Branch data Line data Source code
1 : : /*----------------------------------------------------------------------
2 : : *
3 : : * tableam.c
4 : : * Table access method routines too big to be inline functions.
5 : : *
6 : : * Portions Copyright (c) 1996-2026, PostgreSQL Global Development Group
7 : : * Portions Copyright (c) 1994, Regents of the University of California
8 : : *
9 : : *
10 : : * IDENTIFICATION
11 : : * src/backend/access/table/tableam.c
12 : : *
13 : : * NOTES
14 : : * Note that most functions in here are documented in tableam.h, rather than
15 : : * here. That's because there's a lot of inline functions in tableam.h and
16 : : * it'd be harder to understand if one constantly had to switch between files.
17 : : *
18 : : *----------------------------------------------------------------------
19 : : */
20 : : #include "postgres.h"
21 : :
22 : : #include <math.h>
23 : :
24 : : #include "access/syncscan.h"
25 : : #include "access/tableam.h"
26 : : #include "access/xact.h"
27 : : #include "optimizer/optimizer.h"
28 : : #include "optimizer/plancat.h"
29 : : #include "port/pg_bitutils.h"
30 : : #include "storage/bufmgr.h"
31 : : #include "storage/shmem.h"
32 : : #include "storage/smgr.h"
33 : :
34 : : /*
35 : : * Constants to control the behavior of block allocation to parallel workers
36 : : * during a parallel seqscan. Technically these values do not need to be
37 : : * powers of 2, but having them as powers of 2 makes the math more optimal
38 : : * and makes the ramp-down stepping more even.
39 : : */
40 : :
41 : : /* The number of I/O chunks we try to break a parallel seqscan down into */
42 : : #define PARALLEL_SEQSCAN_NCHUNKS 2048
43 : : /* Ramp down size of allocations when we've only this number of chunks left */
44 : : #define PARALLEL_SEQSCAN_RAMPDOWN_CHUNKS 64
45 : : /* Cap the size of parallel I/O chunks to this number of blocks */
46 : : #define PARALLEL_SEQSCAN_MAX_CHUNK_SIZE 8192
47 : :
48 : : /* GUC variables */
49 : : char *default_table_access_method = DEFAULT_TABLE_ACCESS_METHOD;
50 : : bool synchronize_seqscans = true;
51 : :
52 : :
53 : : /* ----------------------------------------------------------------------------
54 : : * Slot functions.
55 : : * ----------------------------------------------------------------------------
56 : : */
57 : :
58 : : const TupleTableSlotOps *
59 : 18116891 : table_slot_callbacks(Relation relation)
60 : : {
61 : : const TupleTableSlotOps *tts_cb;
62 : :
63 [ + + ]: 18116891 : if (relation->rd_tableam)
64 : 18111083 : tts_cb = relation->rd_tableam->slot_callbacks(relation);
65 [ + + ]: 5808 : else if (relation->rd_rel->relkind == RELKIND_FOREIGN_TABLE)
66 : : {
67 : : /*
68 : : * Historically FDWs expect to store heap tuples in slots. Continue
69 : : * handing them one, to make it less painful to adapt FDWs to new
70 : : * versions. The cost of a heap slot over a virtual slot is pretty
71 : : * small.
72 : : */
73 : 227 : tts_cb = &TTSOpsHeapTuple;
74 : : }
75 : : else
76 : : {
77 : : /*
78 : : * These need to be supported, as some parts of the code (like COPY)
79 : : * need to create slots for such relations too. It seems better to
80 : : * centralize the knowledge that a heap slot is the right thing in
81 : : * that case here.
82 : : */
83 : : Assert(relation->rd_rel->relkind == RELKIND_VIEW ||
84 : : relation->rd_rel->relkind == RELKIND_PARTITIONED_TABLE);
85 : 5581 : tts_cb = &TTSOpsVirtual;
86 : : }
87 : :
88 : 18116891 : return tts_cb;
89 : : }
90 : :
91 : : TupleTableSlot *
92 : 17805876 : table_slot_create(Relation relation, List **reglist)
93 : : {
94 : : const TupleTableSlotOps *tts_cb;
95 : : TupleTableSlot *slot;
96 : :
97 : 17805876 : tts_cb = table_slot_callbacks(relation);
98 : 17805876 : slot = MakeSingleTupleTableSlot(RelationGetDescr(relation), tts_cb);
99 : :
100 [ + + ]: 17805876 : if (reglist)
101 : 169411 : *reglist = lappend(*reglist, slot);
102 : :
103 : 17805876 : return slot;
104 : : }
105 : :
106 : :
107 : : /* ----------------------------------------------------------------------------
108 : : * Table scan functions.
109 : : * ----------------------------------------------------------------------------
110 : : */
111 : :
112 : : TableScanDesc
113 : 46751 : table_beginscan_catalog(Relation relation, int nkeys, ScanKeyData *key)
114 : : {
115 : 46751 : uint32 flags = SO_TYPE_SEQSCAN |
116 : : SO_ALLOW_STRAT | SO_ALLOW_SYNC | SO_ALLOW_PAGEMODE | SO_TEMP_SNAPSHOT;
117 : 46751 : Oid relid = RelationGetRelid(relation);
118 : 46751 : Snapshot snapshot = RegisterSnapshot(GetCatalogSnapshot(relid));
119 : :
120 : 46751 : return table_beginscan_common(relation, snapshot, nkeys, key,
121 : : NULL, flags, SO_NONE);
122 : : }
123 : :
124 : :
125 : : /* ----------------------------------------------------------------------------
126 : : * Parallel table scan related functions.
127 : : * ----------------------------------------------------------------------------
128 : : */
129 : :
130 : : Size
131 : 1337 : table_parallelscan_estimate(Relation rel, Snapshot snapshot)
132 : : {
133 : 1337 : Size sz = 0;
134 : :
135 [ + + ]: 1337 : if (IsMVCCSnapshot(snapshot))
136 : 1210 : sz = add_size(sz, EstimateSnapshotSpace(snapshot));
137 : : else
138 : : Assert(snapshot == SnapshotAny);
139 : :
140 : 1337 : sz = add_size(sz, rel->rd_tableam->parallelscan_estimate(rel));
141 : :
142 : 1337 : return sz;
143 : : }
144 : :
145 : : void
146 : 1337 : table_parallelscan_initialize(Relation rel, ParallelTableScanDesc pscan,
147 : : Snapshot snapshot)
148 : : {
149 : 1337 : Size snapshot_off = rel->rd_tableam->parallelscan_initialize(rel, pscan);
150 : :
151 : 1337 : pscan->phs_snapshot_off = snapshot_off;
152 : :
153 [ + + ]: 1337 : if (IsMVCCSnapshot(snapshot))
154 : : {
155 : 1210 : SerializeSnapshot(snapshot, (char *) pscan + pscan->phs_snapshot_off);
156 : 1210 : pscan->phs_snapshot_any = false;
157 : : }
158 : : else
159 : : {
160 : : Assert(snapshot == SnapshotAny);
161 : 127 : pscan->phs_snapshot_any = true;
162 : : }
163 : 1337 : }
164 : :
165 : : TableScanDesc
166 : 4441 : table_beginscan_parallel(Relation relation, ParallelTableScanDesc pscan,
167 : : uint32 flags)
168 : : {
169 : : Snapshot snapshot;
170 : 4441 : uint32 internal_flags = SO_TYPE_SEQSCAN |
171 : : SO_ALLOW_STRAT | SO_ALLOW_SYNC | SO_ALLOW_PAGEMODE;
172 : :
173 : : Assert(RelFileLocatorEquals(relation->rd_locator, pscan->phs_locator));
174 : :
175 [ + + ]: 4441 : if (!pscan->phs_snapshot_any)
176 : : {
177 : : /* Snapshot was serialized -- restore it */
178 : 4170 : snapshot = RestoreSnapshot((char *) pscan + pscan->phs_snapshot_off);
179 : 4170 : RegisterSnapshot(snapshot);
180 : 4170 : internal_flags |= SO_TEMP_SNAPSHOT;
181 : : }
182 : : else
183 : : {
184 : : /* SnapshotAny passed by caller (not serialized) */
185 : 271 : snapshot = SnapshotAny;
186 : : }
187 : :
188 : 4441 : return table_beginscan_common(relation, snapshot, 0, NULL,
189 : : pscan, internal_flags, flags);
190 : : }
191 : :
192 : : TableScanDesc
193 : 80 : table_beginscan_parallel_tidrange(Relation relation,
194 : : ParallelTableScanDesc pscan,
195 : : uint32 flags)
196 : : {
197 : : Snapshot snapshot;
198 : : TableScanDesc sscan;
199 : 80 : uint32 internal_flags = SO_TYPE_TIDRANGESCAN | SO_ALLOW_PAGEMODE;
200 : :
201 : : Assert(RelFileLocatorEquals(relation->rd_locator, pscan->phs_locator));
202 : :
203 : : /* disable syncscan in parallel tid range scan. */
204 : 80 : pscan->phs_syncscan = false;
205 : :
206 [ + - ]: 80 : if (!pscan->phs_snapshot_any)
207 : : {
208 : : /* Snapshot was serialized -- restore it */
209 : 80 : snapshot = RestoreSnapshot((char *) pscan + pscan->phs_snapshot_off);
210 : 80 : RegisterSnapshot(snapshot);
211 : 80 : internal_flags |= SO_TEMP_SNAPSHOT;
212 : : }
213 : : else
214 : : {
215 : : /* SnapshotAny passed by caller (not serialized) */
216 : 0 : snapshot = SnapshotAny;
217 : : }
218 : :
219 : 80 : sscan = table_beginscan_common(relation, snapshot, 0, NULL,
220 : : pscan, internal_flags, flags);
221 : 80 : return sscan;
222 : : }
223 : :
224 : :
225 : : /* ----------------------------------------------------------------------------
226 : : * Index scan related functions.
227 : : * ----------------------------------------------------------------------------
228 : : */
229 : :
230 : : /*
231 : : * To perform that check simply start an index scan, create the necessary
232 : : * slot, do the heap lookup, and shut everything down again. This could be
233 : : * optimized, but is unlikely to matter from a performance POV. If there
234 : : * frequently are live index pointers also matching a unique index key, the
235 : : * CPU overhead of this routine is unlikely to matter.
236 : : *
237 : : * Note that *tid may be modified when we return true if the AM supports
238 : : * storing multiple row versions reachable via a single index entry (like
239 : : * heap's HOT).
240 : : */
241 : : bool
242 : 7589190 : table_index_fetch_tuple_check(Relation rel,
243 : : ItemPointer tid,
244 : : Snapshot snapshot,
245 : : bool *all_dead)
246 : : {
247 : : IndexFetchTableData *scan;
248 : : TupleTableSlot *slot;
249 : 7589190 : bool call_again = false;
250 : : bool found;
251 : :
252 : 7589190 : slot = table_slot_create(rel, NULL);
253 : 7589190 : scan = table_index_fetch_begin(rel, SO_NONE);
254 : 7589190 : found = table_index_fetch_tuple(scan, tid, snapshot, slot, &call_again,
255 : : all_dead);
256 : 7589190 : table_index_fetch_end(scan);
257 : 7589190 : ExecDropSingleTupleTableSlot(slot);
258 : :
259 : 7589190 : return found;
260 : : }
261 : :
262 : :
263 : : /* ------------------------------------------------------------------------
264 : : * Functions for non-modifying operations on individual tuples
265 : : * ------------------------------------------------------------------------
266 : : */
267 : :
268 : : void
269 : 215 : table_tuple_get_latest_tid(TableScanDesc scan, ItemPointer tid)
270 : : {
271 : 215 : Relation rel = scan->rs_rd;
272 : 215 : const TableAmRoutine *tableam = rel->rd_tableam;
273 : :
274 : : /*
275 : : * Since this can be called with user-supplied TID, don't trust the input
276 : : * too much.
277 : : */
278 [ + + ]: 215 : if (!tableam->tuple_tid_valid(scan, tid))
279 [ + - ]: 8 : ereport(ERROR,
280 : : (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
281 : : errmsg("tid (%u, %u) is not valid for relation \"%s\"",
282 : : ItemPointerGetBlockNumberNoCheck(tid),
283 : : ItemPointerGetOffsetNumberNoCheck(tid),
284 : : RelationGetRelationName(rel))));
285 : :
286 : 207 : tableam->tuple_get_latest_tid(scan, tid);
287 : 207 : }
288 : :
289 : :
290 : : /* ----------------------------------------------------------------------------
291 : : * Functions to make modifications a bit simpler.
292 : : * ----------------------------------------------------------------------------
293 : : */
294 : :
295 : : /*
296 : : * simple_table_tuple_insert - insert a tuple
297 : : *
298 : : * Currently, this routine differs from table_tuple_insert only in supplying a
299 : : * default command ID and not allowing access to the speedup options.
300 : : */
301 : : void
302 : 96661 : simple_table_tuple_insert(Relation rel, TupleTableSlot *slot)
303 : : {
304 : 96661 : table_tuple_insert(rel, slot, GetCurrentCommandId(true), 0, NULL);
305 : 96661 : }
306 : :
307 : : /*
308 : : * simple_table_tuple_delete - delete a tuple
309 : : *
310 : : * This routine may be used to delete a tuple when concurrent updates of
311 : : * the target tuple are not expected (for example, because we have a lock
312 : : * on the relation associated with the tuple). Any failure is reported
313 : : * via ereport().
314 : : */
315 : : void
316 : 40318 : simple_table_tuple_delete(Relation rel, ItemPointer tid, Snapshot snapshot)
317 : : {
318 : : TM_Result result;
319 : : TM_FailureData tmfd;
320 : :
321 : 40318 : result = table_tuple_delete(rel, tid,
322 : : GetCurrentCommandId(true),
323 : : 0, snapshot, InvalidSnapshot,
324 : : true /* wait for commit */ ,
325 : : &tmfd);
326 : :
327 [ - + - - : 40318 : switch (result)
- ]
328 : : {
329 : 0 : case TM_SelfModified:
330 : : /* Tuple was already updated in current command? */
331 [ # # ]: 0 : elog(ERROR, "tuple already updated by self");
332 : : break;
333 : :
334 : 40318 : case TM_Ok:
335 : : /* done successfully */
336 : 40318 : break;
337 : :
338 : 0 : case TM_Updated:
339 [ # # ]: 0 : elog(ERROR, "tuple concurrently updated");
340 : : break;
341 : :
342 : 0 : case TM_Deleted:
343 [ # # ]: 0 : elog(ERROR, "tuple concurrently deleted");
344 : : break;
345 : :
346 : 0 : default:
347 [ # # ]: 0 : elog(ERROR, "unrecognized table_tuple_delete status: %u", result);
348 : : break;
349 : : }
350 : 40318 : }
351 : :
352 : : /*
353 : : * simple_table_tuple_update - replace a tuple
354 : : *
355 : : * This routine may be used to update a tuple when concurrent updates of
356 : : * the target tuple are not expected (for example, because we have a lock
357 : : * on the relation associated with the tuple). Any failure is reported
358 : : * via ereport().
359 : : */
360 : : void
361 : 31927 : simple_table_tuple_update(Relation rel, ItemPointer otid,
362 : : TupleTableSlot *slot,
363 : : Snapshot snapshot,
364 : : TU_UpdateIndexes *update_indexes)
365 : : {
366 : : TM_Result result;
367 : : TM_FailureData tmfd;
368 : : LockTupleMode lockmode;
369 : :
370 : 31927 : result = table_tuple_update(rel, otid, slot,
371 : : GetCurrentCommandId(true),
372 : : 0, snapshot, InvalidSnapshot,
373 : : true /* wait for commit */ ,
374 : : &tmfd, &lockmode, update_indexes);
375 : :
376 [ - + - - : 31927 : switch (result)
- ]
377 : : {
378 : 0 : case TM_SelfModified:
379 : : /* Tuple was already updated in current command? */
380 [ # # ]: 0 : elog(ERROR, "tuple already updated by self");
381 : : break;
382 : :
383 : 31927 : case TM_Ok:
384 : : /* done successfully */
385 : 31927 : break;
386 : :
387 : 0 : case TM_Updated:
388 [ # # ]: 0 : elog(ERROR, "tuple concurrently updated");
389 : : break;
390 : :
391 : 0 : case TM_Deleted:
392 [ # # ]: 0 : elog(ERROR, "tuple concurrently deleted");
393 : : break;
394 : :
395 : 0 : default:
396 [ # # ]: 0 : elog(ERROR, "unrecognized table_tuple_update status: %u", result);
397 : : break;
398 : : }
399 : 31927 : }
400 : :
401 : :
402 : : /* ----------------------------------------------------------------------------
403 : : * Helper functions to implement parallel scans for block oriented AMs.
404 : : * ----------------------------------------------------------------------------
405 : : */
406 : :
407 : : Size
408 : 1337 : table_block_parallelscan_estimate(Relation rel)
409 : : {
410 : 1337 : return sizeof(ParallelBlockTableScanDescData);
411 : : }
412 : :
413 : : Size
414 : 1337 : table_block_parallelscan_initialize(Relation rel, ParallelTableScanDesc pscan)
415 : : {
416 : 1337 : ParallelBlockTableScanDesc bpscan = (ParallelBlockTableScanDesc) pscan;
417 : :
418 : 1337 : bpscan->base.phs_locator = rel->rd_locator;
419 : 1337 : bpscan->phs_nblocks = RelationGetNumberOfBlocks(rel);
420 : : /* compare phs_syncscan initialization to similar logic in initscan */
421 : 3681 : bpscan->base.phs_syncscan = synchronize_seqscans &&
422 [ + + + - ]: 2344 : !RelationUsesLocalBuffers(rel) &&
423 [ + + ]: 1007 : bpscan->phs_nblocks > NBuffers / 4;
424 : 1337 : pg_atomic_init_u32(&bpscan->phs_startblock, InvalidBlockNumber);
425 : 1337 : pg_atomic_init_u32(&bpscan->phs_numblock, InvalidBlockNumber);
426 : 1337 : pg_atomic_init_u64(&bpscan->phs_nallocated, 0);
427 : :
428 : 1337 : return sizeof(ParallelBlockTableScanDescData);
429 : : }
430 : :
431 : : void
432 : 152 : table_block_parallelscan_reinitialize(Relation rel, ParallelTableScanDesc pscan)
433 : : {
434 : 152 : ParallelBlockTableScanDesc bpscan = (ParallelBlockTableScanDesc) pscan;
435 : :
436 : 152 : pg_atomic_write_u64(&bpscan->phs_nallocated, 0);
437 : 152 : }
438 : :
439 : : /*
440 : : * find and set the scan's startblock
441 : : *
442 : : * Determine where the parallel seq scan should start. This function may be
443 : : * called many times, once by each parallel worker. We must be careful only
444 : : * to set the phs_startblock and phs_numblock fields once.
445 : : *
446 : : * Callers may optionally specify a non-InvalidBlockNumber value for
447 : : * 'startblock' to force the scan to start at the given page. Likewise,
448 : : * 'numblocks' can be specified as a non-InvalidBlockNumber to limit the
449 : : * number of blocks to scan to that many blocks.
450 : : */
451 : : void
452 : 2793 : table_block_parallelscan_startblock_init(Relation rel,
453 : : ParallelBlockTableScanWorker pbscanwork,
454 : : ParallelBlockTableScanDesc pbscan,
455 : : BlockNumber startblock,
456 : : BlockNumber numblocks)
457 : : {
458 : : StaticAssertDecl(MaxBlockNumber <= 0xFFFFFFFE,
459 : : "pg_nextpower2_32 may be too small for non-standard BlockNumber width");
460 : :
461 : : BlockNumber scan_nblocks;
462 : :
463 : : /* Reset the state we use for controlling allocation size. */
464 : 2793 : memset(pbscanwork, 0, sizeof(*pbscanwork));
465 : :
466 : : /*
467 : : * When the caller specified a limit on the number of blocks to scan, set
468 : : * that in the ParallelBlockTableScanDesc, if it's not been done by
469 : : * another worker already.
470 : : */
471 [ + + ]: 2793 : if (numblocks != InvalidBlockNumber)
472 : : {
473 : 80 : uint32 expected = InvalidBlockNumber;
474 : :
475 : 80 : pg_atomic_compare_exchange_u32(&pbscan->phs_numblock, &expected,
476 : : numblocks);
477 : : }
478 : :
479 : : /*
480 : : * If the scan's phs_startblock has not yet been initialized, we must do
481 : : * so now. If a startblock was specified, start there, otherwise if this
482 : : * is not a synchronized scan, we just start at block 0, but if it is a
483 : : * synchronized scan, we must get the starting position from the
484 : : * synchronized scan machinery.
485 : : *
486 : : * If another worker initializes phs_startblock concurrently, just use
487 : : * their value.
488 : : */
489 [ + + ]: 2793 : if (pg_atomic_read_u32(&pbscan->phs_startblock) == InvalidBlockNumber)
490 : : {
491 : : BlockNumber newstartblock;
492 : 1326 : uint32 expected = InvalidBlockNumber;
493 : :
494 [ + + ]: 1326 : if (startblock != InvalidBlockNumber)
495 : 16 : newstartblock = startblock;
496 [ + + ]: 1310 : else if (!pbscan->base.phs_syncscan)
497 : 1308 : newstartblock = 0;
498 : : else
499 : 2 : newstartblock = ss_get_location(rel, pbscan->phs_nblocks);
500 : :
501 : 1326 : pg_atomic_compare_exchange_u32(&pbscan->phs_startblock, &expected,
502 : : newstartblock);
503 : : }
504 : :
505 : : /*
506 : : * Figure out how many blocks we're going to scan; either all of them, or
507 : : * just phs_numblock's worth, if a limit has been imposed.
508 : : */
509 [ + + ]: 2793 : if (pg_atomic_read_u32(&pbscan->phs_numblock) == InvalidBlockNumber)
510 : 2713 : scan_nblocks = pbscan->phs_nblocks;
511 : : else
512 : 80 : scan_nblocks = pg_atomic_read_u32(&pbscan->phs_numblock);
513 : :
514 : : /*
515 : : * We determine the chunk size based on scan_nblocks. First we split
516 : : * scan_nblocks into PARALLEL_SEQSCAN_NCHUNKS chunks then we calculate the
517 : : * next highest power of 2 number of the result. This means we split the
518 : : * blocks we're scanning into somewhere between PARALLEL_SEQSCAN_NCHUNKS
519 : : * and PARALLEL_SEQSCAN_NCHUNKS / 2 chunks.
520 : : */
521 [ + + ]: 2793 : pbscanwork->phsw_chunk_size = pg_nextpower2_32(Max(scan_nblocks /
522 : : PARALLEL_SEQSCAN_NCHUNKS, 1));
523 : :
524 : : /*
525 : : * Ensure we don't go over the maximum chunk size with larger tables. This
526 : : * means we may get much more than PARALLEL_SEQSCAN_NCHUNKS for larger
527 : : * tables. Too large a chunk size has been shown to be detrimental to
528 : : * sequential scan performance.
529 : : */
530 : 2793 : pbscanwork->phsw_chunk_size = Min(pbscanwork->phsw_chunk_size,
531 : : PARALLEL_SEQSCAN_MAX_CHUNK_SIZE);
532 : 2793 : }
533 : :
534 : : /*
535 : : * get the next page to scan
536 : : *
537 : : * Get the next page to scan. Even if there are no pages left to scan,
538 : : * another backend could have grabbed a page to scan and not yet finished
539 : : * looking at it, so it doesn't follow that the scan is done when the first
540 : : * backend gets an InvalidBlockNumber return.
541 : : */
542 : : BlockNumber
543 : 146270 : table_block_parallelscan_nextpage(Relation rel,
544 : : ParallelBlockTableScanWorker pbscanwork,
545 : : ParallelBlockTableScanDesc pbscan)
546 : : {
547 : : BlockNumber scan_nblocks;
548 : : BlockNumber page;
549 : : uint64 nallocated;
550 : :
551 : : /*
552 : : * The logic below allocates block numbers out to parallel workers in a
553 : : * way that each worker will receive a set of consecutive block numbers to
554 : : * scan. Earlier versions of this would allocate the next highest block
555 : : * number to the next worker to call this function. This would generally
556 : : * result in workers never receiving consecutive block numbers. Some
557 : : * operating systems would not detect the sequential I/O pattern due to
558 : : * each backend being a different process which could result in poor
559 : : * performance due to inefficient or no readahead. To work around this
560 : : * issue, we now allocate a range of block numbers for each worker and
561 : : * when they come back for another block, we give them the next one in
562 : : * that range until the range is complete. When the worker completes the
563 : : * range of blocks we then allocate another range for it and return the
564 : : * first block number from that range.
565 : : *
566 : : * Here we name these ranges of blocks "chunks". The initial size of
567 : : * these chunks is determined in table_block_parallelscan_startblock_init
568 : : * based on the number of blocks to scan. Towards the end of the scan, we
569 : : * start making reductions in the size of the chunks in order to attempt
570 : : * to divide the remaining work over all the workers as evenly as
571 : : * possible.
572 : : *
573 : : * Here pbscanwork is local worker memory. phsw_chunk_remaining tracks
574 : : * the number of blocks remaining in the chunk. When that reaches 0 then
575 : : * we must allocate a new chunk for the worker.
576 : : *
577 : : * phs_nallocated tracks how many blocks have been allocated to workers
578 : : * already. When phs_nallocated >= rs_nblocks, all blocks have been
579 : : * allocated.
580 : : *
581 : : * Because we use an atomic fetch-and-add to fetch the current value, the
582 : : * phs_nallocated counter will exceed rs_nblocks, because workers will
583 : : * still increment the value, when they try to allocate the next block but
584 : : * all blocks have been allocated already. The counter must be 64 bits
585 : : * wide because of that, to avoid wrapping around when scan_nblocks is
586 : : * close to 2^32.
587 : : *
588 : : * The actual block to return is calculated by adding the counter to the
589 : : * starting block number, modulo phs_nblocks.
590 : : */
591 : :
592 : : /* First, figure out how many blocks we're planning on scanning */
593 [ + + ]: 146270 : if (pg_atomic_read_u32(&pbscan->phs_numblock) == InvalidBlockNumber)
594 : 145858 : scan_nblocks = pbscan->phs_nblocks;
595 : : else
596 : 412 : scan_nblocks = pg_atomic_read_u32(&pbscan->phs_numblock);
597 : :
598 : : /*
599 : : * Now check if we have any remaining blocks in a previous chunk for this
600 : : * worker. We must consume all of the blocks from that before we allocate
601 : : * a new chunk to the worker.
602 : : */
603 [ + + ]: 146270 : if (pbscanwork->phsw_chunk_remaining > 0)
604 : : {
605 : : /*
606 : : * Give them the next block in the range and update the remaining
607 : : * number of blocks.
608 : : */
609 : 17946 : nallocated = ++pbscanwork->phsw_nallocated;
610 : 17946 : pbscanwork->phsw_chunk_remaining--;
611 : : }
612 : : else
613 : : {
614 : : /*
615 : : * When we've only got PARALLEL_SEQSCAN_RAMPDOWN_CHUNKS chunks
616 : : * remaining in the scan, we half the chunk size. Since we reduce the
617 : : * chunk size here, we'll hit this again after doing
618 : : * PARALLEL_SEQSCAN_RAMPDOWN_CHUNKS at the new size. After a few
619 : : * iterations of this, we'll end up doing the last few blocks with the
620 : : * chunk size set to 1.
621 : : */
622 [ + + ]: 128324 : if (pbscanwork->phsw_chunk_size > 1 &&
623 : 3939 : pbscanwork->phsw_nallocated > scan_nblocks -
624 [ + + ]: 3939 : (pbscanwork->phsw_chunk_size * PARALLEL_SEQSCAN_RAMPDOWN_CHUNKS))
625 : 13 : pbscanwork->phsw_chunk_size >>= 1;
626 : :
627 : 128324 : nallocated = pbscanwork->phsw_nallocated =
628 : 128324 : pg_atomic_fetch_add_u64(&pbscan->phs_nallocated,
629 : 128324 : pbscanwork->phsw_chunk_size);
630 : :
631 : : /*
632 : : * Set the remaining number of blocks in this chunk so that subsequent
633 : : * calls from this worker continue on with this chunk until it's done.
634 : : */
635 : 128324 : pbscanwork->phsw_chunk_remaining = pbscanwork->phsw_chunk_size - 1;
636 : : }
637 : :
638 : : /* Check if we've run out of blocks to scan */
639 [ + + ]: 146270 : if (nallocated >= scan_nblocks)
640 : 2793 : page = InvalidBlockNumber; /* all blocks have been allocated */
641 : : else
642 : 143477 : page = (nallocated + pg_atomic_read_u32(&pbscan->phs_startblock)) % pbscan->phs_nblocks;
643 : :
644 : : /*
645 : : * Report scan location. Normally, we report the current page number.
646 : : * When we reach the end of the scan, though, we report the starting page,
647 : : * not the ending page, just so the starting positions for later scans
648 : : * doesn't slew backwards. We only report the position at the end of the
649 : : * scan once, though: subsequent callers will report nothing.
650 : : */
651 [ + + ]: 146270 : if (pbscan->base.phs_syncscan)
652 : : {
653 [ + + ]: 22130 : if (page != InvalidBlockNumber)
654 : 22125 : ss_report_location(rel, page);
655 [ + + ]: 5 : else if (nallocated == pbscan->phs_nblocks)
656 : 2 : ss_report_location(rel, pg_atomic_read_u32(&pbscan->phs_startblock));
657 : : }
658 : :
659 : 146270 : return page;
660 : : }
661 : :
662 : : /* ----------------------------------------------------------------------------
663 : : * Helper functions to implement relation sizing for block oriented AMs.
664 : : * ----------------------------------------------------------------------------
665 : : */
666 : :
667 : : /*
668 : : * table_block_relation_size
669 : : *
670 : : * If a table AM uses the various relation forks as the sole place where data
671 : : * is stored, and if it uses them in the expected manner (e.g. the actual data
672 : : * is in the main fork rather than some other), it can use this implementation
673 : : * of the relation_size callback rather than implementing its own.
674 : : */
675 : : uint64
676 : 1730193 : table_block_relation_size(Relation rel, ForkNumber forkNumber)
677 : : {
678 : 1730193 : uint64 nblocks = 0;
679 : :
680 : : /* InvalidForkNumber indicates returning the size for all forks */
681 [ - + ]: 1730193 : if (forkNumber == InvalidForkNumber)
682 : : {
683 [ # # ]: 0 : for (int i = 0; i < MAX_FORKNUM; i++)
684 : 0 : nblocks += smgrnblocks(RelationGetSmgr(rel), i);
685 : : }
686 : : else
687 : 1730193 : nblocks = smgrnblocks(RelationGetSmgr(rel), forkNumber);
688 : :
689 : 1730174 : return nblocks * BLCKSZ;
690 : : }
691 : :
692 : : /*
693 : : * table_block_relation_estimate_size
694 : : *
695 : : * This function can't be directly used as the implementation of the
696 : : * relation_estimate_size callback, because it has a few additional parameters.
697 : : * Instead, it is intended to be used as a helper function; the caller can
698 : : * pass through the arguments to its relation_estimate_size function plus the
699 : : * additional values required here.
700 : : *
701 : : * overhead_bytes_per_tuple should contain the approximate number of bytes
702 : : * of storage required to store a tuple above and beyond what is required for
703 : : * the tuple data proper. Typically, this would include things like the
704 : : * size of the tuple header and item pointer. This is only used for query
705 : : * planning, so a table AM where the value is not constant could choose to
706 : : * pass a "best guess".
707 : : *
708 : : * usable_bytes_per_page should contain the approximate number of bytes per
709 : : * page usable for tuple data, excluding the page header and any anticipated
710 : : * special space.
711 : : */
712 : : void
713 : 361499 : table_block_relation_estimate_size(Relation rel, int32 *attr_widths,
714 : : BlockNumber *pages, double *tuples,
715 : : double *allvisfrac,
716 : : Size overhead_bytes_per_tuple,
717 : : Size usable_bytes_per_page)
718 : : {
719 : : BlockNumber curpages;
720 : : BlockNumber relpages;
721 : : double reltuples;
722 : : BlockNumber relallvisible;
723 : : double density;
724 : :
725 : : /* it should have storage, so we can call the smgr */
726 : 361499 : curpages = RelationGetNumberOfBlocks(rel);
727 : :
728 : : /* coerce values in pg_class to more desirable types */
729 : 361499 : relpages = (BlockNumber) rel->rd_rel->relpages;
730 : 361499 : reltuples = (double) rel->rd_rel->reltuples;
731 : 361499 : relallvisible = (BlockNumber) rel->rd_rel->relallvisible;
732 : :
733 : : /*
734 : : * HACK: if the relation has never yet been vacuumed, use a minimum size
735 : : * estimate of 10 pages. The idea here is to avoid assuming a
736 : : * newly-created table is really small, even if it currently is, because
737 : : * that may not be true once some data gets loaded into it. Once a vacuum
738 : : * or analyze cycle has been done on it, it's more reasonable to believe
739 : : * the size is somewhat stable.
740 : : *
741 : : * (Note that this is only an issue if the plan gets cached and used again
742 : : * after the table has been filled. What we're trying to avoid is using a
743 : : * nestloop-type plan on a table that has grown substantially since the
744 : : * plan was made. Normally, autovacuum/autoanalyze will occur once enough
745 : : * inserts have happened and cause cached-plan invalidation; but that
746 : : * doesn't happen instantaneously, and it won't happen at all for cases
747 : : * such as temporary tables.)
748 : : *
749 : : * We test "never vacuumed" by seeing whether reltuples < 0.
750 : : *
751 : : * If the table has inheritance children, we don't apply this heuristic.
752 : : * Totally empty parent tables are quite common, so we should be willing
753 : : * to believe that they are empty.
754 : : */
755 [ + + + + ]: 361499 : if (curpages < 10 &&
756 : 91906 : reltuples < 0 &&
757 [ + + ]: 91906 : !rel->rd_rel->relhassubclass)
758 : 89633 : curpages = 10;
759 : :
760 : : /* report estimated # pages */
761 : 361499 : *pages = curpages;
762 : : /* quick exit if rel is clearly empty */
763 [ + + ]: 361499 : if (curpages == 0)
764 : : {
765 : 17381 : *tuples = 0;
766 : 17381 : *allvisfrac = 0;
767 : 17381 : return;
768 : : }
769 : :
770 : : /* estimate number of tuples from previous tuple density */
771 [ + + + + ]: 344118 : if (reltuples >= 0 && relpages > 0)
772 : 213931 : density = reltuples / (double) relpages;
773 : : else
774 : : {
775 : : /*
776 : : * When we have no data because the relation was never yet vacuumed,
777 : : * estimate tuple width from attribute datatypes. We assume here that
778 : : * the pages are completely full, which is OK for tables but is
779 : : * probably an overestimate for indexes. Fortunately
780 : : * get_relation_info() can clamp the overestimate to the parent
781 : : * table's size.
782 : : *
783 : : * Note: this code intentionally disregards alignment considerations,
784 : : * because (a) that would be gilding the lily considering how crude
785 : : * the estimate is, (b) it creates platform dependencies in the
786 : : * default plans which are kind of a headache for regression testing,
787 : : * and (c) different table AMs might use different padding schemes.
788 : : */
789 : : int32 tuple_width;
790 : : int fillfactor;
791 : :
792 : : /*
793 : : * Without reltuples/relpages, we also need to consider fillfactor.
794 : : * The other branch considers it implicitly by calculating density
795 : : * from actual relpages/reltuples statistics.
796 : : */
797 [ + + ]: 130187 : fillfactor = RelationGetFillFactor(rel, HEAP_DEFAULT_FILLFACTOR);
798 : :
799 : 130187 : tuple_width = get_rel_data_width(rel, attr_widths);
800 : 130187 : tuple_width += overhead_bytes_per_tuple;
801 : : /* note: integer division is intentional here */
802 : 130187 : density = (usable_bytes_per_page * fillfactor / 100) / tuple_width;
803 : : /* There's at least one row on the page, even with low fillfactor. */
804 : 130187 : density = clamp_row_est(density);
805 : : }
806 : 344118 : *tuples = rint(density * (double) curpages);
807 : :
808 : : /*
809 : : * We use relallvisible as-is, rather than scaling it up like we do for
810 : : * the pages and tuples counts, on the theory that any pages added since
811 : : * the last VACUUM are most likely not marked all-visible. But costsize.c
812 : : * wants it converted to a fraction.
813 : : */
814 [ + + - + ]: 344118 : if (relallvisible == 0 || curpages <= 0)
815 : 172424 : *allvisfrac = 0;
816 [ + + ]: 171694 : else if ((double) relallvisible >= curpages)
817 : 100405 : *allvisfrac = 1;
818 : : else
819 : 71289 : *allvisfrac = (double) relallvisible / curpages;
820 : : }
|
