| 1 |
/*------------------------------------------------------------------------- |
| 2 |
* |
| 3 |
* gist.c |
| 4 |
* interface routines for the postgres GiST index access method. |
| 5 |
* |
| 6 |
* |
| 7 |
* Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group |
| 8 |
* Portions Copyright (c) 1994, Regents of the University of California |
| 9 |
* |
| 10 |
* IDENTIFICATION |
| 11 |
* src/backend/access/gist/gist.c |
| 12 |
* |
| 13 |
*------------------------------------------------------------------------- |
| 14 |
*/ |
| 15 |
#include "postgres.h" |
| 16 |
|
| 17 |
#include "access/gist_private.h" |
| 18 |
#include "access/gistscan.h" |
| 19 |
#include "access/xloginsert.h" |
| 20 |
#include "catalog/pg_collation.h" |
| 21 |
#include "commands/vacuum.h" |
| 22 |
#include "miscadmin.h" |
| 23 |
#include "nodes/execnodes.h" |
| 24 |
#include "storage/predicate.h" |
| 25 |
#include "utils/fmgrprotos.h" |
| 26 |
#include "utils/index_selfuncs.h" |
| 27 |
#include "utils/memutils.h" |
| 28 |
#include "utils/rel.h" |
| 29 |
|
| 30 |
/* non-export function prototypes */ |
| 31 |
static void gistfixsplit(GISTInsertState *state, GISTSTATE *giststate); |
| 32 |
static bool gistinserttuple(GISTInsertState *state, GISTInsertStack *stack, |
| 33 |
GISTSTATE *giststate, IndexTuple tuple, OffsetNumber oldoffnum); |
| 34 |
static bool gistinserttuples(GISTInsertState *state, GISTInsertStack *stack, |
| 35 |
GISTSTATE *giststate, |
| 36 |
IndexTuple *tuples, int ntup, OffsetNumber oldoffnum, |
| 37 |
Buffer leftchild, Buffer rightchild, |
| 38 |
bool unlockbuf, bool unlockleftchild); |
| 39 |
static void gistfinishsplit(GISTInsertState *state, GISTInsertStack *stack, |
| 40 |
GISTSTATE *giststate, List *splitinfo, bool unlockbuf); |
| 41 |
static void gistprunepage(Relation rel, Page page, Buffer buffer, |
| 42 |
Relation heapRel); |
| 43 |
|
| 44 |
|
| 45 |
#define ROTATEDIST(d) do { \ |
| 46 |
SplitPageLayout *tmp = (SplitPageLayout *) palloc0(sizeof(SplitPageLayout)); \ |
| 47 |
tmp->block.blkno = InvalidBlockNumber; \ |
| 48 |
tmp->buffer = InvalidBuffer; \ |
| 49 |
tmp->next = (d); \ |
| 50 |
(d)=tmp; \ |
| 51 |
} while(0) |
| 52 |
|
| 53 |
|
| 54 |
/* |
| 55 |
* GiST handler function: return IndexAmRoutine with access method parameters |
| 56 |
* and callbacks. |
| 57 |
*/ |
| 58 |
Datum |
| 59 |
gisthandler(PG_FUNCTION_ARGS) |
| 60 |
{ |
| 61 |
IndexAmRoutine *amroutine = makeNode(IndexAmRoutine); |
| 62 |
|
| 63 |
amroutine->amstrategies = 0; |
| 64 |
amroutine->amsupport = GISTNProcs; |
| 65 |
amroutine->amoptsprocnum = GIST_OPTIONS_PROC; |
| 66 |
amroutine->amcanorder = false; |
| 67 |
amroutine->amcanorderbyop = true; |
| 68 |
amroutine->amcanhash = false; |
| 69 |
amroutine->amconsistentequality = false; |
| 70 |
amroutine->amconsistentordering = false; |
| 71 |
amroutine->amcanbackward = false; |
| 72 |
amroutine->amcanunique = false; |
| 73 |
amroutine->amcanmulticol = true; |
| 74 |
amroutine->amoptionalkey = true; |
| 75 |
amroutine->amsearcharray = false; |
| 76 |
amroutine->amsearchnulls = true; |
| 77 |
amroutine->amstorage = true; |
| 78 |
amroutine->amclusterable = true; |
| 79 |
amroutine->ampredlocks = true; |
| 80 |
amroutine->amcanparallel = false; |
| 81 |
amroutine->amcanbuildparallel = false; |
| 82 |
amroutine->amcaninclude = true; |
| 83 |
amroutine->amusemaintenanceworkmem = false; |
| 84 |
amroutine->amsummarizing = false; |
| 85 |
amroutine->amparallelvacuumoptions = |
| 86 |
VACUUM_OPTION_PARALLEL_BULKDEL | VACUUM_OPTION_PARALLEL_COND_CLEANUP; |
| 87 |
amroutine->amkeytype = InvalidOid; |
| 88 |
|
| 89 |
amroutine->ambuild = gistbuild; |
| 90 |
amroutine->ambuildempty = gistbuildempty; |
| 91 |
amroutine->aminsert = gistinsert; |
| 92 |
amroutine->aminsertcleanup = NULL; |
| 93 |
amroutine->ambulkdelete = gistbulkdelete; |
| 94 |
amroutine->amvacuumcleanup = gistvacuumcleanup; |
| 95 |
amroutine->amcanreturn = gistcanreturn; |
| 96 |
amroutine->amcostestimate = gistcostestimate; |
| 97 |
amroutine->amgettreeheight = NULL; |
| 98 |
amroutine->amoptions = gistoptions; |
| 99 |
amroutine->amproperty = gistproperty; |
| 100 |
amroutine->ambuildphasename = NULL; |
| 101 |
amroutine->amvalidate = gistvalidate; |
| 102 |
amroutine->amadjustmembers = gistadjustmembers; |
| 103 |
amroutine->ambeginscan = gistbeginscan; |
| 104 |
amroutine->amrescan = gistrescan; |
| 105 |
amroutine->amgettuple = gistgettuple; |
| 106 |
amroutine->amgetbitmap = gistgetbitmap; |
| 107 |
amroutine->amendscan = gistendscan; |
| 108 |
amroutine->ammarkpos = NULL; |
| 109 |
amroutine->amrestrpos = NULL; |
| 110 |
amroutine->amestimateparallelscan = NULL; |
| 111 |
amroutine->aminitparallelscan = NULL; |
| 112 |
amroutine->amparallelrescan = NULL; |
| 113 |
amroutine->amtranslatestrategy = NULL; |
| 114 |
amroutine->amtranslatecmptype = gisttranslatecmptype; |
| 115 |
|
| 116 |
PG_RETURN_POINTER(amroutine); |
| 117 |
} |
| 118 |
|
| 119 |
/* |
| 120 |
* Create and return a temporary memory context for use by GiST. We |
| 121 |
* _always_ invoke user-provided methods in a temporary memory |
| 122 |
* context, so that memory leaks in those functions cannot cause |
| 123 |
* problems. Also, we use some additional temporary contexts in the |
| 124 |
* GiST code itself, to avoid the need to do some awkward manual |
| 125 |
* memory management. |
| 126 |
*/ |
| 127 |
MemoryContext |
| 128 |
createTempGistContext(void) |
| 129 |
{ |
| 130 |
return AllocSetContextCreate(CurrentMemoryContext, |
| 131 |
"GiST temporary context", |
| 132 |
ALLOCSET_DEFAULT_SIZES); |
| 133 |
} |
| 134 |
|
| 135 |
/* |
| 136 |
* gistbuildempty() -- build an empty gist index in the initialization fork |
| 137 |
*/ |
| 138 |
void |
| 139 |
gistbuildempty(Relation index) |
| 140 |
{ |
| 141 |
Buffer buffer; |
| 142 |
|
| 143 |
/* Initialize the root page */ |
| 144 |
buffer = ExtendBufferedRel(BMR_REL(index), INIT_FORKNUM, NULL, |
| 145 |
EB_SKIP_EXTENSION_LOCK | EB_LOCK_FIRST); |
| 146 |
|
| 147 |
/* Initialize and xlog buffer */ |
| 148 |
START_CRIT_SECTION(); |
| 149 |
GISTInitBuffer(buffer, F_LEAF); |
| 150 |
MarkBufferDirty(buffer); |
| 151 |
log_newpage_buffer(buffer, true); |
| 152 |
END_CRIT_SECTION(); |
| 153 |
|
| 154 |
/* Unlock and release the buffer */ |
| 155 |
UnlockReleaseBuffer(buffer); |
| 156 |
} |
| 157 |
|
| 158 |
/* |
| 159 |
* gistinsert -- wrapper for GiST tuple insertion. |
| 160 |
* |
| 161 |
* This is the public interface routine for tuple insertion in GiSTs. |
| 162 |
* It doesn't do any work; just locks the relation and passes the buck. |
| 163 |
*/ |
| 164 |
bool |
| 165 |
gistinsert(Relation r, Datum *values, bool *isnull, |
| 166 |
ItemPointer ht_ctid, Relation heapRel, |
| 167 |
IndexUniqueCheck checkUnique, |
| 168 |
bool indexUnchanged, |
| 169 |
IndexInfo *indexInfo) |
| 170 |
{ |
| 171 |
GISTSTATE *giststate = (GISTSTATE *) indexInfo->ii_AmCache; |
| 172 |
IndexTuple itup; |
| 173 |
MemoryContext oldCxt; |
| 174 |
|
| 175 |
/* Initialize GISTSTATE cache if first call in this statement */ |
| 176 |
if (giststate == NULL) |
| 177 |
{ |
| 178 |
oldCxt = MemoryContextSwitchTo(indexInfo->ii_Context); |
| 179 |
giststate = initGISTstate(r); |
| 180 |
giststate->tempCxt = createTempGistContext(); |
| 181 |
indexInfo->ii_AmCache = giststate; |
| 182 |
MemoryContextSwitchTo(oldCxt); |
| 183 |
} |
| 184 |
|
| 185 |
oldCxt = MemoryContextSwitchTo(giststate->tempCxt); |
| 186 |
|
| 187 |
itup = gistFormTuple(giststate, r, values, isnull, true); |
| 188 |
itup->t_tid = *ht_ctid; |
| 189 |
|
| 190 |
gistdoinsert(r, itup, 0, giststate, heapRel, false); |
| 191 |
|
| 192 |
/* cleanup */ |
| 193 |
MemoryContextSwitchTo(oldCxt); |
| 194 |
MemoryContextReset(giststate->tempCxt); |
| 195 |
|
| 196 |
return false; |
| 197 |
} |
| 198 |
|
| 199 |
|
| 200 |
/* |
| 201 |
* Place tuples from 'itup' to 'buffer'. If 'oldoffnum' is valid, the tuple |
| 202 |
* at that offset is atomically removed along with inserting the new tuples. |
| 203 |
* This is used to replace a tuple with a new one. |
| 204 |
* |
| 205 |
* If 'leftchildbuf' is valid, we're inserting the downlink for the page |
| 206 |
* to the right of 'leftchildbuf', or updating the downlink for 'leftchildbuf'. |
| 207 |
* F_FOLLOW_RIGHT flag on 'leftchildbuf' is cleared and NSN is set. |
| 208 |
* |
| 209 |
* If 'markfollowright' is true and the page is split, the left child is |
| 210 |
* marked with F_FOLLOW_RIGHT flag. That is the normal case. During buffered |
| 211 |
* index build, however, there is no concurrent access and the page splitting |
| 212 |
* is done in a slightly simpler fashion, and false is passed. |
| 213 |
* |
| 214 |
* If there is not enough room on the page, it is split. All the split |
| 215 |
* pages are kept pinned and locked and returned in *splitinfo, the caller |
| 216 |
* is responsible for inserting the downlinks for them. However, if |
| 217 |
* 'buffer' is the root page and it needs to be split, gistplacetopage() |
| 218 |
* performs the split as one atomic operation, and *splitinfo is set to NIL. |
| 219 |
* In that case, we continue to hold the root page locked, and the child |
| 220 |
* pages are released; note that new tuple(s) are *not* on the root page |
| 221 |
* but in one of the new child pages. |
| 222 |
* |
| 223 |
* If 'newblkno' is not NULL, returns the block number of page the first |
| 224 |
* new/updated tuple was inserted to. Usually it's the given page, but could |
| 225 |
* be its right sibling if the page was split. |
| 226 |
* |
| 227 |
* Returns 'true' if the page was split, 'false' otherwise. |
| 228 |
*/ |
| 229 |
bool |
| 230 |
gistplacetopage(Relation rel, Size freespace, GISTSTATE *giststate, |
| 231 |
Buffer buffer, |
| 232 |
IndexTuple *itup, int ntup, OffsetNumber oldoffnum, |
| 233 |
BlockNumber *newblkno, |
| 234 |
Buffer leftchildbuf, |
| 235 |
List **splitinfo, |
| 236 |
bool markfollowright, |
| 237 |
Relation heapRel, |
| 238 |
bool is_build) |
| 239 |
{ |
| 240 |
BlockNumber blkno = BufferGetBlockNumber(buffer); |
| 241 |
Page page = BufferGetPage(buffer); |
| 242 |
bool is_leaf = (GistPageIsLeaf(page)) ? true : false; |
| 243 |
XLogRecPtr recptr; |
| 244 |
bool is_split; |
| 245 |
|
| 246 |
/* |
| 247 |
* Refuse to modify a page that's incompletely split. This should not |
| 248 |
* happen because we finish any incomplete splits while we walk down the |
| 249 |
* tree. However, it's remotely possible that another concurrent inserter |
| 250 |
* splits a parent page, and errors out before completing the split. We |
| 251 |
* will just throw an error in that case, and leave any split we had in |
| 252 |
* progress unfinished too. The next insert that comes along will clean up |
| 253 |
* the mess. |
| 254 |
*/ |
| 255 |
if (GistFollowRight(page)) |
| 256 |
elog(ERROR, "concurrent GiST page split was incomplete"); |
| 257 |
|
| 258 |
/* should never try to insert to a deleted page */ |
| 259 |
Assert(!GistPageIsDeleted(page)); |
| 260 |
|
| 261 |
*splitinfo = NIL; |
| 262 |
|
| 263 |
/* |
| 264 |
* if isupdate, remove old key: This node's key has been modified, either |
| 265 |
* because a child split occurred or because we needed to adjust our key |
| 266 |
* for an insert in a child node. Therefore, remove the old version of |
| 267 |
* this node's key. |
| 268 |
* |
| 269 |
* for WAL replay, in the non-split case we handle this by setting up a |
| 270 |
* one-element todelete array; in the split case, it's handled implicitly |
| 271 |
* because the tuple vector passed to gistSplit won't include this tuple. |
| 272 |
*/ |
| 273 |
is_split = gistnospace(page, itup, ntup, oldoffnum, freespace); |
| 274 |
|
| 275 |
/* |
| 276 |
* If leaf page is full, try at first to delete dead tuples. And then |
| 277 |
* check again. |
| 278 |
*/ |
| 279 |
if (is_split && GistPageIsLeaf(page) && GistPageHasGarbage(page)) |
| 280 |
{ |
| 281 |
gistprunepage(rel, page, buffer, heapRel); |
| 282 |
is_split = gistnospace(page, itup, ntup, oldoffnum, freespace); |
| 283 |
} |
| 284 |
|
| 285 |
if (is_split) |
| 286 |
{ |
| 287 |
/* no space for insertion */ |
| 288 |
IndexTuple *itvec; |
| 289 |
int tlen; |
| 290 |
SplitPageLayout *dist = NULL, |
| 291 |
*ptr; |
| 292 |
BlockNumber oldrlink = InvalidBlockNumber; |
| 293 |
GistNSN oldnsn = 0; |
| 294 |
SplitPageLayout rootpg; |
| 295 |
bool is_rootsplit; |
| 296 |
int npage; |
| 297 |
|
| 298 |
is_rootsplit = (blkno == GIST_ROOT_BLKNO); |
| 299 |
|
| 300 |
/* |
| 301 |
* Form index tuples vector to split. If we're replacing an old tuple, |
| 302 |
* remove the old version from the vector. |
| 303 |
*/ |
| 304 |
itvec = gistextractpage(page, &tlen); |
| 305 |
if (OffsetNumberIsValid(oldoffnum)) |
| 306 |
{ |
| 307 |
/* on inner page we should remove old tuple */ |
| 308 |
int pos = oldoffnum - FirstOffsetNumber; |
| 309 |
|
| 310 |
tlen--; |
| 311 |
if (pos != tlen) |
| 312 |
memmove(itvec + pos, itvec + pos + 1, sizeof(IndexTuple) * (tlen - pos)); |
| 313 |
} |
| 314 |
itvec = gistjoinvector(itvec, &tlen, itup, ntup); |
| 315 |
dist = gistSplit(rel, page, itvec, tlen, giststate); |
| 316 |
|
| 317 |
/* |
| 318 |
* Check that split didn't produce too many pages. |
| 319 |
*/ |
| 320 |
npage = 0; |
| 321 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 322 |
npage++; |
| 323 |
/* in a root split, we'll add one more page to the list below */ |
| 324 |
if (is_rootsplit) |
| 325 |
npage++; |
| 326 |
if (npage > GIST_MAX_SPLIT_PAGES) |
| 327 |
elog(ERROR, "GiST page split into too many halves (%d, maximum %d)", |
| 328 |
npage, GIST_MAX_SPLIT_PAGES); |
| 329 |
|
| 330 |
/* |
| 331 |
* Set up pages to work with. Allocate new buffers for all but the |
| 332 |
* leftmost page. The original page becomes the new leftmost page, and |
| 333 |
* is just replaced with the new contents. |
| 334 |
* |
| 335 |
* For a root-split, allocate new buffers for all child pages, the |
| 336 |
* original page is overwritten with new root page containing |
| 337 |
* downlinks to the new child pages. |
| 338 |
*/ |
| 339 |
ptr = dist; |
| 340 |
if (!is_rootsplit) |
| 341 |
{ |
| 342 |
/* save old rightlink and NSN */ |
| 343 |
oldrlink = GistPageGetOpaque(page)->rightlink; |
| 344 |
oldnsn = GistPageGetNSN(page); |
| 345 |
|
| 346 |
dist->buffer = buffer; |
| 347 |
dist->block.blkno = BufferGetBlockNumber(buffer); |
| 348 |
dist->page = PageGetTempPageCopySpecial(BufferGetPage(buffer)); |
| 349 |
|
| 350 |
/* clean all flags except F_LEAF */ |
| 351 |
GistPageGetOpaque(dist->page)->flags = (is_leaf) ? F_LEAF : 0; |
| 352 |
|
| 353 |
ptr = ptr->next; |
| 354 |
} |
| 355 |
for (; ptr; ptr = ptr->next) |
| 356 |
{ |
| 357 |
/* Allocate new page */ |
| 358 |
ptr->buffer = gistNewBuffer(rel, heapRel); |
| 359 |
GISTInitBuffer(ptr->buffer, (is_leaf) ? F_LEAF : 0); |
| 360 |
ptr->page = BufferGetPage(ptr->buffer); |
| 361 |
ptr->block.blkno = BufferGetBlockNumber(ptr->buffer); |
| 362 |
PredicateLockPageSplit(rel, |
| 363 |
BufferGetBlockNumber(buffer), |
| 364 |
BufferGetBlockNumber(ptr->buffer)); |
| 365 |
} |
| 366 |
|
| 367 |
/* |
| 368 |
* Now that we know which blocks the new pages go to, set up downlink |
| 369 |
* tuples to point to them. |
| 370 |
*/ |
| 371 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 372 |
{ |
| 373 |
ItemPointerSetBlockNumber(&(ptr->itup->t_tid), ptr->block.blkno); |
| 374 |
GistTupleSetValid(ptr->itup); |
| 375 |
} |
| 376 |
|
| 377 |
/* |
| 378 |
* If this is a root split, we construct the new root page with the |
| 379 |
* downlinks here directly, instead of requiring the caller to insert |
| 380 |
* them. Add the new root page to the list along with the child pages. |
| 381 |
*/ |
| 382 |
if (is_rootsplit) |
| 383 |
{ |
| 384 |
IndexTuple *downlinks; |
| 385 |
int ndownlinks = 0; |
| 386 |
int i; |
| 387 |
|
| 388 |
rootpg.buffer = buffer; |
| 389 |
rootpg.page = PageGetTempPageCopySpecial(BufferGetPage(rootpg.buffer)); |
| 390 |
GistPageGetOpaque(rootpg.page)->flags = 0; |
| 391 |
|
| 392 |
/* Prepare a vector of all the downlinks */ |
| 393 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 394 |
ndownlinks++; |
| 395 |
downlinks = palloc(sizeof(IndexTuple) * ndownlinks); |
| 396 |
for (i = 0, ptr = dist; ptr; ptr = ptr->next) |
| 397 |
downlinks[i++] = ptr->itup; |
| 398 |
|
| 399 |
rootpg.block.blkno = GIST_ROOT_BLKNO; |
| 400 |
rootpg.block.num = ndownlinks; |
| 401 |
rootpg.list = gistfillitupvec(downlinks, ndownlinks, |
| 402 |
&(rootpg.lenlist)); |
| 403 |
rootpg.itup = NULL; |
| 404 |
|
| 405 |
rootpg.next = dist; |
| 406 |
dist = &rootpg; |
| 407 |
} |
| 408 |
else |
| 409 |
{ |
| 410 |
/* Prepare split-info to be returned to caller */ |
| 411 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 412 |
{ |
| 413 |
GISTPageSplitInfo *si = palloc(sizeof(GISTPageSplitInfo)); |
| 414 |
|
| 415 |
si->buf = ptr->buffer; |
| 416 |
si->downlink = ptr->itup; |
| 417 |
*splitinfo = lappend(*splitinfo, si); |
| 418 |
} |
| 419 |
} |
| 420 |
|
| 421 |
/* |
| 422 |
* Fill all pages. All the pages are new, ie. freshly allocated empty |
| 423 |
* pages, or a temporary copy of the old page. |
| 424 |
*/ |
| 425 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 426 |
{ |
| 427 |
char *data = (char *) (ptr->list); |
| 428 |
|
| 429 |
for (int i = 0; i < ptr->block.num; i++) |
| 430 |
{ |
| 431 |
IndexTuple thistup = (IndexTuple) data; |
| 432 |
|
| 433 |
if (PageAddItem(ptr->page, (Item) data, IndexTupleSize(thistup), i + FirstOffsetNumber, false, false) == InvalidOffsetNumber) |
| 434 |
elog(ERROR, "failed to add item to index page in \"%s\"", RelationGetRelationName(rel)); |
| 435 |
|
| 436 |
/* |
| 437 |
* If this is the first inserted/updated tuple, let the caller |
| 438 |
* know which page it landed on. |
| 439 |
*/ |
| 440 |
if (newblkno && ItemPointerEquals(&thistup->t_tid, &(*itup)->t_tid)) |
| 441 |
*newblkno = ptr->block.blkno; |
| 442 |
|
| 443 |
data += IndexTupleSize(thistup); |
| 444 |
} |
| 445 |
|
| 446 |
/* Set up rightlinks */ |
| 447 |
if (ptr->next && ptr->block.blkno != GIST_ROOT_BLKNO) |
| 448 |
GistPageGetOpaque(ptr->page)->rightlink = |
| 449 |
ptr->next->block.blkno; |
| 450 |
else |
| 451 |
GistPageGetOpaque(ptr->page)->rightlink = oldrlink; |
| 452 |
|
| 453 |
/* |
| 454 |
* Mark the all but the right-most page with the follow-right |
| 455 |
* flag. It will be cleared as soon as the downlink is inserted |
| 456 |
* into the parent, but this ensures that if we error out before |
| 457 |
* that, the index is still consistent. (in buffering build mode, |
| 458 |
* any error will abort the index build anyway, so this is not |
| 459 |
* needed.) |
| 460 |
*/ |
| 461 |
if (ptr->next && !is_rootsplit && markfollowright) |
| 462 |
GistMarkFollowRight(ptr->page); |
| 463 |
else |
| 464 |
GistClearFollowRight(ptr->page); |
| 465 |
|
| 466 |
/* |
| 467 |
* Copy the NSN of the original page to all pages. The |
| 468 |
* F_FOLLOW_RIGHT flags ensure that scans will follow the |
| 469 |
* rightlinks until the downlinks are inserted. |
| 470 |
*/ |
| 471 |
GistPageSetNSN(ptr->page, oldnsn); |
| 472 |
} |
| 473 |
|
| 474 |
/* |
| 475 |
* gistXLogSplit() needs to WAL log a lot of pages, prepare WAL |
| 476 |
* insertion for that. NB: The number of pages and data segments |
| 477 |
* specified here must match the calculations in gistXLogSplit()! |
| 478 |
*/ |
| 479 |
if (!is_build && RelationNeedsWAL(rel)) |
| 480 |
XLogEnsureRecordSpace(npage, 1 + npage * 2); |
| 481 |
|
| 482 |
START_CRIT_SECTION(); |
| 483 |
|
| 484 |
/* |
| 485 |
* Must mark buffers dirty before XLogInsert, even though we'll still |
| 486 |
* be changing their opaque fields below. |
| 487 |
*/ |
| 488 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 489 |
MarkBufferDirty(ptr->buffer); |
| 490 |
if (BufferIsValid(leftchildbuf)) |
| 491 |
MarkBufferDirty(leftchildbuf); |
| 492 |
|
| 493 |
/* |
| 494 |
* The first page in the chain was a temporary working copy meant to |
| 495 |
* replace the old page. Copy it over the old page. |
| 496 |
*/ |
| 497 |
PageRestoreTempPage(dist->page, BufferGetPage(dist->buffer)); |
| 498 |
dist->page = BufferGetPage(dist->buffer); |
| 499 |
|
| 500 |
/* |
| 501 |
* Write the WAL record. |
| 502 |
* |
| 503 |
* If we're building a new index, however, we don't WAL-log changes |
| 504 |
* yet. The LSN-NSN interlock between parent and child requires that |
| 505 |
* LSNs never move backwards, so set the LSNs to a value that's |
| 506 |
* smaller than any real or fake unlogged LSN that might be generated |
| 507 |
* later. (There can't be any concurrent scans during index build, so |
| 508 |
* we don't need to be able to detect concurrent splits yet.) |
| 509 |
*/ |
| 510 |
if (is_build) |
| 511 |
recptr = GistBuildLSN; |
| 512 |
else |
| 513 |
{ |
| 514 |
if (RelationNeedsWAL(rel)) |
| 515 |
recptr = gistXLogSplit(is_leaf, |
| 516 |
dist, oldrlink, oldnsn, leftchildbuf, |
| 517 |
markfollowright); |
| 518 |
else |
| 519 |
recptr = gistGetFakeLSN(rel); |
| 520 |
} |
| 521 |
|
| 522 |
for (ptr = dist; ptr; ptr = ptr->next) |
| 523 |
PageSetLSN(ptr->page, recptr); |
| 524 |
|
| 525 |
/* |
| 526 |
* Return the new child buffers to the caller. |
| 527 |
* |
| 528 |
* If this was a root split, we've already inserted the downlink |
| 529 |
* pointers, in the form of a new root page. Therefore we can release |
| 530 |
* all the new buffers, and keep just the root page locked. |
| 531 |
*/ |
| 532 |
if (is_rootsplit) |
| 533 |
{ |
| 534 |
for (ptr = dist->next; ptr; ptr = ptr->next) |
| 535 |
UnlockReleaseBuffer(ptr->buffer); |
| 536 |
} |
| 537 |
} |
| 538 |
else |
| 539 |
{ |
| 540 |
/* |
| 541 |
* Enough space. We always get here if ntup==0. |
| 542 |
*/ |
| 543 |
START_CRIT_SECTION(); |
| 544 |
|
| 545 |
/* |
| 546 |
* Delete old tuple if any, then insert new tuple(s) if any. If |
| 547 |
* possible, use the fast path of PageIndexTupleOverwrite. |
| 548 |
*/ |
| 549 |
if (OffsetNumberIsValid(oldoffnum)) |
| 550 |
{ |
| 551 |
if (ntup == 1) |
| 552 |
{ |
| 553 |
/* One-for-one replacement, so use PageIndexTupleOverwrite */ |
| 554 |
if (!PageIndexTupleOverwrite(page, oldoffnum, (Item) *itup, |
| 555 |
IndexTupleSize(*itup))) |
| 556 |
elog(ERROR, "failed to add item to index page in \"%s\"", |
| 557 |
RelationGetRelationName(rel)); |
| 558 |
} |
| 559 |
else |
| 560 |
{ |
| 561 |
/* Delete old, then append new tuple(s) to page */ |
| 562 |
PageIndexTupleDelete(page, oldoffnum); |
| 563 |
gistfillbuffer(page, itup, ntup, InvalidOffsetNumber); |
| 564 |
} |
| 565 |
} |
| 566 |
else |
| 567 |
{ |
| 568 |
/* Just append new tuples at the end of the page */ |
| 569 |
gistfillbuffer(page, itup, ntup, InvalidOffsetNumber); |
| 570 |
} |
| 571 |
|
| 572 |
MarkBufferDirty(buffer); |
| 573 |
|
| 574 |
if (BufferIsValid(leftchildbuf)) |
| 575 |
MarkBufferDirty(leftchildbuf); |
| 576 |
|
| 577 |
if (is_build) |
| 578 |
recptr = GistBuildLSN; |
| 579 |
else |
| 580 |
{ |
| 581 |
if (RelationNeedsWAL(rel)) |
| 582 |
{ |
| 583 |
OffsetNumber ndeloffs = 0, |
| 584 |
deloffs[1]; |
| 585 |
|
| 586 |
if (OffsetNumberIsValid(oldoffnum)) |
| 587 |
{ |
| 588 |
deloffs[0] = oldoffnum; |
| 589 |
ndeloffs = 1; |
| 590 |
} |
| 591 |
|
| 592 |
recptr = gistXLogUpdate(buffer, |
| 593 |
deloffs, ndeloffs, itup, ntup, |
| 594 |
leftchildbuf); |
| 595 |
} |
| 596 |
else |
| 597 |
recptr = gistGetFakeLSN(rel); |
| 598 |
} |
| 599 |
PageSetLSN(page, recptr); |
| 600 |
|
| 601 |
if (newblkno) |
| 602 |
*newblkno = blkno; |
| 603 |
} |
| 604 |
|
| 605 |
/* |
| 606 |
* If we inserted the downlink for a child page, set NSN and clear |
| 607 |
* F_FOLLOW_RIGHT flag on the left child, so that concurrent scans know to |
| 608 |
* follow the rightlink if and only if they looked at the parent page |
| 609 |
* before we inserted the downlink. |
| 610 |
* |
| 611 |
* Note that we do this *after* writing the WAL record. That means that |
| 612 |
* the possible full page image in the WAL record does not include these |
| 613 |
* changes, and they must be replayed even if the page is restored from |
| 614 |
* the full page image. There's a chicken-and-egg problem: if we updated |
| 615 |
* the child pages first, we wouldn't know the recptr of the WAL record |
| 616 |
* we're about to write. |
| 617 |
*/ |
| 618 |
if (BufferIsValid(leftchildbuf)) |
| 619 |
{ |
| 620 |
Page leftpg = BufferGetPage(leftchildbuf); |
| 621 |
|
| 622 |
GistPageSetNSN(leftpg, recptr); |
| 623 |
GistClearFollowRight(leftpg); |
| 624 |
|
| 625 |
PageSetLSN(leftpg, recptr); |
| 626 |
} |
| 627 |
|
| 628 |
END_CRIT_SECTION(); |
| 629 |
|
| 630 |
return is_split; |
| 631 |
} |
| 632 |
|
| 633 |
/* |
| 634 |
* Workhorse routine for doing insertion into a GiST index. Note that |
| 635 |
* this routine assumes it is invoked in a short-lived memory context, |
| 636 |
* so it does not bother releasing palloc'd allocations. |
| 637 |
*/ |
| 638 |
void |
| 639 |
gistdoinsert(Relation r, IndexTuple itup, Size freespace, |
| 640 |
GISTSTATE *giststate, Relation heapRel, bool is_build) |
| 641 |
{ |
| 642 |
ItemId iid; |
| 643 |
IndexTuple idxtuple; |
| 644 |
GISTInsertStack firststack; |
| 645 |
GISTInsertStack *stack; |
| 646 |
GISTInsertState state; |
| 647 |
bool xlocked = false; |
| 648 |
|
| 649 |
memset(&state, 0, sizeof(GISTInsertState)); |
| 650 |
state.freespace = freespace; |
| 651 |
state.r = r; |
| 652 |
state.heapRel = heapRel; |
| 653 |
state.is_build = is_build; |
| 654 |
|
| 655 |
/* Start from the root */ |
| 656 |
firststack.blkno = GIST_ROOT_BLKNO; |
| 657 |
firststack.lsn = 0; |
| 658 |
firststack.retry_from_parent = false; |
| 659 |
firststack.parent = NULL; |
| 660 |
firststack.downlinkoffnum = InvalidOffsetNumber; |
| 661 |
state.stack = stack = &firststack; |
| 662 |
|
| 663 |
/* |
| 664 |
* Walk down along the path of smallest penalty, updating the parent |
| 665 |
* pointers with the key we're inserting as we go. If we crash in the |
| 666 |
* middle, the tree is consistent, although the possible parent updates |
| 667 |
* were a waste. |
| 668 |
*/ |
| 669 |
for (;;) |
| 670 |
{ |
| 671 |
/* |
| 672 |
* If we split an internal page while descending the tree, we have to |
| 673 |
* retry at the parent. (Normally, the LSN-NSN interlock below would |
| 674 |
* also catch this and cause us to retry. But LSNs are not updated |
| 675 |
* during index build.) |
| 676 |
*/ |
| 677 |
while (stack->retry_from_parent) |
| 678 |
{ |
| 679 |
if (xlocked) |
| 680 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 681 |
xlocked = false; |
| 682 |
ReleaseBuffer(stack->buffer); |
| 683 |
state.stack = stack = stack->parent; |
| 684 |
} |
| 685 |
|
| 686 |
if (XLogRecPtrIsInvalid(stack->lsn)) |
| 687 |
stack->buffer = ReadBuffer(state.r, stack->blkno); |
| 688 |
|
| 689 |
/* |
| 690 |
* Be optimistic and grab shared lock first. Swap it for an exclusive |
| 691 |
* lock later if we need to update the page. |
| 692 |
*/ |
| 693 |
if (!xlocked) |
| 694 |
{ |
| 695 |
LockBuffer(stack->buffer, GIST_SHARE); |
| 696 |
gistcheckpage(state.r, stack->buffer); |
| 697 |
} |
| 698 |
|
| 699 |
stack->page = (Page) BufferGetPage(stack->buffer); |
| 700 |
stack->lsn = xlocked ? |
| 701 |
PageGetLSN(stack->page) : BufferGetLSNAtomic(stack->buffer); |
| 702 |
Assert(!RelationNeedsWAL(state.r) || !XLogRecPtrIsInvalid(stack->lsn)); |
| 703 |
|
| 704 |
/* |
| 705 |
* If this page was split but the downlink was never inserted to the |
| 706 |
* parent because the inserting backend crashed before doing that, fix |
| 707 |
* that now. |
| 708 |
*/ |
| 709 |
if (GistFollowRight(stack->page)) |
| 710 |
{ |
| 711 |
if (!xlocked) |
| 712 |
{ |
| 713 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 714 |
LockBuffer(stack->buffer, GIST_EXCLUSIVE); |
| 715 |
xlocked = true; |
| 716 |
/* someone might've completed the split when we unlocked */ |
| 717 |
if (!GistFollowRight(stack->page)) |
| 718 |
continue; |
| 719 |
} |
| 720 |
gistfixsplit(&state, giststate); |
| 721 |
|
| 722 |
UnlockReleaseBuffer(stack->buffer); |
| 723 |
xlocked = false; |
| 724 |
state.stack = stack = stack->parent; |
| 725 |
continue; |
| 726 |
} |
| 727 |
|
| 728 |
if ((stack->blkno != GIST_ROOT_BLKNO && |
| 729 |
stack->parent->lsn < GistPageGetNSN(stack->page)) || |
| 730 |
GistPageIsDeleted(stack->page)) |
| 731 |
{ |
| 732 |
/* |
| 733 |
* Concurrent split or page deletion detected. There's no |
| 734 |
* guarantee that the downlink for this page is consistent with |
| 735 |
* the tuple we're inserting anymore, so go back to parent and |
| 736 |
* rechoose the best child. |
| 737 |
*/ |
| 738 |
UnlockReleaseBuffer(stack->buffer); |
| 739 |
xlocked = false; |
| 740 |
state.stack = stack = stack->parent; |
| 741 |
continue; |
| 742 |
} |
| 743 |
|
| 744 |
if (!GistPageIsLeaf(stack->page)) |
| 745 |
{ |
| 746 |
/* |
| 747 |
* This is an internal page so continue to walk down the tree. |
| 748 |
* Find the child node that has the minimum insertion penalty. |
| 749 |
*/ |
| 750 |
BlockNumber childblkno; |
| 751 |
IndexTuple newtup; |
| 752 |
GISTInsertStack *item; |
| 753 |
OffsetNumber downlinkoffnum; |
| 754 |
|
| 755 |
downlinkoffnum = gistchoose(state.r, stack->page, itup, giststate); |
| 756 |
iid = PageGetItemId(stack->page, downlinkoffnum); |
| 757 |
idxtuple = (IndexTuple) PageGetItem(stack->page, iid); |
| 758 |
childblkno = ItemPointerGetBlockNumber(&(idxtuple->t_tid)); |
| 759 |
|
| 760 |
/* |
| 761 |
* Check that it's not a leftover invalid tuple from pre-9.1 |
| 762 |
*/ |
| 763 |
if (GistTupleIsInvalid(idxtuple)) |
| 764 |
ereport(ERROR, |
| 765 |
(errmsg("index \"%s\" contains an inner tuple marked as invalid", |
| 766 |
RelationGetRelationName(r)), |
| 767 |
errdetail("This is caused by an incomplete page split at crash recovery before upgrading to PostgreSQL 9.1."), |
| 768 |
errhint("Please REINDEX it."))); |
| 769 |
|
| 770 |
/* |
| 771 |
* Check that the key representing the target child node is |
| 772 |
* consistent with the key we're inserting. Update it if it's not. |
| 773 |
*/ |
| 774 |
newtup = gistgetadjusted(state.r, idxtuple, itup, giststate); |
| 775 |
if (newtup) |
| 776 |
{ |
| 777 |
/* |
| 778 |
* Swap shared lock for an exclusive one. Beware, the page may |
| 779 |
* change while we unlock/lock the page... |
| 780 |
*/ |
| 781 |
if (!xlocked) |
| 782 |
{ |
| 783 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 784 |
LockBuffer(stack->buffer, GIST_EXCLUSIVE); |
| 785 |
xlocked = true; |
| 786 |
stack->page = (Page) BufferGetPage(stack->buffer); |
| 787 |
|
| 788 |
if (PageGetLSN(stack->page) != stack->lsn) |
| 789 |
{ |
| 790 |
/* the page was changed while we unlocked it, retry */ |
| 791 |
continue; |
| 792 |
} |
| 793 |
} |
| 794 |
|
| 795 |
/* |
| 796 |
* Update the tuple. |
| 797 |
* |
| 798 |
* We still hold the lock after gistinserttuple(), but it |
| 799 |
* might have to split the page to make the updated tuple fit. |
| 800 |
* In that case the updated tuple might migrate to the other |
| 801 |
* half of the split, so we have to go back to the parent and |
| 802 |
* descend back to the half that's a better fit for the new |
| 803 |
* tuple. |
| 804 |
*/ |
| 805 |
if (gistinserttuple(&state, stack, giststate, newtup, |
| 806 |
downlinkoffnum)) |
| 807 |
{ |
| 808 |
/* |
| 809 |
* If this was a root split, the root page continues to be |
| 810 |
* the parent and the updated tuple went to one of the |
| 811 |
* child pages, so we just need to retry from the root |
| 812 |
* page. |
| 813 |
*/ |
| 814 |
if (stack->blkno != GIST_ROOT_BLKNO) |
| 815 |
{ |
| 816 |
UnlockReleaseBuffer(stack->buffer); |
| 817 |
xlocked = false; |
| 818 |
state.stack = stack = stack->parent; |
| 819 |
} |
| 820 |
continue; |
| 821 |
} |
| 822 |
} |
| 823 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 824 |
xlocked = false; |
| 825 |
|
| 826 |
/* descend to the chosen child */ |
| 827 |
item = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack)); |
| 828 |
item->blkno = childblkno; |
| 829 |
item->parent = stack; |
| 830 |
item->downlinkoffnum = downlinkoffnum; |
| 831 |
state.stack = stack = item; |
| 832 |
} |
| 833 |
else |
| 834 |
{ |
| 835 |
/* |
| 836 |
* Leaf page. Insert the new key. We've already updated all the |
| 837 |
* parents on the way down, but we might have to split the page if |
| 838 |
* it doesn't fit. gistinserttuple() will take care of that. |
| 839 |
*/ |
| 840 |
|
| 841 |
/* |
| 842 |
* Swap shared lock for an exclusive one. Be careful, the page may |
| 843 |
* change while we unlock/lock the page... |
| 844 |
*/ |
| 845 |
if (!xlocked) |
| 846 |
{ |
| 847 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 848 |
LockBuffer(stack->buffer, GIST_EXCLUSIVE); |
| 849 |
xlocked = true; |
| 850 |
stack->page = (Page) BufferGetPage(stack->buffer); |
| 851 |
stack->lsn = PageGetLSN(stack->page); |
| 852 |
|
| 853 |
if (stack->blkno == GIST_ROOT_BLKNO) |
| 854 |
{ |
| 855 |
/* |
| 856 |
* the only page that can become inner instead of leaf is |
| 857 |
* the root page, so for root we should recheck it |
| 858 |
*/ |
| 859 |
if (!GistPageIsLeaf(stack->page)) |
| 860 |
{ |
| 861 |
/* |
| 862 |
* very rare situation: during unlock/lock index with |
| 863 |
* number of pages = 1 was increased |
| 864 |
*/ |
| 865 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 866 |
xlocked = false; |
| 867 |
continue; |
| 868 |
} |
| 869 |
|
| 870 |
/* |
| 871 |
* we don't need to check root split, because checking |
| 872 |
* leaf/inner is enough to recognize split for root |
| 873 |
*/ |
| 874 |
} |
| 875 |
else if ((GistFollowRight(stack->page) || |
| 876 |
stack->parent->lsn < GistPageGetNSN(stack->page)) || |
| 877 |
GistPageIsDeleted(stack->page)) |
| 878 |
{ |
| 879 |
/* |
| 880 |
* The page was split or deleted while we momentarily |
| 881 |
* unlocked the page. Go back to parent. |
| 882 |
*/ |
| 883 |
UnlockReleaseBuffer(stack->buffer); |
| 884 |
xlocked = false; |
| 885 |
state.stack = stack = stack->parent; |
| 886 |
continue; |
| 887 |
} |
| 888 |
} |
| 889 |
|
| 890 |
/* now state.stack->(page, buffer and blkno) points to leaf page */ |
| 891 |
|
| 892 |
gistinserttuple(&state, stack, giststate, itup, |
| 893 |
InvalidOffsetNumber); |
| 894 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 895 |
|
| 896 |
/* Release any pins we might still hold before exiting */ |
| 897 |
for (; stack; stack = stack->parent) |
| 898 |
ReleaseBuffer(stack->buffer); |
| 899 |
break; |
| 900 |
} |
| 901 |
} |
| 902 |
} |
| 903 |
|
| 904 |
/* |
| 905 |
* Traverse the tree to find path from root page to specified "child" block. |
| 906 |
* |
| 907 |
* returns a new insertion stack, starting from the parent of "child", up |
| 908 |
* to the root. *downlinkoffnum is set to the offset of the downlink in the |
| 909 |
* direct parent of child. |
| 910 |
* |
| 911 |
* To prevent deadlocks, this should lock only one page at a time. |
| 912 |
*/ |
| 913 |
static GISTInsertStack * |
| 914 |
gistFindPath(Relation r, BlockNumber child, OffsetNumber *downlinkoffnum) |
| 915 |
{ |
| 916 |
Page page; |
| 917 |
Buffer buffer; |
| 918 |
OffsetNumber i, |
| 919 |
maxoff; |
| 920 |
ItemId iid; |
| 921 |
IndexTuple idxtuple; |
| 922 |
List *fifo; |
| 923 |
GISTInsertStack *top, |
| 924 |
*ptr; |
| 925 |
BlockNumber blkno; |
| 926 |
|
| 927 |
top = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack)); |
| 928 |
top->blkno = GIST_ROOT_BLKNO; |
| 929 |
top->downlinkoffnum = InvalidOffsetNumber; |
| 930 |
|
| 931 |
fifo = list_make1(top); |
| 932 |
while (fifo != NIL) |
| 933 |
{ |
| 934 |
/* Get next page to visit */ |
| 935 |
top = linitial(fifo); |
| 936 |
fifo = list_delete_first(fifo); |
| 937 |
|
| 938 |
buffer = ReadBuffer(r, top->blkno); |
| 939 |
LockBuffer(buffer, GIST_SHARE); |
| 940 |
gistcheckpage(r, buffer); |
| 941 |
page = (Page) BufferGetPage(buffer); |
| 942 |
|
| 943 |
if (GistPageIsLeaf(page)) |
| 944 |
{ |
| 945 |
/* |
| 946 |
* Because we scan the index top-down, all the rest of the pages |
| 947 |
* in the queue must be leaf pages as well. |
| 948 |
*/ |
| 949 |
UnlockReleaseBuffer(buffer); |
| 950 |
break; |
| 951 |
} |
| 952 |
|
| 953 |
/* currently, internal pages are never deleted */ |
| 954 |
Assert(!GistPageIsDeleted(page)); |
| 955 |
|
| 956 |
top->lsn = BufferGetLSNAtomic(buffer); |
| 957 |
|
| 958 |
/* |
| 959 |
* If F_FOLLOW_RIGHT is set, the page to the right doesn't have a |
| 960 |
* downlink. This should not normally happen.. |
| 961 |
*/ |
| 962 |
if (GistFollowRight(page)) |
| 963 |
elog(ERROR, "concurrent GiST page split was incomplete"); |
| 964 |
|
| 965 |
if (top->parent && top->parent->lsn < GistPageGetNSN(page) && |
| 966 |
GistPageGetOpaque(page)->rightlink != InvalidBlockNumber /* sanity check */ ) |
| 967 |
{ |
| 968 |
/* |
| 969 |
* Page was split while we looked elsewhere. We didn't see the |
| 970 |
* downlink to the right page when we scanned the parent, so add |
| 971 |
* it to the queue now. |
| 972 |
* |
| 973 |
* Put the right page ahead of the queue, so that we visit it |
| 974 |
* next. That's important, because if this is the lowest internal |
| 975 |
* level, just above leaves, we might already have queued up some |
| 976 |
* leaf pages, and we assume that there can't be any non-leaf |
| 977 |
* pages behind leaf pages. |
| 978 |
*/ |
| 979 |
ptr = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack)); |
| 980 |
ptr->blkno = GistPageGetOpaque(page)->rightlink; |
| 981 |
ptr->downlinkoffnum = InvalidOffsetNumber; |
| 982 |
ptr->parent = top->parent; |
| 983 |
|
| 984 |
fifo = lcons(ptr, fifo); |
| 985 |
} |
| 986 |
|
| 987 |
maxoff = PageGetMaxOffsetNumber(page); |
| 988 |
|
| 989 |
for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i)) |
| 990 |
{ |
| 991 |
iid = PageGetItemId(page, i); |
| 992 |
idxtuple = (IndexTuple) PageGetItem(page, iid); |
| 993 |
blkno = ItemPointerGetBlockNumber(&(idxtuple->t_tid)); |
| 994 |
if (blkno == child) |
| 995 |
{ |
| 996 |
/* Found it! */ |
| 997 |
UnlockReleaseBuffer(buffer); |
| 998 |
*downlinkoffnum = i; |
| 999 |
return top; |
| 1000 |
} |
| 1001 |
else |
| 1002 |
{ |
| 1003 |
/* Append this child to the list of pages to visit later */ |
| 1004 |
ptr = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack)); |
| 1005 |
ptr->blkno = blkno; |
| 1006 |
ptr->downlinkoffnum = i; |
| 1007 |
ptr->parent = top; |
| 1008 |
|
| 1009 |
fifo = lappend(fifo, ptr); |
| 1010 |
} |
| 1011 |
} |
| 1012 |
|
| 1013 |
UnlockReleaseBuffer(buffer); |
| 1014 |
} |
| 1015 |
|
| 1016 |
elog(ERROR, "failed to re-find parent of a page in index \"%s\", block %u", |
| 1017 |
RelationGetRelationName(r), child); |
| 1018 |
return NULL; /* keep compiler quiet */ |
| 1019 |
} |
| 1020 |
|
| 1021 |
/* |
| 1022 |
* Updates the stack so that child->parent is the correct parent of the |
| 1023 |
* child. child->parent must be exclusively locked on entry, and will |
| 1024 |
* remain so at exit, but it might not be the same page anymore. |
| 1025 |
*/ |
| 1026 |
static void |
| 1027 |
gistFindCorrectParent(Relation r, GISTInsertStack *child, bool is_build) |
| 1028 |
{ |
| 1029 |
GISTInsertStack *parent = child->parent; |
| 1030 |
ItemId iid; |
| 1031 |
IndexTuple idxtuple; |
| 1032 |
OffsetNumber maxoff; |
| 1033 |
GISTInsertStack *ptr; |
| 1034 |
|
| 1035 |
gistcheckpage(r, parent->buffer); |
| 1036 |
parent->page = (Page) BufferGetPage(parent->buffer); |
| 1037 |
maxoff = PageGetMaxOffsetNumber(parent->page); |
| 1038 |
|
| 1039 |
/* Check if the downlink is still where it was before */ |
| 1040 |
if (child->downlinkoffnum != InvalidOffsetNumber && child->downlinkoffnum <= maxoff) |
| 1041 |
{ |
| 1042 |
iid = PageGetItemId(parent->page, child->downlinkoffnum); |
| 1043 |
idxtuple = (IndexTuple) PageGetItem(parent->page, iid); |
| 1044 |
if (ItemPointerGetBlockNumber(&(idxtuple->t_tid)) == child->blkno) |
| 1045 |
return; /* still there */ |
| 1046 |
} |
| 1047 |
|
| 1048 |
/* |
| 1049 |
* The page has changed since we looked. During normal operation, every |
| 1050 |
* update of a page changes its LSN, so the LSN we memorized should have |
| 1051 |
* changed too. |
| 1052 |
* |
| 1053 |
* During index build, however, we don't WAL-log the changes until we have |
| 1054 |
* built the index, so the LSN doesn't change. There is no concurrent |
| 1055 |
* activity during index build, but we might have changed the parent |
| 1056 |
* ourselves. |
| 1057 |
* |
| 1058 |
* We will also get here if child->downlinkoffnum is invalid. That happens |
| 1059 |
* if 'parent' had been updated by an earlier call to this function on its |
| 1060 |
* grandchild, which had to move right. |
| 1061 |
*/ |
| 1062 |
Assert(parent->lsn != PageGetLSN(parent->page) || is_build || |
| 1063 |
child->downlinkoffnum == InvalidOffsetNumber); |
| 1064 |
|
| 1065 |
/* |
| 1066 |
* Scan the page to re-find the downlink. If the page was split, it might |
| 1067 |
* have moved to a different page, so follow the right links until we find |
| 1068 |
* it. |
| 1069 |
*/ |
| 1070 |
while (true) |
| 1071 |
{ |
| 1072 |
OffsetNumber i; |
| 1073 |
|
| 1074 |
maxoff = PageGetMaxOffsetNumber(parent->page); |
| 1075 |
for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i)) |
| 1076 |
{ |
| 1077 |
iid = PageGetItemId(parent->page, i); |
| 1078 |
idxtuple = (IndexTuple) PageGetItem(parent->page, iid); |
| 1079 |
if (ItemPointerGetBlockNumber(&(idxtuple->t_tid)) == child->blkno) |
| 1080 |
{ |
| 1081 |
/* yes!!, found */ |
| 1082 |
child->downlinkoffnum = i; |
| 1083 |
return; |
| 1084 |
} |
| 1085 |
} |
| 1086 |
|
| 1087 |
parent->blkno = GistPageGetOpaque(parent->page)->rightlink; |
| 1088 |
parent->downlinkoffnum = InvalidOffsetNumber; |
| 1089 |
UnlockReleaseBuffer(parent->buffer); |
| 1090 |
if (parent->blkno == InvalidBlockNumber) |
| 1091 |
{ |
| 1092 |
/* |
| 1093 |
* End of chain and still didn't find parent. It's a very-very |
| 1094 |
* rare situation when the root was split. |
| 1095 |
*/ |
| 1096 |
break; |
| 1097 |
} |
| 1098 |
parent->buffer = ReadBuffer(r, parent->blkno); |
| 1099 |
LockBuffer(parent->buffer, GIST_EXCLUSIVE); |
| 1100 |
gistcheckpage(r, parent->buffer); |
| 1101 |
parent->page = (Page) BufferGetPage(parent->buffer); |
| 1102 |
} |
| 1103 |
|
| 1104 |
/* |
| 1105 |
* awful!!, we need search tree to find parent ... , but before we should |
| 1106 |
* release all old parent |
| 1107 |
*/ |
| 1108 |
|
| 1109 |
ptr = child->parent->parent; /* child->parent already released above */ |
| 1110 |
while (ptr) |
| 1111 |
{ |
| 1112 |
ReleaseBuffer(ptr->buffer); |
| 1113 |
ptr = ptr->parent; |
| 1114 |
} |
| 1115 |
|
| 1116 |
/* ok, find new path */ |
| 1117 |
ptr = parent = gistFindPath(r, child->blkno, &child->downlinkoffnum); |
| 1118 |
|
| 1119 |
/* read all buffers as expected by caller */ |
| 1120 |
/* note we don't lock them or gistcheckpage them here! */ |
| 1121 |
while (ptr) |
| 1122 |
{ |
| 1123 |
ptr->buffer = ReadBuffer(r, ptr->blkno); |
| 1124 |
ptr->page = (Page) BufferGetPage(ptr->buffer); |
| 1125 |
ptr = ptr->parent; |
| 1126 |
} |
| 1127 |
|
| 1128 |
/* install new chain of parents to stack */ |
| 1129 |
child->parent = parent; |
| 1130 |
|
| 1131 |
/* make recursive call to normal processing */ |
| 1132 |
LockBuffer(child->parent->buffer, GIST_EXCLUSIVE); |
| 1133 |
gistFindCorrectParent(r, child, is_build); |
| 1134 |
} |
| 1135 |
|
| 1136 |
/* |
| 1137 |
* Form a downlink pointer for the page in 'buf'. |
| 1138 |
*/ |
| 1139 |
static IndexTuple |
| 1140 |
gistformdownlink(Relation rel, Buffer buf, GISTSTATE *giststate, |
| 1141 |
GISTInsertStack *stack, bool is_build) |
| 1142 |
{ |
| 1143 |
Page page = BufferGetPage(buf); |
| 1144 |
OffsetNumber maxoff; |
| 1145 |
OffsetNumber offset; |
| 1146 |
IndexTuple downlink = NULL; |
| 1147 |
|
| 1148 |
maxoff = PageGetMaxOffsetNumber(page); |
| 1149 |
for (offset = FirstOffsetNumber; offset <= maxoff; offset = OffsetNumberNext(offset)) |
| 1150 |
{ |
| 1151 |
IndexTuple ituple = (IndexTuple) |
| 1152 |
PageGetItem(page, PageGetItemId(page, offset)); |
| 1153 |
|
| 1154 |
if (downlink == NULL) |
| 1155 |
downlink = CopyIndexTuple(ituple); |
| 1156 |
else |
| 1157 |
{ |
| 1158 |
IndexTuple newdownlink; |
| 1159 |
|
| 1160 |
newdownlink = gistgetadjusted(rel, downlink, ituple, |
| 1161 |
giststate); |
| 1162 |
if (newdownlink) |
| 1163 |
downlink = newdownlink; |
| 1164 |
} |
| 1165 |
} |
| 1166 |
|
| 1167 |
/* |
| 1168 |
* If the page is completely empty, we can't form a meaningful downlink |
| 1169 |
* for it. But we have to insert a downlink for the page. Any key will do, |
| 1170 |
* as long as its consistent with the downlink of parent page, so that we |
| 1171 |
* can legally insert it to the parent. A minimal one that matches as few |
| 1172 |
* scans as possible would be best, to keep scans from doing useless work, |
| 1173 |
* but we don't know how to construct that. So we just use the downlink of |
| 1174 |
* the original page that was split - that's as far from optimal as it can |
| 1175 |
* get but will do.. |
| 1176 |
*/ |
| 1177 |
if (!downlink) |
| 1178 |
{ |
| 1179 |
ItemId iid; |
| 1180 |
|
| 1181 |
LockBuffer(stack->parent->buffer, GIST_EXCLUSIVE); |
| 1182 |
gistFindCorrectParent(rel, stack, is_build); |
| 1183 |
iid = PageGetItemId(stack->parent->page, stack->downlinkoffnum); |
| 1184 |
downlink = (IndexTuple) PageGetItem(stack->parent->page, iid); |
| 1185 |
downlink = CopyIndexTuple(downlink); |
| 1186 |
LockBuffer(stack->parent->buffer, GIST_UNLOCK); |
| 1187 |
} |
| 1188 |
|
| 1189 |
ItemPointerSetBlockNumber(&(downlink->t_tid), BufferGetBlockNumber(buf)); |
| 1190 |
GistTupleSetValid(downlink); |
| 1191 |
|
| 1192 |
return downlink; |
| 1193 |
} |
| 1194 |
|
| 1195 |
|
| 1196 |
/* |
| 1197 |
* Complete the incomplete split of state->stack->page. |
| 1198 |
*/ |
| 1199 |
static void |
| 1200 |
gistfixsplit(GISTInsertState *state, GISTSTATE *giststate) |
| 1201 |
{ |
| 1202 |
GISTInsertStack *stack = state->stack; |
| 1203 |
Buffer buf; |
| 1204 |
Page page; |
| 1205 |
List *splitinfo = NIL; |
| 1206 |
|
| 1207 |
ereport(LOG, |
| 1208 |
(errmsg("fixing incomplete split in index \"%s\", block %u", |
| 1209 |
RelationGetRelationName(state->r), stack->blkno))); |
| 1210 |
|
| 1211 |
Assert(GistFollowRight(stack->page)); |
| 1212 |
Assert(OffsetNumberIsValid(stack->downlinkoffnum)); |
| 1213 |
|
| 1214 |
buf = stack->buffer; |
| 1215 |
|
| 1216 |
/* |
| 1217 |
* Read the chain of split pages, following the rightlinks. Construct a |
| 1218 |
* downlink tuple for each page. |
| 1219 |
*/ |
| 1220 |
for (;;) |
| 1221 |
{ |
| 1222 |
GISTPageSplitInfo *si = palloc(sizeof(GISTPageSplitInfo)); |
| 1223 |
IndexTuple downlink; |
| 1224 |
|
| 1225 |
page = BufferGetPage(buf); |
| 1226 |
|
| 1227 |
/* Form the new downlink tuples to insert to parent */ |
| 1228 |
downlink = gistformdownlink(state->r, buf, giststate, stack, state->is_build); |
| 1229 |
|
| 1230 |
si->buf = buf; |
| 1231 |
si->downlink = downlink; |
| 1232 |
|
| 1233 |
splitinfo = lappend(splitinfo, si); |
| 1234 |
|
| 1235 |
if (GistFollowRight(page)) |
| 1236 |
{ |
| 1237 |
/* lock next page */ |
| 1238 |
buf = ReadBuffer(state->r, GistPageGetOpaque(page)->rightlink); |
| 1239 |
LockBuffer(buf, GIST_EXCLUSIVE); |
| 1240 |
} |
| 1241 |
else |
| 1242 |
break; |
| 1243 |
} |
| 1244 |
|
| 1245 |
/* Insert the downlinks */ |
| 1246 |
gistfinishsplit(state, stack, giststate, splitinfo, false); |
| 1247 |
} |
| 1248 |
|
| 1249 |
/* |
| 1250 |
* Insert or replace a tuple in stack->buffer. If 'oldoffnum' is valid, the |
| 1251 |
* tuple at 'oldoffnum' is replaced, otherwise the tuple is inserted as new. |
| 1252 |
* 'stack' represents the path from the root to the page being updated. |
| 1253 |
* |
| 1254 |
* The caller must hold an exclusive lock on stack->buffer. The lock is still |
| 1255 |
* held on return, but the page might not contain the inserted tuple if the |
| 1256 |
* page was split. The function returns true if the page was split, false |
| 1257 |
* otherwise. |
| 1258 |
*/ |
| 1259 |
static bool |
| 1260 |
gistinserttuple(GISTInsertState *state, GISTInsertStack *stack, |
| 1261 |
GISTSTATE *giststate, IndexTuple tuple, OffsetNumber oldoffnum) |
| 1262 |
{ |
| 1263 |
return gistinserttuples(state, stack, giststate, &tuple, 1, oldoffnum, |
| 1264 |
InvalidBuffer, InvalidBuffer, false, false); |
| 1265 |
} |
| 1266 |
|
| 1267 |
/* ---------------- |
| 1268 |
* An extended workhorse version of gistinserttuple(). This version allows |
| 1269 |
* inserting multiple tuples, or replacing a single tuple with multiple tuples. |
| 1270 |
* This is used to recursively update the downlinks in the parent when a page |
| 1271 |
* is split. |
| 1272 |
* |
| 1273 |
* If leftchild and rightchild are valid, we're inserting/replacing the |
| 1274 |
* downlink for rightchild, and leftchild is its left sibling. We clear the |
| 1275 |
* F_FOLLOW_RIGHT flag and update NSN on leftchild, atomically with the |
| 1276 |
* insertion of the downlink. |
| 1277 |
* |
| 1278 |
* To avoid holding locks for longer than necessary, when recursing up the |
| 1279 |
* tree to update the parents, the locking is a bit peculiar here. On entry, |
| 1280 |
* the caller must hold an exclusive lock on stack->buffer, as well as |
| 1281 |
* leftchild and rightchild if given. On return: |
| 1282 |
* |
| 1283 |
* - Lock on stack->buffer is released, if 'unlockbuf' is true. The page is |
| 1284 |
* always kept pinned, however. |
| 1285 |
* - Lock on 'leftchild' is released, if 'unlockleftchild' is true. The page |
| 1286 |
* is kept pinned. |
| 1287 |
* - Lock and pin on 'rightchild' are always released. |
| 1288 |
* |
| 1289 |
* Returns 'true' if the page had to be split. Note that if the page was |
| 1290 |
* split, the inserted/updated tuples might've been inserted to a right |
| 1291 |
* sibling of stack->buffer instead of stack->buffer itself. |
| 1292 |
*/ |
| 1293 |
static bool |
| 1294 |
gistinserttuples(GISTInsertState *state, GISTInsertStack *stack, |
| 1295 |
GISTSTATE *giststate, |
| 1296 |
IndexTuple *tuples, int ntup, OffsetNumber oldoffnum, |
| 1297 |
Buffer leftchild, Buffer rightchild, |
| 1298 |
bool unlockbuf, bool unlockleftchild) |
| 1299 |
{ |
| 1300 |
List *splitinfo; |
| 1301 |
bool is_split; |
| 1302 |
|
| 1303 |
/* |
| 1304 |
* Check for any rw conflicts (in serializable isolation level) just |
| 1305 |
* before we intend to modify the page |
| 1306 |
*/ |
| 1307 |
CheckForSerializableConflictIn(state->r, NULL, BufferGetBlockNumber(stack->buffer)); |
| 1308 |
|
| 1309 |
/* Insert the tuple(s) to the page, splitting the page if necessary */ |
| 1310 |
is_split = gistplacetopage(state->r, state->freespace, giststate, |
| 1311 |
stack->buffer, |
| 1312 |
tuples, ntup, |
| 1313 |
oldoffnum, NULL, |
| 1314 |
leftchild, |
| 1315 |
&splitinfo, |
| 1316 |
true, |
| 1317 |
state->heapRel, |
| 1318 |
state->is_build); |
| 1319 |
|
| 1320 |
/* |
| 1321 |
* Before recursing up in case the page was split, release locks on the |
| 1322 |
* child pages. We don't need to keep them locked when updating the |
| 1323 |
* parent. |
| 1324 |
*/ |
| 1325 |
if (BufferIsValid(rightchild)) |
| 1326 |
UnlockReleaseBuffer(rightchild); |
| 1327 |
if (BufferIsValid(leftchild) && unlockleftchild) |
| 1328 |
LockBuffer(leftchild, GIST_UNLOCK); |
| 1329 |
|
| 1330 |
/* |
| 1331 |
* If we had to split, insert/update the downlinks in the parent. If the |
| 1332 |
* caller requested us to release the lock on stack->buffer, tell |
| 1333 |
* gistfinishsplit() to do that as soon as it's safe to do so. If we |
| 1334 |
* didn't have to split, release it ourselves. |
| 1335 |
*/ |
| 1336 |
if (splitinfo) |
| 1337 |
gistfinishsplit(state, stack, giststate, splitinfo, unlockbuf); |
| 1338 |
else if (unlockbuf) |
| 1339 |
LockBuffer(stack->buffer, GIST_UNLOCK); |
| 1340 |
|
| 1341 |
return is_split; |
| 1342 |
} |
| 1343 |
|
| 1344 |
/* |
| 1345 |
* Finish an incomplete split by inserting/updating the downlinks in parent |
| 1346 |
* page. 'splitinfo' contains all the child pages involved in the split, |
| 1347 |
* from left-to-right. |
| 1348 |
* |
| 1349 |
* On entry, the caller must hold a lock on stack->buffer and all the child |
| 1350 |
* pages in 'splitinfo'. If 'unlockbuf' is true, the lock on stack->buffer is |
| 1351 |
* released on return. The child pages are always unlocked and unpinned. |
| 1352 |
*/ |
| 1353 |
static void |
| 1354 |
gistfinishsplit(GISTInsertState *state, GISTInsertStack *stack, |
| 1355 |
GISTSTATE *giststate, List *splitinfo, bool unlockbuf) |
| 1356 |
{ |
| 1357 |
GISTPageSplitInfo *right; |
| 1358 |
GISTPageSplitInfo *left; |
| 1359 |
IndexTuple tuples[2]; |
| 1360 |
|
| 1361 |
/* A split always contains at least two halves */ |
| 1362 |
Assert(list_length(splitinfo) >= 2); |
| 1363 |
|
| 1364 |
/* |
| 1365 |
* We need to insert downlinks for each new page, and update the downlink |
| 1366 |
* for the original (leftmost) page in the split. Begin at the rightmost |
| 1367 |
* page, inserting one downlink at a time until there's only two pages |
| 1368 |
* left. Finally insert the downlink for the last new page and update the |
| 1369 |
* downlink for the original page as one operation. |
| 1370 |
*/ |
| 1371 |
LockBuffer(stack->parent->buffer, GIST_EXCLUSIVE); |
| 1372 |
|
| 1373 |
/* |
| 1374 |
* Insert downlinks for the siblings from right to left, until there are |
| 1375 |
* only two siblings left. |
| 1376 |
*/ |
| 1377 |
for (int pos = list_length(splitinfo) - 1; pos > 1; pos--) |
| 1378 |
{ |
| 1379 |
right = (GISTPageSplitInfo *) list_nth(splitinfo, pos); |
| 1380 |
left = (GISTPageSplitInfo *) list_nth(splitinfo, pos - 1); |
| 1381 |
|
| 1382 |
gistFindCorrectParent(state->r, stack, state->is_build); |
| 1383 |
if (gistinserttuples(state, stack->parent, giststate, |
| 1384 |
&right->downlink, 1, |
| 1385 |
InvalidOffsetNumber, |
| 1386 |
left->buf, right->buf, false, false)) |
| 1387 |
{ |
| 1388 |
/* |
| 1389 |
* If the parent page was split, the existing downlink might have |
| 1390 |
* moved. |
| 1391 |
*/ |
| 1392 |
stack->downlinkoffnum = InvalidOffsetNumber; |
| 1393 |
} |
| 1394 |
/* gistinserttuples() released the lock on right->buf. */ |
| 1395 |
} |
| 1396 |
|
| 1397 |
right = (GISTPageSplitInfo *) lsecond(splitinfo); |
| 1398 |
left = (GISTPageSplitInfo *) linitial(splitinfo); |
| 1399 |
|
| 1400 |
/* |
| 1401 |
* Finally insert downlink for the remaining right page and update the |
| 1402 |
* downlink for the original page to not contain the tuples that were |
| 1403 |
* moved to the new pages. |
| 1404 |
*/ |
| 1405 |
tuples[0] = left->downlink; |
| 1406 |
tuples[1] = right->downlink; |
| 1407 |
gistFindCorrectParent(state->r, stack, state->is_build); |
| 1408 |
(void) gistinserttuples(state, stack->parent, giststate, |
| 1409 |
tuples, 2, |
| 1410 |
stack->downlinkoffnum, |
| 1411 |
left->buf, right->buf, |
| 1412 |
true, /* Unlock parent */ |
| 1413 |
unlockbuf /* Unlock stack->buffer if caller |
| 1414 |
* wants that */ |
| 1415 |
); |
| 1416 |
|
| 1417 |
/* |
| 1418 |
* The downlink might have moved when we updated it. Even if the page |
| 1419 |
* wasn't split, because gistinserttuples() implements updating the old |
| 1420 |
* tuple by removing and re-inserting it! |
| 1421 |
*/ |
| 1422 |
stack->downlinkoffnum = InvalidOffsetNumber; |
| 1423 |
|
| 1424 |
Assert(left->buf == stack->buffer); |
| 1425 |
|
| 1426 |
/* |
| 1427 |
* If we split the page because we had to adjust the downlink on an |
| 1428 |
* internal page, while descending the tree for inserting a new tuple, |
| 1429 |
* then this might no longer be the correct page for the new tuple. The |
| 1430 |
* downlink to this page might not cover the new tuple anymore, it might |
| 1431 |
* need to go to the newly-created right sibling instead. Tell the caller |
| 1432 |
* to walk back up the stack, to re-check at the parent which page to |
| 1433 |
* insert to. |
| 1434 |
* |
| 1435 |
* Normally, the LSN-NSN interlock during the tree descend would also |
| 1436 |
* detect that a concurrent split happened (by ourselves), and cause us to |
| 1437 |
* retry at the parent. But that mechanism doesn't work during index |
| 1438 |
* build, because we don't do WAL-logging, and don't update LSNs, during |
| 1439 |
* index build. |
| 1440 |
*/ |
| 1441 |
stack->retry_from_parent = true; |
| 1442 |
} |
| 1443 |
|
| 1444 |
/* |
| 1445 |
* gistSplit -- split a page in the tree and fill struct |
| 1446 |
* used for XLOG and real writes buffers. Function is recursive, ie |
| 1447 |
* it will split page until keys will fit in every page. |
| 1448 |
*/ |
| 1449 |
SplitPageLayout * |
| 1450 |
gistSplit(Relation r, |
| 1451 |
Page page, |
| 1452 |
IndexTuple *itup, /* contains compressed entry */ |
| 1453 |
int len, |
| 1454 |
GISTSTATE *giststate) |
| 1455 |
{ |
| 1456 |
IndexTuple *lvectup, |
| 1457 |
*rvectup; |
| 1458 |
GistSplitVector v; |
| 1459 |
int i; |
| 1460 |
SplitPageLayout *res = NULL; |
| 1461 |
|
| 1462 |
/* this should never recurse very deeply, but better safe than sorry */ |
| 1463 |
check_stack_depth(); |
| 1464 |
|
| 1465 |
/* there's no point in splitting an empty page */ |
| 1466 |
Assert(len > 0); |
| 1467 |
|
| 1468 |
/* |
| 1469 |
* If a single tuple doesn't fit on a page, no amount of splitting will |
| 1470 |
* help. |
| 1471 |
*/ |
| 1472 |
if (len == 1) |
| 1473 |
ereport(ERROR, |
| 1474 |
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED), |
| 1475 |
errmsg("index row size %zu exceeds maximum %zu for index \"%s\"", |
| 1476 |
IndexTupleSize(itup[0]), GiSTPageSize, |
| 1477 |
RelationGetRelationName(r)))); |
| 1478 |
|
| 1479 |
memset(v.spl_lisnull, true, |
| 1480 |
sizeof(bool) * giststate->nonLeafTupdesc->natts); |
| 1481 |
memset(v.spl_risnull, true, |
| 1482 |
sizeof(bool) * giststate->nonLeafTupdesc->natts); |
| 1483 |
gistSplitByKey(r, page, itup, len, giststate, &v, 0); |
| 1484 |
|
| 1485 |
/* form left and right vector */ |
| 1486 |
lvectup = (IndexTuple *) palloc(sizeof(IndexTuple) * (len + 1)); |
| 1487 |
rvectup = (IndexTuple *) palloc(sizeof(IndexTuple) * (len + 1)); |
| 1488 |
|
| 1489 |
for (i = 0; i < v.splitVector.spl_nleft; i++) |
| 1490 |
lvectup[i] = itup[v.splitVector.spl_left[i] - 1]; |
| 1491 |
|
| 1492 |
for (i = 0; i < v.splitVector.spl_nright; i++) |
| 1493 |
rvectup[i] = itup[v.splitVector.spl_right[i] - 1]; |
| 1494 |
|
| 1495 |
/* finalize splitting (may need another split) */ |
| 1496 |
if (!gistfitpage(rvectup, v.splitVector.spl_nright)) |
| 1497 |
{ |
| 1498 |
res = gistSplit(r, page, rvectup, v.splitVector.spl_nright, giststate); |
| 1499 |
} |
| 1500 |
else |
| 1501 |
{ |
| 1502 |
ROTATEDIST(res); |
| 1503 |
res->block.num = v.splitVector.spl_nright; |
| 1504 |
res->list = gistfillitupvec(rvectup, v.splitVector.spl_nright, &(res->lenlist)); |
| 1505 |
res->itup = gistFormTuple(giststate, r, v.spl_rattr, v.spl_risnull, false); |
| 1506 |
} |
| 1507 |
|
| 1508 |
if (!gistfitpage(lvectup, v.splitVector.spl_nleft)) |
| 1509 |
{ |
| 1510 |
SplitPageLayout *resptr, |
| 1511 |
*subres; |
| 1512 |
|
| 1513 |
resptr = subres = gistSplit(r, page, lvectup, v.splitVector.spl_nleft, giststate); |
| 1514 |
|
| 1515 |
/* install on list's tail */ |
| 1516 |
while (resptr->next) |
| 1517 |
resptr = resptr->next; |
| 1518 |
|
| 1519 |
resptr->next = res; |
| 1520 |
res = subres; |
| 1521 |
} |
| 1522 |
else |
| 1523 |
{ |
| 1524 |
ROTATEDIST(res); |
| 1525 |
res->block.num = v.splitVector.spl_nleft; |
| 1526 |
res->list = gistfillitupvec(lvectup, v.splitVector.spl_nleft, &(res->lenlist)); |
| 1527 |
res->itup = gistFormTuple(giststate, r, v.spl_lattr, v.spl_lisnull, false); |
| 1528 |
} |
| 1529 |
|
| 1530 |
return res; |
| 1531 |
} |
| 1532 |
|
| 1533 |
/* |
| 1534 |
* Create a GISTSTATE and fill it with information about the index |
| 1535 |
*/ |
| 1536 |
GISTSTATE * |
| 1537 |
initGISTstate(Relation index) |
| 1538 |
{ |
| 1539 |
GISTSTATE *giststate; |
| 1540 |
MemoryContext scanCxt; |
| 1541 |
MemoryContext oldCxt; |
| 1542 |
int i; |
| 1543 |
|
| 1544 |
/* safety check to protect fixed-size arrays in GISTSTATE */ |
| 1545 |
if (index->rd_att->natts > INDEX_MAX_KEYS) |
| 1546 |
elog(ERROR, "numberOfAttributes %d > %d", |
| 1547 |
index->rd_att->natts, INDEX_MAX_KEYS); |
| 1548 |
|
| 1549 |
/* Create the memory context that will hold the GISTSTATE */ |
| 1550 |
scanCxt = AllocSetContextCreate(CurrentMemoryContext, |
| 1551 |
"GiST scan context", |
| 1552 |
ALLOCSET_DEFAULT_SIZES); |
| 1553 |
oldCxt = MemoryContextSwitchTo(scanCxt); |
| 1554 |
|
| 1555 |
/* Create and fill in the GISTSTATE */ |
| 1556 |
giststate = (GISTSTATE *) palloc(sizeof(GISTSTATE)); |
| 1557 |
|
| 1558 |
giststate->scanCxt = scanCxt; |
| 1559 |
giststate->tempCxt = scanCxt; /* caller must change this if needed */ |
| 1560 |
giststate->leafTupdesc = index->rd_att; |
| 1561 |
|
| 1562 |
/* |
| 1563 |
* The truncated tupdesc for non-leaf index tuples, which doesn't contain |
| 1564 |
* the INCLUDE attributes. |
| 1565 |
* |
| 1566 |
* It is used to form tuples during tuple adjustment and page split. |
| 1567 |
* B-tree creates shortened tuple descriptor for every truncated tuple, |
| 1568 |
* because it is doing this less often: it does not have to form truncated |
| 1569 |
* tuples during page split. Also, B-tree is not adjusting tuples on |
| 1570 |
* internal pages the way GiST does. |
| 1571 |
*/ |
| 1572 |
giststate->nonLeafTupdesc = CreateTupleDescTruncatedCopy(index->rd_att, |
| 1573 |
IndexRelationGetNumberOfKeyAttributes(index)); |
| 1574 |
|
| 1575 |
for (i = 0; i < IndexRelationGetNumberOfKeyAttributes(index); i++) |
| 1576 |
{ |
| 1577 |
fmgr_info_copy(&(giststate->consistentFn[i]), |
| 1578 |
index_getprocinfo(index, i + 1, GIST_CONSISTENT_PROC), |
| 1579 |
scanCxt); |
| 1580 |
fmgr_info_copy(&(giststate->unionFn[i]), |
| 1581 |
index_getprocinfo(index, i + 1, GIST_UNION_PROC), |
| 1582 |
scanCxt); |
| 1583 |
|
| 1584 |
/* opclasses are not required to provide a Compress method */ |
| 1585 |
if (OidIsValid(index_getprocid(index, i + 1, GIST_COMPRESS_PROC))) |
| 1586 |
fmgr_info_copy(&(giststate->compressFn[i]), |
| 1587 |
index_getprocinfo(index, i + 1, GIST_COMPRESS_PROC), |
| 1588 |
scanCxt); |
| 1589 |
else |
| 1590 |
giststate->compressFn[i].fn_oid = InvalidOid; |
| 1591 |
|
| 1592 |
/* opclasses are not required to provide a Decompress method */ |
| 1593 |
if (OidIsValid(index_getprocid(index, i + 1, GIST_DECOMPRESS_PROC))) |
| 1594 |
fmgr_info_copy(&(giststate->decompressFn[i]), |
| 1595 |
index_getprocinfo(index, i + 1, GIST_DECOMPRESS_PROC), |
| 1596 |
scanCxt); |
| 1597 |
else |
| 1598 |
giststate->decompressFn[i].fn_oid = InvalidOid; |
| 1599 |
|
| 1600 |
fmgr_info_copy(&(giststate->penaltyFn[i]), |
| 1601 |
index_getprocinfo(index, i + 1, GIST_PENALTY_PROC), |
| 1602 |
scanCxt); |
| 1603 |
fmgr_info_copy(&(giststate->picksplitFn[i]), |
| 1604 |
index_getprocinfo(index, i + 1, GIST_PICKSPLIT_PROC), |
| 1605 |
scanCxt); |
| 1606 |
fmgr_info_copy(&(giststate->equalFn[i]), |
| 1607 |
index_getprocinfo(index, i + 1, GIST_EQUAL_PROC), |
| 1608 |
scanCxt); |
| 1609 |
|
| 1610 |
/* opclasses are not required to provide a Distance method */ |
| 1611 |
if (OidIsValid(index_getprocid(index, i + 1, GIST_DISTANCE_PROC))) |
| 1612 |
fmgr_info_copy(&(giststate->distanceFn[i]), |
| 1613 |
index_getprocinfo(index, i + 1, GIST_DISTANCE_PROC), |
| 1614 |
scanCxt); |
| 1615 |
else |
| 1616 |
giststate->distanceFn[i].fn_oid = InvalidOid; |
| 1617 |
|
| 1618 |
/* opclasses are not required to provide a Fetch method */ |
| 1619 |
if (OidIsValid(index_getprocid(index, i + 1, GIST_FETCH_PROC))) |
| 1620 |
fmgr_info_copy(&(giststate->fetchFn[i]), |
| 1621 |
index_getprocinfo(index, i + 1, GIST_FETCH_PROC), |
| 1622 |
scanCxt); |
| 1623 |
else |
| 1624 |
giststate->fetchFn[i].fn_oid = InvalidOid; |
| 1625 |
|
| 1626 |
/* |
| 1627 |
* If the index column has a specified collation, we should honor that |
| 1628 |
* while doing comparisons. However, we may have a collatable storage |
| 1629 |
* type for a noncollatable indexed data type. If there's no index |
| 1630 |
* collation then specify default collation in case the support |
| 1631 |
* functions need collation. This is harmless if the support |
| 1632 |
* functions don't care about collation, so we just do it |
| 1633 |
* unconditionally. (We could alternatively call get_typcollation, |
| 1634 |
* but that seems like expensive overkill --- there aren't going to be |
| 1635 |
* any cases where a GiST storage type has a nondefault collation.) |
| 1636 |
*/ |
| 1637 |
if (OidIsValid(index->rd_indcollation[i])) |
| 1638 |
giststate->supportCollation[i] = index->rd_indcollation[i]; |
| 1639 |
else |
| 1640 |
giststate->supportCollation[i] = DEFAULT_COLLATION_OID; |
| 1641 |
} |
| 1642 |
|
| 1643 |
/* No opclass information for INCLUDE attributes */ |
| 1644 |
for (; i < index->rd_att->natts; i++) |
| 1645 |
{ |
| 1646 |
giststate->consistentFn[i].fn_oid = InvalidOid; |
| 1647 |
giststate->unionFn[i].fn_oid = InvalidOid; |
| 1648 |
giststate->compressFn[i].fn_oid = InvalidOid; |
| 1649 |
giststate->decompressFn[i].fn_oid = InvalidOid; |
| 1650 |
giststate->penaltyFn[i].fn_oid = InvalidOid; |
| 1651 |
giststate->picksplitFn[i].fn_oid = InvalidOid; |
| 1652 |
giststate->equalFn[i].fn_oid = InvalidOid; |
| 1653 |
giststate->distanceFn[i].fn_oid = InvalidOid; |
| 1654 |
giststate->fetchFn[i].fn_oid = InvalidOid; |
| 1655 |
giststate->supportCollation[i] = InvalidOid; |
| 1656 |
} |
| 1657 |
|
| 1658 |
MemoryContextSwitchTo(oldCxt); |
| 1659 |
|
| 1660 |
return giststate; |
| 1661 |
} |
| 1662 |
|
| 1663 |
void |
| 1664 |
freeGISTstate(GISTSTATE *giststate) |
| 1665 |
{ |
| 1666 |
/* It's sufficient to delete the scanCxt */ |
| 1667 |
MemoryContextDelete(giststate->scanCxt); |
| 1668 |
} |
| 1669 |
|
| 1670 |
/* |
| 1671 |
* gistprunepage() -- try to remove LP_DEAD items from the given page. |
| 1672 |
* Function assumes that buffer is exclusively locked. |
| 1673 |
*/ |
| 1674 |
static void |
| 1675 |
gistprunepage(Relation rel, Page page, Buffer buffer, Relation heapRel) |
| 1676 |
{ |
| 1677 |
OffsetNumber deletable[MaxIndexTuplesPerPage]; |
| 1678 |
int ndeletable = 0; |
| 1679 |
OffsetNumber offnum, |
| 1680 |
maxoff; |
| 1681 |
|
| 1682 |
Assert(GistPageIsLeaf(page)); |
| 1683 |
|
| 1684 |
/* |
| 1685 |
* Scan over all items to see which ones need to be deleted according to |
| 1686 |
* LP_DEAD flags. |
| 1687 |
*/ |
| 1688 |
maxoff = PageGetMaxOffsetNumber(page); |
| 1689 |
for (offnum = FirstOffsetNumber; |
| 1690 |
offnum <= maxoff; |
| 1691 |
offnum = OffsetNumberNext(offnum)) |
| 1692 |
{ |
| 1693 |
ItemId itemId = PageGetItemId(page, offnum); |
| 1694 |
|
| 1695 |
if (ItemIdIsDead(itemId)) |
| 1696 |
deletable[ndeletable++] = offnum; |
| 1697 |
} |
| 1698 |
|
| 1699 |
if (ndeletable > 0) |
| 1700 |
{ |
| 1701 |
TransactionId snapshotConflictHorizon = InvalidTransactionId; |
| 1702 |
|
| 1703 |
if (XLogStandbyInfoActive() && RelationNeedsWAL(rel)) |
| 1704 |
snapshotConflictHorizon = |
| 1705 |
index_compute_xid_horizon_for_tuples(rel, heapRel, buffer, |
| 1706 |
deletable, ndeletable); |
| 1707 |
|
| 1708 |
START_CRIT_SECTION(); |
| 1709 |
|
| 1710 |
PageIndexMultiDelete(page, deletable, ndeletable); |
| 1711 |
|
| 1712 |
/* |
| 1713 |
* Mark the page as not containing any LP_DEAD items. This is not |
| 1714 |
* certainly true (there might be some that have recently been marked, |
| 1715 |
* but weren't included in our target-item list), but it will almost |
| 1716 |
* always be true and it doesn't seem worth an additional page scan to |
| 1717 |
* check it. Remember that F_HAS_GARBAGE is only a hint anyway. |
| 1718 |
*/ |
| 1719 |
GistClearPageHasGarbage(page); |
| 1720 |
|
| 1721 |
MarkBufferDirty(buffer); |
| 1722 |
|
| 1723 |
/* XLOG stuff */ |
| 1724 |
if (RelationNeedsWAL(rel)) |
| 1725 |
{ |
| 1726 |
XLogRecPtr recptr; |
| 1727 |
|
| 1728 |
recptr = gistXLogDelete(buffer, |
| 1729 |
deletable, ndeletable, |
| 1730 |
snapshotConflictHorizon, |
| 1731 |
heapRel); |
| 1732 |
|
| 1733 |
PageSetLSN(page, recptr); |
| 1734 |
} |
| 1735 |
else |
| 1736 |
PageSetLSN(page, gistGetFakeLSN(rel)); |
| 1737 |
|
| 1738 |
END_CRIT_SECTION(); |
| 1739 |
} |
| 1740 |
|
| 1741 |
/* |
| 1742 |
* Note: if we didn't find any LP_DEAD items, then the page's |
| 1743 |
* F_HAS_GARBAGE hint bit is falsely set. We do not bother expending a |
| 1744 |
* separate write to clear it, however. We will clear it when we split |
| 1745 |
* the page. |
| 1746 |
*/ |
| 1747 |
} |