 :
579 : #define ST_SORT qsort_tuple_signed
580 : #define ST_ELEMENT_TYPE SortTuple
581 : #define ST_COMPARE(a, b, state) qsort_tuple_signed_compare(a, b, state)
582 : #define ST_COMPARE_ARG_TYPE Tuplesortstate
583 : #define ST_CHECK_FOR_INTERRUPTS
584 : #define ST_SCOPE static
585 : #define ST_DEFINE
586 : #include "lib/sort_template.h"
587 :
588 : #define ST_SORT qsort_tuple_int32
589 : #define ST_ELEMENT_TYPE SortTuple
590 : #define ST_COMPARE(a, b, state) qsort_tuple_int32_compare(a, b, state)
591 : #define ST_COMPARE_ARG_TYPE Tuplesortstate
592 : #define ST_CHECK_FOR_INTERRUPTS
593 : #define ST_SCOPE static
594 : #define ST_DEFINE
595 : #include "lib/sort_template.h"
596 :
597 : #define ST_SORT qsort_tuple
598 : #define ST_ELEMENT_TYPE SortTuple
599 : #define ST_COMPARE_RUNTIME_POINTER
600 : #define ST_COMPARE_ARG_TYPE Tuplesortstate
601 : #define ST_CHECK_FOR_INTERRUPTS
602 : #define ST_SCOPE static
603 : #define ST_DECLARE
604 : #define ST_DEFINE
605 : #include "lib/sort_template.h"
606 :
607 : #define ST_SORT qsort_ssup
608 : #define ST_ELEMENT_TYPE SortTuple
609 : #define ST_COMPARE(a, b, ssup) \
610 : ApplySortComparator((a)->datum1, (a)->isnull1, \
611 : (b)->datum1, (b)->isnull1, (ssup))
612 : #define ST_COMPARE_ARG_TYPE SortSupportData
613 : #define ST_CHECK_FOR_INTERRUPTS
614 : #define ST_SCOPE static
615 : #define ST_DEFINE
616 : #include "lib/sort_template.h"
617 :
618 : /*
619 : * tuplesort_begin_xxx
620 : *
621 : * Initialize for a tuple sort operation.
622 : *
623 : * After calling tuplesort_begin, the caller should call tuplesort_putXXX
624 : * zero or more times, then call tuplesort_performsort when all the tuples
625 : * have been supplied. After performsort, retrieve the tuples in sorted
626 : * order by calling tuplesort_getXXX until it returns false/NULL. (If random
627 : * access was requested, rescan, markpos, and restorepos can also be called.)
628 : * Call tuplesort_end to terminate the operation and release memory/disk space.
629 : *
630 : * Each variant of tuplesort_begin has a workMem parameter specifying the
631 : * maximum number of kilobytes of RAM to use before spilling data to disk.
632 : * (The normal value of this parameter is work_mem, but some callers use
633 : * other values.) Each variant also has a sortopt which is a bitmask of
634 : * sort options. See TUPLESORT_* definitions in tuplesort.h
635 : */
636 :
637 : Tuplesortstate *
638 272646 : tuplesort_begin_common(int workMem, SortCoordinate coordinate, int sortopt)
639 : {
640 : Tuplesortstate *state;
641 : MemoryContext maincontext;
642 : MemoryContext sortcontext;
643 : MemoryContext oldcontext;
644 :
645 : /* See leader_takeover_tapes() remarks on random access support */
646 272646 : if (coordinate && (sortopt & TUPLESORT_RANDOMACCESS))
647 0 : elog(ERROR, "random access disallowed under parallel sort");
648 :
649 : /*
650 : * Memory context surviving tuplesort_reset. This memory context holds
651 : * data which is useful to keep while sorting multiple similar batches.
652 : */
653 272646 : maincontext = AllocSetContextCreate(CurrentMemoryContext,
654 : "TupleSort main",
655 : ALLOCSET_DEFAULT_SIZES);
656 :
657 : /*
658 : * Create a working memory context for one sort operation. The content of
659 : * this context is deleted by tuplesort_reset.
660 : */
661 272646 : sortcontext = AllocSetContextCreate(maincontext,
662 : "TupleSort sort",
663 : ALLOCSET_DEFAULT_SIZES);
664 :
665 : /*
666 : * Additionally a working memory context for tuples is setup in
667 : * tuplesort_begin_batch.
668 : */
669 :
670 : /*
671 : * Make the Tuplesortstate within the per-sortstate context. This way, we
672 : * don't need a separate pfree() operation for it at shutdown.
673 : */
674 272646 : oldcontext = MemoryContextSwitchTo(maincontext);
675 :
676 272646 : state = (Tuplesortstate *) palloc0(sizeof(Tuplesortstate));
677 :
678 272646 : if (trace_sort)
679 0 : pg_rusage_init(&state->ru_start);
680 :
681 272646 : state->base.sortopt = sortopt;
682 272646 : state->base.tuples = true;
683 272646 : state->abbrevNext = 10;
684 :
685 : /*
686 : * workMem is forced to be at least 64KB, the current minimum valid value
687 : * for the work_mem GUC. This is a defense against parallel sort callers
688 : * that divide out memory among many workers in a way that leaves each
689 : * with very little memory.
690 : */
691 272646 : state->allowedMem = Max(workMem, 64) * (int64) 1024;
692 272646 : state->base.sortcontext = sortcontext;
693 272646 : state->base.maincontext = maincontext;
694 :
695 : /*
696 : * Initial size of array must be more than ALLOCSET_SEPARATE_THRESHOLD;
697 : * see comments in grow_memtuples().
698 : */
699 272646 : state->memtupsize = INITIAL_MEMTUPSIZE;
700 272646 : state->memtuples = NULL;
701 :
702 : /*
703 : * After all of the other non-parallel-related state, we setup all of the
704 : * state needed for each batch.
705 : */
706 272646 : tuplesort_begin_batch(state);
707 :
708 : /*
709 : * Initialize parallel-related state based on coordination information
710 : * from caller
711 : */
712 272646 : if (!coordinate)
713 : {
714 : /* Serial sort */
715 271982 : state->shared = NULL;
716 271982 : state->worker = -1;
717 271982 : state->nParticipants = -1;
718 : }
719 664 : else if (coordinate->isWorker)
720 : {
721 : /* Parallel worker produces exactly one final run from all input */
722 444 : state->shared = coordinate->sharedsort;
723 444 : state->worker = worker_get_identifier(state);
724 444 : state->nParticipants = -1;
725 : }
726 : else
727 : {
728 : /* Parallel leader state only used for final merge */
729 220 : state->shared = coordinate->sharedsort;
730 220 : state->worker = -1;
731 220 : state->nParticipants = coordinate->nParticipants;
732 : Assert(state->nParticipants >= 1);
733 : }
734 :
735 272646 : MemoryContextSwitchTo(oldcontext);
736 :
737 272646 : return state;
738 : }
739 :
740 : /*
741 : * tuplesort_begin_batch
742 : *
743 : * Setup, or reset, all state need for processing a new set of tuples with this
744 : * sort state. Called both from tuplesort_begin_common (the first time sorting
745 : * with this sort state) and tuplesort_reset (for subsequent usages).
746 : */
747 : static void
748 275692 : tuplesort_begin_batch(Tuplesortstate *state)
749 : {
750 : MemoryContext oldcontext;
751 :
752 275692 : oldcontext = MemoryContextSwitchTo(state->base.maincontext);
753 :
754 : /*
755 : * Caller tuple (e.g. IndexTuple) memory context.
756 : *
757 : * A dedicated child context used exclusively for caller passed tuples
758 : * eases memory management. Resetting at key points reduces
759 : * fragmentation. Note that the memtuples array of SortTuples is allocated
760 : * in the parent context, not this context, because there is no need to
761 : * free memtuples early. For bounded sorts, tuples may be pfreed in any
762 : * order, so we use a regular aset.c context so that it can make use of
763 : * free'd memory. When the sort is not bounded, we make use of a bump.c
764 : * context as this keeps allocations more compact with less wastage.
765 : * Allocations are also slightly more CPU efficient.
766 : */
767 275692 : if (TupleSortUseBumpTupleCxt(state->base.sortopt))
768 274322 : state->base.tuplecontext = BumpContextCreate(state->base.sortcontext,
769 : "Caller tuples",
770 : ALLOCSET_DEFAULT_SIZES);
771 : else
772 1370 : state->base.tuplecontext = AllocSetContextCreate(state->base.sortcontext,
773 : "Caller tuples",
774 : ALLOCSET_DEFAULT_SIZES);
775 :
776 :
777 275692 : state->status = TSS_INITIAL;
778 275692 : state->bounded = false;
779 275692 : state->boundUsed = false;
780 :
781 275692 : state->availMem = state->allowedMem;
782 :
783 275692 : state->tapeset = NULL;
784 :
785 275692 : state->memtupcount = 0;
786 :
787 : /*
788 : * Initial size of array must be more than ALLOCSET_SEPARATE_THRESHOLD;
789 : * see comments in grow_memtuples().
790 : */
791 275692 : state->growmemtuples = true;
792 275692 : state->slabAllocatorUsed = false;
793 275692 : if (state->memtuples != NULL && state->memtupsize != INITIAL_MEMTUPSIZE)
794 : {
795 0 : pfree(state->memtuples);
796 0 : state->memtuples = NULL;
797 0 : state->memtupsize = INITIAL_MEMTUPSIZE;
798 : }
799 275692 : if (state->memtuples == NULL)
800 : {
801 272646 : state->memtuples = (SortTuple *) palloc(state->memtupsize * sizeof(SortTuple));
802 272646 : USEMEM(state, GetMemoryChunkSpace(state->memtuples));
803 : }
804 :
805 : /* workMem must be large enough for the minimal memtuples array */
806 275692 : if (LACKMEM(state))
807 0 : elog(ERROR, "insufficient memory allowed for sort");
808 :
809 275692 : state->currentRun = 0;
810 :
811 : /*
812 : * Tape variables (inputTapes, outputTapes, etc.) will be initialized by
813 : * inittapes(), if needed.
814 : */
815 :
816 275692 : state->result_tape = NULL; /* flag that result tape has not been formed */
817 :
818 275692 : MemoryContextSwitchTo(oldcontext);
819 275692 : }
820 :
821 : /*
822 : * tuplesort_set_bound
823 : *
824 : * Advise tuplesort that at most the first N result tuples are required.
825 : *
826 : * Must be called before inserting any tuples. (Actually, we could allow it
827 : * as long as the sort hasn't spilled to disk, but there seems no need for
828 : * delayed calls at the moment.)
829 : *
830 : * This is a hint only. The tuplesort may still return more tuples than
831 : * requested. Parallel leader tuplesorts will always ignore the hint.
832 : */
833 : void
834 1236 : tuplesort_set_bound(Tuplesortstate *state, int64 bound)
835 : {
836 : /* Assert we're called before loading any tuples */
837 : Assert(state->status == TSS_INITIAL && state->memtupcount == 0);
838 : /* Assert we allow bounded sorts */
839 : Assert(state->base.sortopt & TUPLESORT_ALLOWBOUNDED);
840 : /* Can't set the bound twice, either */
841 : Assert(!state->bounded);
842 : /* Also, this shouldn't be called in a parallel worker */
843 : Assert(!WORKER(state));
844 :
845 : /* Parallel leader allows but ignores hint */
846 1236 : if (LEADER(state))
847 0 : return;
848 :
849 : #ifdef DEBUG_BOUNDED_SORT
850 : /* Honor GUC setting that disables the feature (for easy testing) */
851 : if (!optimize_bounded_sort)
852 : return;
853 : #endif
854 :
855 : /* We want to be able to compute bound * 2, so limit the setting */
856 1236 : if (bound > (int64) (INT_MAX / 2))
857 0 : return;
858 :
859 1236 : state->bounded = true;
860 1236 : state->bound = (int) bound;
861 :
862 : /*
863 : * Bounded sorts are not an effective target for abbreviated key
864 : * optimization. Disable by setting state to be consistent with no
865 : * abbreviation support.
866 : */
867 1236 : state->base.sortKeys->abbrev_converter = NULL;
868 1236 : if (state->base.sortKeys->abbrev_full_comparator)
869 16 : state->base.sortKeys->comparator = state->base.sortKeys->abbrev_full_comparator;
870 :
871 : /* Not strictly necessary, but be tidy */
872 1236 : state->base.sortKeys->abbrev_abort = NULL;
873 1236 : state->base.sortKeys->abbrev_full_comparator = NULL;
874 : }
875 :
876 : /*
877 : * tuplesort_used_bound
878 : *
879 : * Allow callers to find out if the sort state was able to use a bound.
880 : */
881 : bool
882 290 : tuplesort_used_bound(Tuplesortstate *state)
883 : {
884 290 : return state->boundUsed;
885 : }
886 :
887 : /*
888 : * tuplesort_free
889 : *
890 : * Internal routine for freeing resources of tuplesort.
891 : */
892 : static void
893 275430 : tuplesort_free(Tuplesortstate *state)
894 : {
895 : /* context swap probably not needed, but let's be safe */
896 275430 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
897 : int64 spaceUsed;
898 :
899 275430 : if (state->tapeset)
900 722 : spaceUsed = LogicalTapeSetBlocks(state->tapeset);
901 : else
902 274708 : spaceUsed = (state->allowedMem - state->availMem + 1023) / 1024;
903 :
904 : /*
905 : * Delete temporary "tape" files, if any.
906 : *
907 : * We don't bother to destroy the individual tapes here. They will go away
908 : * with the sortcontext. (In TSS_FINALMERGE state, we have closed
909 : * finished tapes already.)
910 : */
911 275430 : if (state->tapeset)
912 722 : LogicalTapeSetClose(state->tapeset);
913 :
914 275430 : if (trace_sort)
915 : {
916 0 : if (state->tapeset)
917 0 : elog(LOG, "%s of worker %d ended, %" PRId64 " disk blocks used: %s",
918 : SERIAL(state) ? "external sort" : "parallel external sort",
919 : state->worker, spaceUsed, pg_rusage_show(&state->ru_start));
920 : else
921 0 : elog(LOG, "%s of worker %d ended, %" PRId64 " KB used: %s",
922 : SERIAL(state) ? "internal sort" : "unperformed parallel sort",
923 : state->worker, spaceUsed, pg_rusage_show(&state->ru_start));
924 : }
925 :
926 : TRACE_POSTGRESQL_SORT_DONE(state->tapeset != NULL, spaceUsed);
927 :
928 275430 : FREESTATE(state);
929 275430 : MemoryContextSwitchTo(oldcontext);
930 :
931 : /*
932 : * Free the per-sort memory context, thereby releasing all working memory.
933 : */
934 275430 : MemoryContextReset(state->base.sortcontext);
935 275430 : }
936 :
937 : /*
938 : * tuplesort_end
939 : *
940 : * Release resources and clean up.
941 : *
942 : * NOTE: after calling this, any pointers returned by tuplesort_getXXX are
943 : * pointing to garbage. Be careful not to attempt to use or free such
944 : * pointers afterwards!
945 : */
946 : void
947 272384 : tuplesort_end(Tuplesortstate *state)
948 : {
949 272384 : tuplesort_free(state);
950 :
951 : /*
952 : * Free the main memory context, including the Tuplesortstate struct
953 : * itself.
954 : */
955 272384 : MemoryContextDelete(state->base.maincontext);
956 272384 : }
957 :
958 : /*
959 : * tuplesort_updatemax
960 : *
961 : * Update maximum resource usage statistics.
962 : */
963 : static void
964 3442 : tuplesort_updatemax(Tuplesortstate *state)
965 : {
966 : int64 spaceUsed;
967 : bool isSpaceDisk;
968 :
969 : /*
970 : * Note: it might seem we should provide both memory and disk usage for a
971 : * disk-based sort. However, the current code doesn't track memory space
972 : * accurately once we have begun to return tuples to the caller (since we
973 : * don't account for pfree's the caller is expected to do), so we cannot
974 : * rely on availMem in a disk sort. This does not seem worth the overhead
975 : * to fix. Is it worth creating an API for the memory context code to
976 : * tell us how much is actually used in sortcontext?
977 : */
978 3442 : if (state->tapeset)
979 : {
980 6 : isSpaceDisk = true;
981 6 : spaceUsed = LogicalTapeSetBlocks(state->tapeset) * BLCKSZ;
982 : }
983 : else
984 : {
985 3436 : isSpaceDisk = false;
986 3436 : spaceUsed = state->allowedMem - state->availMem;
987 : }
988 :
989 : /*
990 : * Sort evicts data to the disk when it wasn't able to fit that data into
991 : * main memory. This is why we assume space used on the disk to be more
992 : * important for tracking resource usage than space used in memory. Note
993 : * that the amount of space occupied by some tupleset on the disk might be
994 : * less than amount of space occupied by the same tupleset in memory due
995 : * to more compact representation.
996 : */
997 3442 : if ((isSpaceDisk && !state->isMaxSpaceDisk) ||
998 3436 : (isSpaceDisk == state->isMaxSpaceDisk && spaceUsed > state->maxSpace))
999 : {
1000 460 : state->maxSpace = spaceUsed;
1001 460 : state->isMaxSpaceDisk = isSpaceDisk;
1002 460 : state->maxSpaceStatus = state->status;
1003 : }
1004 3442 : }
1005 :
1006 : /*
1007 : * tuplesort_reset
1008 : *
1009 : * Reset the tuplesort. Reset all the data in the tuplesort, but leave the
1010 : * meta-information in. After tuplesort_reset, tuplesort is ready to start
1011 : * a new sort. This allows avoiding recreation of tuple sort states (and
1012 : * save resources) when sorting multiple small batches.
1013 : */
1014 : void
1015 3046 : tuplesort_reset(Tuplesortstate *state)
1016 : {
1017 3046 : tuplesort_updatemax(state);
1018 3046 : tuplesort_free(state);
1019 :
1020 : /*
1021 : * After we've freed up per-batch memory, re-setup all of the state common
1022 : * to both the first batch and any subsequent batch.
1023 : */
1024 3046 : tuplesort_begin_batch(state);
1025 :
1026 3046 : state->lastReturnedTuple = NULL;
1027 3046 : state->slabMemoryBegin = NULL;
1028 3046 : state->slabMemoryEnd = NULL;
1029 3046 : state->slabFreeHead = NULL;
1030 3046 : }
1031 :
1032 : /*
1033 : * Grow the memtuples[] array, if possible within our memory constraint. We
1034 : * must not exceed INT_MAX tuples in memory or the caller-provided memory
1035 : * limit. Return true if we were able to enlarge the array, false if not.
1036 : *
1037 : * Normally, at each increment we double the size of the array. When doing
1038 : * that would exceed a limit, we attempt one last, smaller increase (and then
1039 : * clear the growmemtuples flag so we don't try any more). That allows us to
1040 : * use memory as fully as permitted; sticking to the pure doubling rule could
1041 : * result in almost half going unused. Because availMem moves around with
1042 : * tuple addition/removal, we need some rule to prevent making repeated small
1043 : * increases in memtupsize, which would just be useless thrashing. The
1044 : * growmemtuples flag accomplishes that and also prevents useless
1045 : * recalculations in this function.
1046 : */
1047 : static bool
1048 7976 : grow_memtuples(Tuplesortstate *state)
1049 : {
1050 : int newmemtupsize;
1051 7976 : int memtupsize = state->memtupsize;
1052 7976 : int64 memNowUsed = state->allowedMem - state->availMem;
1053 :
1054 : /* Forget it if we've already maxed out memtuples, per comment above */
1055 7976 : if (!state->growmemtuples)
1056 122 : return false;
1057 :
1058 : /* Select new value of memtupsize */
1059 7854 : if (memNowUsed <= state->availMem)
1060 : {
1061 : /*
1062 : * We've used no more than half of allowedMem; double our usage,
1063 : * clamping at INT_MAX tuples.
1064 : */
1065 7724 : if (memtupsize < INT_MAX / 2)
1066 7724 : newmemtupsize = memtupsize * 2;
1067 : else
1068 : {
1069 0 : newmemtupsize = INT_MAX;
1070 0 : state->growmemtuples = false;
1071 : }
1072 : }
1073 : else
1074 : {
1075 : /*
1076 : * This will be the last increment of memtupsize. Abandon doubling
1077 : * strategy and instead increase as much as we safely can.
1078 : *
1079 : * To stay within allowedMem, we can't increase memtupsize by more
1080 : * than availMem / sizeof(SortTuple) elements. In practice, we want
1081 : * to increase it by considerably less, because we need to leave some
1082 : * space for the tuples to which the new array slots will refer. We
1083 : * assume the new tuples will be about the same size as the tuples
1084 : * we've already seen, and thus we can extrapolate from the space
1085 : * consumption so far to estimate an appropriate new size for the
1086 : * memtuples array. The optimal value might be higher or lower than
1087 : * this estimate, but it's hard to know that in advance. We again
1088 : * clamp at INT_MAX tuples.
1089 : *
1090 : * This calculation is safe against enlarging the array so much that
1091 : * LACKMEM becomes true, because the memory currently used includes
1092 : * the present array; thus, there would be enough allowedMem for the
1093 : * new array elements even if no other memory were currently used.
1094 : *
1095 : * We do the arithmetic in float8, because otherwise the product of
1096 : * memtupsize and allowedMem could overflow. Any inaccuracy in the
1097 : * result should be insignificant; but even if we computed a
1098 : * completely insane result, the checks below will prevent anything
1099 : * really bad from happening.
1100 : */
1101 : double grow_ratio;
1102 :
1103 130 : grow_ratio = (double) state->allowedMem / (double) memNowUsed;
1104 130 : if (memtupsize * grow_ratio < INT_MAX)
1105 130 : newmemtupsize = (int) (memtupsize * grow_ratio);
1106 : else
1107 0 : newmemtupsize = INT_MAX;
1108 :
1109 : /* We won't make any further enlargement attempts */
1110 130 : state->growmemtuples = false;
1111 : }
1112 :
1113 : /* Must enlarge array by at least one element, else report failure */
1114 7854 : if (newmemtupsize <= memtupsize)
1115 0 : goto noalloc;
1116 :
1117 : /*
1118 : * On a 32-bit machine, allowedMem could exceed MaxAllocHugeSize. Clamp
1119 : * to ensure our request won't be rejected. Note that we can easily
1120 : * exhaust address space before facing this outcome. (This is presently
1121 : * impossible due to guc.c's MAX_KILOBYTES limitation on work_mem, but
1122 : * don't rely on that at this distance.)
1123 : */
1124 7854 : if ((Size) newmemtupsize >= MaxAllocHugeSize / sizeof(SortTuple))
1125 : {
1126 0 : newmemtupsize = (int) (MaxAllocHugeSize / sizeof(SortTuple));
1127 0 : state->growmemtuples = false; /* can't grow any more */
1128 : }
1129 :
1130 : /*
1131 : * We need to be sure that we do not cause LACKMEM to become true, else
1132 : * the space management algorithm will go nuts. The code above should
1133 : * never generate a dangerous request, but to be safe, check explicitly
1134 : * that the array growth fits within availMem. (We could still cause
1135 : * LACKMEM if the memory chunk overhead associated with the memtuples
1136 : * array were to increase. That shouldn't happen because we chose the
1137 : * initial array size large enough to ensure that palloc will be treating
1138 : * both old and new arrays as separate chunks. But we'll check LACKMEM
1139 : * explicitly below just in case.)
1140 : */
1141 7854 : if (state->availMem < (int64) ((newmemtupsize - memtupsize) * sizeof(SortTuple)))
1142 0 : goto noalloc;
1143 :
1144 : /* OK, do it */
1145 7854 : FREEMEM(state, GetMemoryChunkSpace(state->memtuples));
1146 7854 : state->memtupsize = newmemtupsize;
1147 7854 : state->memtuples = (SortTuple *)
1148 7854 : repalloc_huge(state->memtuples,
1149 7854 : state->memtupsize * sizeof(SortTuple));
1150 7854 : USEMEM(state, GetMemoryChunkSpace(state->memtuples));
1151 7854 : if (LACKMEM(state))
1152 0 : elog(ERROR, "unexpected out-of-memory situation in tuplesort");
1153 7854 : return true;
1154 :
1155 0 : noalloc:
1156 : /* If for any reason we didn't realloc, shut off future attempts */
1157 0 : state->growmemtuples = false;
1158 0 : return false;
1159 : }
1160 :
1161 : /*
1162 : * Shared code for tuple and datum cases.
1163 : */
1164 : void
1165 30211304 : tuplesort_puttuple_common(Tuplesortstate *state, SortTuple *tuple,
1166 : bool useAbbrev, Size tuplen)
1167 : {
1168 30211304 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
1169 :
1170 : Assert(!LEADER(state));
1171 :
1172 : /* account for the memory used for this tuple */
1173 30211304 : USEMEM(state, tuplen);
1174 30211304 : state->tupleMem += tuplen;
1175 :
1176 30211304 : if (!useAbbrev)
1177 : {
1178 : /*
1179 : * Leave ordinary Datum representation, or NULL value. If there is a
1180 : * converter it won't expect NULL values, and cost model is not
1181 : * required to account for NULL, so in that case we avoid calling
1182 : * converter and just set datum1 to zeroed representation (to be
1183 : * consistent, and to support cheap inequality tests for NULL
1184 : * abbreviated keys).
1185 : */
1186 : }
1187 4434304 : else if (!consider_abort_common(state))
1188 : {
1189 : /* Store abbreviated key representation */
1190 4434208 : tuple->datum1 = state->base.sortKeys->abbrev_converter(tuple->datum1,
1191 : state->base.sortKeys);
1192 : }
1193 : else
1194 : {
1195 : /*
1196 : * Set state to be consistent with never trying abbreviation.
1197 : *
1198 : * Alter datum1 representation in already-copied tuples, so as to
1199 : * ensure a consistent representation (current tuple was just
1200 : * handled). It does not matter if some dumped tuples are already
1201 : * sorted on tape, since serialized tuples lack abbreviated keys
1202 : * (TSS_BUILDRUNS state prevents control reaching here in any case).
1203 : */
1204 96 : REMOVEABBREV(state, state->memtuples, state->memtupcount);
1205 : }
1206 :
1207 30211304 : switch (state->status)
1208 : {
1209 25392818 : case TSS_INITIAL:
1210 :
1211 : /*
1212 : * Save the tuple into the unsorted array. First, grow the array
1213 : * as needed. Note that we try to grow the array when there is
1214 : * still one free slot remaining --- if we fail, there'll still be
1215 : * room to store the incoming tuple, and then we'll switch to
1216 : * tape-based operation.
1217 : */
1218 25392818 : if (state->memtupcount >= state->memtupsize - 1)
1219 : {
1220 7976 : (void) grow_memtuples(state);
1221 : Assert(state->memtupcount < state->memtupsize);
1222 : }
1223 25392818 : state->memtuples[state->memtupcount++] = *tuple;
1224 :
1225 : /*
1226 : * Check if it's time to switch over to a bounded heapsort. We do
1227 : * so if the input tuple count exceeds twice the desired tuple
1228 : * count (this is a heuristic for where heapsort becomes cheaper
1229 : * than a quicksort), or if we've just filled workMem and have
1230 : * enough tuples to meet the bound.
1231 : *
1232 : * Note that once we enter TSS_BOUNDED state we will always try to
1233 : * complete the sort that way. In the worst case, if later input
1234 : * tuples are larger than earlier ones, this might cause us to
1235 : * exceed workMem significantly.
1236 : */
1237 25392818 : if (state->bounded &&
1238 51318 : (state->memtupcount > state->bound * 2 ||
1239 50910 : (state->memtupcount > state->bound && LACKMEM(state))))
1240 : {
1241 408 : if (trace_sort)
1242 0 : elog(LOG, "switching to bounded heapsort at %d tuples: %s",
1243 : state->memtupcount,
1244 : pg_rusage_show(&state->ru_start));
1245 408 : make_bounded_heap(state);
1246 408 : MemoryContextSwitchTo(oldcontext);
1247 408 : return;
1248 : }
1249 :
1250 : /*
1251 : * Done if we still fit in available memory and have array slots.
1252 : */
1253 25392410 : if (state->memtupcount < state->memtupsize && !LACKMEM(state))
1254 : {
1255 25392288 : MemoryContextSwitchTo(oldcontext);
1256 25392288 : return;
1257 : }
1258 :
1259 : /*
1260 : * Nope; time to switch to tape-based operation.
1261 : */
1262 122 : inittapes(state, true);
1263 :
1264 : /*
1265 : * Dump all tuples.
1266 : */
1267 122 : dumptuples(state, false);
1268 122 : break;
1269 :
1270 3750180 : case TSS_BOUNDED:
1271 :
1272 : /*
1273 : * We don't want to grow the array here, so check whether the new
1274 : * tuple can be discarded before putting it in. This should be a
1275 : * good speed optimization, too, since when there are many more
1276 : * input tuples than the bound, most input tuples can be discarded
1277 : * with just this one comparison. Note that because we currently
1278 : * have the sort direction reversed, we must check for <= not >=.
1279 : */
1280 3750180 : if (COMPARETUP(state, tuple, &state->memtuples[0]) <= 0)
1281 : {
1282 : /* new tuple <= top of the heap, so we can discard it */
1283 3247202 : free_sort_tuple(state, tuple);
1284 3247202 : CHECK_FOR_INTERRUPTS();
1285 : }
1286 : else
1287 : {
1288 : /* discard top of heap, replacing it with the new tuple */
1289 502978 : free_sort_tuple(state, &state->memtuples[0]);
1290 502978 : tuplesort_heap_replace_top(state, tuple);
1291 : }
1292 3750180 : break;
1293 :
1294 1068306 : case TSS_BUILDRUNS:
1295 :
1296 : /*
1297 : * Save the tuple into the unsorted array (there must be space)
1298 : */
1299 1068306 : state->memtuples[state->memtupcount++] = *tuple;
1300 :
1301 : /*
1302 : * If we are over the memory limit, dump all tuples.
1303 : */
1304 1068306 : dumptuples(state, false);
1305 1068306 : break;
1306 :
1307 0 : default:
1308 0 : elog(ERROR, "invalid tuplesort state");
1309 : break;
1310 : }
1311 4818608 : MemoryContextSwitchTo(oldcontext);
1312 : }
1313 :
1314 : static bool
1315 4434304 : consider_abort_common(Tuplesortstate *state)
1316 : {
1317 : Assert(state->base.sortKeys[0].abbrev_converter != NULL);
1318 : Assert(state->base.sortKeys[0].abbrev_abort != NULL);
1319 : Assert(state->base.sortKeys[0].abbrev_full_comparator != NULL);
1320 :
1321 : /*
1322 : * Check effectiveness of abbreviation optimization. Consider aborting
1323 : * when still within memory limit.
1324 : */
1325 4434304 : if (state->status == TSS_INITIAL &&
1326 3961874 : state->memtupcount >= state->abbrevNext)
1327 : {
1328 5020 : state->abbrevNext *= 2;
1329 :
1330 : /*
1331 : * Check opclass-supplied abbreviation abort routine. It may indicate
1332 : * that abbreviation should not proceed.
1333 : */
1334 5020 : if (!state->base.sortKeys->abbrev_abort(state->memtupcount,
1335 : state->base.sortKeys))
1336 4924 : return false;
1337 :
1338 : /*
1339 : * Finally, restore authoritative comparator, and indicate that
1340 : * abbreviation is not in play by setting abbrev_converter to NULL
1341 : */
1342 96 : state->base.sortKeys[0].comparator = state->base.sortKeys[0].abbrev_full_comparator;
1343 96 : state->base.sortKeys[0].abbrev_converter = NULL;
1344 : /* Not strictly necessary, but be tidy */
1345 96 : state->base.sortKeys[0].abbrev_abort = NULL;
1346 96 : state->base.sortKeys[0].abbrev_full_comparator = NULL;
1347 :
1348 : /* Give up - expect original pass-by-value representation */
1349 96 : return true;
1350 : }
1351 :
1352 4429284 : return false;
1353 : }
1354 :
1355 : /*
1356 : * All tuples have been provided; finish the sort.
1357 : */
1358 : void
1359 234106 : tuplesort_performsort(Tuplesortstate *state)
1360 : {
1361 234106 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
1362 :
1363 234106 : if (trace_sort)
1364 0 : elog(LOG, "performsort of worker %d starting: %s",
1365 : state->worker, pg_rusage_show(&state->ru_start));
1366 :
1367 234106 : switch (state->status)
1368 : {
1369 233576 : case TSS_INITIAL:
1370 :
1371 : /*
1372 : * We were able to accumulate all the tuples within the allowed
1373 : * amount of memory, or leader to take over worker tapes
1374 : */
1375 233576 : if (SERIAL(state))
1376 : {
1377 : /* Just qsort 'em and we're done */
1378 232976 : tuplesort_sort_memtuples(state);
1379 232886 : state->status = TSS_SORTEDINMEM;
1380 : }
1381 600 : else if (WORKER(state))
1382 : {
1383 : /*
1384 : * Parallel workers must still dump out tuples to tape. No
1385 : * merge is required to produce single output run, though.
1386 : */
1387 444 : inittapes(state, false);
1388 444 : dumptuples(state, true);
1389 444 : worker_nomergeruns(state);
1390 444 : state->status = TSS_SORTEDONTAPE;
1391 : }
1392 : else
1393 : {
1394 : /*
1395 : * Leader will take over worker tapes and merge worker runs.
1396 : * Note that mergeruns sets the correct state->status.
1397 : */
1398 156 : leader_takeover_tapes(state);
1399 156 : mergeruns(state);
1400 : }
1401 233486 : state->current = 0;
1402 233486 : state->eof_reached = false;
1403 233486 : state->markpos_block = 0L;
1404 233486 : state->markpos_offset = 0;
1405 233486 : state->markpos_eof = false;
1406 233486 : break;
1407 :
1408 408 : case TSS_BOUNDED:
1409 :
1410 : /*
1411 : * We were able to accumulate all the tuples required for output
1412 : * in memory, using a heap to eliminate excess tuples. Now we
1413 : * have to transform the heap to a properly-sorted array. Note
1414 : * that sort_bounded_heap sets the correct state->status.
1415 : */
1416 408 : sort_bounded_heap(state);
1417 408 : state->current = 0;
1418 408 : state->eof_reached = false;
1419 408 : state->markpos_offset = 0;
1420 408 : state->markpos_eof = false;
1421 408 : break;
1422 :
1423 122 : case TSS_BUILDRUNS:
1424 :
1425 : /*
1426 : * Finish tape-based sort. First, flush all tuples remaining in
1427 : * memory out to tape; then merge until we have a single remaining
1428 : * run (or, if !randomAccess and !WORKER(), one run per tape).
1429 : * Note that mergeruns sets the correct state->status.
1430 : */
1431 122 : dumptuples(state, true);
1432 122 : mergeruns(state);
1433 122 : state->eof_reached = false;
1434 122 : state->markpos_block = 0L;
1435 122 : state->markpos_offset = 0;
1436 122 : state->markpos_eof = false;
1437 122 : break;
1438 :
1439 0 : default:
1440 0 : elog(ERROR, "invalid tuplesort state");
1441 : break;
1442 : }
1443 :
1444 234016 : if (trace_sort)
1445 : {
1446 0 : if (state->status == TSS_FINALMERGE)
1447 0 : elog(LOG, "performsort of worker %d done (except %d-way final merge): %s",
1448 : state->worker, state->nInputTapes,
1449 : pg_rusage_show(&state->ru_start));
1450 : else
1451 0 : elog(LOG, "performsort of worker %d done: %s",
1452 : state->worker, pg_rusage_show(&state->ru_start));
1453 : }
1454 :
1455 234016 : MemoryContextSwitchTo(oldcontext);
1456 234016 : }
1457 :
1458 : /*
1459 : * Internal routine to fetch the next tuple in either forward or back
1460 : * direction into *stup. Returns false if no more tuples.
1461 : * Returned tuple belongs to tuplesort memory context, and must not be freed
1462 : * by caller. Note that fetched tuple is stored in memory that may be
1463 : * recycled by any future fetch.
1464 : */
1465 : bool
1466 27443524 : tuplesort_gettuple_common(Tuplesortstate *state, bool forward,
1467 : SortTuple *stup)
1468 : {
1469 : unsigned int tuplen;
1470 : size_t nmoved;
1471 :
1472 : Assert(!WORKER(state));
1473 :
1474 27443524 : switch (state->status)
1475 : {
1476 22595944 : case TSS_SORTEDINMEM:
1477 : Assert(forward || state->base.sortopt & TUPLESORT_RANDOMACCESS);
1478 : Assert(!state->slabAllocatorUsed);
1479 22595944 : if (forward)
1480 : {
1481 22595878 : if (state->current < state->memtupcount)
1482 : {
1483 22364826 : *stup = state->memtuples[state->current++];
1484 22364826 : return true;
1485 : }
1486 231052 : state->eof_reached = true;
1487 :
1488 : /*
1489 : * Complain if caller tries to retrieve more tuples than
1490 : * originally asked for in a bounded sort. This is because
1491 : * returning EOF here might be the wrong thing.
1492 : */
1493 231052 : if (state->bounded && state->current >= state->bound)
1494 0 : elog(ERROR, "retrieved too many tuples in a bounded sort");
1495 :
1496 231052 : return false;
1497 : }
1498 : else
1499 : {
1500 66 : if (state->current <= 0)
1501 0 : return false;
1502 :
1503 : /*
1504 : * if all tuples are fetched already then we return last
1505 : * tuple, else - tuple before last returned.
1506 : */
1507 66 : if (state->eof_reached)
1508 12 : state->eof_reached = false;
1509 : else
1510 : {
1511 54 : state->current--; /* last returned tuple */
1512 54 : if (state->current <= 0)
1513 6 : return false;
1514 : }
1515 60 : *stup = state->memtuples[state->current - 1];
1516 60 : return true;
1517 : }
1518 : break;
1519 :
1520 272994 : case TSS_SORTEDONTAPE:
1521 : Assert(forward || state->base.sortopt & TUPLESORT_RANDOMACCESS);
1522 : Assert(state->slabAllocatorUsed);
1523 :
1524 : /*
1525 : * The slot that held the tuple that we returned in previous
1526 : * gettuple call can now be reused.
1527 : */
1528 272994 : if (state->lastReturnedTuple)
1529 : {
1530 152850 : RELEASE_SLAB_SLOT(state, state->lastReturnedTuple);
1531 152850 : state->lastReturnedTuple = NULL;
1532 : }
1533 :
1534 272994 : if (forward)
1535 : {
1536 272964 : if (state->eof_reached)
1537 0 : return false;
1538 :
1539 272964 : if ((tuplen = getlen(state->result_tape, true)) != 0)
1540 : {
1541 272940 : READTUP(state, stup, state->result_tape, tuplen);
1542 :
1543 : /*
1544 : * Remember the tuple we return, so that we can recycle
1545 : * its memory on next call. (This can be NULL, in the
1546 : * !state->tuples case).
1547 : */
1548 272940 : state->lastReturnedTuple = stup->tuple;
1549 :
1550 272940 : return true;
1551 : }
1552 : else
1553 : {
1554 24 : state->eof_reached = true;
1555 24 : return false;
1556 : }
1557 : }
1558 :
1559 : /*
1560 : * Backward.
1561 : *
1562 : * if all tuples are fetched already then we return last tuple,
1563 : * else - tuple before last returned.
1564 : */
1565 30 : if (state->eof_reached)
1566 : {
1567 : /*
1568 : * Seek position is pointing just past the zero tuplen at the
1569 : * end of file; back up to fetch last tuple's ending length
1570 : * word. If seek fails we must have a completely empty file.
1571 : */
1572 12 : nmoved = LogicalTapeBackspace(state->result_tape,
1573 : 2 * sizeof(unsigned int));
1574 12 : if (nmoved == 0)
1575 0 : return false;
1576 12 : else if (nmoved != 2 * sizeof(unsigned int))
1577 0 : elog(ERROR, "unexpected tape position");
1578 12 : state->eof_reached = false;
1579 : }
1580 : else
1581 : {
1582 : /*
1583 : * Back up and fetch previously-returned tuple's ending length
1584 : * word. If seek fails, assume we are at start of file.
1585 : */
1586 18 : nmoved = LogicalTapeBackspace(state->result_tape,
1587 : sizeof(unsigned int));
1588 18 : if (nmoved == 0)
1589 0 : return false;
1590 18 : else if (nmoved != sizeof(unsigned int))
1591 0 : elog(ERROR, "unexpected tape position");
1592 18 : tuplen = getlen(state->result_tape, false);
1593 :
1594 : /*
1595 : * Back up to get ending length word of tuple before it.
1596 : */
1597 18 : nmoved = LogicalTapeBackspace(state->result_tape,
1598 : tuplen + 2 * sizeof(unsigned int));
1599 18 : if (nmoved == tuplen + sizeof(unsigned int))
1600 : {
1601 : /*
1602 : * We backed up over the previous tuple, but there was no
1603 : * ending length word before it. That means that the prev
1604 : * tuple is the first tuple in the file. It is now the
1605 : * next to read in forward direction (not obviously right,
1606 : * but that is what in-memory case does).
1607 : */
1608 6 : return false;
1609 : }
1610 12 : else if (nmoved != tuplen + 2 * sizeof(unsigned int))
1611 0 : elog(ERROR, "bogus tuple length in backward scan");
1612 : }
1613 :
1614 24 : tuplen = getlen(state->result_tape, false);
1615 :
1616 : /*
1617 : * Now we have the length of the prior tuple, back up and read it.
1618 : * Note: READTUP expects we are positioned after the initial
1619 : * length word of the tuple, so back up to that point.
1620 : */
1621 24 : nmoved = LogicalTapeBackspace(state->result_tape,
1622 : tuplen);
1623 24 : if (nmoved != tuplen)
1624 0 : elog(ERROR, "bogus tuple length in backward scan");
1625 24 : READTUP(state, stup, state->result_tape, tuplen);
1626 :
1627 : /*
1628 : * Remember the tuple we return, so that we can recycle its memory
1629 : * on next call. (This can be NULL, in the Datum case).
1630 : */
1631 24 : state->lastReturnedTuple = stup->tuple;
1632 :
1633 24 : return true;
1634 :
1635 4574586 : case TSS_FINALMERGE:
1636 : Assert(forward);
1637 : /* We are managing memory ourselves, with the slab allocator. */
1638 : Assert(state->slabAllocatorUsed);
1639 :
1640 : /*
1641 : * The slab slot holding the tuple that we returned in previous
1642 : * gettuple call can now be reused.
1643 : */
1644 4574586 : if (state->lastReturnedTuple)
1645 : {
1646 4514296 : RELEASE_SLAB_SLOT(state, state->lastReturnedTuple);
1647 4514296 : state->lastReturnedTuple = NULL;
1648 : }
1649 :
1650 : /*
1651 : * This code should match the inner loop of mergeonerun().
1652 : */
1653 4574586 : if (state->memtupcount > 0)
1654 : {
1655 4574344 : int srcTapeIndex = state->memtuples[0].srctape;
1656 4574344 : LogicalTape *srcTape = state->inputTapes[srcTapeIndex];
1657 : SortTuple newtup;
1658 :
1659 4574344 : *stup = state->memtuples[0];
1660 :
1661 : /*
1662 : * Remember the tuple we return, so that we can recycle its
1663 : * memory on next call. (This can be NULL, in the Datum case).
1664 : */
1665 4574344 : state->lastReturnedTuple = stup->tuple;
1666 :
1667 : /*
1668 : * Pull next tuple from tape, and replace the returned tuple
1669 : * at top of the heap with it.
1670 : */
1671 4574344 : if (!mergereadnext(state, srcTape, &newtup))
1672 : {
1673 : /*
1674 : * If no more data, we've reached end of run on this tape.
1675 : * Remove the top node from the heap.
1676 : */
1677 354 : tuplesort_heap_delete_top(state);
1678 354 : state->nInputRuns--;
1679 :
1680 : /*
1681 : * Close the tape. It'd go away at the end of the sort
1682 : * anyway, but better to release the memory early.
1683 : */
1684 354 : LogicalTapeClose(srcTape);
1685 354 : return true;
1686 : }
1687 4573990 : newtup.srctape = srcTapeIndex;
1688 4573990 : tuplesort_heap_replace_top(state, &newtup);
1689 4573990 : return true;
1690 : }
1691 242 : return false;
1692 :
1693 0 : default:
1694 0 : elog(ERROR, "invalid tuplesort state");
1695 : return false; /* keep compiler quiet */
1696 : }
1697 : }
1698 :
1699 :
1700 : /*
1701 : * Advance over N tuples in either forward or back direction,
1702 : * without returning any data. N==0 is a no-op.
1703 : * Returns true if successful, false if ran out of tuples.
1704 : */
1705 : bool
1706 392 : tuplesort_skiptuples(Tuplesortstate *state, int64 ntuples, bool forward)
1707 : {
1708 : MemoryContext oldcontext;
1709 :
1710 : /*
1711 : * We don't actually support backwards skip yet, because no callers need
1712 : * it. The API is designed to allow for that later, though.
1713 : */
1714 : Assert(forward);
1715 : Assert(ntuples >= 0);
1716 : Assert(!WORKER(state));
1717 :
1718 392 : switch (state->status)
1719 : {
1720 368 : case TSS_SORTEDINMEM:
1721 368 : if (state->memtupcount - state->current >= ntuples)
1722 : {
1723 368 : state->current += ntuples;
1724 368 : return true;
1725 : }
1726 0 : state->current = state->memtupcount;
1727 0 : state->eof_reached = true;
1728 :
1729 : /*
1730 : * Complain if caller tries to retrieve more tuples than
1731 : * originally asked for in a bounded sort. This is because
1732 : * returning EOF here might be the wrong thing.
1733 : */
1734 0 : if (state->bounded && state->current >= state->bound)
1735 0 : elog(ERROR, "retrieved too many tuples in a bounded sort");
1736 :
1737 0 : return false;
1738 :
1739 24 : case TSS_SORTEDONTAPE:
1740 : case TSS_FINALMERGE:
1741 :
1742 : /*
1743 : * We could probably optimize these cases better, but for now it's
1744 : * not worth the trouble.
1745 : */
1746 24 : oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
1747 240132 : while (ntuples-- > 0)
1748 : {
1749 : SortTuple stup;
1750 :
1751 240108 : if (!tuplesort_gettuple_common(state, forward, &stup))
1752 : {
1753 0 : MemoryContextSwitchTo(oldcontext);
1754 0 : return false;
1755 : }
1756 240108 : CHECK_FOR_INTERRUPTS();
1757 : }
1758 24 : MemoryContextSwitchTo(oldcontext);
1759 24 : return true;
1760 :
1761 0 : default:
1762 0 : elog(ERROR, "invalid tuplesort state");
1763 : return false; /* keep compiler quiet */
1764 : }
1765 : }
1766 :
1767 : /*
1768 : * tuplesort_merge_order - report merge order we'll use for given memory
1769 : * (note: "merge order" just means the number of input tapes in the merge).
1770 : *
1771 : * This is exported for use by the planner. allowedMem is in bytes.
1772 : */
1773 : int
1774 17934 : tuplesort_merge_order(int64 allowedMem)
1775 : {
1776 : int mOrder;
1777 :
1778 : /*----------
1779 : * In the merge phase, we need buffer space for each input and output tape.
1780 : * Each pass in the balanced merge algorithm reads from M input tapes, and
1781 : * writes to N output tapes. Each tape consumes TAPE_BUFFER_OVERHEAD bytes
1782 : * of memory. In addition to that, we want MERGE_BUFFER_SIZE workspace per
1783 : * input tape.
1784 : *
1785 : * totalMem = M * (TAPE_BUFFER_OVERHEAD + MERGE_BUFFER_SIZE) +
1786 : * N * TAPE_BUFFER_OVERHEAD
1787 : *
1788 : * Except for the last and next-to-last merge passes, where there can be
1789 : * fewer tapes left to process, M = N. We choose M so that we have the
1790 : * desired amount of memory available for the input buffers
1791 : * (TAPE_BUFFER_OVERHEAD + MERGE_BUFFER_SIZE), given the total memory
1792 : * available for the tape buffers (allowedMem).
1793 : *
1794 : * Note: you might be thinking we need to account for the memtuples[]
1795 : * array in this calculation, but we effectively treat that as part of the
1796 : * MERGE_BUFFER_SIZE workspace.
1797 : *----------
1798 : */
1799 17934 : mOrder = allowedMem /
1800 : (2 * TAPE_BUFFER_OVERHEAD + MERGE_BUFFER_SIZE);
1801 :
1802 : /*
1803 : * Even in minimum memory, use at least a MINORDER merge. On the other
1804 : * hand, even when we have lots of memory, do not use more than a MAXORDER
1805 : * merge. Tapes are pretty cheap, but they're not entirely free. Each
1806 : * additional tape reduces the amount of memory available to build runs,
1807 : * which in turn can cause the same sort to need more runs, which makes
1808 : * merging slower even if it can still be done in a single pass. Also,
1809 : * high order merges are quite slow due to CPU cache effects; it can be
1810 : * faster to pay the I/O cost of a multi-pass merge than to perform a
1811 : * single merge pass across many hundreds of tapes.
1812 : */
1813 17934 : mOrder = Max(mOrder, MINORDER);
1814 17934 : mOrder = Min(mOrder, MAXORDER);
1815 :
1816 17934 : return mOrder;
1817 : }
1818 :
1819 : /*
1820 : * Helper function to calculate how much memory to allocate for the read buffer
1821 : * of each input tape in a merge pass.
1822 : *
1823 : * 'avail_mem' is the amount of memory available for the buffers of all the
1824 : * tapes, both input and output.
1825 : * 'nInputTapes' and 'nInputRuns' are the number of input tapes and runs.
1826 : * 'maxOutputTapes' is the max. number of output tapes we should produce.
1827 : */
1828 : static int64
1829 308 : merge_read_buffer_size(int64 avail_mem, int nInputTapes, int nInputRuns,
1830 : int maxOutputTapes)
1831 : {
1832 : int nOutputRuns;
1833 : int nOutputTapes;
1834 :
1835 : /*
1836 : * How many output tapes will we produce in this pass?
1837 : *
1838 : * This is nInputRuns / nInputTapes, rounded up.
1839 : */
1840 308 : nOutputRuns = (nInputRuns + nInputTapes - 1) / nInputTapes;
1841 :
1842 308 : nOutputTapes = Min(nOutputRuns, maxOutputTapes);
1843 :
1844 : /*
1845 : * Each output tape consumes TAPE_BUFFER_OVERHEAD bytes of memory. All
1846 : * remaining memory is divided evenly between the input tapes.
1847 : *
1848 : * This also follows from the formula in tuplesort_merge_order, but here
1849 : * we derive the input buffer size from the amount of memory available,
1850 : * and M and N.
1851 : */
1852 308 : return Max((avail_mem - TAPE_BUFFER_OVERHEAD * nOutputTapes) / nInputTapes, 0);
1853 : }
1854 :
1855 : /*
1856 : * inittapes - initialize for tape sorting.
1857 : *
1858 : * This is called only if we have found we won't sort in memory.
1859 : */
1860 : static void
1861 566 : inittapes(Tuplesortstate *state, bool mergeruns)
1862 : {
1863 : Assert(!LEADER(state));
1864 :
1865 566 : if (mergeruns)
1866 : {
1867 : /* Compute number of input tapes to use when merging */
1868 122 : state->maxTapes = tuplesort_merge_order(state->allowedMem);
1869 : }
1870 : else
1871 : {
1872 : /* Workers can sometimes produce single run, output without merge */
1873 : Assert(WORKER(state));
1874 444 : state->maxTapes = MINORDER;
1875 : }
1876 :
1877 566 : if (trace_sort)
1878 0 : elog(LOG, "worker %d switching to external sort with %d tapes: %s",
1879 : state->worker, state->maxTapes, pg_rusage_show(&state->ru_start));
1880 :
1881 : /* Create the tape set */
1882 566 : inittapestate(state, state->maxTapes);
1883 566 : state->tapeset =
1884 566 : LogicalTapeSetCreate(false,
1885 566 : state->shared ? &state->shared->fileset : NULL,
1886 : state->worker);
1887 :
1888 566 : state->currentRun = 0;
1889 :
1890 : /*
1891 : * Initialize logical tape arrays.
1892 : */
1893 566 : state->inputTapes = NULL;
1894 566 : state->nInputTapes = 0;
1895 566 : state->nInputRuns = 0;
1896 :
1897 566 : state->outputTapes = palloc0(state->maxTapes * sizeof(LogicalTape *));
1898 566 : state->nOutputTapes = 0;
1899 566 : state->nOutputRuns = 0;
1900 :
1901 566 : state->status = TSS_BUILDRUNS;
1902 :
1903 566 : selectnewtape(state);
1904 566 : }
1905 :
1906 : /*
1907 : * inittapestate - initialize generic tape management state
1908 : */
1909 : static void
1910 722 : inittapestate(Tuplesortstate *state, int maxTapes)
1911 : {
1912 : int64 tapeSpace;
1913 :
1914 : /*
1915 : * Decrease availMem to reflect the space needed for tape buffers; but
1916 : * don't decrease it to the point that we have no room for tuples. (That
1917 : * case is only likely to occur if sorting pass-by-value Datums; in all
1918 : * other scenarios the memtuples[] array is unlikely to occupy more than
1919 : * half of allowedMem. In the pass-by-value case it's not important to
1920 : * account for tuple space, so we don't care if LACKMEM becomes
1921 : * inaccurate.)
1922 : */
1923 722 : tapeSpace = (int64) maxTapes * TAPE_BUFFER_OVERHEAD;
1924 :
1925 722 : if (tapeSpace + GetMemoryChunkSpace(state->memtuples) < state->allowedMem)
1926 618 : USEMEM(state, tapeSpace);
1927 :
1928 : /*
1929 : * Make sure that the temp file(s) underlying the tape set are created in
1930 : * suitable temp tablespaces. For parallel sorts, this should have been
1931 : * called already, but it doesn't matter if it is called a second time.
1932 : */
1933 722 : PrepareTempTablespaces();
1934 722 : }
1935 :
1936 : /*
1937 : * selectnewtape -- select next tape to output to.
1938 : *
1939 : * This is called after finishing a run when we know another run
1940 : * must be started. This is used both when building the initial
1941 : * runs, and during merge passes.
1942 : */
1943 : static void
1944 1642 : selectnewtape(Tuplesortstate *state)
1945 : {
1946 : /*
1947 : * At the beginning of each merge pass, nOutputTapes and nOutputRuns are
1948 : * both zero. On each call, we create a new output tape to hold the next
1949 : * run, until maxTapes is reached. After that, we assign new runs to the
1950 : * existing tapes in a round robin fashion.
1951 : */
1952 1642 : if (state->nOutputTapes < state->maxTapes)
1953 : {
1954 : /* Create a new tape to hold the next run */
1955 : Assert(state->outputTapes[state->nOutputRuns] == NULL);
1956 : Assert(state->nOutputRuns == state->nOutputTapes);
1957 1060 : state->destTape = LogicalTapeCreate(state->tapeset);
1958 1060 : state->outputTapes[state->nOutputTapes] = state->destTape;
1959 1060 : state->nOutputTapes++;
1960 1060 : state->nOutputRuns++;
1961 : }
1962 : else
1963 : {
1964 : /*
1965 : * We have reached the max number of tapes. Append to an existing
1966 : * tape.
1967 : */
1968 582 : state->destTape = state->outputTapes[state->nOutputRuns % state->nOutputTapes];
1969 582 : state->nOutputRuns++;
1970 : }
1971 1642 : }
1972 :
1973 : /*
1974 : * Initialize the slab allocation arena, for the given number of slots.
1975 : */
1976 : static void
1977 278 : init_slab_allocator(Tuplesortstate *state, int numSlots)
1978 : {
1979 278 : if (numSlots > 0)
1980 : {
1981 : char *p;
1982 : int i;
1983 :
1984 266 : state->slabMemoryBegin = palloc(numSlots * SLAB_SLOT_SIZE);
1985 266 : state->slabMemoryEnd = state->slabMemoryBegin +
1986 266 : numSlots * SLAB_SLOT_SIZE;
1987 266 : state->slabFreeHead = (SlabSlot *) state->slabMemoryBegin;
1988 266 : USEMEM(state, numSlots * SLAB_SLOT_SIZE);
1989 :
1990 266 : p = state->slabMemoryBegin;
1991 1012 : for (i = 0; i < numSlots - 1; i++)
1992 : {
1993 746 : ((SlabSlot *) p)->nextfree = (SlabSlot *) (p + SLAB_SLOT_SIZE);
1994 746 : p += SLAB_SLOT_SIZE;
1995 : }
1996 266 : ((SlabSlot *) p)->nextfree = NULL;
1997 : }
1998 : else
1999 : {
2000 12 : state->slabMemoryBegin = state->slabMemoryEnd = NULL;
2001 12 : state->slabFreeHead = NULL;
2002 : }
2003 278 : state->slabAllocatorUsed = true;
2004 278 : }
2005 :
2006 : /*
2007 : * mergeruns -- merge all the completed initial runs.
2008 : *
2009 : * This implements the Balanced k-Way Merge Algorithm. All input data has
2010 : * already been written to initial runs on tape (see dumptuples).
2011 : */
2012 : static void
2013 278 : mergeruns(Tuplesortstate *state)
2014 : {
2015 : int tapenum;
2016 :
2017 : Assert(state->status == TSS_BUILDRUNS);
2018 : Assert(state->memtupcount == 0);
2019 :
2020 278 : if (state->base.sortKeys != NULL && state->base.sortKeys->abbrev_converter != NULL)
2021 : {
2022 : /*
2023 : * If there are multiple runs to be merged, when we go to read back
2024 : * tuples from disk, abbreviated keys will not have been stored, and
2025 : * we don't care to regenerate them. Disable abbreviation from this
2026 : * point on.
2027 : */
2028 30 : state->base.sortKeys->abbrev_converter = NULL;
2029 30 : state->base.sortKeys->comparator = state->base.sortKeys->abbrev_full_comparator;
2030 :
2031 : /* Not strictly necessary, but be tidy */
2032 30 : state->base.sortKeys->abbrev_abort = NULL;
2033 30 : state->base.sortKeys->abbrev_full_comparator = NULL;
2034 : }
2035 :
2036 : /*
2037 : * Reset tuple memory. We've freed all the tuples that we previously
2038 : * allocated. We will use the slab allocator from now on.
2039 : */
2040 278 : MemoryContextResetOnly(state->base.tuplecontext);
2041 :
2042 : /*
2043 : * We no longer need a large memtuples array. (We will allocate a smaller
2044 : * one for the heap later.)
2045 : */
2046 278 : FREEMEM(state, GetMemoryChunkSpace(state->memtuples));
2047 278 : pfree(state->memtuples);
2048 278 : state->memtuples = NULL;
2049 :
2050 : /*
2051 : * Initialize the slab allocator. We need one slab slot per input tape,
2052 : * for the tuples in the heap, plus one to hold the tuple last returned
2053 : * from tuplesort_gettuple. (If we're sorting pass-by-val Datums,
2054 : * however, we don't need to do allocate anything.)
2055 : *
2056 : * In a multi-pass merge, we could shrink this allocation for the last
2057 : * merge pass, if it has fewer tapes than previous passes, but we don't
2058 : * bother.
2059 : *
2060 : * From this point on, we no longer use the USEMEM()/LACKMEM() mechanism
2061 : * to track memory usage of individual tuples.
2062 : */
2063 278 : if (state->base.tuples)
2064 266 : init_slab_allocator(state, state->nOutputTapes + 1);
2065 : else
2066 12 : init_slab_allocator(state, 0);
2067 :
2068 : /*
2069 : * Allocate a new 'memtuples' array, for the heap. It will hold one tuple
2070 : * from each input tape.
2071 : *
2072 : * We could shrink this, too, between passes in a multi-pass merge, but we
2073 : * don't bother. (The initial input tapes are still in outputTapes. The
2074 : * number of input tapes will not increase between passes.)
2075 : */
2076 278 : state->memtupsize = state->nOutputTapes;
2077 556 : state->memtuples = (SortTuple *) MemoryContextAlloc(state->base.maincontext,
2078 278 : state->nOutputTapes * sizeof(SortTuple));
2079 278 : USEMEM(state, GetMemoryChunkSpace(state->memtuples));
2080 :
2081 : /*
2082 : * Use all the remaining memory we have available for tape buffers among
2083 : * all the input tapes. At the beginning of each merge pass, we will
2084 : * divide this memory between the input and output tapes in the pass.
2085 : */
2086 278 : state->tape_buffer_mem = state->availMem;
2087 278 : USEMEM(state, state->tape_buffer_mem);
2088 278 : if (trace_sort)
2089 0 : elog(LOG, "worker %d using %zu KB of memory for tape buffers",
2090 : state->worker, state->tape_buffer_mem / 1024);
2091 :
2092 : for (;;)
2093 : {
2094 : /*
2095 : * On the first iteration, or if we have read all the runs from the
2096 : * input tapes in a multi-pass merge, it's time to start a new pass.
2097 : * Rewind all the output tapes, and make them inputs for the next
2098 : * pass.
2099 : */
2100 416 : if (state->nInputRuns == 0)
2101 : {
2102 : int64 input_buffer_size;
2103 :
2104 : /* Close the old, emptied, input tapes */
2105 308 : if (state->nInputTapes > 0)
2106 : {
2107 210 : for (tapenum = 0; tapenum < state->nInputTapes; tapenum++)
2108 180 : LogicalTapeClose(state->inputTapes[tapenum]);
2109 30 : pfree(state->inputTapes);
2110 : }
2111 :
2112 : /* Previous pass's outputs become next pass's inputs. */
2113 308 : state->inputTapes = state->outputTapes;
2114 308 : state->nInputTapes = state->nOutputTapes;
2115 308 : state->nInputRuns = state->nOutputRuns;
2116 :
2117 : /*
2118 : * Reset output tape variables. The actual LogicalTapes will be
2119 : * created as needed, here we only allocate the array to hold
2120 : * them.
2121 : */
2122 308 : state->outputTapes = palloc0(state->nInputTapes * sizeof(LogicalTape *));
2123 308 : state->nOutputTapes = 0;
2124 308 : state->nOutputRuns = 0;
2125 :
2126 : /*
2127 : * Redistribute the memory allocated for tape buffers, among the
2128 : * new input and output tapes.
2129 : */
2130 308 : input_buffer_size = merge_read_buffer_size(state->tape_buffer_mem,
2131 : state->nInputTapes,
2132 : state->nInputRuns,
2133 : state->maxTapes);
2134 :
2135 308 : if (trace_sort)
2136 0 : elog(LOG, "starting merge pass of %d input runs on %d tapes, " INT64_FORMAT " KB of memory for each input tape: %s",
2137 : state->nInputRuns, state->nInputTapes, input_buffer_size / 1024,
2138 : pg_rusage_show(&state->ru_start));
2139 :
2140 : /* Prepare the new input tapes for merge pass. */
2141 1222 : for (tapenum = 0; tapenum < state->nInputTapes; tapenum++)
2142 914 : LogicalTapeRewindForRead(state->inputTapes[tapenum], input_buffer_size);
2143 :
2144 : /*
2145 : * If there's just one run left on each input tape, then only one
2146 : * merge pass remains. If we don't have to produce a materialized
2147 : * sorted tape, we can stop at this point and do the final merge
2148 : * on-the-fly.
2149 : */
2150 308 : if ((state->base.sortopt & TUPLESORT_RANDOMACCESS) == 0
2151 290 : && state->nInputRuns <= state->nInputTapes
2152 260 : && !WORKER(state))
2153 : {
2154 : /* Tell logtape.c we won't be writing anymore */
2155 260 : LogicalTapeSetForgetFreeSpace(state->tapeset);
2156 : /* Initialize for the final merge pass */
2157 260 : beginmerge(state);
2158 260 : state->status = TSS_FINALMERGE;
2159 260 : return;
2160 : }
2161 : }
2162 :
2163 : /* Select an output tape */
2164 156 : selectnewtape(state);
2165 :
2166 : /* Merge one run from each input tape. */
2167 156 : mergeonerun(state);
2168 :
2169 : /*
2170 : * If the input tapes are empty, and we output only one output run,
2171 : * we're done. The current output tape contains the final result.
2172 : */
2173 156 : if (state->nInputRuns == 0 && state->nOutputRuns <= 1)
2174 18 : break;
2175 : }
2176 :
2177 : /*
2178 : * Done. The result is on a single run on a single tape.
2179 : */
2180 18 : state->result_tape = state->outputTapes[0];
2181 18 : if (!WORKER(state))
2182 18 : LogicalTapeFreeze(state->result_tape, NULL);
2183 : else
2184 0 : worker_freeze_result_tape(state);
2185 18 : state->status = TSS_SORTEDONTAPE;
2186 :
2187 : /* Close all the now-empty input tapes, to release their read buffers. */
2188 102 : for (tapenum = 0; tapenum < state->nInputTapes; tapenum++)
2189 84 : LogicalTapeClose(state->inputTapes[tapenum]);
2190 : }
2191 :
2192 : /*
2193 : * Merge one run from each input tape.
2194 : */
2195 : static void
2196 156 : mergeonerun(Tuplesortstate *state)
2197 : {
2198 : int srcTapeIndex;
2199 : LogicalTape *srcTape;
2200 :
2201 : /*
2202 : * Start the merge by loading one tuple from each active source tape into
2203 : * the heap.
2204 : */
2205 156 : beginmerge(state);
2206 :
2207 : Assert(state->slabAllocatorUsed);
2208 :
2209 : /*
2210 : * Execute merge by repeatedly extracting lowest tuple in heap, writing it
2211 : * out, and replacing it with next tuple from same tape (if there is
2212 : * another one).
2213 : */
2214 855588 : while (state->memtupcount > 0)
2215 : {
2216 : SortTuple stup;
2217 :
2218 : /* write the tuple to destTape */
2219 855432 : srcTapeIndex = state->memtuples[0].srctape;
2220 855432 : srcTape = state->inputTapes[srcTapeIndex];
2221 855432 : WRITETUP(state, state->destTape, &state->memtuples[0]);
2222 :
2223 : /* recycle the slot of the tuple we just wrote out, for the next read */
2224 855432 : if (state->memtuples[0].tuple)
2225 735348 : RELEASE_SLAB_SLOT(state, state->memtuples[0].tuple);
2226 :
2227 : /*
2228 : * pull next tuple from the tape, and replace the written-out tuple in
2229 : * the heap with it.
2230 : */
2231 855432 : if (mergereadnext(state, srcTape, &stup))
2232 : {
2233 854586 : stup.srctape = srcTapeIndex;
2234 854586 : tuplesort_heap_replace_top(state, &stup);
2235 : }
2236 : else
2237 : {
2238 846 : tuplesort_heap_delete_top(state);
2239 846 : state->nInputRuns--;
2240 : }
2241 : }
2242 :
2243 : /*
2244 : * When the heap empties, we're done. Write an end-of-run marker on the
2245 : * output tape.
2246 : */
2247 156 : markrunend(state->destTape);
2248 156 : }
2249 :
2250 : /*
2251 : * beginmerge - initialize for a merge pass
2252 : *
2253 : * Fill the merge heap with the first tuple from each input tape.
2254 : */
2255 : static void
2256 416 : beginmerge(Tuplesortstate *state)
2257 : {
2258 : int activeTapes;
2259 : int srcTapeIndex;
2260 :
2261 : /* Heap should be empty here */
2262 : Assert(state->memtupcount == 0);
2263 :
2264 416 : activeTapes = Min(state->nInputTapes, state->nInputRuns);
2265 :
2266 1912 : for (srcTapeIndex = 0; srcTapeIndex < activeTapes; srcTapeIndex++)
2267 : {
2268 : SortTuple tup;
2269 :
2270 1496 : if (mergereadnext(state, state->inputTapes[srcTapeIndex], &tup))
2271 : {
2272 1248 : tup.srctape = srcTapeIndex;
2273 1248 : tuplesort_heap_insert(state, &tup);
2274 : }
2275 : }
2276 416 : }
2277 :
2278 : /*
2279 : * mergereadnext - read next tuple from one merge input tape
2280 : *
2281 : * Returns false on EOF.
2282 : */
2283 : static bool
2284 5431272 : mergereadnext(Tuplesortstate *state, LogicalTape *srcTape, SortTuple *stup)
2285 : {
2286 : unsigned int tuplen;
2287 :
2288 : /* read next tuple, if any */
2289 5431272 : if ((tuplen = getlen(srcTape, true)) == 0)
2290 1448 : return false;
2291 5429824 : READTUP(state, stup, srcTape, tuplen);
2292 :
2293 5429824 : return true;
2294 : }
2295 :
2296 : /*
2297 : * dumptuples - remove tuples from memtuples and write initial run to tape
2298 : *
2299 : * When alltuples = true, dump everything currently in memory. (This case is
2300 : * only used at end of input data.)
2301 : */
2302 : static void
2303 1068994 : dumptuples(Tuplesortstate *state, bool alltuples)
2304 : {
2305 : int memtupwrite;
2306 : int i;
2307 :
2308 : /*
2309 : * Nothing to do if we still fit in available memory and have array slots,
2310 : * unless this is the final call during initial run generation.
2311 : */
2312 1068994 : if (state->memtupcount < state->memtupsize && !LACKMEM(state) &&
2313 1068074 : !alltuples)
2314 1067508 : return;
2315 :
2316 : /*
2317 : * Final call might require no sorting, in rare cases where we just so
2318 : * happen to have previously LACKMEM()'d at the point where exactly all
2319 : * remaining tuples are loaded into memory, just before input was
2320 : * exhausted. In general, short final runs are quite possible, but avoid
2321 : * creating a completely empty run. In a worker, though, we must produce
2322 : * at least one tape, even if it's empty.
2323 : */
2324 1486 : if (state->memtupcount == 0 && state->currentRun > 0)
2325 0 : return;
2326 :
2327 : Assert(state->status == TSS_BUILDRUNS);
2328 :
2329 : /*
2330 : * It seems unlikely that this limit will ever be exceeded, but take no
2331 : * chances
2332 : */
2333 1486 : if (state->currentRun == INT_MAX)
2334 0 : ereport(ERROR,
2335 : (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
2336 : errmsg("cannot have more than %d runs for an external sort",
2337 : INT_MAX)));
2338 :
2339 1486 : if (state->currentRun > 0)
2340 920 : selectnewtape(state);
2341 :
2342 1486 : state->currentRun++;
2343 :
2344 1486 : if (trace_sort)
2345 0 : elog(LOG, "worker %d starting quicksort of run %d: %s",
2346 : state->worker, state->currentRun,
2347 : pg_rusage_show(&state->ru_start));
2348 :
2349 : /*
2350 : * Sort all tuples accumulated within the allowed amount of memory for
2351 : * this run using quicksort
2352 : */
2353 1486 : tuplesort_sort_memtuples(state);
2354 :
2355 1486 : if (trace_sort)
2356 0 : elog(LOG, "worker %d finished quicksort of run %d: %s",
2357 : state->worker, state->currentRun,
2358 : pg_rusage_show(&state->ru_start));
2359 :
2360 1486 : memtupwrite = state->memtupcount;
2361 5071082 : for (i = 0; i < memtupwrite; i++)
2362 : {
2363 5069596 : SortTuple *stup = &state->memtuples[i];
2364 :
2365 5069596 : WRITETUP(state, state->destTape, stup);
2366 : }
2367 :
2368 1486 : state->memtupcount = 0;
2369 :
2370 : /*
2371 : * Reset tuple memory. We've freed all of the tuples that we previously
2372 : * allocated. It's important to avoid fragmentation when there is a stark
2373 : * change in the sizes of incoming tuples. In bounded sorts,
2374 : * fragmentation due to AllocSetFree's bucketing by size class might be
2375 : * particularly bad if this step wasn't taken.
2376 : */
2377 1486 : MemoryContextReset(state->base.tuplecontext);
2378 :
2379 : /*
2380 : * Now update the memory accounting to subtract the memory used by the
2381 : * tuple.
2382 : */
2383 1486 : FREEMEM(state, state->tupleMem);
2384 1486 : state->tupleMem = 0;
2385 :
2386 1486 : markrunend(state->destTape);
2387 :
2388 1486 : if (trace_sort)
2389 0 : elog(LOG, "worker %d finished writing run %d to tape %d: %s",
2390 : state->worker, state->currentRun, (state->currentRun - 1) % state->nOutputTapes + 1,
2391 : pg_rusage_show(&state->ru_start));
2392 : }
2393 :
2394 : /*
2395 : * tuplesort_rescan - rewind and replay the scan
2396 : */
2397 : void
2398 58 : tuplesort_rescan(Tuplesortstate *state)
2399 : {
2400 58 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
2401 :
2402 : Assert(state->base.sortopt & TUPLESORT_RANDOMACCESS);
2403 :
2404 58 : switch (state->status)
2405 : {
2406 52 : case TSS_SORTEDINMEM:
2407 52 : state->current = 0;
2408 52 : state->eof_reached = false;
2409 52 : state->markpos_offset = 0;
2410 52 : state->markpos_eof = false;
2411 52 : break;
2412 6 : case TSS_SORTEDONTAPE:
2413 6 : LogicalTapeRewindForRead(state->result_tape, 0);
2414 6 : state->eof_reached = false;
2415 6 : state->markpos_block = 0L;
2416 6 : state->markpos_offset = 0;
2417 6 : state->markpos_eof = false;
2418 6 : break;
2419 0 : default:
2420 0 : elog(ERROR, "invalid tuplesort state");
2421 : break;
2422 : }
2423 :
2424 58 : MemoryContextSwitchTo(oldcontext);
2425 58 : }
2426 :
2427 : /*
2428 : * tuplesort_markpos - saves current position in the merged sort file
2429 : */
2430 : void
2431 586202 : tuplesort_markpos(Tuplesortstate *state)
2432 : {
2433 586202 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
2434 :
2435 : Assert(state->base.sortopt & TUPLESORT_RANDOMACCESS);
2436 :
2437 586202 : switch (state->status)
2438 : {
2439 577394 : case TSS_SORTEDINMEM:
2440 577394 : state->markpos_offset = state->current;
2441 577394 : state->markpos_eof = state->eof_reached;
2442 577394 : break;
2443 8808 : case TSS_SORTEDONTAPE:
2444 8808 : LogicalTapeTell(state->result_tape,
2445 : &state->markpos_block,
2446 : &state->markpos_offset);
2447 8808 : state->markpos_eof = state->eof_reached;
2448 8808 : break;
2449 0 : default:
2450 0 : elog(ERROR, "invalid tuplesort state");
2451 : break;
2452 : }
2453 :
2454 586202 : MemoryContextSwitchTo(oldcontext);
2455 586202 : }
2456 :
2457 : /*
2458 : * tuplesort_restorepos - restores current position in merged sort file to
2459 : * last saved position
2460 : */
2461 : void
2462 32786 : tuplesort_restorepos(Tuplesortstate *state)
2463 : {
2464 32786 : MemoryContext oldcontext = MemoryContextSwitchTo(state->base.sortcontext);
2465 :
2466 : Assert(state->base.sortopt & TUPLESORT_RANDOMACCESS);
2467 :
2468 32786 : switch (state->status)
2469 : {
2470 26594 : case TSS_SORTEDINMEM:
2471 26594 : state->current = state->markpos_offset;
2472 26594 : state->eof_reached = state->markpos_eof;
2473 26594 : break;
2474 6192 : case TSS_SORTEDONTAPE:
2475 6192 : LogicalTapeSeek(state->result_tape,
2476 : state->markpos_block,
2477 : state->markpos_offset);
2478 6192 : state->eof_reached = state->markpos_eof;
2479 6192 : break;
2480 0 : default:
2481 0 : elog(ERROR, "invalid tuplesort state");
2482 : break;
2483 : }
2484 :
2485 32786 : MemoryContextSwitchTo(oldcontext);
2486 32786 : }
2487 :
2488 : /*
2489 : * tuplesort_get_stats - extract summary statistics
2490 : *
2491 : * This can be called after tuplesort_performsort() finishes to obtain
2492 : * printable summary information about how the sort was performed.
2493 : */
2494 : void
2495 396 : tuplesort_get_stats(Tuplesortstate *state,
2496 : TuplesortInstrumentation *stats)
2497 : {
2498 : /*
2499 : * Note: it might seem we should provide both memory and disk usage for a
2500 : * disk-based sort. However, the current code doesn't track memory space
2501 : * accurately once we have begun to return tuples to the caller (since we
2502 : * don't account for pfree's the caller is expected to do), so we cannot
2503 : * rely on availMem in a disk sort. This does not seem worth the overhead
2504 : * to fix. Is it worth creating an API for the memory context code to
2505 : * tell us how much is actually used in sortcontext?
2506 : */
2507 396 : tuplesort_updatemax(state);
2508 :
2509 396 : if (state->isMaxSpaceDisk)
2510 6 : stats->spaceType = SORT_SPACE_TYPE_DISK;
2511 : else
2512 390 : stats->spaceType = SORT_SPACE_TYPE_MEMORY;
2513 396 : stats->spaceUsed = (state->maxSpace + 1023) / 1024;
2514 :
2515 396 : switch (state->maxSpaceStatus)
2516 : {
2517 390 : case TSS_SORTEDINMEM:
2518 390 : if (state->boundUsed)
2519 42 : stats->sortMethod = SORT_TYPE_TOP_N_HEAPSORT;
2520 : else
2521 348 : stats->sortMethod = SORT_TYPE_QUICKSORT;
2522 390 : break;
2523 0 : case TSS_SORTEDONTAPE:
2524 0 : stats->sortMethod = SORT_TYPE_EXTERNAL_SORT;
2525 0 : break;
2526 6 : case TSS_FINALMERGE:
2527 6 : stats->sortMethod = SORT_TYPE_EXTERNAL_MERGE;
2528 6 : break;
2529 0 : default:
2530 0 : stats->sortMethod = SORT_TYPE_STILL_IN_PROGRESS;
2531 0 : break;
2532 : }
2533 396 : }
2534 :
2535 : /*
2536 : * Convert TuplesortMethod to a string.
2537 : */
2538 : const char *
2539 294 : tuplesort_method_name(TuplesortMethod m)
2540 : {
2541 294 : switch (m)
2542 : {
2543 0 : case SORT_TYPE_STILL_IN_PROGRESS:
2544 0 : return "still in progress";
2545 42 : case SORT_TYPE_TOP_N_HEAPSORT:
2546 42 : return "top-N heapsort";
2547 246 : case SORT_TYPE_QUICKSORT:
2548 246 : return "quicksort";
2549 0 : case SORT_TYPE_EXTERNAL_SORT:
2550 0 : return "external sort";
2551 6 : case SORT_TYPE_EXTERNAL_MERGE:
2552 6 : return "external merge";
2553 : }
2554 :
2555 0 : return "unknown";
2556 : }
2557 :
2558 : /*
2559 : * Convert TuplesortSpaceType to a string.
2560 : */
2561 : const char *
2562 258 : tuplesort_space_type_name(TuplesortSpaceType t)
2563 : {
2564 : Assert(t == SORT_SPACE_TYPE_DISK || t == SORT_SPACE_TYPE_MEMORY);
2565 258 : return t == SORT_SPACE_TYPE_DISK ? "Disk" : "Memory";
2566 : }
2567 :
2568 :
2569 : /*
2570 : * Heap manipulation routines, per Knuth's Algorithm 5.2.3H.
2571 : */
2572 :
2573 : /*
2574 : * Convert the existing unordered array of SortTuples to a bounded heap,
2575 : * discarding all but the smallest "state->bound" tuples.
2576 : *
2577 : * When working with a bounded heap, we want to keep the largest entry
2578 : * at the root (array entry zero), instead of the smallest as in the normal
2579 : * sort case. This allows us to discard the largest entry cheaply.
2580 : * Therefore, we temporarily reverse the sort direction.
2581 : */
2582 : static void
2583 408 : make_bounded_heap(Tuplesortstate *state)
2584 : {
2585 408 : int tupcount = state->memtupcount;
2586 : int i;
2587 :
2588 : Assert(state->status == TSS_INITIAL);
2589 : Assert(state->bounded);
2590 : Assert(tupcount >= state->bound);
2591 : Assert(SERIAL(state));
2592 :
2593 : /* Reverse sort direction so largest entry will be at root */
2594 408 : reversedirection(state);
2595 :
2596 408 : state->memtupcount = 0; /* make the heap empty */
2597 37980 : for (i = 0; i < tupcount; i++)
2598 : {
2599 37572 : if (state->memtupcount < state->bound)
2600 : {
2601 : /* Insert next tuple into heap */
2602 : /* Must copy source tuple to avoid possible overwrite */
2603 18582 : SortTuple stup = state->memtuples[i];
2604 :
2605 18582 : tuplesort_heap_insert(state, &stup);
2606 : }
2607 : else
2608 : {
2609 : /*
2610 : * The heap is full. Replace the largest entry with the new
2611 : * tuple, or just discard it, if it's larger than anything already
2612 : * in the heap.
2613 : */
2614 18990 : if (COMPARETUP(state, &state->memtuples[i], &state->memtuples[0]) <= 0)
2615 : {
2616 9802 : free_sort_tuple(state, &state->memtuples[i]);
2617 9802 : CHECK_FOR_INTERRUPTS();
2618 : }
2619 : else
2620 9188 : tuplesort_heap_replace_top(state, &state->memtuples[i]);
2621 : }
2622 : }
2623 :
2624 : Assert(state->memtupcount == state->bound);
2625 408 : state->status = TSS_BOUNDED;
2626 408 : }
2627 :
2628 : /*
2629 : * Convert the bounded heap to a properly-sorted array
2630 : */
2631 : static void
2632 408 : sort_bounded_heap(Tuplesortstate *state)
2633 : {
2634 408 : int tupcount = state->memtupcount;
2635 :
2636 : Assert(state->status == TSS_BOUNDED);
2637 : Assert(state->bounded);
2638 : Assert(tupcount == state->bound);
2639 : Assert(SERIAL(state));
2640 :
2641 : /*
2642 : * We can unheapify in place because each delete-top call will remove the
2643 : * largest entry, which we can promptly store in the newly freed slot at
2644 : * the end. Once we're down to a single-entry heap, we're done.
2645 : */
2646 18582 : while (state->memtupcount > 1)
2647 : {
2648 18174 : SortTuple stup = state->memtuples[0];
2649 :
2650 : /* this sifts-up the next-largest entry and decreases memtupcount */
2651 18174 : tuplesort_heap_delete_top(state);
2652 18174 : state->memtuples[state->memtupcount] = stup;
2653 : }
2654 408 : state->memtupcount = tupcount;
2655 :
2656 : /*
2657 : * Reverse sort direction back to the original state. This is not
2658 : * actually necessary but seems like a good idea for tidiness.
2659 : */
2660 408 : reversedirection(state);
2661 :
2662 408 : state->status = TSS_SORTEDINMEM;
2663 408 : state->boundUsed = true;
2664 408 : }
2665 :
2666 : /*
2667 : * Sort all memtuples using specialized qsort() routines.
2668 : *
2669 : * Quicksort is used for small in-memory sorts, and external sort runs.
2670 : */
2671 : static void
2672 234462 : tuplesort_sort_memtuples(Tuplesortstate *state)
2673 : {
2674 : Assert(!LEADER(state));
2675 :
2676 234462 : if (state->memtupcount > 1)
2677 : {
2678 : /*
2679 : * Do we have the leading column's value or abbreviation in datum1,
2680 : * and is there a specialization for its comparator?
2681 : */
2682 67062 : if (state->base.haveDatum1 && state->base.sortKeys)
2683 : {
2684 67026 : if (state->base.sortKeys[0].comparator == ssup_datum_unsigned_cmp)
2685 : {
2686 3336 : qsort_tuple_unsigned(state->memtuples,
2687 3336 : state->memtupcount,
2688 : state);
2689 3320 : return;
2690 : }
2691 63690 : else if (state->base.sortKeys[0].comparator == ssup_datum_signed_cmp)
2692 : {
2693 1312 : qsort_tuple_signed(state->memtuples,
2694 1312 : state->memtupcount,
2695 : state);
2696 1312 : return;
2697 : }
2698 62378 : else if (state->base.sortKeys[0].comparator == ssup_datum_int32_cmp)
2699 : {
2700 39292 : qsort_tuple_int32(state->memtuples,
2701 39292 : state->memtupcount,
2702 : state);
2703 39232 : return;
2704 : }
2705 : }
2706 :
2707 : /* Can we use the single-key sort function? */
2708 23122 : if (state->base.onlyKey != NULL)
2709 : {
2710 9942 : qsort_ssup(state->memtuples, state->memtupcount,
2711 9942 : state->base.onlyKey);
2712 : }
2713 : else
2714 : {
2715 13180 : qsort_tuple(state->memtuples,
2716 13180 : state->memtupcount,
2717 : state->base.comparetup,
2718 : state);
2719 : }
2720 : }
2721 : }
2722 :
2723 : /*
2724 : * Insert a new tuple into an empty or existing heap, maintaining the
2725 : * heap invariant. Caller is responsible for ensuring there's room.
2726 : *
2727 : * Note: For some callers, tuple points to a memtuples[] entry above the
2728 : * end of the heap. This is safe as long as it's not immediately adjacent
2729 : * to the end of the heap (ie, in the [memtupcount] array entry) --- if it
2730 : * is, it might get overwritten before being moved into the heap!
2731 : */
2732 : static void
2733 19830 : tuplesort_heap_insert(Tuplesortstate *state, SortTuple *tuple)
2734 : {
2735 : SortTuple *memtuples;
2736 : int j;
2737 :
2738 19830 : memtuples = state->memtuples;
2739 : Assert(state->memtupcount < state->memtupsize);
2740 :
2741 19830 : CHECK_FOR_INTERRUPTS();
2742 :
2743 : /*
2744 : * Sift-up the new entry, per Knuth 5.2.3 exercise 16. Note that Knuth is
2745 : * using 1-based array indexes, not 0-based.
2746 : */
2747 19830 : j = state->memtupcount++;
2748 57704 : while (j > 0)
2749 : {
2750 50916 : int i = (j - 1) >> 1;
2751 :
2752 50916 : if (COMPARETUP(state, tuple, &memtuples[i]) >= 0)
2753 13042 : break;
2754 37874 : memtuples[j] = memtuples[i];
2755 37874 : j = i;
2756 : }
2757 19830 : memtuples[j] = *tuple;
2758 19830 : }
2759 :
2760 : /*
2761 : * Remove the tuple at state->memtuples[0] from the heap. Decrement
2762 : * memtupcount, and sift up to maintain the heap invariant.
2763 : *
2764 : * The caller has already free'd the tuple the top node points to,
2765 : * if necessary.
2766 : */
2767 : static void
2768 19374 : tuplesort_heap_delete_top(Tuplesortstate *state)
2769 : {
2770 19374 : SortTuple *memtuples = state->memtuples;
2771 : SortTuple *tuple;
2772 :
2773 19374 : if (--state->memtupcount <= 0)
2774 284 : return;
2775 :
2776 : /*
2777 : * Remove the last tuple in the heap, and re-insert it, by replacing the
2778 : * current top node with it.
2779 : */
2780 19090 : tuple = &memtuples[state->memtupcount];
2781 19090 : tuplesort_heap_replace_top(state, tuple);
2782 : }
2783 :
2784 : /*
2785 : * Replace the tuple at state->memtuples[0] with a new tuple. Sift up to
2786 : * maintain the heap invariant.
2787 : *
2788 : * This corresponds to Knuth's "sift-up" algorithm (Algorithm 5.2.3H,
2789 : * Heapsort, steps H3-H8).
2790 : */
2791 : static void
2792 5959832 : tuplesort_heap_replace_top(Tuplesortstate *state, SortTuple *tuple)
2793 : {
2794 5959832 : SortTuple *memtuples = state->memtuples;
2795 : unsigned int i,
2796 : n;
2797 :
2798 : Assert(state->memtupcount >= 1);
2799 :
2800 5959832 : CHECK_FOR_INTERRUPTS();
2801 :
2802 : /*
2803 : * state->memtupcount is "int", but we use "unsigned int" for i, j, n.
2804 : * This prevents overflow in the "2 * i + 1" calculation, since at the top
2805 : * of the loop we must have i < n <= INT_MAX <= UINT_MAX/2.
2806 : */
2807 5959832 : n = state->memtupcount;
2808 5959832 : i = 0; /* i is where the "hole" is */
2809 : for (;;)
2810 1806798 : {
2811 7766630 : unsigned int j = 2 * i + 1;
2812 :
2813 7766630 : if (j >= n)
2814 1399062 : break;
2815 8800938 : if (j + 1 < n &&
2816 2433370 : COMPARETUP(state, &memtuples[j], &memtuples[j + 1]) > 0)
2817 960626 : j++;
2818 6367568 : if (COMPARETUP(state, tuple, &memtuples[j]) <= 0)
2819 4560770 : break;
2820 1806798 : memtuples[i] = memtuples[j];
2821 1806798 : i = j;
2822 : }
2823 5959832 : memtuples[i] = *tuple;
2824 5959832 : }
2825 :
2826 : /*
2827 : * Function to reverse the sort direction from its current state
2828 : *
2829 : * It is not safe to call this when performing hash tuplesorts
2830 : */
2831 : static void
2832 816 : reversedirection(Tuplesortstate *state)
2833 : {
2834 816 : SortSupport sortKey = state->base.sortKeys;
2835 : int nkey;
2836 :
2837 1992 : for (nkey = 0; nkey < state->base.nKeys; nkey++, sortKey++)
2838 : {
2839 1176 : sortKey->ssup_reverse = !sortKey->ssup_reverse;
2840 1176 : sortKey->ssup_nulls_first = !sortKey->ssup_nulls_first;
2841 : }
2842 816 : }
2843 :
2844 :
2845 : /*
2846 : * Tape interface routines
2847 : */
2848 :
2849 : static unsigned int
2850 5704278 : getlen(LogicalTape *tape, bool eofOK)
2851 : {
2852 : unsigned int len;
2853 :
2854 5704278 : if (LogicalTapeRead(tape,
2855 : &len, sizeof(len)) != sizeof(len))
2856 0 : elog(ERROR, "unexpected end of tape");
2857 5704278 : if (len == 0 && !eofOK)
2858 0 : elog(ERROR, "unexpected end of data");
2859 5704278 : return len;
2860 : }
2861 :
2862 : static void
2863 1642 : markrunend(LogicalTape *tape)
2864 : {
2865 1642 : unsigned int len = 0;
2866 :
2867 1642 : LogicalTapeWrite(tape, &len, sizeof(len));
2868 1642 : }
2869 :
2870 : /*
2871 : * Get memory for tuple from within READTUP() routine.
2872 : *
2873 : * We use next free slot from the slab allocator, or palloc() if the tuple
2874 : * is too large for that.
2875 : */
2876 : void *
2877 5402560 : tuplesort_readtup_alloc(Tuplesortstate *state, Size tuplen)
2878 : {
2879 : SlabSlot *buf;
2880 :
2881 : /*
2882 : * We pre-allocate enough slots in the slab arena that we should never run
2883 : * out.
2884 : */
2885 : Assert(state->slabFreeHead);
2886 :
2887 5402560 : if (tuplen > SLAB_SLOT_SIZE || !state->slabFreeHead)
2888 0 : return MemoryContextAlloc(state->base.sortcontext, tuplen);
2889 : else
2890 : {
2891 5402560 : buf = state->slabFreeHead;
2892 : /* Reuse this slot */
2893 5402560 : state->slabFreeHead = buf->nextfree;
2894 :
2895 5402560 : return buf;
2896 : }
2897 : }
2898 :
2899 :
2900 : /*
2901 : * Parallel sort routines
2902 : */
2903 :
2904 : /*
2905 : * tuplesort_estimate_shared - estimate required shared memory allocation
2906 : *
2907 : * nWorkers is an estimate of the number of workers (it's the number that
2908 : * will be requested).
2909 : */
2910 : Size
2911 158 : tuplesort_estimate_shared(int nWorkers)
2912 : {
2913 : Size tapesSize;
2914 :
2915 : Assert(nWorkers > 0);
2916 :
2917 : /* Make sure that BufFile shared state is MAXALIGN'd */
2918 158 : tapesSize = mul_size(sizeof(TapeShare), nWorkers);
2919 158 : tapesSize = MAXALIGN(add_size(tapesSize, offsetof(Sharedsort, tapes)));
2920 :
2921 158 : return tapesSize;
2922 : }
2923 :
2924 : /*
2925 : * tuplesort_initialize_shared - initialize shared tuplesort state
2926 : *
2927 : * Must be called from leader process before workers are launched, to
2928 : * establish state needed up-front for worker tuplesortstates. nWorkers
2929 : * should match the argument passed to tuplesort_estimate_shared().
2930 : */
2931 : void
2932 222 : tuplesort_initialize_shared(Sharedsort *shared, int nWorkers, dsm_segment *seg)
2933 : {
2934 : int i;
2935 :
2936 : Assert(nWorkers > 0);
2937 :
2938 222 : SpinLockInit(&shared->mutex);
2939 222 : shared->currentWorker = 0;
2940 222 : shared->workersFinished = 0;
2941 222 : SharedFileSetInit(&shared->fileset, seg);
2942 222 : shared->nTapes = nWorkers;
2943 674 : for (i = 0; i < nWorkers; i++)
2944 : {
2945 452 : shared->tapes[i].firstblocknumber = 0L;
2946 : }
2947 222 : }
2948 :
2949 : /*
2950 : * tuplesort_attach_shared - attach to shared tuplesort state
2951 : *
2952 : * Must be called by all worker processes.
2953 : */
2954 : void
2955 224 : tuplesort_attach_shared(Sharedsort *shared, dsm_segment *seg)
2956 : {
2957 : /* Attach to SharedFileSet */
2958 224 : SharedFileSetAttach(&shared->fileset, seg);
2959 224 : }
2960 :
2961 : /*
2962 : * worker_get_identifier - Assign and return ordinal identifier for worker
2963 : *
2964 : * The order in which these are assigned is not well defined, and should not
2965 : * matter; worker numbers across parallel sort participants need only be
2966 : * distinct and gapless. logtape.c requires this.
2967 : *
2968 : * Note that the identifiers assigned from here have no relation to
2969 : * ParallelWorkerNumber number, to avoid making any assumption about
2970 : * caller's requirements. However, we do follow the ParallelWorkerNumber
2971 : * convention of representing a non-worker with worker number -1. This
2972 : * includes the leader, as well as serial Tuplesort processes.
2973 : */
2974 : static int
2975 444 : worker_get_identifier(Tuplesortstate *state)
2976 : {
2977 444 : Sharedsort *shared = state->shared;
2978 : int worker;
2979 :
2980 : Assert(WORKER(state));
2981 :
2982 444 : SpinLockAcquire(&shared->mutex);
2983 444 : worker = shared->currentWorker++;
2984 444 : SpinLockRelease(&shared->mutex);
2985 :
2986 444 : return worker;
2987 : }
2988 :
2989 : /*
2990 : * worker_freeze_result_tape - freeze worker's result tape for leader
2991 : *
2992 : * This is called by workers just after the result tape has been determined,
2993 : * instead of calling LogicalTapeFreeze() directly. They do so because
2994 : * workers require a few additional steps over similar serial
2995 : * TSS_SORTEDONTAPE external sort cases, which also happen here. The extra
2996 : * steps are around freeing now unneeded resources, and representing to
2997 : * leader that worker's input run is available for its merge.
2998 : *
2999 : * There should only be one final output run for each worker, which consists
3000 : * of all tuples that were originally input into worker.
3001 : */
3002 : static void
3003 444 : worker_freeze_result_tape(Tuplesortstate *state)
3004 : {
3005 444 : Sharedsort *shared = state->shared;
3006 : TapeShare output;
3007 :
3008 : Assert(WORKER(state));
3009 : Assert(state->result_tape != NULL);
3010 : Assert(state->memtupcount == 0);
3011 :
3012 : /*
3013 : * Free most remaining memory, in case caller is sensitive to our holding
3014 : * on to it. memtuples may not be a tiny merge heap at this point.
3015 : */
3016 444 : pfree(state->memtuples);
3017 : /* Be tidy */
3018 444 : state->memtuples = NULL;
3019 444 : state->memtupsize = 0;
3020 :
3021 : /*
3022 : * Parallel worker requires result tape metadata, which is to be stored in
3023 : * shared memory for leader
3024 : */
3025 444 : LogicalTapeFreeze(state->result_tape, &output);
3026 :
3027 : /* Store properties of output tape, and update finished worker count */
3028 444 : SpinLockAcquire(&shared->mutex);
3029 444 : shared->tapes[state->worker] = output;
3030 444 : shared->workersFinished++;
3031 444 : SpinLockRelease(&shared->mutex);
3032 444 : }
3033 :
3034 : /*
3035 : * worker_nomergeruns - dump memtuples in worker, without merging
3036 : *
3037 : * This called as an alternative to mergeruns() with a worker when no
3038 : * merging is required.
3039 : */
3040 : static void
3041 444 : worker_nomergeruns(Tuplesortstate *state)
3042 : {
3043 : Assert(WORKER(state));
3044 : Assert(state->result_tape == NULL);
3045 : Assert(state->nOutputRuns == 1);
3046 :
3047 444 : state->result_tape = state->destTape;
3048 444 : worker_freeze_result_tape(state);
3049 444 : }
3050 :
3051 : /*
3052 : * leader_takeover_tapes - create tapeset for leader from worker tapes
3053 : *
3054 : * So far, leader Tuplesortstate has performed no actual sorting. By now, all
3055 : * sorting has occurred in workers, all of which must have already returned
3056 : * from tuplesort_performsort().
3057 : *
3058 : * When this returns, leader process is left in a state that is virtually
3059 : * indistinguishable from it having generated runs as a serial external sort
3060 : * might have.
3061 : */
3062 : static void
3063 156 : leader_takeover_tapes(Tuplesortstate *state)
3064 : {
3065 156 : Sharedsort *shared = state->shared;
3066 156 : int nParticipants = state->nParticipants;
3067 : int workersFinished;
3068 : int j;
3069 :
3070 : Assert(LEADER(state));
3071 : Assert(nParticipants >= 1);
3072 :
3073 156 : SpinLockAcquire(&shared->mutex);
3074 156 : workersFinished = shared->workersFinished;
3075 156 : SpinLockRelease(&shared->mutex);
3076 :
3077 156 : if (nParticipants != workersFinished)
3078 0 : elog(ERROR, "cannot take over tapes before all workers finish");
3079 :
3080 : /*
3081 : * Create the tapeset from worker tapes, including a leader-owned tape at
3082 : * the end. Parallel workers are far more expensive than logical tapes,
3083 : * so the number of tapes allocated here should never be excessive.
3084 : */
3085 156 : inittapestate(state, nParticipants);
3086 156 : state->tapeset = LogicalTapeSetCreate(false, &shared->fileset, -1);
3087 :
3088 : /*
3089 : * Set currentRun to reflect the number of runs we will merge (it's not
3090 : * used for anything, this is just pro forma)
3091 : */
3092 156 : state->currentRun = nParticipants;
3093 :
3094 : /*
3095 : * Initialize the state to look the same as after building the initial
3096 : * runs.
3097 : *
3098 : * There will always be exactly 1 run per worker, and exactly one input
3099 : * tape per run, because workers always output exactly 1 run, even when
3100 : * there were no input tuples for workers to sort.
3101 : */
3102 156 : state->inputTapes = NULL;
3103 156 : state->nInputTapes = 0;
3104 156 : state->nInputRuns = 0;
3105 :
3106 156 : state->outputTapes = palloc0(nParticipants * sizeof(LogicalTape *));
3107 156 : state->nOutputTapes = nParticipants;
3108 156 : state->nOutputRuns = nParticipants;
3109 :
3110 472 : for (j = 0; j < nParticipants; j++)
3111 : {
3112 316 : state->outputTapes[j] = LogicalTapeImport(state->tapeset, j, &shared->tapes[j]);
3113 : }
3114 :
3115 156 : state->status = TSS_BUILDRUNS;
3116 156 : }
3117 :
3118 : /*
3119 : * Convenience routine to free a tuple previously loaded into sort memory
3120 : */
3121 : static void
3122 3759982 : free_sort_tuple(Tuplesortstate *state, SortTuple *stup)
3123 : {
3124 3759982 : if (stup->tuple)
3125 : {
3126 3599704 : FREEMEM(state, GetMemoryChunkSpace(stup->tuple));
3127 3599704 : pfree(stup->tuple);
3128 3599704 : stup->tuple = NULL;
3129 : }
3130 3759982 : }
3131 :
3132 : int
3133 0 : ssup_datum_unsigned_cmp(Datum x, Datum y, SortSupport ssup)
3134 : {
3135 0 : if (x < y)
3136 0 : return -1;
3137 0 : else if (x > y)
3138 0 : return 1;
3139 : else
3140 0 : return 0;
3141 : }
3142 :
3143 : int
3144 1139832 : ssup_datum_signed_cmp(Datum x, Datum y, SortSupport ssup)
3145 : {
3146 1139832 : int64 xx = DatumGetInt64(x);
3147 1139832 : int64 yy = DatumGetInt64(y);
3148 :
3149 1139832 : if (xx < yy)
3150 426530 : return -1;
3151 713302 : else if (xx > yy)
3152 358468 : return 1;
3153 : else
3154 354834 : return 0;
3155 : }
3156 :
3157 : int
3158 180533632 : ssup_datum_int32_cmp(Datum x, Datum y, SortSupport ssup)
3159 : {
3160 180533632 : int32 xx = DatumGetInt32(x);
3161 180533632 : int32 yy = DatumGetInt32(y);
3162 :
3163 180533632 : if (xx < yy)
3164 44679588 : return -1;
3165 135854044 : else if (xx > yy)
3166 41681734 : return 1;
3167 : else
3168 94172310 : return 0;
3169 : }
|
