Line data Source code
1 : /* -------------------------------------------------------------------------
2 : *
3 : * decode.c
4 : * This module decodes WAL records read using xlogreader.h's APIs for the
5 : * purpose of logical decoding by passing information to the
6 : * reorderbuffer module (containing the actual changes) and to the
7 : * snapbuild module to build a fitting catalog snapshot (to be able to
8 : * properly decode the changes in the reorderbuffer).
9 : *
10 : * NOTE:
11 : * This basically tries to handle all low level xlog stuff for
12 : * reorderbuffer.c and snapbuild.c. There's some minor leakage where a
13 : * specific record's struct is used to pass data along, but those just
14 : * happen to contain the right amount of data in a convenient
15 : * format. There isn't and shouldn't be much intelligence about the
16 : * contents of records in here except turning them into a more usable
17 : * format.
18 : *
19 : * Portions Copyright (c) 1996-2026, PostgreSQL Global Development Group
20 : * Portions Copyright (c) 1994, Regents of the University of California
21 : *
22 : * IDENTIFICATION
23 : * src/backend/replication/logical/decode.c
24 : *
25 : * -------------------------------------------------------------------------
26 : */
27 : #include "postgres.h"
28 :
29 : #include "access/heapam_xlog.h"
30 : #include "access/transam.h"
31 : #include "access/xact.h"
32 : #include "access/xlog_internal.h"
33 : #include "access/xlogreader.h"
34 : #include "access/xlogrecord.h"
35 : #include "catalog/pg_control.h"
36 : #include "replication/decode.h"
37 : #include "replication/logical.h"
38 : #include "replication/message.h"
39 : #include "replication/reorderbuffer.h"
40 : #include "replication/snapbuild.h"
41 : #include "storage/standbydefs.h"
42 :
43 : /* individual record(group)'s handlers */
44 : static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
45 : static void DecodeUpdate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
46 : static void DecodeDelete(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
47 : static void DecodeTruncate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
48 : static void DecodeMultiInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
49 : static void DecodeSpecConfirm(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
50 :
51 : static void DecodeCommit(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
52 : xl_xact_parsed_commit *parsed, TransactionId xid,
53 : bool two_phase);
54 : static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
55 : xl_xact_parsed_abort *parsed, TransactionId xid,
56 : bool two_phase);
57 : static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
58 : xl_xact_parsed_prepare *parsed);
59 :
60 :
61 : /* common function to decode tuples */
62 : static void DecodeXLogTuple(char *data, Size len, HeapTuple tuple);
63 :
64 : /* helper functions for decoding transactions */
65 : static inline bool FilterPrepare(LogicalDecodingContext *ctx,
66 : TransactionId xid, const char *gid);
67 : static bool DecodeTXNNeedSkip(LogicalDecodingContext *ctx,
68 : XLogRecordBuffer *buf, Oid txn_dbid,
69 : ReplOriginId origin_id);
70 :
71 : /*
72 : * Take every XLogReadRecord()ed record and perform the actions required to
73 : * decode it using the output plugin already setup in the logical decoding
74 : * context.
75 : *
76 : * NB: Note that every record's xid needs to be processed by reorderbuffer
77 : * (xids contained in the content of records are not relevant for this rule).
78 : * That means that for records which'd otherwise not go through the
79 : * reorderbuffer ReorderBufferProcessXid() has to be called. We don't want to
80 : * call ReorderBufferProcessXid for each record type by default, because
81 : * e.g. empty xacts can be handled more efficiently if there's no previous
82 : * state for them.
83 : *
84 : * We also support the ability to fast forward thru records, skipping some
85 : * record types completely - see individual record types for details.
86 : */
87 : void
88 4145702 : LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *record)
89 : {
90 : XLogRecordBuffer buf;
91 : TransactionId txid;
92 : RmgrData rmgr;
93 :
94 4145702 : buf.origptr = ctx->reader->ReadRecPtr;
95 4145702 : buf.endptr = ctx->reader->EndRecPtr;
96 4145702 : buf.record = record;
97 :
98 4145702 : txid = XLogRecGetTopXid(record);
99 :
100 : /*
101 : * If the top-level xid is valid, we need to assign the subxact to the
102 : * top-level xact. We need to do this for all records, hence we do it
103 : * before the switch.
104 : */
105 4145702 : if (TransactionIdIsValid(txid))
106 : {
107 1310 : ReorderBufferAssignChild(ctx->reorder,
108 : txid,
109 1310 : XLogRecGetXid(record),
110 : buf.origptr);
111 : }
112 :
113 4145702 : rmgr = GetRmgr(XLogRecGetRmid(record));
114 :
115 4145702 : if (rmgr.rm_decode != NULL)
116 3210612 : rmgr.rm_decode(ctx, &buf);
117 : else
118 : {
119 : /* just deal with xid, and done */
120 935090 : ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(record),
121 : buf.origptr);
122 : }
123 4145646 : }
124 :
125 : /*
126 : * Handle rmgr XLOG_ID records for LogicalDecodingProcessRecord().
127 : */
128 : void
129 12216 : xlog_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
130 : {
131 12216 : SnapBuild *builder = ctx->snapshot_builder;
132 12216 : uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
133 :
134 12216 : ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(buf->record),
135 : buf->origptr);
136 :
137 12216 : switch (info)
138 : {
139 : /* this is also used in END_OF_RECOVERY checkpoints */
140 122 : case XLOG_CHECKPOINT_SHUTDOWN:
141 : case XLOG_END_OF_RECOVERY:
142 122 : SnapBuildSerializationPoint(builder, buf->origptr);
143 :
144 122 : break;
145 148 : case XLOG_CHECKPOINT_ONLINE:
146 :
147 : /*
148 : * a RUNNING_XACTS record will have been logged near to this, we
149 : * can restart from there.
150 : */
151 148 : break;
152 0 : case XLOG_LOGICAL_DECODING_STATUS_CHANGE:
153 : {
154 : bool logical_decoding;
155 :
156 0 : memcpy(&logical_decoding, XLogRecGetData(buf->record), sizeof(bool));
157 :
158 : /*
159 : * Error out as we should not decode this WAL record.
160 : *
161 : * Logical decoding is disabled, and existing logical slots on
162 : * the standby are invalidated when this WAL record is
163 : * replayed. No logical decoder can process this WAL record
164 : * until replay completes, and by then the slots are already
165 : * invalidated. Furthermore, no new logical slots can be
166 : * created while logical decoding is disabled. This cannot
167 : * occur even on primary either, since it will not restart
168 : * with wal_level < replica if any logical slots exist.
169 : */
170 0 : elog(ERROR, "unexpected logical decoding status change %d",
171 : logical_decoding);
172 :
173 : break;
174 : }
175 11946 : case XLOG_NOOP:
176 : case XLOG_NEXTOID:
177 : case XLOG_SWITCH:
178 : case XLOG_BACKUP_END:
179 : case XLOG_PARAMETER_CHANGE:
180 : case XLOG_RESTORE_POINT:
181 : case XLOG_FPW_CHANGE:
182 : case XLOG_FPI_FOR_HINT:
183 : case XLOG_FPI:
184 : case XLOG_OVERWRITE_CONTRECORD:
185 : case XLOG_CHECKPOINT_REDO:
186 11946 : break;
187 0 : default:
188 0 : elog(ERROR, "unexpected RM_XLOG_ID record type: %u", info);
189 : }
190 12216 : }
191 :
192 : /*
193 : * Handle rmgr XACT_ID records for LogicalDecodingProcessRecord().
194 : */
195 : void
196 18158 : xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
197 : {
198 18158 : SnapBuild *builder = ctx->snapshot_builder;
199 18158 : ReorderBuffer *reorder = ctx->reorder;
200 18158 : XLogReaderState *r = buf->record;
201 18158 : uint8 info = XLogRecGetInfo(r) & XLOG_XACT_OPMASK;
202 :
203 : /*
204 : * If the snapshot isn't yet fully built, we cannot decode anything, so
205 : * bail out.
206 : */
207 18158 : if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
208 32 : return;
209 :
210 18126 : switch (info)
211 : {
212 6504 : case XLOG_XACT_COMMIT:
213 : case XLOG_XACT_COMMIT_PREPARED:
214 : {
215 : xl_xact_commit *xlrec;
216 : xl_xact_parsed_commit parsed;
217 : TransactionId xid;
218 6504 : bool two_phase = false;
219 :
220 6504 : xlrec = (xl_xact_commit *) XLogRecGetData(r);
221 6504 : ParseCommitRecord(XLogRecGetInfo(buf->record), xlrec, &parsed);
222 :
223 6504 : if (!TransactionIdIsValid(parsed.twophase_xid))
224 6266 : xid = XLogRecGetXid(r);
225 : else
226 238 : xid = parsed.twophase_xid;
227 :
228 : /*
229 : * We would like to process the transaction in a two-phase
230 : * manner iff output plugin supports two-phase commits and
231 : * doesn't filter the transaction at prepare time.
232 : */
233 6504 : if (info == XLOG_XACT_COMMIT_PREPARED)
234 238 : two_phase = !(FilterPrepare(ctx, xid,
235 238 : parsed.twophase_gid));
236 :
237 6504 : DecodeCommit(ctx, buf, &parsed, xid, two_phase);
238 6452 : break;
239 : }
240 386 : case XLOG_XACT_ABORT:
241 : case XLOG_XACT_ABORT_PREPARED:
242 : {
243 : xl_xact_abort *xlrec;
244 : xl_xact_parsed_abort parsed;
245 : TransactionId xid;
246 386 : bool two_phase = false;
247 :
248 386 : xlrec = (xl_xact_abort *) XLogRecGetData(r);
249 386 : ParseAbortRecord(XLogRecGetInfo(buf->record), xlrec, &parsed);
250 :
251 386 : if (!TransactionIdIsValid(parsed.twophase_xid))
252 296 : xid = XLogRecGetXid(r);
253 : else
254 90 : xid = parsed.twophase_xid;
255 :
256 : /*
257 : * We would like to process the transaction in a two-phase
258 : * manner iff output plugin supports two-phase commits and
259 : * doesn't filter the transaction at prepare time.
260 : */
261 386 : if (info == XLOG_XACT_ABORT_PREPARED)
262 90 : two_phase = !(FilterPrepare(ctx, xid,
263 90 : parsed.twophase_gid));
264 :
265 386 : DecodeAbort(ctx, buf, &parsed, xid, two_phase);
266 386 : break;
267 : }
268 232 : case XLOG_XACT_ASSIGNMENT:
269 :
270 : /*
271 : * We assign subxact to the toplevel xact while processing each
272 : * record if required. So, we don't need to do anything here. See
273 : * LogicalDecodingProcessRecord.
274 : */
275 232 : break;
276 10636 : case XLOG_XACT_INVALIDATIONS:
277 : {
278 : TransactionId xid;
279 : xl_xact_invals *invals;
280 :
281 10636 : xid = XLogRecGetXid(r);
282 10636 : invals = (xl_xact_invals *) XLogRecGetData(r);
283 :
284 : /*
285 : * Execute the invalidations for xid-less transactions,
286 : * otherwise, accumulate them so that they can be processed at
287 : * the commit time.
288 : */
289 10636 : if (TransactionIdIsValid(xid))
290 : {
291 10620 : if (!ctx->fast_forward)
292 10552 : ReorderBufferAddInvalidations(reorder, xid,
293 : buf->origptr,
294 10552 : invals->nmsgs,
295 10552 : invals->msgs);
296 10620 : ReorderBufferXidSetCatalogChanges(ctx->reorder, xid,
297 : buf->origptr);
298 : }
299 16 : else if (!ctx->fast_forward)
300 16 : ReorderBufferImmediateInvalidation(ctx->reorder,
301 16 : invals->nmsgs,
302 16 : invals->msgs);
303 :
304 10636 : break;
305 : }
306 368 : case XLOG_XACT_PREPARE:
307 : {
308 : xl_xact_parsed_prepare parsed;
309 : xl_xact_prepare *xlrec;
310 :
311 : /* ok, parse it */
312 368 : xlrec = (xl_xact_prepare *) XLogRecGetData(r);
313 368 : ParsePrepareRecord(XLogRecGetInfo(buf->record),
314 : xlrec, &parsed);
315 :
316 : /*
317 : * We would like to process the transaction in a two-phase
318 : * manner iff output plugin supports two-phase commits and
319 : * doesn't filter the transaction at prepare time.
320 : */
321 368 : if (FilterPrepare(ctx, parsed.twophase_xid,
322 : parsed.twophase_gid))
323 : {
324 40 : ReorderBufferProcessXid(reorder, parsed.twophase_xid,
325 : buf->origptr);
326 40 : break;
327 : }
328 :
329 : /*
330 : * Note that if the prepared transaction has locked [user]
331 : * catalog tables exclusively then decoding prepare can block
332 : * till the main transaction is committed because it needs to
333 : * lock the catalog tables.
334 : *
335 : * XXX Now, this can even lead to a deadlock if the prepare
336 : * transaction is waiting to get it logically replicated for
337 : * distributed 2PC. This can be avoided by disallowing
338 : * preparing transactions that have locked [user] catalog
339 : * tables exclusively but as of now, we ask users not to do
340 : * such an operation.
341 : */
342 328 : DecodePrepare(ctx, buf, &parsed);
343 328 : break;
344 : }
345 0 : default:
346 0 : elog(ERROR, "unexpected RM_XACT_ID record type: %u", info);
347 : }
348 : }
349 :
350 : /*
351 : * Handle rmgr STANDBY_ID records for LogicalDecodingProcessRecord().
352 : */
353 : void
354 7768 : standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
355 : {
356 7768 : SnapBuild *builder = ctx->snapshot_builder;
357 7768 : XLogReaderState *r = buf->record;
358 7768 : uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
359 :
360 7768 : ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
361 :
362 7768 : switch (info)
363 : {
364 2888 : case XLOG_RUNNING_XACTS:
365 : {
366 2888 : xl_running_xacts *running = (xl_running_xacts *) XLogRecGetData(r);
367 :
368 2888 : SnapBuildProcessRunningXacts(builder, buf->origptr, running);
369 :
370 : /*
371 : * Abort all transactions that we keep track of, that are
372 : * older than the record's oldestRunningXid. This is the most
373 : * convenient spot for doing so since, in contrast to shutdown
374 : * or end-of-recovery checkpoints, we have information about
375 : * all running transactions which includes prepared ones,
376 : * while shutdown checkpoints just know that no non-prepared
377 : * transactions are in progress.
378 : */
379 2884 : ReorderBufferAbortOld(ctx->reorder, running->oldestRunningXid);
380 : }
381 2884 : break;
382 4864 : case XLOG_STANDBY_LOCK:
383 4864 : break;
384 16 : case XLOG_INVALIDATIONS:
385 :
386 : /*
387 : * We are processing the invalidations at the command level via
388 : * XLOG_XACT_INVALIDATIONS. So we don't need to do anything here.
389 : */
390 16 : break;
391 0 : default:
392 0 : elog(ERROR, "unexpected RM_STANDBY_ID record type: %u", info);
393 : }
394 7764 : }
395 :
396 : /*
397 : * Handle rmgr HEAP2_ID records for LogicalDecodingProcessRecord().
398 : */
399 : void
400 65980 : heap2_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
401 : {
402 65980 : uint8 info = XLogRecGetInfo(buf->record) & XLOG_HEAP_OPMASK;
403 65980 : TransactionId xid = XLogRecGetXid(buf->record);
404 65980 : SnapBuild *builder = ctx->snapshot_builder;
405 :
406 65980 : ReorderBufferProcessXid(ctx->reorder, xid, buf->origptr);
407 :
408 : /*
409 : * If we don't have snapshot or we are just fast-forwarding, there is no
410 : * point in decoding data changes. However, it's crucial to build the base
411 : * snapshot during fast-forward mode (as is done in
412 : * SnapBuildProcessChange()) because we require the snapshot's xmin when
413 : * determining the candidate catalog_xmin for the replication slot. See
414 : * SnapBuildProcessRunningXacts().
415 : */
416 65980 : if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
417 20 : return;
418 :
419 65960 : switch (info)
420 : {
421 12302 : case XLOG_HEAP2_MULTI_INSERT:
422 12302 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
423 12302 : !ctx->fast_forward)
424 12166 : DecodeMultiInsert(ctx, buf);
425 12302 : break;
426 50098 : case XLOG_HEAP2_NEW_CID:
427 50098 : if (!ctx->fast_forward)
428 : {
429 : xl_heap_new_cid *xlrec;
430 :
431 49710 : xlrec = (xl_heap_new_cid *) XLogRecGetData(buf->record);
432 49710 : SnapBuildProcessNewCid(builder, xid, buf->origptr, xlrec);
433 :
434 49710 : break;
435 : }
436 : case XLOG_HEAP2_REWRITE:
437 :
438 : /*
439 : * Although these records only exist to serve the needs of logical
440 : * decoding, all the work happens as part of crash or archive
441 : * recovery, so we don't need to do anything here.
442 : */
443 568 : break;
444 :
445 : /*
446 : * Everything else here is just low level physical stuff we're not
447 : * interested in.
448 : */
449 3380 : case XLOG_HEAP2_PRUNE_ON_ACCESS:
450 : case XLOG_HEAP2_PRUNE_VACUUM_SCAN:
451 : case XLOG_HEAP2_PRUNE_VACUUM_CLEANUP:
452 : case XLOG_HEAP2_VISIBLE:
453 : case XLOG_HEAP2_LOCK_UPDATED:
454 3380 : break;
455 0 : default:
456 0 : elog(ERROR, "unexpected RM_HEAP2_ID record type: %u", info);
457 : }
458 : }
459 :
460 : /*
461 : * Handle rmgr HEAP_ID records for LogicalDecodingProcessRecord().
462 : */
463 : void
464 3106302 : heap_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
465 : {
466 3106302 : uint8 info = XLogRecGetInfo(buf->record) & XLOG_HEAP_OPMASK;
467 3106302 : TransactionId xid = XLogRecGetXid(buf->record);
468 3106302 : SnapBuild *builder = ctx->snapshot_builder;
469 :
470 3106302 : ReorderBufferProcessXid(ctx->reorder, xid, buf->origptr);
471 :
472 : /*
473 : * If we don't have snapshot or we are just fast-forwarding, there is no
474 : * point in decoding data changes. However, it's crucial to build the base
475 : * snapshot during fast-forward mode (as is done in
476 : * SnapBuildProcessChange()) because we require the snapshot's xmin when
477 : * determining the candidate catalog_xmin for the replication slot. See
478 : * SnapBuildProcessRunningXacts().
479 : */
480 3106302 : if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
481 10 : return;
482 :
483 3106292 : switch (info)
484 : {
485 1964004 : case XLOG_HEAP_INSERT:
486 1964004 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
487 1964004 : !ctx->fast_forward)
488 1963750 : DecodeInsert(ctx, buf);
489 1964004 : break;
490 :
491 : /*
492 : * Treat HOT update as normal updates. There is no useful
493 : * information in the fact that we could make it a HOT update
494 : * locally and the WAL layout is compatible.
495 : */
496 334854 : case XLOG_HEAP_HOT_UPDATE:
497 : case XLOG_HEAP_UPDATE:
498 334854 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
499 334854 : !ctx->fast_forward)
500 334840 : DecodeUpdate(ctx, buf);
501 334854 : break;
502 :
503 442016 : case XLOG_HEAP_DELETE:
504 442016 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
505 442016 : !ctx->fast_forward)
506 442010 : DecodeDelete(ctx, buf);
507 442016 : break;
508 :
509 126 : case XLOG_HEAP_TRUNCATE:
510 126 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
511 126 : !ctx->fast_forward)
512 124 : DecodeTruncate(ctx, buf);
513 126 : break;
514 :
515 2080 : case XLOG_HEAP_INPLACE:
516 :
517 : /*
518 : * Inplace updates are only ever performed on catalog tuples and
519 : * can, per definition, not change tuple visibility. Since we
520 : * also don't decode catalog tuples, we're not interested in the
521 : * record's contents.
522 : */
523 2080 : break;
524 :
525 35832 : case XLOG_HEAP_CONFIRM:
526 35832 : if (SnapBuildProcessChange(builder, xid, buf->origptr) &&
527 35832 : !ctx->fast_forward)
528 35832 : DecodeSpecConfirm(ctx, buf);
529 35832 : break;
530 :
531 327380 : case XLOG_HEAP_LOCK:
532 : /* we don't care about row level locks for now */
533 327380 : break;
534 :
535 0 : default:
536 0 : elog(ERROR, "unexpected RM_HEAP_ID record type: %u", info);
537 : break;
538 : }
539 : }
540 :
541 : /*
542 : * Ask output plugin whether we want to skip this PREPARE and send
543 : * this transaction as a regular commit later.
544 : */
545 : static inline bool
546 696 : FilterPrepare(LogicalDecodingContext *ctx, TransactionId xid,
547 : const char *gid)
548 : {
549 : /*
550 : * Skip if decoding of two-phase transactions at PREPARE time is not
551 : * enabled. In that case, all two-phase transactions are considered
552 : * filtered out and will be applied as regular transactions at COMMIT
553 : * PREPARED.
554 : */
555 696 : if (!ctx->twophase)
556 46 : return true;
557 :
558 : /*
559 : * The filter_prepare callback is optional. When not supplied, all
560 : * prepared transactions should go through.
561 : */
562 650 : if (ctx->callbacks.filter_prepare_cb == NULL)
563 354 : return false;
564 :
565 296 : return filter_prepare_cb_wrapper(ctx, xid, gid);
566 : }
567 :
568 : static inline bool
569 2771948 : FilterByOrigin(LogicalDecodingContext *ctx, ReplOriginId origin_id)
570 : {
571 2771948 : if (ctx->callbacks.filter_by_origin_cb == NULL)
572 54 : return false;
573 :
574 2771894 : return filter_by_origin_cb_wrapper(ctx, origin_id);
575 : }
576 :
577 : /*
578 : * Handle rmgr LOGICALMSG_ID records for LogicalDecodingProcessRecord().
579 : */
580 : void
581 188 : logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
582 : {
583 188 : SnapBuild *builder = ctx->snapshot_builder;
584 188 : XLogReaderState *r = buf->record;
585 188 : TransactionId xid = XLogRecGetXid(r);
586 188 : uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
587 188 : ReplOriginId origin_id = XLogRecGetOrigin(r);
588 188 : Snapshot snapshot = NULL;
589 : xl_logical_message *message;
590 :
591 188 : if (info != XLOG_LOGICAL_MESSAGE)
592 0 : elog(ERROR, "unexpected RM_LOGICALMSG_ID record type: %u", info);
593 :
594 188 : ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
595 :
596 : /* If we don't have snapshot, there is no point in decoding messages */
597 188 : if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
598 0 : return;
599 :
600 188 : message = (xl_logical_message *) XLogRecGetData(r);
601 :
602 372 : if (message->dbId != ctx->slot->data.database ||
603 184 : FilterByOrigin(ctx, origin_id))
604 8 : return;
605 :
606 180 : if (message->transactional &&
607 78 : !SnapBuildProcessChange(builder, xid, buf->origptr))
608 0 : return;
609 282 : else if (!message->transactional &&
610 204 : (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
611 102 : SnapBuildXactNeedsSkip(builder, buf->origptr)))
612 80 : return;
613 :
614 : /*
615 : * We also skip decoding in fast_forward mode. This check must be last
616 : * because we don't want to set the processing_required flag unless we
617 : * have a decodable message.
618 : */
619 100 : if (ctx->fast_forward)
620 : {
621 : /*
622 : * We need to set processing_required flag to notify the message's
623 : * existence to the caller. Usually, the flag is set when either the
624 : * COMMIT or ABORT records are decoded, but this must be turned on
625 : * here because the non-transactional logical message is decoded
626 : * without waiting for these records.
627 : */
628 6 : if (!message->transactional)
629 6 : ctx->processing_required = true;
630 :
631 6 : return;
632 : }
633 :
634 : /*
635 : * If this is a non-transactional change, get the snapshot we're expected
636 : * to use. We only get here when the snapshot is consistent, and the
637 : * change is not meant to be skipped.
638 : *
639 : * For transactional changes we don't need a snapshot, we'll use the
640 : * regular snapshot maintained by ReorderBuffer. We just leave it NULL.
641 : */
642 94 : if (!message->transactional)
643 16 : snapshot = SnapBuildGetOrBuildSnapshot(builder);
644 :
645 94 : ReorderBufferQueueMessage(ctx->reorder, xid, snapshot, buf->endptr,
646 94 : message->transactional,
647 94 : message->message, /* first part of message is
648 : * prefix */
649 : message->message_size,
650 94 : message->message + message->prefix_size);
651 : }
652 :
653 : /*
654 : * Consolidated commit record handling between the different form of commit
655 : * records.
656 : *
657 : * 'two_phase' indicates that caller wants to process the transaction in two
658 : * phases, first process prepare if not already done and then process
659 : * commit_prepared.
660 : */
661 : static void
662 6504 : DecodeCommit(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
663 : xl_xact_parsed_commit *parsed, TransactionId xid,
664 : bool two_phase)
665 : {
666 6504 : XLogRecPtr origin_lsn = InvalidXLogRecPtr;
667 6504 : TimestampTz commit_time = parsed->xact_time;
668 6504 : ReplOriginId origin_id = XLogRecGetOrigin(buf->record);
669 : int i;
670 :
671 6504 : if (parsed->xinfo & XACT_XINFO_HAS_ORIGIN)
672 : {
673 256 : origin_lsn = parsed->origin_lsn;
674 256 : commit_time = parsed->origin_timestamp;
675 : }
676 :
677 6504 : SnapBuildCommitTxn(ctx->snapshot_builder, buf->origptr, xid,
678 : parsed->nsubxacts, parsed->subxacts,
679 : parsed->xinfo);
680 :
681 : /* ----
682 : * Check whether we are interested in this specific transaction, and tell
683 : * the reorderbuffer to forget the content of the (sub-)transactions
684 : * if not.
685 : *
686 : * We can't just use ReorderBufferAbort() here, because we need to execute
687 : * the transaction's invalidations. This currently won't be needed if
688 : * we're just skipping over the transaction because currently we only do
689 : * so during startup, to get to the first transaction the client needs. As
690 : * we have reset the catalog caches before starting to read WAL, and we
691 : * haven't yet touched any catalogs, there can't be anything to invalidate.
692 : * But if we're "forgetting" this commit because it happened in another
693 : * database, the invalidations might be important, because they could be
694 : * for shared catalogs and we might have loaded data into the relevant
695 : * syscaches.
696 : * ---
697 : */
698 6504 : if (DecodeTXNNeedSkip(ctx, buf, parsed->dbId, origin_id))
699 : {
700 5498 : for (i = 0; i < parsed->nsubxacts; i++)
701 : {
702 1910 : ReorderBufferForget(ctx->reorder, parsed->subxacts[i], buf->origptr);
703 : }
704 3588 : ReorderBufferForget(ctx->reorder, xid, buf->origptr);
705 :
706 3588 : return;
707 : }
708 :
709 : /* tell the reorderbuffer about the surviving subtransactions */
710 3448 : for (i = 0; i < parsed->nsubxacts; i++)
711 : {
712 532 : ReorderBufferCommitChild(ctx->reorder, xid, parsed->subxacts[i],
713 : buf->origptr, buf->endptr);
714 : }
715 :
716 : /*
717 : * Send the final commit record if the transaction data is already
718 : * decoded, otherwise, process the entire transaction.
719 : */
720 2916 : if (two_phase)
721 : {
722 66 : ReorderBufferFinishPrepared(ctx->reorder, xid, buf->origptr, buf->endptr,
723 66 : SnapBuildGetTwoPhaseAt(ctx->snapshot_builder),
724 : commit_time, origin_id, origin_lsn,
725 66 : parsed->twophase_gid, true);
726 : }
727 : else
728 : {
729 2850 : ReorderBufferCommit(ctx->reorder, xid, buf->origptr, buf->endptr,
730 : commit_time, origin_id, origin_lsn);
731 : }
732 :
733 : /*
734 : * Update the decoding stats at transaction prepare/commit/abort.
735 : * Additionally we send the stats when we spill or stream the changes to
736 : * avoid losing them in case the decoding is interrupted. It is not clear
737 : * that sending more or less frequently than this would be better.
738 : */
739 2864 : UpdateDecodingStats(ctx);
740 : }
741 :
742 : /*
743 : * Decode PREPARE record. Similar logic as in DecodeCommit.
744 : *
745 : * Note that we don't skip prepare even if have detected concurrent abort
746 : * because it is quite possible that we had already sent some changes before we
747 : * detect abort in which case we need to abort those changes in the subscriber.
748 : * To abort such changes, we do send the prepare and then the rollback prepared
749 : * which is what happened on the publisher-side as well. Now, we can invent a
750 : * new abort API wherein in such cases we send abort and skip sending prepared
751 : * and rollback prepared but then it is not that straightforward because we
752 : * might have streamed this transaction by that time in which case it is
753 : * handled when the rollback is encountered. It is not impossible to optimize
754 : * the concurrent abort case but it can introduce design complexity w.r.t
755 : * handling different cases so leaving it for now as it doesn't seem worth it.
756 : */
757 : static void
758 328 : DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
759 : xl_xact_parsed_prepare *parsed)
760 : {
761 328 : SnapBuild *builder = ctx->snapshot_builder;
762 328 : XLogRecPtr origin_lsn = parsed->origin_lsn;
763 328 : TimestampTz prepare_time = parsed->xact_time;
764 328 : ReplOriginId origin_id = XLogRecGetOrigin(buf->record);
765 : int i;
766 328 : TransactionId xid = parsed->twophase_xid;
767 :
768 328 : if (parsed->origin_timestamp != 0)
769 16 : prepare_time = parsed->origin_timestamp;
770 :
771 : /*
772 : * Remember the prepare info for a txn so that it can be used later in
773 : * commit prepared if required. See ReorderBufferFinishPrepared.
774 : */
775 328 : if (!ReorderBufferRememberPrepareInfo(ctx->reorder, xid, buf->origptr,
776 : buf->endptr, prepare_time, origin_id,
777 : origin_lsn))
778 0 : return;
779 :
780 : /* We can't start streaming unless a consistent state is reached. */
781 328 : if (SnapBuildCurrentState(builder) < SNAPBUILD_CONSISTENT)
782 : {
783 6 : ReorderBufferSkipPrepare(ctx->reorder, xid);
784 6 : return;
785 : }
786 :
787 : /*
788 : * Check whether we need to process this transaction. See
789 : * DecodeTXNNeedSkip for the reasons why we sometimes want to skip the
790 : * transaction.
791 : *
792 : * We can't call ReorderBufferForget as we did in DecodeCommit as the txn
793 : * hasn't yet been committed, removing this txn before a commit might
794 : * result in the computation of an incorrect restart_lsn. See
795 : * SnapBuildProcessRunningXacts. But we need to process cache
796 : * invalidations if there are any for the reasons mentioned in
797 : * DecodeCommit.
798 : */
799 322 : if (DecodeTXNNeedSkip(ctx, buf, parsed->dbId, origin_id))
800 : {
801 246 : ReorderBufferSkipPrepare(ctx->reorder, xid);
802 246 : ReorderBufferInvalidate(ctx->reorder, xid, buf->origptr);
803 246 : return;
804 : }
805 :
806 : /* Tell the reorderbuffer about the surviving subtransactions. */
807 78 : for (i = 0; i < parsed->nsubxacts; i++)
808 : {
809 2 : ReorderBufferCommitChild(ctx->reorder, xid, parsed->subxacts[i],
810 : buf->origptr, buf->endptr);
811 : }
812 :
813 : /* replay actions of all transaction + subtransactions in order */
814 76 : ReorderBufferPrepare(ctx->reorder, xid, parsed->twophase_gid);
815 :
816 : /*
817 : * Update the decoding stats at transaction prepare/commit/abort.
818 : * Additionally we send the stats when we spill or stream the changes to
819 : * avoid losing them in case the decoding is interrupted. It is not clear
820 : * that sending more or less frequently than this would be better.
821 : */
822 76 : UpdateDecodingStats(ctx);
823 : }
824 :
825 :
826 : /*
827 : * Get the data from the various forms of abort records and pass it on to
828 : * snapbuild.c and reorderbuffer.c.
829 : *
830 : * 'two_phase' indicates to finish prepared transaction.
831 : */
832 : static void
833 386 : DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
834 : xl_xact_parsed_abort *parsed, TransactionId xid,
835 : bool two_phase)
836 : {
837 : int i;
838 386 : XLogRecPtr origin_lsn = InvalidXLogRecPtr;
839 386 : TimestampTz abort_time = parsed->xact_time;
840 386 : ReplOriginId origin_id = XLogRecGetOrigin(buf->record);
841 : bool skip_xact;
842 :
843 386 : if (parsed->xinfo & XACT_XINFO_HAS_ORIGIN)
844 : {
845 8 : origin_lsn = parsed->origin_lsn;
846 8 : abort_time = parsed->origin_timestamp;
847 : }
848 :
849 : /*
850 : * Check whether we need to process this transaction. See
851 : * DecodeTXNNeedSkip for the reasons why we sometimes want to skip the
852 : * transaction.
853 : */
854 386 : skip_xact = DecodeTXNNeedSkip(ctx, buf, parsed->dbId, origin_id);
855 :
856 : /*
857 : * Send the final rollback record for a prepared transaction unless we
858 : * need to skip it. For non-two-phase xacts, simply forget the xact.
859 : */
860 386 : if (two_phase && !skip_xact)
861 : {
862 18 : ReorderBufferFinishPrepared(ctx->reorder, xid, buf->origptr, buf->endptr,
863 : InvalidXLogRecPtr,
864 : abort_time, origin_id, origin_lsn,
865 18 : parsed->twophase_gid, false);
866 : }
867 : else
868 : {
869 380 : for (i = 0; i < parsed->nsubxacts; i++)
870 : {
871 12 : ReorderBufferAbort(ctx->reorder, parsed->subxacts[i],
872 12 : buf->record->EndRecPtr, abort_time);
873 : }
874 :
875 368 : ReorderBufferAbort(ctx->reorder, xid, buf->record->EndRecPtr,
876 : abort_time);
877 : }
878 :
879 : /* update the decoding stats */
880 386 : UpdateDecodingStats(ctx);
881 386 : }
882 :
883 : /*
884 : * Parse XLOG_HEAP_INSERT (not MULTI_INSERT!) records into tuplebufs.
885 : *
886 : * Inserts can contain the new tuple.
887 : */
888 : static void
889 1963750 : DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
890 : {
891 : Size datalen;
892 : char *tupledata;
893 : Size tuplelen;
894 1963750 : XLogReaderState *r = buf->record;
895 : xl_heap_insert *xlrec;
896 : ReorderBufferChange *change;
897 : RelFileLocator target_locator;
898 :
899 1963750 : xlrec = (xl_heap_insert *) XLogRecGetData(r);
900 :
901 : /*
902 : * Ignore insert records without new tuples (this does happen when
903 : * raw_heap_insert marks the TOAST record as HEAP_INSERT_NO_LOGICAL).
904 : */
905 1963750 : if (!(xlrec->flags & XLH_INSERT_CONTAINS_NEW_TUPLE))
906 8064 : return;
907 :
908 : /* only interested in our database */
909 1955892 : XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
910 1955892 : if (target_locator.dbOid != ctx->slot->data.database)
911 0 : return;
912 :
913 : /* output plugin doesn't look for this origin, no need to queue */
914 1955892 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
915 206 : return;
916 :
917 1955686 : change = ReorderBufferAllocChange(ctx->reorder);
918 1955686 : if (!(xlrec->flags & XLH_INSERT_IS_SPECULATIVE))
919 1919854 : change->action = REORDER_BUFFER_CHANGE_INSERT;
920 : else
921 35832 : change->action = REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT;
922 1955686 : change->origin_id = XLogRecGetOrigin(r);
923 :
924 1955686 : memcpy(&change->data.tp.rlocator, &target_locator, sizeof(RelFileLocator));
925 :
926 1955686 : tupledata = XLogRecGetBlockData(r, 0, &datalen);
927 1955686 : tuplelen = datalen - SizeOfHeapHeader;
928 :
929 1955686 : change->data.tp.newtuple =
930 1955686 : ReorderBufferAllocTupleBuf(ctx->reorder, tuplelen);
931 :
932 1955686 : DecodeXLogTuple(tupledata, datalen, change->data.tp.newtuple);
933 :
934 1955686 : change->data.tp.clear_toast_afterwards = true;
935 :
936 1955686 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r), buf->origptr,
937 : change,
938 1955686 : xlrec->flags & XLH_INSERT_ON_TOAST_RELATION);
939 : }
940 :
941 : /*
942 : * Parse XLOG_HEAP_UPDATE and XLOG_HEAP_HOT_UPDATE, which have the same layout
943 : * in the record, from wal into proper tuplebufs.
944 : *
945 : * Updates can possibly contain a new tuple and the old primary key.
946 : */
947 : static void
948 334840 : DecodeUpdate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
949 : {
950 334840 : XLogReaderState *r = buf->record;
951 : xl_heap_update *xlrec;
952 : ReorderBufferChange *change;
953 : char *data;
954 : RelFileLocator target_locator;
955 :
956 334840 : xlrec = (xl_heap_update *) XLogRecGetData(r);
957 :
958 : /* only interested in our database */
959 334840 : XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
960 334840 : if (target_locator.dbOid != ctx->slot->data.database)
961 244 : return;
962 :
963 : /* output plugin doesn't look for this origin, no need to queue */
964 334702 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
965 106 : return;
966 :
967 334596 : change = ReorderBufferAllocChange(ctx->reorder);
968 334596 : change->action = REORDER_BUFFER_CHANGE_UPDATE;
969 334596 : change->origin_id = XLogRecGetOrigin(r);
970 334596 : memcpy(&change->data.tp.rlocator, &target_locator, sizeof(RelFileLocator));
971 :
972 334596 : if (xlrec->flags & XLH_UPDATE_CONTAINS_NEW_TUPLE)
973 : {
974 : Size datalen;
975 : Size tuplelen;
976 :
977 330978 : data = XLogRecGetBlockData(r, 0, &datalen);
978 :
979 330978 : tuplelen = datalen - SizeOfHeapHeader;
980 :
981 330978 : change->data.tp.newtuple =
982 330978 : ReorderBufferAllocTupleBuf(ctx->reorder, tuplelen);
983 :
984 330978 : DecodeXLogTuple(data, datalen, change->data.tp.newtuple);
985 : }
986 :
987 334596 : if (xlrec->flags & XLH_UPDATE_CONTAINS_OLD)
988 : {
989 : Size datalen;
990 : Size tuplelen;
991 :
992 : /* caution, remaining data in record is not aligned */
993 624 : data = XLogRecGetData(r) + SizeOfHeapUpdate;
994 624 : datalen = XLogRecGetDataLen(r) - SizeOfHeapUpdate;
995 624 : tuplelen = datalen - SizeOfHeapHeader;
996 :
997 624 : change->data.tp.oldtuple =
998 624 : ReorderBufferAllocTupleBuf(ctx->reorder, tuplelen);
999 :
1000 624 : DecodeXLogTuple(data, datalen, change->data.tp.oldtuple);
1001 : }
1002 :
1003 334596 : change->data.tp.clear_toast_afterwards = true;
1004 :
1005 334596 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r), buf->origptr,
1006 : change, false);
1007 : }
1008 :
1009 : /*
1010 : * Parse XLOG_HEAP_DELETE from wal into proper tuplebufs.
1011 : *
1012 : * Deletes can possibly contain the old primary key.
1013 : */
1014 : static void
1015 442010 : DecodeDelete(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
1016 : {
1017 442010 : XLogReaderState *r = buf->record;
1018 : xl_heap_delete *xlrec;
1019 : ReorderBufferChange *change;
1020 : RelFileLocator target_locator;
1021 :
1022 442010 : xlrec = (xl_heap_delete *) XLogRecGetData(r);
1023 :
1024 : /* only interested in our database */
1025 442010 : XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
1026 442010 : if (target_locator.dbOid != ctx->slot->data.database)
1027 128 : return;
1028 :
1029 : /* output plugin doesn't look for this origin, no need to queue */
1030 441916 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
1031 34 : return;
1032 :
1033 441882 : change = ReorderBufferAllocChange(ctx->reorder);
1034 :
1035 441882 : if (xlrec->flags & XLH_DELETE_IS_SUPER)
1036 0 : change->action = REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT;
1037 : else
1038 441882 : change->action = REORDER_BUFFER_CHANGE_DELETE;
1039 :
1040 441882 : change->origin_id = XLogRecGetOrigin(r);
1041 :
1042 441882 : memcpy(&change->data.tp.rlocator, &target_locator, sizeof(RelFileLocator));
1043 :
1044 : /* old primary key stored */
1045 441882 : if (xlrec->flags & XLH_DELETE_CONTAINS_OLD)
1046 : {
1047 318310 : Size datalen = XLogRecGetDataLen(r) - SizeOfHeapDelete;
1048 318310 : Size tuplelen = datalen - SizeOfHeapHeader;
1049 :
1050 : Assert(XLogRecGetDataLen(r) > (SizeOfHeapDelete + SizeOfHeapHeader));
1051 :
1052 318310 : change->data.tp.oldtuple =
1053 318310 : ReorderBufferAllocTupleBuf(ctx->reorder, tuplelen);
1054 :
1055 318310 : DecodeXLogTuple((char *) xlrec + SizeOfHeapDelete,
1056 : datalen, change->data.tp.oldtuple);
1057 : }
1058 :
1059 441882 : change->data.tp.clear_toast_afterwards = true;
1060 :
1061 441882 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r), buf->origptr,
1062 : change, false);
1063 : }
1064 :
1065 : /*
1066 : * Parse XLOG_HEAP_TRUNCATE from wal
1067 : */
1068 : static void
1069 124 : DecodeTruncate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
1070 : {
1071 124 : XLogReaderState *r = buf->record;
1072 : xl_heap_truncate *xlrec;
1073 : ReorderBufferChange *change;
1074 :
1075 124 : xlrec = (xl_heap_truncate *) XLogRecGetData(r);
1076 :
1077 : /* only interested in our database */
1078 124 : if (xlrec->dbId != ctx->slot->data.database)
1079 0 : return;
1080 :
1081 : /* output plugin doesn't look for this origin, no need to queue */
1082 124 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
1083 2 : return;
1084 :
1085 122 : change = ReorderBufferAllocChange(ctx->reorder);
1086 122 : change->action = REORDER_BUFFER_CHANGE_TRUNCATE;
1087 122 : change->origin_id = XLogRecGetOrigin(r);
1088 122 : if (xlrec->flags & XLH_TRUNCATE_CASCADE)
1089 2 : change->data.truncate.cascade = true;
1090 122 : if (xlrec->flags & XLH_TRUNCATE_RESTART_SEQS)
1091 4 : change->data.truncate.restart_seqs = true;
1092 122 : change->data.truncate.nrelids = xlrec->nrelids;
1093 244 : change->data.truncate.relids = ReorderBufferAllocRelids(ctx->reorder,
1094 122 : xlrec->nrelids);
1095 122 : memcpy(change->data.truncate.relids, xlrec->relids,
1096 122 : xlrec->nrelids * sizeof(Oid));
1097 122 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r),
1098 : buf->origptr, change, false);
1099 : }
1100 :
1101 : /*
1102 : * Decode XLOG_HEAP2_MULTI_INSERT record into multiple tuplebufs.
1103 : *
1104 : * Currently MULTI_INSERT will always contain the full tuples.
1105 : */
1106 : static void
1107 12166 : DecodeMultiInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
1108 : {
1109 12166 : XLogReaderState *r = buf->record;
1110 : xl_heap_multi_insert *xlrec;
1111 : int i;
1112 : char *data;
1113 : char *tupledata;
1114 : Size tuplelen;
1115 : RelFileLocator rlocator;
1116 :
1117 12166 : xlrec = (xl_heap_multi_insert *) XLogRecGetData(r);
1118 :
1119 : /*
1120 : * Ignore insert records without new tuples. This happens when a
1121 : * multi_insert is done on a catalog or on a non-persistent relation.
1122 : */
1123 12166 : if (!(xlrec->flags & XLH_INSERT_CONTAINS_NEW_TUPLE))
1124 12122 : return;
1125 :
1126 : /* only interested in our database */
1127 140 : XLogRecGetBlockTag(r, 0, &rlocator, NULL, NULL);
1128 140 : if (rlocator.dbOid != ctx->slot->data.database)
1129 96 : return;
1130 :
1131 : /* output plugin doesn't look for this origin, no need to queue */
1132 44 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
1133 0 : return;
1134 :
1135 : /*
1136 : * We know that this multi_insert isn't for a catalog, so the block should
1137 : * always have data even if a full-page write of it is taken.
1138 : */
1139 44 : tupledata = XLogRecGetBlockData(r, 0, &tuplelen);
1140 : Assert(tupledata != NULL);
1141 :
1142 44 : data = tupledata;
1143 2202 : for (i = 0; i < xlrec->ntuples; i++)
1144 : {
1145 : ReorderBufferChange *change;
1146 : xl_multi_insert_tuple *xlhdr;
1147 : int datalen;
1148 : HeapTuple tuple;
1149 : HeapTupleHeader header;
1150 :
1151 2158 : change = ReorderBufferAllocChange(ctx->reorder);
1152 2158 : change->action = REORDER_BUFFER_CHANGE_INSERT;
1153 2158 : change->origin_id = XLogRecGetOrigin(r);
1154 :
1155 2158 : memcpy(&change->data.tp.rlocator, &rlocator, sizeof(RelFileLocator));
1156 :
1157 2158 : xlhdr = (xl_multi_insert_tuple *) SHORTALIGN(data);
1158 2158 : data = ((char *) xlhdr) + SizeOfMultiInsertTuple;
1159 2158 : datalen = xlhdr->datalen;
1160 :
1161 2158 : change->data.tp.newtuple =
1162 2158 : ReorderBufferAllocTupleBuf(ctx->reorder, datalen);
1163 :
1164 2158 : tuple = change->data.tp.newtuple;
1165 2158 : header = tuple->t_data;
1166 :
1167 : /* not a disk based tuple */
1168 2158 : ItemPointerSetInvalid(&tuple->t_self);
1169 :
1170 : /*
1171 : * We can only figure this out after reassembling the transactions.
1172 : */
1173 2158 : tuple->t_tableOid = InvalidOid;
1174 :
1175 2158 : tuple->t_len = datalen + SizeofHeapTupleHeader;
1176 :
1177 2158 : memset(header, 0, SizeofHeapTupleHeader);
1178 :
1179 2158 : memcpy((char *) tuple->t_data + SizeofHeapTupleHeader, data, datalen);
1180 2158 : header->t_infomask = xlhdr->t_infomask;
1181 2158 : header->t_infomask2 = xlhdr->t_infomask2;
1182 2158 : header->t_hoff = xlhdr->t_hoff;
1183 :
1184 : /*
1185 : * Reset toast reassembly state only after the last row in the last
1186 : * xl_multi_insert_tuple record emitted by one heap_multi_insert()
1187 : * call.
1188 : */
1189 2158 : if (xlrec->flags & XLH_INSERT_LAST_IN_MULTI &&
1190 438 : (i + 1) == xlrec->ntuples)
1191 34 : change->data.tp.clear_toast_afterwards = true;
1192 : else
1193 2124 : change->data.tp.clear_toast_afterwards = false;
1194 :
1195 2158 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r),
1196 : buf->origptr, change, false);
1197 :
1198 : /* move to the next xl_multi_insert_tuple entry */
1199 2158 : data += datalen;
1200 : }
1201 : Assert(data == tupledata + tuplelen);
1202 : }
1203 :
1204 : /*
1205 : * Parse XLOG_HEAP_CONFIRM from wal into a confirmation change.
1206 : *
1207 : * This is pretty trivial, all the state essentially already setup by the
1208 : * speculative insertion.
1209 : */
1210 : static void
1211 35832 : DecodeSpecConfirm(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
1212 : {
1213 35832 : XLogReaderState *r = buf->record;
1214 : ReorderBufferChange *change;
1215 : RelFileLocator target_locator;
1216 :
1217 : /* only interested in our database */
1218 35832 : XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
1219 35832 : if (target_locator.dbOid != ctx->slot->data.database)
1220 0 : return;
1221 :
1222 : /* output plugin doesn't look for this origin, no need to queue */
1223 35832 : if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
1224 0 : return;
1225 :
1226 35832 : change = ReorderBufferAllocChange(ctx->reorder);
1227 35832 : change->action = REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM;
1228 35832 : change->origin_id = XLogRecGetOrigin(r);
1229 :
1230 35832 : memcpy(&change->data.tp.rlocator, &target_locator, sizeof(RelFileLocator));
1231 :
1232 35832 : change->data.tp.clear_toast_afterwards = true;
1233 :
1234 35832 : ReorderBufferQueueChange(ctx->reorder, XLogRecGetXid(r), buf->origptr,
1235 : change, false);
1236 : }
1237 :
1238 :
1239 : /*
1240 : * Read a HeapTuple as WAL logged by heap_insert, heap_update and heap_delete
1241 : * (but not by heap_multi_insert) into a tuplebuf.
1242 : *
1243 : * The size 'len' and the pointer 'data' in the record need to be
1244 : * computed outside as they are record specific.
1245 : */
1246 : static void
1247 2605598 : DecodeXLogTuple(char *data, Size len, HeapTuple tuple)
1248 : {
1249 : xl_heap_header xlhdr;
1250 2605598 : int datalen = len - SizeOfHeapHeader;
1251 : HeapTupleHeader header;
1252 :
1253 : Assert(datalen >= 0);
1254 :
1255 2605598 : tuple->t_len = datalen + SizeofHeapTupleHeader;
1256 2605598 : header = tuple->t_data;
1257 :
1258 : /* not a disk based tuple */
1259 2605598 : ItemPointerSetInvalid(&tuple->t_self);
1260 :
1261 : /* we can only figure this out after reassembling the transactions */
1262 2605598 : tuple->t_tableOid = InvalidOid;
1263 :
1264 : /* data is not stored aligned, copy to aligned storage */
1265 2605598 : memcpy(&xlhdr, data, SizeOfHeapHeader);
1266 :
1267 2605598 : memset(header, 0, SizeofHeapTupleHeader);
1268 :
1269 2605598 : memcpy(((char *) tuple->t_data) + SizeofHeapTupleHeader,
1270 2605598 : data + SizeOfHeapHeader,
1271 : datalen);
1272 :
1273 2605598 : header->t_infomask = xlhdr.t_infomask;
1274 2605598 : header->t_infomask2 = xlhdr.t_infomask2;
1275 2605598 : header->t_hoff = xlhdr.t_hoff;
1276 2605598 : }
1277 :
1278 : /*
1279 : * Check whether we are interested in this specific transaction.
1280 : *
1281 : * There can be several reasons we might not be interested in this
1282 : * transaction:
1283 : * 1) We might not be interested in decoding transactions up to this
1284 : * LSN. This can happen because we previously decoded it and now just
1285 : * are restarting or if we haven't assembled a consistent snapshot yet.
1286 : * 2) The transaction happened in another database.
1287 : * 3) The output plugin is not interested in the origin.
1288 : * 4) We are doing fast-forwarding
1289 : */
1290 : static bool
1291 7212 : DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
1292 : Oid txn_dbid, ReplOriginId origin_id)
1293 : {
1294 7212 : if (SnapBuildXactNeedsSkip(ctx->snapshot_builder, buf->origptr) ||
1295 6436 : (txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
1296 3254 : FilterByOrigin(ctx, origin_id))
1297 4040 : return true;
1298 :
1299 : /*
1300 : * We also skip decoding in fast_forward mode. In passing set the
1301 : * processing_required flag to indicate that if it were not for
1302 : * fast_forward mode, processing would have been required.
1303 : */
1304 3172 : if (ctx->fast_forward)
1305 : {
1306 48 : ctx->processing_required = true;
1307 48 : return true;
1308 : }
1309 :
1310 3124 : return false;
1311 : }
|
