/*
* replay addition of overflow page for hash index
*/
static void
hash_xlog_add_ovfl_page(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_add_ovfl_page *xlrec = (xl_hash_add_ovfl_page *) XLogRecGetData(record);
Buffer leftbuf;
Buffer ovflbuf;
Buffer metabuf;
BlockNumber leftblk;
BlockNumber rightblk;
BlockNumber newmapblk = InvalidBlockNumber;
Page ovflpage;
HashPageOpaque ovflopaque;
uint32 *num_bucket;
char *data;
Size datalen PG_USED_FOR_ASSERTS_ONLY;
bool new_bmpage = false;
XLogRecGetBlockTag(record, 0, NULL, NULL, &rightblk);
XLogRecGetBlockTag(record, 1, NULL, NULL, &leftblk);
ovflbuf = XLogInitBufferForRedo(record, 0);
Assert(BufferIsValid(ovflbuf));
data = XLogRecGetBlockData(record, 0, &datalen);
num_bucket = (uint32 *) data;
Assert(datalen == sizeof(uint32));
_hash_initbuf(ovflbuf, InvalidBlockNumber, *num_bucket, LH_OVERFLOW_PAGE,
true);
/* update backlink */
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = leftblk;
PageSetLSN(ovflpage, lsn);
MarkBufferDirty(ovflbuf);
if (XLogReadBufferForRedo(record, 1, &leftbuf) == BLK_NEEDS_REDO)
{
Page leftpage;
HashPageOpaque leftopaque;
leftpage = BufferGetPage(leftbuf);
leftopaque = (HashPageOpaque) PageGetSpecialPointer(leftpage);
leftopaque->hasho_nextblkno = rightblk;
PageSetLSN(leftpage, lsn);
MarkBufferDirty(leftbuf);
}
if (BufferIsValid(leftbuf))
UnlockReleaseBuffer(leftbuf);
UnlockReleaseBuffer(ovflbuf);
/*
* Note: in normal operation, we'd update the bitmap and meta page while
* still holding lock on the overflow pages. But during replay it's not
* necessary to hold those locks, since no other index updates can be
* happening concurrently.
*/
if (XLogRecHasBlockRef(record, 2))
{
Buffer mapbuffer;
if (XLogReadBufferForRedo(record, 2, &mapbuffer) == BLK_NEEDS_REDO)
{
Page mappage = (Page) BufferGetPage(mapbuffer);
uint32 *freep = NULL;
char *data;
uint32 *bitmap_page_bit;
freep = HashPageGetBitmap(mappage);
data = XLogRecGetBlockData(record, 2, &datalen);
bitmap_page_bit = (uint32 *) data;
SETBIT(freep, *bitmap_page_bit);
PageSetLSN(mappage, lsn);
MarkBufferDirty(mapbuffer);
}
if (BufferIsValid(mapbuffer))
UnlockReleaseBuffer(mapbuffer);
}
if (XLogRecHasBlockRef(record, 3))
{
Buffer newmapbuf;
newmapbuf = XLogInitBufferForRedo(record, 3);
_hash_initbitmapbuffer(newmapbuf, xlrec->bmsize, true);
new_bmpage = true;
newmapblk = BufferGetBlockNumber(newmapbuf);
MarkBufferDirty(newmapbuf);
PageSetLSN(BufferGetPage(newmapbuf), lsn);
UnlockReleaseBuffer(newmapbuf);
}
if (XLogReadBufferForRedo(record, 4, &metabuf) == BLK_NEEDS_REDO)
{
HashMetaPage metap;
Page page;
uint32 *firstfree_ovflpage;
data = XLogRecGetBlockData(record, 4, &datalen);
firstfree_ovflpage = (uint32 *) data;
page = BufferGetPage(metabuf);
metap = HashPageGetMeta(page);
metap->hashm_firstfree = *firstfree_ovflpage;
if (!xlrec->bmpage_found)
{
metap->hashm_spares[metap->hashm_ovflpoint]++;
if (new_bmpage)
{
Assert(BlockNumberIsValid(newmapblk));
metap->hashm_mapp[metap->hashm_nmaps] = newmapblk;
metap->hashm_nmaps++;
metap->hashm_spares[metap->hashm_ovflpoint]++;
}
}
PageSetLSN(page, lsn);
MarkBufferDirty(metabuf);
}
if (BufferIsValid(metabuf))
UnlockReleaseBuffer(metabuf);
}
/*
* _hash_freeovflpage() -
*
* Remove this overflow page from its bucket's chain, and mark the page as
* free. On entry, ovflbuf is write-locked; it is released before exiting.
*
* Since this function is invoked in VACUUM, we provide an access strategy
* parameter that controls fetches of the bucket pages.
*
* Returns the block number of the page that followed the given page
* in the bucket, or InvalidBlockNumber if no following page.
*
* NB: caller must not hold lock on metapage, nor on either page that's
* adjacent in the bucket chain. The caller had better hold exclusive lock
* on the bucket, too.
*/
BlockNumber
_hash_freeovflpage(Relation rel, Buffer ovflbuf,
BufferAccessStrategy bstrategy)
{
HashMetaPage metap;
Buffer metabuf;
Buffer mapbuf;
BlockNumber ovflblkno;
BlockNumber prevblkno;
BlockNumber blkno;
BlockNumber nextblkno;
HashPageOpaque ovflopaque;
Page ovflpage;
Page mappage;
uint32 *freep;
uint32 ovflbitno;
int32 bitmappage,
bitmapbit;
Bucket bucket PG_USED_FOR_ASSERTS_ONLY;
/* Get information from the doomed page */
_hash_checkpage(rel, ovflbuf, LH_OVERFLOW_PAGE);
ovflblkno = BufferGetBlockNumber(ovflbuf);
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
nextblkno = ovflopaque->hasho_nextblkno;
prevblkno = ovflopaque->hasho_prevblkno;
bucket = ovflopaque->hasho_bucket;
/*
* Zero the page for debugging's sake; then write and release it. (Note:
* if we failed to zero the page here, we'd have problems with the Assert
* in _hash_pageinit() when the page is reused.)
*/
MemSet(ovflpage, 0, BufferGetPageSize(ovflbuf));
_hash_wrtbuf(rel, ovflbuf);
/*
* Fix up the bucket chain. this is a doubly-linked list, so we must fix
* up the bucket chain members behind and ahead of the overflow page being
* deleted. No concurrency issues since we hold exclusive lock on the
* entire bucket.
*/
if (BlockNumberIsValid(prevblkno))
{
Buffer prevbuf = _hash_getbuf_with_strategy(rel,
prevblkno,
HASH_WRITE,
LH_BUCKET_PAGE | LH_OVERFLOW_PAGE,
bstrategy);
Page prevpage = BufferGetPage(prevbuf);
HashPageOpaque prevopaque = (HashPageOpaque) PageGetSpecialPointer(prevpage);
Assert(prevopaque->hasho_bucket == bucket);
prevopaque->hasho_nextblkno = nextblkno;
_hash_wrtbuf(rel, prevbuf);
}
if (BlockNumberIsValid(nextblkno))
{
Buffer nextbuf = _hash_getbuf_with_strategy(rel,
nextblkno,
HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
Page nextpage = BufferGetPage(nextbuf);
HashPageOpaque nextopaque = (HashPageOpaque) PageGetSpecialPointer(nextpage);
Assert(nextopaque->hasho_bucket == bucket);
nextopaque->hasho_prevblkno = prevblkno;
_hash_wrtbuf(rel, nextbuf);
}
/* Note: bstrategy is intentionally not used for metapage and bitmap */
/* Read the metapage so we can determine which bitmap page to use */
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/* Identify which bit to set */
ovflbitno = blkno_to_bitno(metap, ovflblkno);
bitmappage = ovflbitno >> BMPG_SHIFT(metap);
bitmapbit = ovflbitno & BMPG_MASK(metap);
if (bitmappage >= metap->hashm_nmaps)
elog(ERROR, "invalid overflow bit number %u", ovflbitno);
blkno = metap->hashm_mapp[bitmappage];
/* Release metapage lock while we access the bitmap page */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/* Clear the bitmap bit to indicate that this overflow page is free */
mapbuf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BITMAP_PAGE);
mappage = BufferGetPage(mapbuf);
freep = HashPageGetBitmap(mappage);
Assert(ISSET(freep, bitmapbit));
CLRBIT(freep, bitmapbit);
_hash_wrtbuf(rel, mapbuf);
/* Get write-lock on metapage to update firstfree */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
/* if this is now the first free page, update hashm_firstfree */
if (ovflbitno < metap->hashm_firstfree)
{
metap->hashm_firstfree = ovflbitno;
_hash_wrtbuf(rel, metabuf);
}
else
{
/* no need to change metapage */
_hash_relbuf(rel, metabuf);
}
return nextblkno;
}
/*
* Bulk deletion of all index entries pointing to a set of heap tuples.
* The set of target tuples is specified via a callback routine that tells
* whether any given heap tuple (identified by ItemPointer) is being deleted.
*
* This function also deletes the tuples that are moved by split to other
* bucket.
*
* Result: a palloc'd struct containing statistical info for VACUUM displays.
*/
IndexBulkDeleteResult *
hashbulkdelete(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
IndexBulkDeleteCallback callback, void *callback_state)
{
Relation rel = info->index;
double tuples_removed;
double num_index_tuples;
double orig_ntuples;
Bucket orig_maxbucket;
Bucket cur_maxbucket;
Bucket cur_bucket;
Buffer metabuf = InvalidBuffer;
HashMetaPage metap;
HashMetaPage cachedmetap;
tuples_removed = 0;
num_index_tuples = 0;
/*
* We need a copy of the metapage so that we can use its hashm_spares[]
* values to compute bucket page addresses, but a cached copy should be
* good enough. (If not, we'll detect that further down and refresh the
* cache as necessary.)
*/
cachedmetap = _hash_getcachedmetap(rel, &metabuf, false);
Assert(cachedmetap != NULL);
orig_maxbucket = cachedmetap->hashm_maxbucket;
orig_ntuples = cachedmetap->hashm_ntuples;
/* Scan the buckets that we know exist */
cur_bucket = 0;
cur_maxbucket = orig_maxbucket;
loop_top:
while (cur_bucket <= cur_maxbucket)
{
BlockNumber bucket_blkno;
BlockNumber blkno;
Buffer bucket_buf;
Buffer buf;
HashPageOpaque bucket_opaque;
Page page;
bool split_cleanup = false;
/* Get address of bucket's start page */
bucket_blkno = BUCKET_TO_BLKNO(cachedmetap, cur_bucket);
blkno = bucket_blkno;
/*
* We need to acquire a cleanup lock on the primary bucket page to out
* wait concurrent scans before deleting the dead tuples.
*/
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL, info->strategy);
LockBufferForCleanup(buf);
_hash_checkpage(rel, buf, LH_BUCKET_PAGE);
page = BufferGetPage(buf);
bucket_opaque = (HashPageOpaque) PageGetSpecialPointer(page);
/*
* If the bucket contains tuples that are moved by split, then we need
* to delete such tuples. We can't delete such tuples if the split
* operation on bucket is not finished as those are needed by scans.
*/
if (!H_BUCKET_BEING_SPLIT(bucket_opaque) &&
H_NEEDS_SPLIT_CLEANUP(bucket_opaque))
{
split_cleanup = true;
/*
* This bucket might have been split since we last held a lock on
* the metapage. If so, hashm_maxbucket, hashm_highmask and
* hashm_lowmask might be old enough to cause us to fail to remove
* tuples left behind by the most recent split. To prevent that,
* now that the primary page of the target bucket has been locked
* (and thus can't be further split), check whether we need to
* update our cached metapage data.
*/
Assert(bucket_opaque->hasho_prevblkno != InvalidBlockNumber);
if (bucket_opaque->hasho_prevblkno > cachedmetap->hashm_maxbucket)
{
cachedmetap = _hash_getcachedmetap(rel, &metabuf, true);
Assert(cachedmetap != NULL);
}
}
bucket_buf = buf;
hashbucketcleanup(rel, cur_bucket, bucket_buf, blkno, info->strategy,
cachedmetap->hashm_maxbucket,
cachedmetap->hashm_highmask,
cachedmetap->hashm_lowmask, &tuples_removed,
&num_index_tuples, split_cleanup,
callback, callback_state);
_hash_dropbuf(rel, bucket_buf);
/* Advance to next bucket */
cur_bucket++;
}
if (BufferIsInvalid(metabuf))
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_NOLOCK, LH_META_PAGE);
/* Write-lock metapage and check for split since we started */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
if (cur_maxbucket != metap->hashm_maxbucket)
{
/* There's been a split, so process the additional bucket(s) */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
cachedmetap = _hash_getcachedmetap(rel, &metabuf, true);
Assert(cachedmetap != NULL);
cur_maxbucket = cachedmetap->hashm_maxbucket;
goto loop_top;
}
/* Okay, we're really done. Update tuple count in metapage. */
START_CRIT_SECTION();
if (orig_maxbucket == metap->hashm_maxbucket &&
orig_ntuples == metap->hashm_ntuples)
{
/*
* No one has split or inserted anything since start of scan, so
* believe our count as gospel.
*/
metap->hashm_ntuples = num_index_tuples;
}
else
{
/*
* Otherwise, our count is untrustworthy since we may have
* double-scanned tuples in split buckets. Proceed by dead-reckoning.
* (Note: we still return estimated_count = false, because using this
* count is better than not updating reltuples at all.)
*/
if (metap->hashm_ntuples > tuples_removed)
metap->hashm_ntuples -= tuples_removed;
else
metap->hashm_ntuples = 0;
num_index_tuples = metap->hashm_ntuples;
}
MarkBufferDirty(metabuf);
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
xl_hash_update_meta_page xlrec;
XLogRecPtr recptr;
xlrec.ntuples = metap->hashm_ntuples;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashUpdateMetaPage);
XLogRegisterBuffer(0, metabuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_UPDATE_META_PAGE);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
END_CRIT_SECTION();
_hash_relbuf(rel, metabuf);
/* return statistics */
if (stats == NULL)
stats = (IndexBulkDeleteResult *) palloc0(sizeof(IndexBulkDeleteResult));
stats->estimated_count = false;
stats->num_index_tuples = num_index_tuples;
stats->tuples_removed += tuples_removed;
/* hashvacuumcleanup will fill in num_pages */
return stats;
}
static void
btree_xlog_split(bool onleft, bool isroot, XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_btree_split *xlrec = (xl_btree_split *) XLogRecGetData(record);
bool isleaf = (xlrec->level == 0);
Buffer lbuf;
Buffer rbuf;
Page rpage;
BTPageOpaque ropaque;
char *datapos;
Size datalen;
Item left_hikey = NULL;
Size left_hikeysz = 0;
BlockNumber leftsib;
BlockNumber rightsib;
BlockNumber rnext;
XLogRecGetBlockTag(record, 0, NULL, NULL, &leftsib);
XLogRecGetBlockTag(record, 1, NULL, NULL, &rightsib);
if (!XLogRecGetBlockTag(record, 2, NULL, NULL, &rnext))
rnext = P_NONE;
/*
* Clear the incomplete split flag on the left sibling of the child page
* this is a downlink for. (Like in btree_xlog_insert, this can be done
* before locking the other pages)
*/
if (!isleaf)
_bt_clear_incomplete_split(record, 3);
/* Reconstruct right (new) sibling page from scratch */
rbuf = XLogInitBufferForRedo(record, 1);
datapos = XLogRecGetBlockData(record, 1, &datalen);
rpage = (Page) BufferGetPage(rbuf);
_bt_pageinit(rpage, BufferGetPageSize(rbuf));
ropaque = (BTPageOpaque) PageGetSpecialPointer(rpage);
ropaque->btpo_prev = leftsib;
ropaque->btpo_next = rnext;
ropaque->btpo.level = xlrec->level;
ropaque->btpo_flags = isleaf ? BTP_LEAF : 0;
ropaque->btpo_cycleid = 0;
_bt_restore_page(rpage, datapos, datalen);
/*
* On leaf level, the high key of the left page is equal to the first key
* on the right page.
*/
if (isleaf)
{
ItemId hiItemId = PageGetItemId(rpage, P_FIRSTDATAKEY(ropaque));
left_hikey = PageGetItem(rpage, hiItemId);
left_hikeysz = ItemIdGetLength(hiItemId);
}
PageSetLSN(rpage, lsn);
MarkBufferDirty(rbuf);
/* don't release the buffer yet; we touch right page's first item below */
/* Now reconstruct left (original) sibling page */
if (XLogReadBufferForRedo(record, 0, &lbuf) == BLK_NEEDS_REDO)
{
/*
* To retain the same physical order of the tuples that they had, we
* initialize a temporary empty page for the left page and add all the
* items to that in item number order. This mirrors how _bt_split()
* works. It's not strictly required to retain the same physical
* order, as long as the items are in the correct item number order,
* but it helps debugging. See also _bt_restore_page(), which does
* the same for the right page.
*/
Page lpage = (Page) BufferGetPage(lbuf);
BTPageOpaque lopaque = (BTPageOpaque) PageGetSpecialPointer(lpage);
OffsetNumber off;
Item newitem = NULL;
Size newitemsz = 0;
Page newlpage;
OffsetNumber leftoff;
datapos = XLogRecGetBlockData(record, 0, &datalen);
if (onleft)
{
newitem = (Item) datapos;
newitemsz = MAXALIGN(IndexTupleSize(newitem));
datapos += newitemsz;
datalen -= newitemsz;
}
/* Extract left hikey and its size (assuming 16-bit alignment) */
if (!isleaf)
{
left_hikey = (Item) datapos;
left_hikeysz = MAXALIGN(IndexTupleSize(left_hikey));
datapos += left_hikeysz;
datalen -= left_hikeysz;
}
Assert(datalen == 0);
newlpage = PageGetTempPageCopySpecial(lpage);
/* Set high key */
leftoff = P_HIKEY;
if (PageAddItem(newlpage, left_hikey, left_hikeysz,
P_HIKEY, false, false) == InvalidOffsetNumber)
elog(PANIC, "failed to add high key to left page after split");
leftoff = OffsetNumberNext(leftoff);
for (off = P_FIRSTDATAKEY(lopaque); off < xlrec->firstright; off++)
{
ItemId itemid;
Size itemsz;
Item item;
/* add the new item if it was inserted on left page */
if (onleft && off == xlrec->newitemoff)
{
if (PageAddItem(newlpage, newitem, newitemsz, leftoff,
false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add new item to left page after split");
leftoff = OffsetNumberNext(leftoff);
}
itemid = PageGetItemId(lpage, off);
itemsz = ItemIdGetLength(itemid);
item = PageGetItem(lpage, itemid);
if (PageAddItem(newlpage, item, itemsz, leftoff,
false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add old item to left page after split");
leftoff = OffsetNumberNext(leftoff);
}
/* cope with possibility that newitem goes at the end */
if (onleft && off == xlrec->newitemoff)
{
if (PageAddItem(newlpage, newitem, newitemsz, leftoff,
false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add new item to left page after split");
leftoff = OffsetNumberNext(leftoff);
}
PageRestoreTempPage(newlpage, lpage);
/* Fix opaque fields */
lopaque->btpo_flags = BTP_INCOMPLETE_SPLIT;
if (isleaf)
lopaque->btpo_flags |= BTP_LEAF;
lopaque->btpo_next = rightsib;
lopaque->btpo_cycleid = 0;
PageSetLSN(lpage, lsn);
MarkBufferDirty(lbuf);
}
/* We no longer need the buffers */
if (BufferIsValid(lbuf))
UnlockReleaseBuffer(lbuf);
UnlockReleaseBuffer(rbuf);
/*
* Fix left-link of the page to the right of the new right sibling.
*
* Note: in normal operation, we do this while still holding lock on the
* two split pages. However, that's not necessary for correctness in WAL
* replay, because no other index update can be in progress, and readers
* will cope properly when following an obsolete left-link.
*/
if (rnext != P_NONE)
{
Buffer buffer;
if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
{
Page page = (Page) BufferGetPage(buffer);
BTPageOpaque pageop = (BTPageOpaque) PageGetSpecialPointer(page);
pageop->btpo_prev = rightsib;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
}
}
static void
btree_xlog_mark_page_halfdead(uint8 info, XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_btree_mark_page_halfdead *xlrec = (xl_btree_mark_page_halfdead *) XLogRecGetData(record);
Buffer buffer;
Page page;
BTPageOpaque pageop;
IndexTupleData trunctuple;
/*
* In normal operation, we would lock all the pages this WAL record
* touches before changing any of them. In WAL replay, it should be okay
* to lock just one page at a time, since no concurrent index updates can
* be happening, and readers should not care whether they arrive at the
* target page or not (since it's surely empty).
*/
/* parent page */
if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
OffsetNumber poffset;
ItemId itemid;
IndexTuple itup;
OffsetNumber nextoffset;
BlockNumber rightsib;
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
poffset = xlrec->poffset;
nextoffset = OffsetNumberNext(poffset);
itemid = PageGetItemId(page, nextoffset);
itup = (IndexTuple) PageGetItem(page, itemid);
rightsib = ItemPointerGetBlockNumber(&itup->t_tid);
itemid = PageGetItemId(page, poffset);
itup = (IndexTuple) PageGetItem(page, itemid);
ItemPointerSet(&(itup->t_tid), rightsib, P_HIKEY);
nextoffset = OffsetNumberNext(poffset);
PageIndexTupleDelete(page, nextoffset);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
/* Rewrite the leaf page as a halfdead page */
buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
_bt_pageinit(page, BufferGetPageSize(buffer));
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
pageop->btpo_prev = xlrec->leftblk;
pageop->btpo_next = xlrec->rightblk;
pageop->btpo.level = 0;
pageop->btpo_flags = BTP_HALF_DEAD | BTP_LEAF;
pageop->btpo_cycleid = 0;
/*
* Construct a dummy hikey item that points to the next parent to be
* deleted (if any).
*/
MemSet(&trunctuple, 0, sizeof(IndexTupleData));
trunctuple.t_info = sizeof(IndexTupleData);
if (xlrec->topparent != InvalidBlockNumber)
ItemPointerSet(&trunctuple.t_tid, xlrec->topparent, P_HIKEY);
else
ItemPointerSetInvalid(&trunctuple.t_tid);
if (PageAddItem(page, (Item) &trunctuple, sizeof(IndexTupleData), P_HIKEY,
false, false) == InvalidOffsetNumber)
elog(ERROR, "could not add dummy high key to half-dead page");
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
Datum
bt_page_items(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_P(0);
uint32 blkno = PG_GETARG_UINT32(1);
Datum result;
char *values[6];
HeapTuple tuple;
FuncCallContext *fctx;
MemoryContext mctx;
struct user_args *uargs;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pageinspect functions"))));
if (SRF_IS_FIRSTCALL())
{
RangeVar *relrv;
Relation rel;
Buffer buffer;
BTPageOpaque opaque;
TupleDesc tupleDesc;
fctx = SRF_FIRSTCALL_INIT();
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = relation_openrv(relrv, AccessShareLock);
if (!IS_INDEX(rel) || !IS_BTREE(rel))
elog(ERROR, "relation \"%s\" is not a btree index",
RelationGetRelationName(rel));
/*
* Reject attempts to read non-local temporary relations; we would be
* likely to get wrong data since we have no visibility into the
* owning session's local buffers.
*/
if (RELATION_IS_OTHER_TEMP(rel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
if (blkno == 0)
elog(ERROR, "block 0 is a meta page");
CHECK_RELATION_BLOCK_RANGE(rel, blkno);
buffer = ReadBuffer(rel, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
/*
* We copy the page into local storage to avoid holding pin on the
* buffer longer than we must, and possibly failing to release it at
* all if the calling query doesn't fetch all rows.
*/
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
uargs = palloc(sizeof(struct user_args));
uargs->page = palloc(BLCKSZ);
memcpy(uargs->page, BufferGetPage(buffer), BLCKSZ);
UnlockReleaseBuffer(buffer);
relation_close(rel, AccessShareLock);
uargs->offset = FirstOffsetNumber;
opaque = (BTPageOpaque) PageGetSpecialPointer(uargs->page);
if (P_ISDELETED(opaque))
elog(NOTICE, "page is deleted");
fctx->max_calls = PageGetMaxOffsetNumber(uargs->page);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupleDesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
fctx->attinmeta = TupleDescGetAttInMetadata(tupleDesc);
fctx->user_fctx = uargs;
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
uargs = fctx->user_fctx;
if (fctx->call_cntr < fctx->max_calls)
{
ItemId id;
IndexTuple itup;
int j;
int off;
int dlen;
char *dump;
char *ptr;
id = PageGetItemId(uargs->page, uargs->offset);
if (!ItemIdIsValid(id))
elog(ERROR, "invalid ItemId");
itup = (IndexTuple) PageGetItem(uargs->page, id);
j = 0;
values[j++] = psprintf("%d", uargs->offset);
values[j++] = psprintf("(%u,%u)",
BlockIdGetBlockNumber(&(itup->t_tid.ip_blkid)),
itup->t_tid.ip_posid);
values[j++] = psprintf("%d", (int) IndexTupleSize(itup));
values[j++] = psprintf("%c", IndexTupleHasNulls(itup) ? 't' : 'f');
values[j++] = psprintf("%c", IndexTupleHasVarwidths(itup) ? 't' : 'f');
ptr = (char *) itup + IndexInfoFindDataOffset(itup->t_info);
dlen = IndexTupleSize(itup) - IndexInfoFindDataOffset(itup->t_info);
dump = palloc0(dlen * 3 + 1);
values[j] = dump;
for (off = 0; off < dlen; off++)
{
if (off > 0)
*dump++ = ' ';
sprintf(dump, "%02x", *(ptr + off) & 0xff);
dump += 2;
}
tuple = BuildTupleFromCStrings(fctx->attinmeta, values);
result = HeapTupleGetDatum(tuple);
uargs->offset = uargs->offset + 1;
SRF_RETURN_NEXT(fctx, result);
}
else
{
pfree(uargs->page);
pfree(uargs);
SRF_RETURN_DONE(fctx);
}
}
Datum
bt_page_items_bytea(PG_FUNCTION_ARGS)
{
bytea *raw_page = PG_GETARG_BYTEA_P(0);
Datum result;
FuncCallContext *fctx;
struct user_args *uargs;
int raw_page_size;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pageinspect functions"))));
if (SRF_IS_FIRSTCALL())
{
BTPageOpaque opaque;
MemoryContext mctx;
TupleDesc tupleDesc;
raw_page_size = VARSIZE(raw_page) - VARHDRSZ;
if (raw_page_size < SizeOfPageHeaderData)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("input page too small (%d bytes)", raw_page_size)));
fctx = SRF_FIRSTCALL_INIT();
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
uargs = palloc(sizeof(struct user_args));
uargs->page = VARDATA(raw_page);
uargs->offset = FirstOffsetNumber;
opaque = (BTPageOpaque) PageGetSpecialPointer(uargs->page);
if (P_ISMETA(opaque))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("block is a meta page")));
if (P_ISDELETED(opaque))
elog(NOTICE, "page is deleted");
fctx->max_calls = PageGetMaxOffsetNumber(uargs->page);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupleDesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
fctx->attinmeta = TupleDescGetAttInMetadata(tupleDesc);
fctx->user_fctx = uargs;
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
uargs = fctx->user_fctx;
if (fctx->call_cntr < fctx->max_calls)
{
result = bt_page_print_tuples(fctx, uargs->page, uargs->offset);
uargs->offset++;
SRF_RETURN_NEXT(fctx, result);
}
else
{
pfree(uargs);
SRF_RETURN_DONE(fctx);
}
}
/*
* _hash_metapinit() -- Initialize the metadata page of a hash index,
* the two buckets that we begin with and the initial
* bitmap page.
*
* We are fairly cavalier about locking here, since we know that no one else
* could be accessing this index. In particular the rule about not holding
* multiple buffer locks is ignored.
*/
void
_hash_metapinit(Relation rel)
{
HashMetaPage metap;
HashPageOpaque pageopaque;
Buffer metabuf;
Buffer buf;
Page pg;
int32 data_width;
int32 item_width;
int32 ffactor;
uint16 i;
/* safety check */
if (RelationGetNumberOfBlocks(rel) != 0)
elog(ERROR, "cannot initialize non-empty hash index \"%s\"",
RelationGetRelationName(rel));
/*
* Determine the target fill factor (tuples per bucket) for this index.
* The idea is to make the fill factor correspond to pages about 3/4ths
* full. We can compute it exactly if the index datatype is fixed-width,
* but for var-width there's some guessing involved.
*/
data_width = get_typavgwidth(RelationGetDescr(rel)->attrs[0]->atttypid,
RelationGetDescr(rel)->attrs[0]->atttypmod);
item_width = MAXALIGN(sizeof(HashItemData)) + MAXALIGN(data_width) +
sizeof(ItemIdData); /* include the line pointer */
ffactor = (BLCKSZ * 3 / 4) / item_width;
/* keep to a sane range */
if (ffactor < 10)
ffactor = 10;
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_WRITE);
pg = BufferGetPage(metabuf);
_hash_pageinit(pg, BufferGetPageSize(metabuf));
pageopaque = (HashPageOpaque) PageGetSpecialPointer(pg);
pageopaque->hasho_prevblkno = InvalidBlockNumber;
pageopaque->hasho_nextblkno = InvalidBlockNumber;
pageopaque->hasho_bucket = -1;
pageopaque->hasho_flag = LH_META_PAGE;
pageopaque->hasho_filler = HASHO_FILL;
metap = (HashMetaPage) pg;
metap->hashm_magic = HASH_MAGIC;
metap->hashm_version = HASH_VERSION;
metap->hashm_ntuples = 0;
metap->hashm_nmaps = 0;
metap->hashm_ffactor = ffactor;
metap->hashm_bsize = BufferGetPageSize(metabuf);
/* find largest bitmap array size that will fit in page size */
for (i = _hash_log2(metap->hashm_bsize); i > 0; --i)
{
if ((1 << i) <= (metap->hashm_bsize -
(MAXALIGN(sizeof(PageHeaderData)) +
MAXALIGN(sizeof(HashPageOpaqueData)))))
break;
}
Assert(i > 0);
metap->hashm_bmsize = 1 << i;
metap->hashm_bmshift = i + BYTE_TO_BIT;
Assert((1 << BMPG_SHIFT(metap)) == (BMPG_MASK(metap) + 1));
metap->hashm_procid = index_getprocid(rel, 1, HASHPROC);
/*
* We initialize the index with two buckets, 0 and 1, occupying physical
* blocks 1 and 2. The first freespace bitmap page is in block 3.
*/
metap->hashm_maxbucket = metap->hashm_lowmask = 1; /* nbuckets - 1 */
metap->hashm_highmask = 3; /* (nbuckets << 1) - 1 */
MemSet((char *) metap->hashm_spares, 0, sizeof(metap->hashm_spares));
MemSet((char *) metap->hashm_mapp, 0, sizeof(metap->hashm_mapp));
metap->hashm_spares[1] = 1; /* the first bitmap page is only spare */
metap->hashm_ovflpoint = 1;
metap->hashm_firstfree = 0;
/*
* Initialize the first two buckets
*/
for (i = 0; i <= 1; i++)
{
buf = _hash_getbuf(rel, BUCKET_TO_BLKNO(metap, i), HASH_WRITE);
pg = BufferGetPage(buf);
_hash_pageinit(pg, BufferGetPageSize(buf));
pageopaque = (HashPageOpaque) PageGetSpecialPointer(pg);
pageopaque->hasho_prevblkno = InvalidBlockNumber;
pageopaque->hasho_nextblkno = InvalidBlockNumber;
pageopaque->hasho_bucket = i;
pageopaque->hasho_flag = LH_BUCKET_PAGE;
pageopaque->hasho_filler = HASHO_FILL;
_hash_wrtbuf(rel, buf);
}
/*
* Initialize first bitmap page. Can't do this until we
* create the first two buckets, else smgr will complain.
*/
_hash_initbitmap(rel, metap, 3);
/* all done */
_hash_wrtbuf(rel, metabuf);
}
/*
* _hash_splitbucket -- split 'obucket' into 'obucket' and 'nbucket'
*
* We are splitting a bucket that consists of a base bucket page and zero
* or more overflow (bucket chain) pages. We must relocate tuples that
* belong in the new bucket, and compress out any free space in the old
* bucket.
*
* The caller must hold exclusive locks on both buckets to ensure that
* no one else is trying to access them (see README).
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state. (The metapage is only
* touched if it becomes necessary to add or remove overflow pages.)
*/
static void
_hash_splitbucket(Relation rel,
Buffer metabuf,
Bucket obucket,
Bucket nbucket,
BlockNumber start_oblkno,
BlockNumber start_nblkno,
uint32 maxbucket,
uint32 highmask,
uint32 lowmask)
{
Bucket bucket;
Buffer obuf;
Buffer nbuf;
BlockNumber oblkno;
BlockNumber nblkno;
bool null;
Datum datum;
HashItem hitem;
HashPageOpaque oopaque;
HashPageOpaque nopaque;
IndexTuple itup;
Size itemsz;
OffsetNumber ooffnum;
OffsetNumber noffnum;
OffsetNumber omaxoffnum;
Page opage;
Page npage;
TupleDesc itupdesc = RelationGetDescr(rel);
/*
* It should be okay to simultaneously write-lock pages from each
* bucket, since no one else can be trying to acquire buffer lock
* on pages of either bucket.
*/
oblkno = start_oblkno;
nblkno = start_nblkno;
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE);
nbuf = _hash_getbuf(rel, nblkno, HASH_WRITE);
opage = BufferGetPage(obuf);
npage = BufferGetPage(nbuf);
_hash_checkpage(rel, opage, LH_BUCKET_PAGE);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
/* initialize the new bucket's primary page */
_hash_pageinit(npage, BufferGetPageSize(nbuf));
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
nopaque->hasho_prevblkno = InvalidBlockNumber;
nopaque->hasho_nextblkno = InvalidBlockNumber;
nopaque->hasho_bucket = nbucket;
nopaque->hasho_flag = LH_BUCKET_PAGE;
nopaque->hasho_filler = HASHO_FILL;
/*
* Partition the tuples in the old bucket between the old bucket and the
* new bucket, advancing along the old bucket's overflow bucket chain
* and adding overflow pages to the new bucket as needed.
*/
ooffnum = FirstOffsetNumber;
omaxoffnum = PageGetMaxOffsetNumber(opage);
for (;;)
{
/*
* at each iteration through this loop, each of these variables
* should be up-to-date: obuf opage oopaque ooffnum omaxoffnum
*/
/* check if we're at the end of the page */
if (ooffnum > omaxoffnum)
{
/* at end of page, but check for an(other) overflow page */
oblkno = oopaque->hasho_nextblkno;
if (!BlockNumberIsValid(oblkno))
break;
/*
* we ran out of tuples on this particular page, but we
* have more overflow pages; advance to next page.
*/
_hash_wrtbuf(rel, obuf);
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE);
opage = BufferGetPage(obuf);
_hash_checkpage(rel, opage, LH_OVERFLOW_PAGE);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
ooffnum = FirstOffsetNumber;
omaxoffnum = PageGetMaxOffsetNumber(opage);
continue;
}
/*
* Re-hash the tuple to determine which bucket it now belongs in.
*
* It is annoying to call the hash function while holding locks,
* but releasing and relocking the page for each tuple is unappealing
* too.
*/
hitem = (HashItem) PageGetItem(opage, PageGetItemId(opage, ooffnum));
itup = &(hitem->hash_itup);
datum = index_getattr(itup, 1, itupdesc, &null);
Assert(!null);
bucket = _hash_hashkey2bucket(_hash_datum2hashkey(rel, datum),
maxbucket, highmask, lowmask);
if (bucket == nbucket)
{
/*
* insert the tuple into the new bucket. if it doesn't fit on
* the current page in the new bucket, we must allocate a new
* overflow page and place the tuple on that page instead.
*/
itemsz = IndexTupleDSize(hitem->hash_itup)
+ (sizeof(HashItemData) - sizeof(IndexTupleData));
itemsz = MAXALIGN(itemsz);
if (PageGetFreeSpace(npage) < itemsz)
{
/* write out nbuf and drop lock, but keep pin */
_hash_chgbufaccess(rel, nbuf, HASH_WRITE, HASH_NOLOCK);
/* chain to a new overflow page */
nbuf = _hash_addovflpage(rel, metabuf, nbuf);
npage = BufferGetPage(nbuf);
_hash_checkpage(rel, npage, LH_OVERFLOW_PAGE);
/* we don't need nopaque within the loop */
}
noffnum = OffsetNumberNext(PageGetMaxOffsetNumber(npage));
if (PageAddItem(npage, (Item) hitem, itemsz, noffnum, LP_USED)
== InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(rel));
/*
* now delete the tuple from the old bucket. after this
* section of code, 'ooffnum' will actually point to the
* ItemId to which we would point if we had advanced it before
* the deletion (PageIndexTupleDelete repacks the ItemId
* array). this also means that 'omaxoffnum' is exactly one
* less than it used to be, so we really can just decrement it
* instead of calling PageGetMaxOffsetNumber.
*/
PageIndexTupleDelete(opage, ooffnum);
omaxoffnum = OffsetNumberPrev(omaxoffnum);
}
else
{
/*
* the tuple stays on this page. we didn't move anything, so
* we didn't delete anything and therefore we don't have to
* change 'omaxoffnum'.
*/
Assert(bucket == obucket);
ooffnum = OffsetNumberNext(ooffnum);
}
}
/*
* We're at the end of the old bucket chain, so we're done partitioning
* the tuples. Before quitting, call _hash_squeezebucket to ensure the
* tuples remaining in the old bucket (including the overflow pages) are
* packed as tightly as possible. The new bucket is already tight.
*/
_hash_wrtbuf(rel, obuf);
_hash_wrtbuf(rel, nbuf);
_hash_squeezebucket(rel, obucket, start_oblkno);
}
static void
rtdoinsert(Relation r, IndexTuple itup, RTSTATE *rtstate)
{
Page page;
Buffer buffer;
BlockNumber blk;
IndexTuple which;
OffsetNumber l;
RTSTACK *stack;
RTreePageOpaque opaque;
Datum datum;
blk = P_ROOT;
buffer = InvalidBuffer;
stack = NULL;
do
{
/* release the current buffer, read in the next one */
buffer = ReleaseAndReadBuffer(buffer, r, blk);
page = (Page) BufferGetPage(buffer);
opaque = (RTreePageOpaque) PageGetSpecialPointer(page);
if (!(opaque->flags & F_LEAF))
{
RTSTACK *n;
ItemId iid;
n = (RTSTACK *) palloc(sizeof(RTSTACK));
n->rts_parent = stack;
n->rts_blk = blk;
n->rts_child = choose(r, page, itup, rtstate);
stack = n;
iid = PageGetItemId(page, n->rts_child);
which = (IndexTuple) PageGetItem(page, iid);
blk = ItemPointerGetBlockNumber(&(which->t_tid));
}
} while (!(opaque->flags & F_LEAF));
if (nospace(page, itup))
{
/* need to do a split */
rtdosplit(r, buffer, stack, itup, rtstate);
freestack(stack);
WriteBuffer(buffer); /* don't forget to release buffer! */
return;
}
/* add the item and write the buffer */
if (PageIsEmpty(page))
{
l = PageAddItem(page, (Item) itup, IndexTupleSize(itup),
FirstOffsetNumber,
LP_USED);
}
else
{
l = PageAddItem(page, (Item) itup, IndexTupleSize(itup),
OffsetNumberNext(PageGetMaxOffsetNumber(page)),
LP_USED);
}
if (l == InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(r));
WriteBuffer(buffer);
datum = IndexTupleGetDatum(itup);
/* now expand the page boundary in the parent to include the new child */
rttighten(r, stack, datum, IndexTupleAttSize(itup), rtstate);
freestack(stack);
}
/*
* rtdosplit -- split a page in the tree.
*
* rtpicksplit does the interesting work of choosing the split.
* This routine just does the bit-pushing.
*/
static void
rtdosplit(Relation r,
Buffer buffer,
RTSTACK *stack,
IndexTuple itup,
RTSTATE *rtstate)
{
Page p;
Buffer leftbuf,
rightbuf;
Page left,
right;
ItemId itemid;
IndexTuple item;
IndexTuple ltup,
rtup;
OffsetNumber maxoff;
OffsetNumber i;
OffsetNumber leftoff,
rightoff;
BlockNumber lbknum,
rbknum;
BlockNumber bufblock;
RTreePageOpaque opaque;
bool *isnull;
SPLITVEC v;
OffsetNumber *spl_left,
*spl_right;
TupleDesc tupDesc;
int n;
OffsetNumber newitemoff;
p = (Page) BufferGetPage(buffer);
opaque = (RTreePageOpaque) PageGetSpecialPointer(p);
rtpicksplit(r, p, &v, itup, rtstate);
/*
* The root of the tree is the first block in the relation. If we're
* about to split the root, we need to do some hocus-pocus to enforce this
* guarantee.
*/
if (BufferGetBlockNumber(buffer) == P_ROOT)
{
leftbuf = ReadBuffer(r, P_NEW);
RTInitBuffer(leftbuf, opaque->flags);
lbknum = BufferGetBlockNumber(leftbuf);
left = (Page) BufferGetPage(leftbuf);
}
else
{
leftbuf = buffer;
IncrBufferRefCount(buffer);
lbknum = BufferGetBlockNumber(buffer);
left = (Page) PageGetTempPage(p, sizeof(RTreePageOpaqueData));
}
rightbuf = ReadBuffer(r, P_NEW);
RTInitBuffer(rightbuf, opaque->flags);
rbknum = BufferGetBlockNumber(rightbuf);
right = (Page) BufferGetPage(rightbuf);
spl_left = v.spl_left;
spl_right = v.spl_right;
leftoff = rightoff = FirstOffsetNumber;
maxoff = PageGetMaxOffsetNumber(p);
newitemoff = OffsetNumberNext(maxoff);
/*
* spl_left contains a list of the offset numbers of the tuples that will
* go to the left page. For each offset number, get the tuple item, then
* add the item to the left page. Similarly for the right side.
*/
/* fill left node */
for (n = 0; n < v.spl_nleft; n++)
{
i = *spl_left;
if (i == newitemoff)
item = itup;
else
{
itemid = PageGetItemId(p, i);
item = (IndexTuple) PageGetItem(p, itemid);
}
if (PageAddItem(left, (Item) item, IndexTupleSize(item),
leftoff, LP_USED) == InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(r));
leftoff = OffsetNumberNext(leftoff);
spl_left++; /* advance in left split vector */
}
/* fill right node */
for (n = 0; n < v.spl_nright; n++)
{
i = *spl_right;
if (i == newitemoff)
item = itup;
else
{
itemid = PageGetItemId(p, i);
item = (IndexTuple) PageGetItem(p, itemid);
}
if (PageAddItem(right, (Item) item, IndexTupleSize(item),
rightoff, LP_USED) == InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(r));
rightoff = OffsetNumberNext(rightoff);
spl_right++; /* advance in right split vector */
}
/* Make sure we consumed all of the split vectors, and release 'em */
Assert(*spl_left == InvalidOffsetNumber);
Assert(*spl_right == InvalidOffsetNumber);
pfree(v.spl_left);
pfree(v.spl_right);
if ((bufblock = BufferGetBlockNumber(buffer)) != P_ROOT)
PageRestoreTempPage(left, p);
WriteBuffer(leftbuf);
WriteBuffer(rightbuf);
/*
* Okay, the page is split. We have three things left to do:
*
* 1) Adjust any active scans on this index to cope with changes we
* introduced in its structure by splitting this page.
*
* 2) "Tighten" the bounding box of the pointer to the left page in the
* parent node in the tree, if any. Since we moved a bunch of stuff off
* the left page, we expect it to get smaller. This happens in the
* internal insertion routine.
*
* 3) Insert a pointer to the right page in the parent. This may cause
* the parent to split. If it does, we need to repeat steps one and two
* for each split node in the tree.
*/
/* adjust active scans */
rtadjscans(r, RTOP_SPLIT, bufblock, FirstOffsetNumber);
tupDesc = r->rd_att;
isnull = (bool *) palloc(r->rd_rel->relnatts * sizeof(bool));
memset(isnull, false, r->rd_rel->relnatts * sizeof(bool));
ltup = index_form_tuple(tupDesc, &(v.spl_ldatum), isnull);
rtup = index_form_tuple(tupDesc, &(v.spl_rdatum), isnull);
pfree(isnull);
pfree(DatumGetPointer(v.spl_ldatum));
pfree(DatumGetPointer(v.spl_rdatum));
/* set pointers to new child pages in the internal index tuples */
ItemPointerSet(&(ltup->t_tid), lbknum, 1);
ItemPointerSet(&(rtup->t_tid), rbknum, 1);
rtintinsert(r, stack, ltup, rtup, rtstate);
pfree(ltup);
pfree(rtup);
}
/*
* replay delete operation of hash index
*/
static void
hash_xlog_delete(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_delete *xldata = (xl_hash_delete *) XLogRecGetData(record);
Buffer bucketbuf = InvalidBuffer;
Buffer deletebuf;
Page page;
XLogRedoAction action;
/*
* Ensure we have a cleanup lock on primary bucket page before we start
* with the actual replay operation. This is to ensure that neither a
* scan can start nor a scan can be already-in-progress during the replay
* of this operation. If we allow scans during this operation, then they
* can miss some records or show the same record multiple times.
*/
if (xldata->is_primary_bucket_page)
action = XLogReadBufferForRedoExtended(record, 1, RBM_NORMAL, true, &deletebuf);
else
{
/*
* we don't care for return value as the purpose of reading bucketbuf
* is to ensure a cleanup lock on primary bucket page.
*/
(void) XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &bucketbuf);
action = XLogReadBufferForRedo(record, 1, &deletebuf);
}
/* replay the record for deleting entries in bucket page */
if (action == BLK_NEEDS_REDO)
{
char *ptr;
Size len;
ptr = XLogRecGetBlockData(record, 1, &len);
page = (Page) BufferGetPage(deletebuf);
if (len > 0)
{
OffsetNumber *unused;
OffsetNumber *unend;
unused = (OffsetNumber *) ptr;
unend = (OffsetNumber *) ((char *) ptr + len);
if ((unend - unused) > 0)
PageIndexMultiDelete(page, unused, unend - unused);
}
/*
* Mark the page as not containing any LP_DEAD items only if
* clear_dead_marking flag is set to true. See comments in
* hashbucketcleanup() for details.
*/
if (xldata->clear_dead_marking)
{
HashPageOpaque pageopaque;
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
pageopaque->hasho_flag &= ~LH_PAGE_HAS_DEAD_TUPLES;
}
PageSetLSN(page, lsn);
MarkBufferDirty(deletebuf);
}
if (BufferIsValid(deletebuf))
UnlockReleaseBuffer(deletebuf);
if (BufferIsValid(bucketbuf))
UnlockReleaseBuffer(bucketbuf);
}
/*
* replay squeeze page operation of hash index
*/
static void
hash_xlog_squeeze_page(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_squeeze_page *xldata = (xl_hash_squeeze_page *) XLogRecGetData(record);
Buffer bucketbuf = InvalidBuffer;
Buffer writebuf;
Buffer ovflbuf;
Buffer prevbuf = InvalidBuffer;
Buffer mapbuf;
XLogRedoAction action;
/*
* Ensure we have a cleanup lock on primary bucket page before we start
* with the actual replay operation. This is to ensure that neither a
* scan can start nor a scan can be already-in-progress during the replay
* of this operation. If we allow scans during this operation, then they
* can miss some records or show the same record multiple times.
*/
if (xldata->is_prim_bucket_same_wrt)
action = XLogReadBufferForRedoExtended(record, 1, RBM_NORMAL, true, &writebuf);
else
{
/*
* we don't care for return value as the purpose of reading bucketbuf
* is to ensure a cleanup lock on primary bucket page.
*/
(void) XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &bucketbuf);
action = XLogReadBufferForRedo(record, 1, &writebuf);
}
/* replay the record for adding entries in overflow buffer */
if (action == BLK_NEEDS_REDO)
{
Page writepage;
char *begin;
char *data;
Size datalen;
uint16 ninserted = 0;
data = begin = XLogRecGetBlockData(record, 1, &datalen);
writepage = (Page) BufferGetPage(writebuf);
if (xldata->ntups > 0)
{
OffsetNumber *towrite = (OffsetNumber *) data;
data += sizeof(OffsetNumber) * xldata->ntups;
while (data - begin < datalen)
{
IndexTuple itup = (IndexTuple) data;
Size itemsz;
OffsetNumber l;
itemsz = IndexTupleDSize(*itup);
itemsz = MAXALIGN(itemsz);
data += itemsz;
l = PageAddItem(writepage, (Item) itup, itemsz, towrite[ninserted], false, false);
if (l == InvalidOffsetNumber)
elog(ERROR, "hash_xlog_squeeze_page: failed to add item to hash index page, size %d bytes",
(int) itemsz);
ninserted++;
}
}
/*
* number of tuples inserted must be same as requested in REDO record.
*/
Assert(ninserted == xldata->ntups);
/*
* if the page on which are adding tuples is a page previous to freed
* overflow page, then update its nextblno.
*/
if (xldata->is_prev_bucket_same_wrt)
{
HashPageOpaque writeopaque = (HashPageOpaque) PageGetSpecialPointer(writepage);
writeopaque->hasho_nextblkno = xldata->nextblkno;
}
PageSetLSN(writepage, lsn);
MarkBufferDirty(writebuf);
}
/* replay the record for initializing overflow buffer */
if (XLogReadBufferForRedo(record, 2, &ovflbuf) == BLK_NEEDS_REDO)
{
Page ovflpage;
ovflpage = BufferGetPage(ovflbuf);
_hash_pageinit(ovflpage, BufferGetPageSize(ovflbuf));
PageSetLSN(ovflpage, lsn);
MarkBufferDirty(ovflbuf);
}
if (BufferIsValid(ovflbuf))
UnlockReleaseBuffer(ovflbuf);
/* replay the record for page previous to the freed overflow page */
if (!xldata->is_prev_bucket_same_wrt &&
XLogReadBufferForRedo(record, 3, &prevbuf) == BLK_NEEDS_REDO)
{
Page prevpage = BufferGetPage(prevbuf);
HashPageOpaque prevopaque = (HashPageOpaque) PageGetSpecialPointer(prevpage);
prevopaque->hasho_nextblkno = xldata->nextblkno;
PageSetLSN(prevpage, lsn);
MarkBufferDirty(prevbuf);
}
if (BufferIsValid(prevbuf))
UnlockReleaseBuffer(prevbuf);
/* replay the record for page next to the freed overflow page */
if (XLogRecHasBlockRef(record, 4))
{
Buffer nextbuf;
if (XLogReadBufferForRedo(record, 4, &nextbuf) == BLK_NEEDS_REDO)
{
Page nextpage = BufferGetPage(nextbuf);
HashPageOpaque nextopaque = (HashPageOpaque) PageGetSpecialPointer(nextpage);
nextopaque->hasho_prevblkno = xldata->prevblkno;
PageSetLSN(nextpage, lsn);
MarkBufferDirty(nextbuf);
}
if (BufferIsValid(nextbuf))
UnlockReleaseBuffer(nextbuf);
}
if (BufferIsValid(writebuf))
UnlockReleaseBuffer(writebuf);
if (BufferIsValid(bucketbuf))
UnlockReleaseBuffer(bucketbuf);
/*
* Note: in normal operation, we'd update the bitmap and meta page while
* still holding lock on the primary bucket page and overflow pages. But
* during replay it's not necessary to hold those locks, since no other
* index updates can be happening concurrently.
*/
/* replay the record for bitmap page */
if (XLogReadBufferForRedo(record, 5, &mapbuf) == BLK_NEEDS_REDO)
{
Page mappage = (Page) BufferGetPage(mapbuf);
uint32 *freep = NULL;
char *data;
uint32 *bitmap_page_bit;
Size datalen;
freep = HashPageGetBitmap(mappage);
data = XLogRecGetBlockData(record, 5, &datalen);
bitmap_page_bit = (uint32 *) data;
CLRBIT(freep, *bitmap_page_bit);
PageSetLSN(mappage, lsn);
MarkBufferDirty(mapbuf);
}
if (BufferIsValid(mapbuf))
UnlockReleaseBuffer(mapbuf);
/* replay the record for meta page */
if (XLogRecHasBlockRef(record, 6))
{
Buffer metabuf;
if (XLogReadBufferForRedo(record, 6, &metabuf) == BLK_NEEDS_REDO)
{
HashMetaPage metap;
Page page;
char *data;
uint32 *firstfree_ovflpage;
Size datalen;
data = XLogRecGetBlockData(record, 6, &datalen);
firstfree_ovflpage = (uint32 *) data;
page = BufferGetPage(metabuf);
metap = HashPageGetMeta(page);
metap->hashm_firstfree = *firstfree_ovflpage;
PageSetLSN(page, lsn);
MarkBufferDirty(metabuf);
}
if (BufferIsValid(metabuf))
UnlockReleaseBuffer(metabuf);
}
}
/*
* replay allocation of page for split operation
*/
static void
hash_xlog_split_allocate_page(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_split_allocate_page *xlrec = (xl_hash_split_allocate_page *) XLogRecGetData(record);
Buffer oldbuf;
Buffer newbuf;
Buffer metabuf;
Size datalen PG_USED_FOR_ASSERTS_ONLY;
char *data;
XLogRedoAction action;
/*
* To be consistent with normal operation, here we take cleanup locks on
* both the old and new buckets even though there can't be any concurrent
* inserts.
*/
/* replay the record for old bucket */
action = XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &oldbuf);
/*
* Note that we still update the page even if it was restored from a full
* page image, because the special space is not included in the image.
*/
if (action == BLK_NEEDS_REDO || action == BLK_RESTORED)
{
Page oldpage;
HashPageOpaque oldopaque;
oldpage = BufferGetPage(oldbuf);
oldopaque = (HashPageOpaque) PageGetSpecialPointer(oldpage);
oldopaque->hasho_flag = xlrec->old_bucket_flag;
oldopaque->hasho_prevblkno = xlrec->new_bucket;
PageSetLSN(oldpage, lsn);
MarkBufferDirty(oldbuf);
}
/* replay the record for new bucket */
newbuf = XLogInitBufferForRedo(record, 1);
_hash_initbuf(newbuf, xlrec->new_bucket, xlrec->new_bucket,
xlrec->new_bucket_flag, true);
if (!IsBufferCleanupOK(newbuf))
elog(PANIC, "hash_xlog_split_allocate_page: failed to acquire cleanup lock");
MarkBufferDirty(newbuf);
PageSetLSN(BufferGetPage(newbuf), lsn);
/*
* We can release the lock on old bucket early as well but doing here to
* consistent with normal operation.
*/
if (BufferIsValid(oldbuf))
UnlockReleaseBuffer(oldbuf);
if (BufferIsValid(newbuf))
UnlockReleaseBuffer(newbuf);
/*
* Note: in normal operation, we'd update the meta page while still
* holding lock on the old and new bucket pages. But during replay it's
* not necessary to hold those locks, since no other bucket splits can be
* happening concurrently.
*/
/* replay the record for metapage changes */
if (XLogReadBufferForRedo(record, 2, &metabuf) == BLK_NEEDS_REDO)
{
Page page;
HashMetaPage metap;
page = BufferGetPage(metabuf);
metap = HashPageGetMeta(page);
metap->hashm_maxbucket = xlrec->new_bucket;
data = XLogRecGetBlockData(record, 2, &datalen);
if (xlrec->flags & XLH_SPLIT_META_UPDATE_MASKS)
{
uint32 lowmask;
uint32 *highmask;
/* extract low and high masks. */
memcpy(&lowmask, data, sizeof(uint32));
highmask = (uint32 *) ((char *) data + sizeof(uint32));
/* update metapage */
metap->hashm_lowmask = lowmask;
metap->hashm_highmask = *highmask;
data += sizeof(uint32) * 2;
}
if (xlrec->flags & XLH_SPLIT_META_UPDATE_SPLITPOINT)
{
uint32 ovflpoint;
uint32 *ovflpages;
/* extract information of overflow pages. */
memcpy(&ovflpoint, data, sizeof(uint32));
ovflpages = (uint32 *) ((char *) data + sizeof(uint32));
/* update metapage */
metap->hashm_spares[ovflpoint] = *ovflpages;
metap->hashm_ovflpoint = ovflpoint;
}
MarkBufferDirty(metabuf);
PageSetLSN(BufferGetPage(metabuf), lsn);
}
if (BufferIsValid(metabuf))
UnlockReleaseBuffer(metabuf);
}
/*
* _hash_step() -- step to the next valid item in a scan in the bucket.
*
* If no valid record exists in the requested direction, return
* false. Else, return true and set the hashso_curpos for the
* scan to the right thing.
*
* 'bufP' points to the current buffer, which is pinned and read-locked.
* On success exit, we have pin and read-lock on whichever page
* contains the right item; on failure, we have released all buffers.
*/
bool
_hash_step(IndexScanDesc scan, Buffer *bufP, ScanDirection dir)
{
Relation rel = scan->indexRelation;
HashScanOpaque so = (HashScanOpaque) scan->opaque;
ItemPointer current;
Buffer buf;
Page page;
HashPageOpaque opaque;
OffsetNumber maxoff;
OffsetNumber offnum;
BlockNumber blkno;
IndexTuple itup;
current = &(so->hashso_curpos);
buf = *bufP;
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
page = BufferGetPage(buf);
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
/*
* If _hash_step is called from _hash_first, current will not be valid, so
* we can't dereference it. However, in that case, we presumably want to
* start at the beginning/end of the page...
*/
maxoff = PageGetMaxOffsetNumber(page);
if (ItemPointerIsValid(current))
offnum = ItemPointerGetOffsetNumber(current);
else
offnum = InvalidOffsetNumber;
/*
* 'offnum' now points to the last tuple we examined (if any).
*
* continue to step through tuples until: 1) we get to the end of the
* bucket chain or 2) we find a valid tuple.
*/
do
{
switch (dir)
{
case ForwardScanDirection:
if (offnum != InvalidOffsetNumber)
offnum = OffsetNumberNext(offnum); /* move forward */
else
{
/* new page, locate starting position by binary search */
offnum = _hash_binsearch(page, so->hashso_sk_hash);
}
for (;;)
{
/*
* check if we're still in the range of items with the
* target hash key
*/
if (offnum <= maxoff)
{
Assert(offnum >= FirstOffsetNumber);
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
if (so->hashso_sk_hash == _hash_get_indextuple_hashkey(itup))
break; /* yes, so exit for-loop */
}
/*
* ran off the end of this page, try the next
*/
_hash_readnext(rel, &buf, &page, &opaque);
if (BufferIsValid(buf))
{
maxoff = PageGetMaxOffsetNumber(page);
offnum = _hash_binsearch(page, so->hashso_sk_hash);
}
else
{
/* end of bucket */
itup = NULL;
break; /* exit for-loop */
}
}
break;
case BackwardScanDirection:
if (offnum != InvalidOffsetNumber)
offnum = OffsetNumberPrev(offnum); /* move back */
else
{
/* new page, locate starting position by binary search */
offnum = _hash_binsearch_last(page, so->hashso_sk_hash);
}
for (;;)
{
/*
* check if we're still in the range of items with the
* target hash key
*/
if (offnum >= FirstOffsetNumber)
{
Assert(offnum <= maxoff);
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
if (so->hashso_sk_hash == _hash_get_indextuple_hashkey(itup))
break; /* yes, so exit for-loop */
}
/*
* ran off the end of this page, try the next
*/
_hash_readprev(rel, &buf, &page, &opaque);
if (BufferIsValid(buf))
{
maxoff = PageGetMaxOffsetNumber(page);
offnum = _hash_binsearch_last(page, so->hashso_sk_hash);
}
else
{
/* end of bucket */
itup = NULL;
break; /* exit for-loop */
}
}
break;
default:
/* NoMovementScanDirection */
/* this should not be reached */
itup = NULL;
break;
}
if (itup == NULL)
{
/* we ran off the end of the bucket without finding a match */
*bufP = so->hashso_curbuf = InvalidBuffer;
ItemPointerSetInvalid(current);
return false;
}
/* check the tuple quals, loop around if not met */
} while (!_hash_checkqual(scan, itup));
/* if we made it to here, we've found a valid tuple */
blkno = BufferGetBlockNumber(buf);
*bufP = so->hashso_curbuf = buf;
ItemPointerSet(current, blkno, offnum);
return true;
}
/*
* btvacuumpage --- VACUUM one page
*
* This processes a single page for btvacuumscan(). In some cases we
* must go back and re-examine previously-scanned pages; this routine
* recurses when necessary to handle that case.
*
* blkno is the page to process. orig_blkno is the highest block number
* reached by the outer btvacuumscan loop (the same as blkno, unless we
* are recursing to re-examine a previous page).
*/
static void
btvacuumpage(BTVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)
{
IndexVacuumInfo *info = vstate->info;
IndexBulkDeleteResult *stats = vstate->stats;
IndexBulkDeleteCallback callback = vstate->callback;
void *callback_state = vstate->callback_state;
Relation rel = info->index;
bool delete_now;
BlockNumber recurse_to;
Buffer buf;
Page page;
BTPageOpaque opaque;
restart:
delete_now = false;
recurse_to = P_NONE;
/* call vacuum_delay_point while not holding any buffer lock */
vacuum_delay_point();
/*
* We can't use _bt_getbuf() here because it always applies
* _bt_checkpage(), which will barf on an all-zero page. We want to
* recycle all-zero pages, not fail. Also, we want to use a nondefault
* buffer access strategy.
*/
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
info->strategy);
LockBuffer(buf, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (!PageIsNew(page))
_bt_checkpage(rel, buf);
/*
* If we are recursing, the only case we want to do anything with is a
* live leaf page having the current vacuum cycle ID. Any other state
* implies we already saw the page (eg, deleted it as being empty).
*/
if (blkno != orig_blkno)
{
if (_bt_page_recyclable(page) ||
P_IGNORE(opaque) ||
!P_ISLEAF(opaque) ||
opaque->btpo_cycleid != vstate->cycleid)
{
_bt_relbuf(rel, buf);
return;
}
}
/* If the page is in use, update lastUsedPage */
if (!_bt_page_recyclable(page) && vstate->lastUsedPage < blkno)
vstate->lastUsedPage = blkno;
/* Page is valid, see what to do with it */
if (_bt_page_recyclable(page))
{
/* Okay to recycle this page */
RecordFreeIndexPage(rel, blkno);
vstate->totFreePages++;
stats->pages_deleted++;
}
else if (P_ISDELETED(opaque))
{
/* Already deleted, but can't recycle yet */
stats->pages_deleted++;
}
else if (P_ISHALFDEAD(opaque))
{
/* Half-dead, try to delete */
delete_now = true;
}
else if (P_ISLEAF(opaque))
{
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable;
OffsetNumber offnum,
minoff,
maxoff;
/*
* Trade in the initial read lock for a super-exclusive write lock on
* this page. We must get such a lock on every leaf page over the
* course of the vacuum scan, whether or not it actually contains any
* deletable tuples --- see nbtree/README.
*/
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBufferForCleanup(buf);
/*
* Check whether we need to recurse back to earlier pages. What we
* are concerned about is a page split that happened since we started
* the vacuum scan. If the split moved some tuples to a lower page
* then we might have missed 'em. If so, set up for tail recursion.
* (Must do this before possibly clearing btpo_cycleid below!)
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid &&
!(opaque->btpo_flags & BTP_SPLIT_END) &&
!P_RIGHTMOST(opaque) &&
opaque->btpo_next < orig_blkno)
recurse_to = opaque->btpo_next;
/*
* Scan over all items to see which ones need deleted according to the
* callback function.
*/
ndeletable = 0;
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(page);
if (callback)
{
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
IndexTuple itup;
ItemPointer htup;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offnum));
htup = &(itup->t_tid);
/*
* During Hot Standby we currently assume that
* XLOG_BTREE_VACUUM records do not produce conflicts. That is
* only true as long as the callback function depends only
* upon whether the index tuple refers to heap tuples removed
* in the initial heap scan. When vacuum starts it derives a
* value of OldestXmin. Backends taking later snapshots could
* have a RecentGlobalXmin with a later xid than the vacuum's
* OldestXmin, so it is possible that row versions deleted
* after OldestXmin could be marked as killed by other
* backends. The callback function *could* look at the index
* tuple state in isolation and decide to delete the index
* tuple, though currently it does not. If it ever did, we
* would need to reconsider whether XLOG_BTREE_VACUUM records
* should cause conflicts. If they did cause conflicts they
* would be fairly harsh conflicts, since we haven't yet
* worked out a way to pass a useful value for
* latestRemovedXid on the XLOG_BTREE_VACUUM records. This
* applies to *any* type of index that marks index tuples as
* killed.
*/
if (callback(htup, callback_state))
deletable[ndeletable++] = offnum;
}
}
/*
* Apply any needed deletes. We issue just one _bt_delitems_vacuum()
* call per page, so as to minimize WAL traffic.
*/
if (ndeletable > 0)
{
BlockNumber lastBlockVacuumed = BufferGetBlockNumber(buf);
_bt_delitems_vacuum(rel, buf, deletable, ndeletable,
vstate->lastBlockVacuumed);
/*
* Keep track of the block number of the lastBlockVacuumed, so we
* can scan those blocks as well during WAL replay. This then
* provides concurrency protection and allows btrees to be used
* while in recovery.
*/
if (lastBlockVacuumed > vstate->lastBlockVacuumed)
vstate->lastBlockVacuumed = lastBlockVacuumed;
stats->tuples_removed += ndeletable;
/* must recompute maxoff */
maxoff = PageGetMaxOffsetNumber(page);
}
else
{
/*
* If the page has been split during this vacuum cycle, it seems
* worth expending a write to clear btpo_cycleid even if we don't
* have any deletions to do. (If we do, _bt_delitems_vacuum takes
* care of this.) This ensures we won't process the page again.
*
* We treat this like a hint-bit update because there's no need to
* WAL-log it.
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid)
{
opaque->btpo_cycleid = 0;
SetBufferCommitInfoNeedsSave(buf);
}
}
/*
* If it's now empty, try to delete; else count the live tuples. We
* don't delete when recursing, though, to avoid putting entries into
* freePages out-of-order (doesn't seem worth any extra code to handle
* the case).
*/
if (minoff > maxoff)
delete_now = (blkno == orig_blkno);
else
stats->num_index_tuples += maxoff - minoff + 1;
}
if (delete_now)
{
MemoryContext oldcontext;
int ndel;
/* Run pagedel in a temp context to avoid memory leakage */
MemoryContextReset(vstate->pagedelcontext);
oldcontext = MemoryContextSwitchTo(vstate->pagedelcontext);
ndel = _bt_pagedel(rel, buf, NULL);
/* count only this page, else may double-count parent */
if (ndel)
stats->pages_deleted++;
MemoryContextSwitchTo(oldcontext);
/* pagedel released buffer, so we shouldn't */
}
else
_bt_relbuf(rel, buf);
/*
* This is really tail recursion, but if the compiler is too stupid to
* optimize it as such, we'd eat an uncomfortably large amount of stack
* space per recursion level (due to the deletable[] array). A failure is
* improbable since the number of levels isn't likely to be large ... but
* just in case, let's hand-optimize into a loop.
*/
if (recurse_to != P_NONE)
{
blkno = recurse_to;
goto restart;
}
}
/*
* Redo recompression of posting list. Doing all the changes in-place is not
* always possible, because it might require more space than we've on the page.
* Instead, once modification is required we copy unprocessed tail of the page
* into separately allocated chunk of memory for further reading original
* versions of segments. Thanks to that we don't bother about moving page data
* in-place.
*/
static void
ginRedoRecompress(Page page, ginxlogRecompressDataLeaf *data)
{
int actionno;
int segno;
GinPostingList *oldseg;
Pointer segmentend;
char *walbuf;
int totalsize;
Pointer tailCopy = NULL;
Pointer writePtr;
Pointer segptr;
/*
* If the page is in pre-9.4 format, convert to new format first.
*/
if (!GinPageIsCompressed(page))
{
ItemPointer uncompressed = (ItemPointer) GinDataPageGetData(page);
int nuncompressed = GinPageGetOpaque(page)->maxoff;
int npacked;
/*
* Empty leaf pages are deleted as part of vacuum, but leftmost and
* rightmost pages are never deleted. So, pg_upgrade'd from pre-9.4
* instances might contain empty leaf pages, and we need to handle
* them correctly.
*/
if (nuncompressed > 0)
{
GinPostingList *plist;
plist = ginCompressPostingList(uncompressed, nuncompressed,
BLCKSZ, &npacked);
totalsize = SizeOfGinPostingList(plist);
Assert(npacked == nuncompressed);
memcpy(GinDataLeafPageGetPostingList(page), plist, totalsize);
}
else
{
totalsize = 0;
}
GinDataPageSetDataSize(page, totalsize);
GinPageSetCompressed(page);
GinPageGetOpaque(page)->maxoff = InvalidOffsetNumber;
}
oldseg = GinDataLeafPageGetPostingList(page);
writePtr = (Pointer) oldseg;
segmentend = (Pointer) oldseg + GinDataLeafPageGetPostingListSize(page);
segno = 0;
walbuf = ((char *) data) + sizeof(ginxlogRecompressDataLeaf);
for (actionno = 0; actionno < data->nactions; actionno++)
{
uint8 a_segno = *((uint8 *) (walbuf++));
uint8 a_action = *((uint8 *) (walbuf++));
GinPostingList *newseg = NULL;
int newsegsize = 0;
ItemPointerData *items = NULL;
uint16 nitems = 0;
ItemPointerData *olditems;
int nolditems;
ItemPointerData *newitems;
int nnewitems;
int segsize;
/* Extract all the information we need from the WAL record */
if (a_action == GIN_SEGMENT_INSERT ||
a_action == GIN_SEGMENT_REPLACE)
{
newseg = (GinPostingList *) walbuf;
newsegsize = SizeOfGinPostingList(newseg);
walbuf += SHORTALIGN(newsegsize);
}
if (a_action == GIN_SEGMENT_ADDITEMS)
{
memcpy(&nitems, walbuf, sizeof(uint16));
walbuf += sizeof(uint16);
items = (ItemPointerData *) walbuf;
walbuf += nitems * sizeof(ItemPointerData);
}
/* Skip to the segment that this action concerns */
Assert(segno <= a_segno);
while (segno < a_segno)
{
/*
* Once modification is started and page tail is copied, we've
* to copy unmodified segments.
*/
segsize = SizeOfGinPostingList(oldseg);
if (tailCopy)
{
Assert(writePtr + segsize < PageGetSpecialPointer(page));
memcpy(writePtr, (Pointer) oldseg, segsize);
}
writePtr += segsize;
oldseg = GinNextPostingListSegment(oldseg);
segno++;
}
/*
* ADDITEMS action is handled like REPLACE, but the new segment to
* replace the old one is reconstructed using the old segment from
* disk and the new items from the WAL record.
*/
if (a_action == GIN_SEGMENT_ADDITEMS)
{
int npacked;
olditems = ginPostingListDecode(oldseg, &nolditems);
newitems = ginMergeItemPointers(items, nitems,
olditems, nolditems,
&nnewitems);
Assert(nnewitems == nolditems + nitems);
newseg = ginCompressPostingList(newitems, nnewitems,
BLCKSZ, &npacked);
Assert(npacked == nnewitems);
newsegsize = SizeOfGinPostingList(newseg);
a_action = GIN_SEGMENT_REPLACE;
}
segptr = (Pointer) oldseg;
if (segptr != segmentend)
segsize = SizeOfGinPostingList(oldseg);
else
{
/*
* Positioned after the last existing segment. Only INSERTs
* expected here.
*/
Assert(a_action == GIN_SEGMENT_INSERT);
segsize = 0;
}
/*
* We're about to start modification of the page. So, copy tail of the
* page if it's not done already.
*/
if (!tailCopy && segptr != segmentend)
{
int tailSize = segmentend - segptr;
tailCopy = (Pointer) palloc(tailSize);
memcpy(tailCopy, segptr, tailSize);
segptr = tailCopy;
oldseg = (GinPostingList *) segptr;
segmentend = segptr + tailSize;
}
switch (a_action)
{
case GIN_SEGMENT_DELETE:
segptr += segsize;
segno++;
break;
case GIN_SEGMENT_INSERT:
/* copy the new segment in place */
Assert(writePtr + newsegsize <= PageGetSpecialPointer(page));
memcpy(writePtr, newseg, newsegsize);
writePtr += newsegsize;
break;
case GIN_SEGMENT_REPLACE:
/* copy the new version of segment in place */
Assert(writePtr + newsegsize <= PageGetSpecialPointer(page));
memcpy(writePtr, newseg, newsegsize);
writePtr += newsegsize;
segptr += segsize;
segno++;
break;
default:
elog(ERROR, "unexpected GIN leaf action: %u", a_action);
}
oldseg = (GinPostingList *) segptr;
}
/* Copy the rest of unmodified segments if any. */
segptr = (Pointer) oldseg;
if (segptr != segmentend && tailCopy)
{
int restSize = segmentend - segptr;
Assert(writePtr + restSize <= PageGetSpecialPointer(page));
memcpy(writePtr, segptr, restSize);
writePtr += restSize;
}
totalsize = writePtr - (Pointer) GinDataLeafPageGetPostingList(page);
GinDataPageSetDataSize(page, totalsize);
}
Datum
bt_page_items(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_P(0);
uint32 blkno = PG_GETARG_UINT32(1);
RangeVar *relrv;
Datum result;
char *values[BTPAGEITEMS_NCOLUMNS];
BTPageOpaque opaque;
HeapTuple tuple;
ItemId id;
FuncCallContext *fctx;
MemoryContext mctx;
struct user_args *uargs = NULL;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pgstattuple functions"))));
if (blkno == 0)
elog(ERROR, "Block 0 is a meta page.");
if (SRF_IS_FIRSTCALL())
{
fctx = SRF_FIRSTCALL_INIT();
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
uargs = palloc(sizeof(struct user_args));
uargs->tupd = RelationNameGetTupleDesc(BTPAGEITEMS_TYPE);
uargs->offset = FirstOffsetNumber;
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
uargs->rel = relation_openrv(relrv, AccessShareLock);
CHECK_RELATION_BLOCK_RANGE(uargs->rel, blkno);
uargs->buffer = ReadBuffer(uargs->rel, blkno);
if (!IS_INDEX(uargs->rel) || !IS_BTREE(uargs->rel))
elog(ERROR, "bt_page_items() can be used only on b-tree index.");
uargs->page = BufferGetPage(uargs->buffer);
opaque = (BTPageOpaque) PageGetSpecialPointer(uargs->page);
if (P_ISDELETED(opaque))
elog(NOTICE, "bt_page_items(): this page is deleted.");
fctx->max_calls = PageGetMaxOffsetNumber(uargs->page);
fctx->user_fctx = uargs;
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
uargs = fctx->user_fctx;
if (fctx->call_cntr < fctx->max_calls)
{
IndexTuple itup;
id = PageGetItemId(uargs->page, uargs->offset);
if (!ItemIdIsValid(id))
elog(ERROR, "Invalid ItemId.");
itup = (IndexTuple) PageGetItem(uargs->page, id);
{
int j = 0;
BlockNumber blkno = BlockIdGetBlockNumber(&(itup->t_tid.ip_blkid));
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", uargs->offset);
values[j] = palloc(32);
snprintf(values[j++], 32, "(%u,%u)", blkno, itup->t_tid.ip_posid);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", (int) IndexTupleSize(itup));
values[j] = palloc(32);
snprintf(values[j++], 32, "%c", IndexTupleHasNulls(itup) ? 't' : 'f');
values[j] = palloc(32);
snprintf(values[j++], 32, "%c", IndexTupleHasVarwidths(itup) ? 't' : 'f');
{
int off;
char *dump;
char *ptr = (char *) itup + IndexInfoFindDataOffset(itup->t_info);
dump = palloc(IndexTupleSize(itup) * 3);
memset(dump, 0, IndexTupleSize(itup) * 3);
for (off = 0;
off < IndexTupleSize(itup) - IndexInfoFindDataOffset(itup->t_info);
off++)
{
if (dump[0] == '\0')
sprintf(dump, "%02x", *(ptr + off) & 0xff);
else
{
char buf[4];
sprintf(buf, " %02x", *(ptr + off) & 0xff);
strcat(dump, buf);
}
}
values[j] = dump;
}
tuple = BuildTupleFromCStrings(TupleDescGetAttInMetadata(uargs->tupd), values);
result = TupleGetDatum(TupleDescGetSlot(uargs->tupd), tuple);
}
uargs->offset = uargs->offset + 1;
SRF_RETURN_NEXT(fctx, result);
}
else
{
ReleaseBuffer(uargs->buffer);
relation_close(uargs->rel, AccessShareLock);
SRF_RETURN_DONE(fctx);
}
}
/*-------------------------------------------------------
* bt_page_items()
*
* Get IndexTupleData set in a btree page
*
* Usage: SELECT * FROM bt_page_items('t1_pkey', 1);
*-------------------------------------------------------
*/
Datum
bt_page_items(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_PP(0);
uint32 blkno = PG_GETARG_UINT32(1);
Datum result;
FuncCallContext *fctx;
MemoryContext mctx;
struct user_args *uargs;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pageinspect functions"))));
if (SRF_IS_FIRSTCALL())
{
RangeVar *relrv;
Relation rel;
Buffer buffer;
BTPageOpaque opaque;
TupleDesc tupleDesc;
fctx = SRF_FIRSTCALL_INIT();
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = relation_openrv(relrv, AccessShareLock);
if (!IS_INDEX(rel) || !IS_BTREE(rel))
elog(ERROR, "relation \"%s\" is not a btree index",
RelationGetRelationName(rel));
/*
* Reject attempts to read non-local temporary relations; we would be
* likely to get wrong data since we have no visibility into the
* owning session's local buffers.
*/
if (RELATION_IS_OTHER_TEMP(rel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
if (blkno == 0)
elog(ERROR, "block 0 is a meta page");
CHECK_RELATION_BLOCK_RANGE(rel, blkno);
buffer = ReadBuffer(rel, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
/*
* We copy the page into local storage to avoid holding pin on the
* buffer longer than we must, and possibly failing to release it at
* all if the calling query doesn't fetch all rows.
*/
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
uargs = palloc(sizeof(struct user_args));
uargs->page = palloc(BLCKSZ);
memcpy(uargs->page, BufferGetPage(buffer), BLCKSZ);
UnlockReleaseBuffer(buffer);
relation_close(rel, AccessShareLock);
uargs->offset = FirstOffsetNumber;
opaque = (BTPageOpaque) PageGetSpecialPointer(uargs->page);
if (P_ISDELETED(opaque))
elog(NOTICE, "page is deleted");
fctx->max_calls = PageGetMaxOffsetNumber(uargs->page);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupleDesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
fctx->attinmeta = TupleDescGetAttInMetadata(tupleDesc);
fctx->user_fctx = uargs;
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
uargs = fctx->user_fctx;
if (fctx->call_cntr < fctx->max_calls)
{
result = bt_page_print_tuples(fctx, uargs->page, uargs->offset);
uargs->offset++;
SRF_RETURN_NEXT(fctx, result);
}
else
{
pfree(uargs->page);
pfree(uargs);
SRF_RETURN_DONE(fctx);
}
}
/*
* Verify that the given bytea contains a HASH page, or die in the attempt.
* A pointer to a palloc'd, properly aligned copy of the page is returned.
*/
static Page
verify_hash_page(bytea *raw_page, int flags)
{
Page page = get_page_from_raw(raw_page);
int pagetype = LH_UNUSED_PAGE;
/* Treat new pages as unused. */
if (!PageIsNew(page))
{
HashPageOpaque pageopaque;
if (PageGetSpecialSize(page) != MAXALIGN(sizeof(HashPageOpaqueData)))
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index table contains corrupted page")));
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
if (pageopaque->hasho_page_id != HASHO_PAGE_ID)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("page is not a hash page"),
errdetail("Expected %08x, got %08x.",
HASHO_PAGE_ID, pageopaque->hasho_page_id)));
pagetype = pageopaque->hasho_flag & LH_PAGE_TYPE;
}
/* Check that page type is sane. */
if (pagetype != LH_OVERFLOW_PAGE && pagetype != LH_BUCKET_PAGE &&
pagetype != LH_BITMAP_PAGE && pagetype != LH_META_PAGE &&
pagetype != LH_UNUSED_PAGE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid hash page type %08x", pagetype)));
/* If requested, verify page type. */
if (flags != 0 && (pagetype & flags) == 0)
{
switch (flags)
{
case LH_META_PAGE:
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("page is not a hash meta page")));
case LH_BUCKET_PAGE | LH_OVERFLOW_PAGE:
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("page is not a hash bucket or overflow page")));
case LH_OVERFLOW_PAGE:
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("page is not a hash overflow page")));
default:
elog(ERROR,
"hash page of type %08x not in mask %08x",
pagetype, flags);
}
}
/*
* If it is the metapage, also verify magic number and version.
*/
if (pagetype == LH_META_PAGE)
{
HashMetaPage metap = HashPageGetMeta(page);
if (metap->hashm_magic != HASH_MAGIC)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("invalid magic number for metadata"),
errdetail("Expected 0x%08x, got 0x%08x.",
HASH_MAGIC, metap->hashm_magic)));
if (metap->hashm_version != HASH_VERSION)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("invalid version for metadata"),
errdetail("Expected %d, got %d",
HASH_VERSION, metap->hashm_version)));
}
return page;
}
/* -------------------------------------------------
* GetBTPageStatistics()
*
* Collect statistics of single b-tree page
* -------------------------------------------------
*/
static void
GetBTPageStatistics(BlockNumber blkno, Buffer buffer, BTPageStat *stat)
{
Page page = BufferGetPage(buffer);
PageHeader phdr = (PageHeader) page;
OffsetNumber maxoff = PageGetMaxOffsetNumber(page);
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
int item_size = 0;
int off;
stat->blkno = blkno;
stat->max_avail = BLCKSZ - (BLCKSZ - phdr->pd_special + SizeOfPageHeaderData);
stat->dead_items = stat->live_items = 0;
stat->page_size = PageGetPageSize(page);
/* page type (flags) */
if (P_ISDELETED(opaque))
{
stat->type = 'd';
stat->btpo.xact = opaque->btpo.xact;
return;
}
else if (P_IGNORE(opaque))
stat->type = 'e';
else if (P_ISLEAF(opaque))
stat->type = 'l';
else if (P_ISROOT(opaque))
stat->type = 'r';
else
stat->type = 'i';
/* btpage opaque data */
stat->btpo_prev = opaque->btpo_prev;
stat->btpo_next = opaque->btpo_next;
stat->btpo.level = opaque->btpo.level;
stat->btpo_flags = opaque->btpo_flags;
stat->btpo_cycleid = opaque->btpo_cycleid;
/* count live and dead tuples, and free space */
for (off = FirstOffsetNumber; off <= maxoff; off++)
{
IndexTuple itup;
ItemId id = PageGetItemId(page, off);
itup = (IndexTuple) PageGetItem(page, id);
item_size += IndexTupleSize(itup);
if (!ItemIdIsDead(id))
stat->live_items++;
else
stat->dead_items++;
}
stat->free_size = PageGetFreeSpace(page);
if ((stat->live_items + stat->dead_items) > 0)
stat->avg_item_size = item_size / (stat->live_items + stat->dead_items);
else
stat->avg_item_size = 0;
}
/**
* @brief Read the left-most leaf page by walking down on index tree structure
* from root node.
*
* Process flow
* -# Open index file and read meta page
* -# Get block number of root page
* -# Read "fast root" page
* -# Read left child page until reaching left-most leaf page
*
* After calling this function, the members of BTReader are the following:
* - smgr : Smgr relation of the existing index file.
* - blkno : block number of left-most leaf page. If there is no leaf page,
* InvalidBlockNumber is set.
* - offnum : InvalidOffsetNumber is set.
* - page : Left-most leaf page, or undefined if no leaf page.
*
* @param reader [in/out] B-Tree index reader
* @return true iff there are some tuples
*/
static bool
BTReaderInit(BTReader *reader, Relation rel)
{
BTPageOpaque metaopaque;
BTMetaPageData *metad;
BTPageOpaque opaque;
BlockNumber blkno;
/*
* HACK: We cannot use smgropen because smgrs returned from it
* will be closed automatically when we assign a new file node.
*
* XXX: It might be better to open the previous relfilenode with
* smgropen *after* RelationSetNewRelfilenode.
*/
memset(&reader->smgr, 0, sizeof(reader->smgr));
#if PG_VERSION_NUM >= 90100
reader->smgr.smgr_rnode.node = rel->rd_node;
reader->smgr.smgr_rnode.backend =
rel->rd_backend == MyBackendId ? MyBackendId : InvalidBackendId;
#else
reader->smgr.smgr_rnode = rel->rd_node;
#endif
reader->smgr.smgr_which = 0; /* md.c */
reader->blkno = InvalidBlockNumber;
reader->offnum = InvalidOffsetNumber;
reader->page = palloc(BLCKSZ);
/*
* Read meta page and check sanity of it.
*
* XXX: It might be better to do REINDEX against corrupted indexes
* instead of raising errors because we've spent long time for data
* loading...
*/
BTReaderReadPage(reader, BTREE_METAPAGE);
metaopaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
metad = BTPageGetMeta(reader->page);
if (!(metaopaque->btpo_flags & BTP_META) ||
metad->btm_magic != BTREE_MAGIC)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" is not a reader",
RelationGetRelationName(rel))));
if (metad->btm_version != BTREE_VERSION)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("version mismatch in index \"%s\": file version %d,"
" code version %d",
RelationGetRelationName(rel),
metad->btm_version, BTREE_VERSION)));
if (metad->btm_root == P_NONE)
{
/* No root page; We ignore the index in the subsequent build. */
reader->blkno = InvalidBlockNumber;
return false;
}
/* Go to the fast root page. */
blkno = metad->btm_fastroot;
BTReaderReadPage(reader, blkno);
opaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
/* Walk down to the left-most leaf page */
while (!P_ISLEAF(opaque))
{
ItemId firstid;
IndexTuple itup;
/* Get the block number of the left child */
firstid = PageGetItemId(reader->page, P_FIRSTDATAKEY(opaque));
itup = (IndexTuple) PageGetItem(reader->page, firstid);
blkno = ItemPointerGetBlockNumber(&(itup->t_tid));
/* Go down to children */
for (;;)
{
BTReaderReadPage(reader, blkno);
opaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
if (!P_IGNORE(opaque))
break;
if (P_RIGHTMOST(opaque))
{
/* We reach end of the index without any valid leaves. */
reader->blkno = InvalidBlockNumber;
return false;
}
blkno = opaque->btpo_next;
}
}
return true;
}
static void
btree_xlog_vacuum(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_btree_vacuum *xlrec = (xl_btree_vacuum *) XLogRecGetData(record);
Buffer buffer;
Page page;
BTPageOpaque opaque;
/*
* If queries might be active then we need to ensure every leaf page is
* unpinned between the lastBlockVacuumed and the current block, if there
* are any. This prevents replay of the VACUUM from reaching the stage of
* removing heap tuples while there could still be indexscans "in flight"
* to those particular tuples (see nbtree/README).
*
* It might be worth checking if there are actually any backends running;
* if not, we could just skip this.
*
* Since VACUUM can visit leaf pages out-of-order, it might issue records
* with lastBlockVacuumed >= block; that's not an error, it just means
* nothing to do now.
*
* Note: since we touch all pages in the range, we will lock non-leaf
* pages, and also any empty (all-zero) pages that may be in the index. It
* doesn't seem worth the complexity to avoid that. But it's important
* that HotStandbyActiveInReplay() will not return true if the database
* isn't yet consistent; so we need not fear reading still-corrupt blocks
* here during crash recovery.
*/
if (HotStandbyActiveInReplay())
{
RelFileNode thisrnode;
BlockNumber thisblkno;
BlockNumber blkno;
XLogRecGetBlockTag(record, 0, &thisrnode, NULL, &thisblkno);
for (blkno = xlrec->lastBlockVacuumed + 1; blkno < thisblkno; blkno++)
{
/*
* We use RBM_NORMAL_NO_LOG mode because it's not an error
* condition to see all-zero pages. The original btvacuumpage
* scan would have skipped over all-zero pages, noting them in FSM
* but not bothering to initialize them just yet; so we mustn't
* throw an error here. (We could skip acquiring the cleanup lock
* if PageIsNew, but it's probably not worth the cycles to test.)
*
* XXX we don't actually need to read the block, we just need to
* confirm it is unpinned. If we had a special call into the
* buffer manager we could optimise this so that if the block is
* not in shared_buffers we confirm it as unpinned.
*/
buffer = XLogReadBufferExtended(thisrnode, MAIN_FORKNUM, blkno,
RBM_NORMAL_NO_LOG);
if (BufferIsValid(buffer))
{
LockBufferForCleanup(buffer);
UnlockReleaseBuffer(buffer);
}
}
}
/*
* Like in btvacuumpage(), we need to take a cleanup lock on every leaf
* page. See nbtree/README for details.
*/
if (XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &buffer)
== BLK_NEEDS_REDO)
{
char *ptr;
Size len;
ptr = XLogRecGetBlockData(record, 0, &len);
page = (Page) BufferGetPage(buffer);
if (len > 0)
{
OffsetNumber *unused;
OffsetNumber *unend;
unused = (OffsetNumber *) ptr;
unend = (OffsetNumber *) ((char *) ptr + len);
if ((unend - unused) > 0)
PageIndexMultiDelete(page, unused, unend - unused);
}
/*
* Mark the page as not containing any LP_DEAD items --- see comments
* in _bt_delitems_vacuum().
*/
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
opaque->btpo_flags &= ~BTP_HAS_GARBAGE;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
}
/*
* Bulk deletion of all index entries pointing to a set of heap tuples.
* The set of target tuples is specified via a callback routine that tells
* whether any given heap tuple (identified by ItemPointer) is being deleted.
*
* Result: a palloc'd struct containing statistical info for VACUUM displays.
*/
Datum
hashbulkdelete(PG_FUNCTION_ARGS)
{
IndexVacuumInfo *info = (IndexVacuumInfo *) PG_GETARG_POINTER(0);
IndexBulkDeleteResult *stats = (IndexBulkDeleteResult *) PG_GETARG_POINTER(1);
IndexBulkDeleteCallback callback = (IndexBulkDeleteCallback) PG_GETARG_POINTER(2);
void *callback_state = (void *) PG_GETARG_POINTER(3);
Relation rel = info->index;
double tuples_removed;
double num_index_tuples;
double orig_ntuples;
Bucket orig_maxbucket;
Bucket cur_maxbucket;
Bucket cur_bucket;
Buffer metabuf;
HashMetaPage metap;
HashMetaPageData local_metapage;
tuples_removed = 0;
num_index_tuples = 0;
/*
* Read the metapage to fetch original bucket and tuple counts. Also, we
* keep a copy of the last-seen metapage so that we can use its
* hashm_spares[] values to compute bucket page addresses. This is a bit
* hokey but perfectly safe, since the interesting entries in the spares
* array cannot change under us; and it beats rereading the metapage for
* each bucket.
*/
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
orig_maxbucket = metap->hashm_maxbucket;
orig_ntuples = metap->hashm_ntuples;
memcpy(&local_metapage, metap, sizeof(local_metapage));
_hash_relbuf(rel, metabuf);
/* Scan the buckets that we know exist */
cur_bucket = 0;
cur_maxbucket = orig_maxbucket;
loop_top:
while (cur_bucket <= cur_maxbucket)
{
BlockNumber bucket_blkno;
BlockNumber blkno;
bool bucket_dirty = false;
/* Get address of bucket's start page */
bucket_blkno = BUCKET_TO_BLKNO(&local_metapage, cur_bucket);
/* Exclusive-lock the bucket so we can shrink it */
_hash_getlock(rel, bucket_blkno, HASH_EXCLUSIVE);
/* Shouldn't have any active scans locally, either */
if (_hash_has_active_scan(rel, cur_bucket))
elog(ERROR, "hash index has active scan during VACUUM");
/* Scan each page in bucket */
blkno = bucket_blkno;
while (BlockNumberIsValid(blkno))
{
Buffer buf;
Page page;
HashPageOpaque opaque;
OffsetNumber offno;
OffsetNumber maxoffno;
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable = 0;
vacuum_delay_point();
buf = _hash_getbuf_with_strategy(rel, blkno, HASH_WRITE,
LH_BUCKET_PAGE | LH_OVERFLOW_PAGE,
info->strategy);
page = BufferGetPage(buf);
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
Assert(opaque->hasho_bucket == cur_bucket);
/* Scan each tuple in page */
maxoffno = PageGetMaxOffsetNumber(page);
for (offno = FirstOffsetNumber;
offno <= maxoffno;
offno = OffsetNumberNext(offno))
{
IndexTuple itup;
ItemPointer htup;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offno));
htup = &(itup->t_tid);
if (callback(htup, callback_state))
{
/* mark the item for deletion */
deletable[ndeletable++] = offno;
tuples_removed += 1;
}
else
num_index_tuples += 1;
}
/*
* Apply deletions and write page if needed, advance to next page.
*/
blkno = opaque->hasho_nextblkno;
if (ndeletable > 0)
{
PageIndexMultiDelete(page, deletable, ndeletable);
_hash_wrtbuf(rel, buf);
bucket_dirty = true;
}
else
_hash_relbuf(rel, buf);
}
/* If we deleted anything, try to compact free space */
if (bucket_dirty)
_hash_squeezebucket(rel, cur_bucket, bucket_blkno,
info->strategy);
/* Release bucket lock */
_hash_droplock(rel, bucket_blkno, HASH_EXCLUSIVE);
/* Advance to next bucket */
cur_bucket++;
}
/* Write-lock metapage and check for split since we started */
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_WRITE, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
if (cur_maxbucket != metap->hashm_maxbucket)
{
/* There's been a split, so process the additional bucket(s) */
cur_maxbucket = metap->hashm_maxbucket;
memcpy(&local_metapage, metap, sizeof(local_metapage));
_hash_relbuf(rel, metabuf);
goto loop_top;
}
/* Okay, we're really done. Update tuple count in metapage. */
if (orig_maxbucket == metap->hashm_maxbucket &&
orig_ntuples == metap->hashm_ntuples)
{
/*
* No one has split or inserted anything since start of scan, so
* believe our count as gospel.
*/
metap->hashm_ntuples = num_index_tuples;
}
else
{
/*
* Otherwise, our count is untrustworthy since we may have
* double-scanned tuples in split buckets. Proceed by dead-reckoning.
* (Note: we still return estimated_count = false, because using this
* count is better than not updating reltuples at all.)
*/
if (metap->hashm_ntuples > tuples_removed)
metap->hashm_ntuples -= tuples_removed;
else
metap->hashm_ntuples = 0;
num_index_tuples = metap->hashm_ntuples;
}
_hash_wrtbuf(rel, metabuf);
/* return statistics */
if (stats == NULL)
stats = (IndexBulkDeleteResult *) palloc0(sizeof(IndexBulkDeleteResult));
stats->estimated_count = false;
stats->num_index_tuples = num_index_tuples;
stats->tuples_removed += tuples_removed;
/* hashvacuumcleanup will fill in num_pages */
PG_RETURN_POINTER(stats);
}
static void
btree_xlog_unlink_page(uint8 info, XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_btree_unlink_page *xlrec = (xl_btree_unlink_page *) XLogRecGetData(record);
BlockNumber leftsib;
BlockNumber rightsib;
Buffer buffer;
Page page;
BTPageOpaque pageop;
leftsib = xlrec->leftsib;
rightsib = xlrec->rightsib;
/*
* In normal operation, we would lock all the pages this WAL record
* touches before changing any of them. In WAL replay, it should be okay
* to lock just one page at a time, since no concurrent index updates can
* be happening, and readers should not care whether they arrive at the
* target page or not (since it's surely empty).
*/
/* Fix left-link of right sibling */
if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
pageop->btpo_prev = leftsib;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
/* Fix right-link of left sibling, if any */
if (leftsib != P_NONE)
{
if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
pageop->btpo_next = rightsib;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
}
/* Rewrite target page as empty deleted page */
buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
_bt_pageinit(page, BufferGetPageSize(buffer));
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
pageop->btpo_prev = leftsib;
pageop->btpo_next = rightsib;
pageop->btpo.xact = xlrec->btpo_xact;
pageop->btpo_flags = BTP_DELETED;
pageop->btpo_cycleid = 0;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
/*
* If we deleted a parent of the targeted leaf page, instead of the leaf
* itself, update the leaf to point to the next remaining child in the
* branch.
*/
if (XLogRecHasBlockRef(record, 3))
{
/*
* There is no real data on the page, so we just re-create it from
* scratch using the information from the WAL record.
*/
IndexTupleData trunctuple;
buffer = XLogInitBufferForRedo(record, 3);
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
_bt_pageinit(page, BufferGetPageSize(buffer));
pageop->btpo_flags = BTP_HALF_DEAD | BTP_LEAF;
pageop->btpo_prev = xlrec->leafleftsib;
pageop->btpo_next = xlrec->leafrightsib;
pageop->btpo.level = 0;
pageop->btpo_cycleid = 0;
/* Add a dummy hikey item */
MemSet(&trunctuple, 0, sizeof(IndexTupleData));
trunctuple.t_info = sizeof(IndexTupleData);
if (xlrec->topparent != InvalidBlockNumber)
ItemPointerSet(&trunctuple.t_tid, xlrec->topparent, P_HIKEY);
else
ItemPointerSetInvalid(&trunctuple.t_tid);
if (PageAddItem(page, (Item) &trunctuple, sizeof(IndexTupleData), P_HIKEY,
false, false) == InvalidOffsetNumber)
elog(ERROR, "could not add dummy high key to half-dead page");
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
/* Update metapage if needed */
if (info == XLOG_BTREE_UNLINK_PAGE_META)
_bt_restore_meta(record, 4);
}
/*
* _hash_checkpage -- sanity checks on the format of all hash pages
*
* If flags is not zero, it is a bitwise OR of the acceptable values of
* hasho_flag.
*/
void
_hash_checkpage(Relation rel, Buffer buf, int flags)
{
Page page = BufferGetPage(buf);
/*
* ReadBuffer verifies that every newly-read page passes
* PageHeaderIsValid, which means it either contains a reasonably sane
* page header or is all-zero. We have to defend against the all-zero
* case, however.
*/
if (PageIsNew(page))
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" contains unexpected zero page at block %u",
RelationGetRelationName(rel),
BufferGetBlockNumber(buf)),
errhint("Please REINDEX it.")));
/*
* Additionally check that the special area looks sane.
*/
if (PageGetSpecialSize(page) != MAXALIGN(sizeof(HashPageOpaqueData)))
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" contains corrupted page at block %u",
RelationGetRelationName(rel),
BufferGetBlockNumber(buf)),
errhint("Please REINDEX it.")));
if (flags)
{
HashPageOpaque opaque = (HashPageOpaque) PageGetSpecialPointer(page);
if ((opaque->hasho_flag & flags) == 0)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" contains corrupted page at block %u",
RelationGetRelationName(rel),
BufferGetBlockNumber(buf)),
errhint("Please REINDEX it.")));
}
/*
* When checking the metapage, also verify magic number and version.
*/
if (flags == LH_META_PAGE)
{
HashMetaPage metap = HashPageGetMeta(page);
if (metap->hashm_magic != HASH_MAGIC)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" is not a hash index",
RelationGetRelationName(rel))));
if (metap->hashm_version != HASH_VERSION)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" has wrong hash version",
RelationGetRelationName(rel)),
errhint("Please REINDEX it.")));
}
}
/*
* _hash_squeezebucket(rel, bucket)
*
* Try to squeeze the tuples onto pages occurring earlier in the
* bucket chain in an attempt to free overflow pages. When we start
* the "squeezing", the page from which we start taking tuples (the
* "read" page) is the last bucket in the bucket chain and the page
* onto which we start squeezing tuples (the "write" page) is the
* first page in the bucket chain. The read page works backward and
* the write page works forward; the procedure terminates when the
* read page and write page are the same page.
*
* At completion of this procedure, it is guaranteed that all pages in
* the bucket are nonempty, unless the bucket is totally empty (in
* which case all overflow pages will be freed). The original implementation
* required that to be true on entry as well, but it's a lot easier for
* callers to leave empty overflow pages and let this guy clean it up.
*
* Caller must hold exclusive lock on the target bucket. This allows
* us to safely lock multiple pages in the bucket.
*
* Since this function is invoked in VACUUM, we provide an access strategy
* parameter that controls fetches of the bucket pages.
*/
void
_hash_squeezebucket(Relation rel,
Bucket bucket,
BlockNumber bucket_blkno,
BufferAccessStrategy bstrategy)
{
BlockNumber wblkno;
BlockNumber rblkno;
Buffer wbuf;
Buffer rbuf;
Page wpage;
Page rpage;
HashPageOpaque wopaque;
HashPageOpaque ropaque;
bool wbuf_dirty;
/*
* start squeezing into the base bucket page.
*/
wblkno = bucket_blkno;
wbuf = _hash_getbuf_with_strategy(rel,
wblkno,
HASH_WRITE,
LH_BUCKET_PAGE,
bstrategy);
wpage = BufferGetPage(wbuf);
wopaque = (HashPageOpaque) PageGetSpecialPointer(wpage);
/*
* if there aren't any overflow pages, there's nothing to squeeze.
*/
if (!BlockNumberIsValid(wopaque->hasho_nextblkno))
{
_hash_relbuf(rel, wbuf);
return;
}
/*
* Find the last page in the bucket chain by starting at the base bucket
* page and working forward. Note: we assume that a hash bucket chain is
* usually smaller than the buffer ring being used by VACUUM, else using
* the access strategy here would be counterproductive.
*/
rbuf = InvalidBuffer;
ropaque = wopaque;
do
{
rblkno = ropaque->hasho_nextblkno;
if (rbuf != InvalidBuffer)
_hash_relbuf(rel, rbuf);
rbuf = _hash_getbuf_with_strategy(rel,
rblkno,
HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
rpage = BufferGetPage(rbuf);
ropaque = (HashPageOpaque) PageGetSpecialPointer(rpage);
Assert(ropaque->hasho_bucket == bucket);
} while (BlockNumberIsValid(ropaque->hasho_nextblkno));
/*
* squeeze the tuples.
*/
wbuf_dirty = false;
for (;;)
{
OffsetNumber roffnum;
OffsetNumber maxroffnum;
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable = 0;
/* Scan each tuple in "read" page */
maxroffnum = PageGetMaxOffsetNumber(rpage);
for (roffnum = FirstOffsetNumber;
roffnum <= maxroffnum;
roffnum = OffsetNumberNext(roffnum))
{
IndexTuple itup;
Size itemsz;
itup = (IndexTuple) PageGetItem(rpage,
PageGetItemId(rpage, roffnum));
itemsz = IndexTupleDSize(*itup);
itemsz = MAXALIGN(itemsz);
/*
* Walk up the bucket chain, looking for a page big enough for
* this item. Exit if we reach the read page.
*/
while (PageGetFreeSpace(wpage) < itemsz)
{
Assert(!PageIsEmpty(wpage));
wblkno = wopaque->hasho_nextblkno;
Assert(BlockNumberIsValid(wblkno));
if (wbuf_dirty)
_hash_wrtbuf(rel, wbuf);
else
_hash_relbuf(rel, wbuf);
/* nothing more to do if we reached the read page */
if (rblkno == wblkno)
{
if (ndeletable > 0)
{
/* Delete tuples we already moved off read page */
PageIndexMultiDelete(rpage, deletable, ndeletable);
_hash_wrtbuf(rel, rbuf);
}
else
_hash_relbuf(rel, rbuf);
return;
}
wbuf = _hash_getbuf_with_strategy(rel,
wblkno,
HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
wpage = BufferGetPage(wbuf);
wopaque = (HashPageOpaque) PageGetSpecialPointer(wpage);
Assert(wopaque->hasho_bucket == bucket);
wbuf_dirty = false;
}
/*
* we have found room so insert on the "write" page, being careful
* to preserve hashkey ordering. (If we insert many tuples into
* the same "write" page it would be worth qsort'ing instead of
* doing repeated _hash_pgaddtup.)
*/
(void) _hash_pgaddtup(rel, wbuf, itemsz, itup);
wbuf_dirty = true;
/* remember tuple for deletion from "read" page */
deletable[ndeletable++] = roffnum;
}
/*
* If we reach here, there are no live tuples on the "read" page ---
* it was empty when we got to it, or we moved them all. So we can
* just free the page without bothering with deleting tuples
* individually. Then advance to the previous "read" page.
*
* Tricky point here: if our read and write pages are adjacent in the
* bucket chain, our write lock on wbuf will conflict with
* _hash_freeovflpage's attempt to update the sibling links of the
* removed page. However, in that case we are done anyway, so we can
* simply drop the write lock before calling _hash_freeovflpage.
*/
rblkno = ropaque->hasho_prevblkno;
Assert(BlockNumberIsValid(rblkno));
/* are we freeing the page adjacent to wbuf? */
if (rblkno == wblkno)
{
/* yes, so release wbuf lock first */
if (wbuf_dirty)
_hash_wrtbuf(rel, wbuf);
else
_hash_relbuf(rel, wbuf);
/* free this overflow page (releases rbuf) */
_hash_freeovflpage(rel, rbuf, bstrategy);
/* done */
return;
}
/* free this overflow page, then get the previous one */
_hash_freeovflpage(rel, rbuf, bstrategy);
rbuf = _hash_getbuf_with_strategy(rel,
rblkno,
HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
rpage = BufferGetPage(rbuf);
ropaque = (HashPageOpaque) PageGetSpecialPointer(rpage);
Assert(ropaque->hasho_bucket == bucket);
}
/* NOTREACHED */
}
/*
* _hash_first() -- Find the first item in a scan.
*
* Find the first item in the index that
* satisfies the qualification associated with the scan descriptor. On
* success, the page containing the current index tuple is read locked
* and pinned, and the scan's opaque data entry is updated to
* include the buffer.
*/
bool
_hash_first(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
HashScanOpaque so = (HashScanOpaque) scan->opaque;
ScanKey cur;
uint32 hashkey;
Bucket bucket;
BlockNumber blkno;
Buffer buf;
Buffer metabuf;
Page page;
HashPageOpaque opaque;
HashMetaPage metap;
IndexTuple itup;
ItemPointer current;
OffsetNumber offnum;
pgstat_count_index_scan(rel);
current = &(so->hashso_curpos);
ItemPointerSetInvalid(current);
/*
* We do not support hash scans with no index qualification, because we
* would have to read the whole index rather than just one bucket. That
* creates a whole raft of problems, since we haven't got a practical way
* to lock all the buckets against splits or compactions.
*/
if (scan->numberOfKeys < 1)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("hash indexes do not support whole-index scans")));
/* There may be more than one index qual, but we hash only the first */
cur = &scan->keyData[0];
/* We support only single-column hash indexes */
Assert(cur->sk_attno == 1);
/* And there's only one operator strategy, too */
Assert(cur->sk_strategy == HTEqualStrategyNumber);
/*
* If the constant in the index qual is NULL, assume it cannot match any
* items in the index.
*/
if (cur->sk_flags & SK_ISNULL)
return false;
/*
* Okay to compute the hash key. We want to do this before acquiring any
* locks, in case a user-defined hash function happens to be slow.
*
* If scankey operator is not a cross-type comparison, we can use the
* cached hash function; otherwise gotta look it up in the catalogs.
*
* We support the convention that sk_subtype == InvalidOid means the
* opclass input type; this is a hack to simplify life for ScanKeyInit().
*/
if (cur->sk_subtype == rel->rd_opcintype[0] ||
cur->sk_subtype == InvalidOid)
hashkey = _hash_datum2hashkey(rel, cur->sk_argument);
else
hashkey = _hash_datum2hashkey_type(rel, cur->sk_argument,
cur->sk_subtype);
so->hashso_sk_hash = hashkey;
/*
* Acquire shared split lock so we can compute the target bucket safely
* (see README).
*/
_hash_getlock(rel, 0, HASH_SHARE);
/* Read the metapage */
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* Compute the target bucket number, and convert to block number.
*/
bucket = _hash_hashkey2bucket(hashkey,
metap->hashm_maxbucket,
metap->hashm_highmask,
metap->hashm_lowmask);
blkno = BUCKET_TO_BLKNO(metap, bucket);
/* done with the metapage */
_hash_relbuf(rel, metabuf);
/*
* Acquire share lock on target bucket; then we can release split lock.
*/
_hash_getlock(rel, blkno, HASH_SHARE);
_hash_droplock(rel, 0, HASH_SHARE);
/* Update scan opaque state to show we have lock on the bucket */
so->hashso_bucket = bucket;
so->hashso_bucket_valid = true;
so->hashso_bucket_blkno = blkno;
/* Fetch the primary bucket page for the bucket */
buf = _hash_getbuf(rel, blkno, HASH_READ, LH_BUCKET_PAGE);
page = BufferGetPage(buf);
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
Assert(opaque->hasho_bucket == bucket);
/* If a backwards scan is requested, move to the end of the chain */
if (ScanDirectionIsBackward(dir))
{
while (BlockNumberIsValid(opaque->hasho_nextblkno))
_hash_readnext(rel, &buf, &page, &opaque);
}
/* Now find the first tuple satisfying the qualification */
if (!_hash_step(scan, &buf, dir))
return false;
/* if we're here, _hash_step found a valid tuple */
offnum = ItemPointerGetOffsetNumber(current);
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
page = BufferGetPage(buf);
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
so->hashso_heappos = itup->t_tid;
return true;
}
/*
* Helper function to perform deletion of index entries from a bucket.
*
* This function expects that the caller has acquired a cleanup lock on the
* primary bucket page, and will return with a write lock again held on the
* primary bucket page. The lock won't necessarily be held continuously,
* though, because we'll release it when visiting overflow pages.
*
* It would be very bad if this function cleaned a page while some other
* backend was in the midst of scanning it, because hashgettuple assumes
* that the next valid TID will be greater than or equal to the current
* valid TID. There can't be any concurrent scans in progress when we first
* enter this function because of the cleanup lock we hold on the primary
* bucket page, but as soon as we release that lock, there might be. We
* handle that by conspiring to prevent those scans from passing our cleanup
* scan. To do that, we lock the next page in the bucket chain before
* releasing the lock on the previous page. (This type of lock chaining is
* not ideal, so we might want to look for a better solution at some point.)
*
* We need to retain a pin on the primary bucket to ensure that no concurrent
* split can start.
*/
void
hashbucketcleanup(Relation rel, Bucket cur_bucket, Buffer bucket_buf,
BlockNumber bucket_blkno, BufferAccessStrategy bstrategy,
uint32 maxbucket, uint32 highmask, uint32 lowmask,
double *tuples_removed, double *num_index_tuples,
bool split_cleanup,
IndexBulkDeleteCallback callback, void *callback_state)
{
BlockNumber blkno;
Buffer buf;
Bucket new_bucket PG_USED_FOR_ASSERTS_ONLY = InvalidBucket;
bool bucket_dirty = false;
blkno = bucket_blkno;
buf = bucket_buf;
if (split_cleanup)
new_bucket = _hash_get_newbucket_from_oldbucket(rel, cur_bucket,
lowmask, maxbucket);
/* Scan each page in bucket */
for (;;)
{
HashPageOpaque opaque;
OffsetNumber offno;
OffsetNumber maxoffno;
Buffer next_buf;
Page page;
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable = 0;
bool retain_pin = false;
bool clear_dead_marking = false;
vacuum_delay_point();
page = BufferGetPage(buf);
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
/* Scan each tuple in page */
maxoffno = PageGetMaxOffsetNumber(page);
for (offno = FirstOffsetNumber;
offno <= maxoffno;
offno = OffsetNumberNext(offno))
{
ItemPointer htup;
IndexTuple itup;
Bucket bucket;
bool kill_tuple = false;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offno));
htup = &(itup->t_tid);
/*
* To remove the dead tuples, we strictly want to rely on results
* of callback function. refer btvacuumpage for detailed reason.
*/
if (callback && callback(htup, callback_state))
{
kill_tuple = true;
if (tuples_removed)
*tuples_removed += 1;
}
else if (split_cleanup)
{
/* delete the tuples that are moved by split. */
bucket = _hash_hashkey2bucket(_hash_get_indextuple_hashkey(itup),
maxbucket,
highmask,
lowmask);
/* mark the item for deletion */
if (bucket != cur_bucket)
{
/*
* We expect tuples to either belong to current bucket or
* new_bucket. This is ensured because we don't allow
* further splits from bucket that contains garbage. See
* comments in _hash_expandtable.
*/
Assert(bucket == new_bucket);
kill_tuple = true;
}
}
if (kill_tuple)
{
/* mark the item for deletion */
deletable[ndeletable++] = offno;
}
else
{
/* we're keeping it, so count it */
if (num_index_tuples)
*num_index_tuples += 1;
}
}
/* retain the pin on primary bucket page till end of bucket scan */
if (blkno == bucket_blkno)
retain_pin = true;
else
retain_pin = false;
blkno = opaque->hasho_nextblkno;
/*
* Apply deletions, advance to next page and write page if needed.
*/
if (ndeletable > 0)
{
/* No ereport(ERROR) until changes are logged */
START_CRIT_SECTION();
PageIndexMultiDelete(page, deletable, ndeletable);
bucket_dirty = true;
/*
* Let us mark the page as clean if vacuum removes the DEAD tuples
* from an index page. We do this by clearing
* LH_PAGE_HAS_DEAD_TUPLES flag.
*/
if (tuples_removed && *tuples_removed > 0 &&
H_HAS_DEAD_TUPLES(opaque))
{
opaque->hasho_flag &= ~LH_PAGE_HAS_DEAD_TUPLES;
clear_dead_marking = true;
}
MarkBufferDirty(buf);
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
xl_hash_delete xlrec;
XLogRecPtr recptr;
xlrec.clear_dead_marking = clear_dead_marking;
xlrec.is_primary_bucket_page = (buf == bucket_buf) ? true : false;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashDelete);
/*
* bucket buffer needs to be registered to ensure that we can
* acquire a cleanup lock on it during replay.
*/
if (!xlrec.is_primary_bucket_page)
XLogRegisterBuffer(0, bucket_buf, REGBUF_STANDARD | REGBUF_NO_IMAGE);
XLogRegisterBuffer(1, buf, REGBUF_STANDARD);
XLogRegisterBufData(1, (char *) deletable,
ndeletable * sizeof(OffsetNumber));
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_DELETE);
PageSetLSN(BufferGetPage(buf), recptr);
}
END_CRIT_SECTION();
}
/* bail out if there are no more pages to scan. */
if (!BlockNumberIsValid(blkno))
break;
next_buf = _hash_getbuf_with_strategy(rel, blkno, HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
/*
* release the lock on previous page after acquiring the lock on next
* page
*/
if (retain_pin)
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
else
_hash_relbuf(rel, buf);
buf = next_buf;
}
/*
* lock the bucket page to clear the garbage flag and squeeze the bucket.
* if the current buffer is same as bucket buffer, then we already have
* lock on bucket page.
*/
if (buf != bucket_buf)
{
_hash_relbuf(rel, buf);
LockBuffer(bucket_buf, BUFFER_LOCK_EXCLUSIVE);
}
/*
* Clear the garbage flag from bucket after deleting the tuples that are
* moved by split. We purposefully clear the flag before squeeze bucket,
* so that after restart, vacuum shouldn't again try to delete the moved
* by split tuples.
*/
if (split_cleanup)
{
HashPageOpaque bucket_opaque;
Page page;
page = BufferGetPage(bucket_buf);
bucket_opaque = (HashPageOpaque) PageGetSpecialPointer(page);
/* No ereport(ERROR) until changes are logged */
START_CRIT_SECTION();
bucket_opaque->hasho_flag &= ~LH_BUCKET_NEEDS_SPLIT_CLEANUP;
MarkBufferDirty(bucket_buf);
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
XLogRecPtr recptr;
XLogBeginInsert();
XLogRegisterBuffer(0, bucket_buf, REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_SPLIT_CLEANUP);
PageSetLSN(page, recptr);
}
END_CRIT_SECTION();
}
/*
* If we have deleted anything, try to compact free space. For squeezing
* the bucket, we must have a cleanup lock, else it can impact the
* ordering of tuples for a scan that has started before it.
*/
if (bucket_dirty && IsBufferCleanupOK(bucket_buf))
_hash_squeezebucket(rel, cur_bucket, bucket_blkno, bucket_buf,
bstrategy);
else
LockBuffer(bucket_buf, BUFFER_LOCK_UNLOCK);
}
/*
* replay delete operation in hash index to remove
* tuples marked as DEAD during index tuple insertion.
*/
static void
hash_xlog_vacuum_one_page(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_vacuum_one_page *xldata;
Buffer buffer;
Buffer metabuf;
Page page;
XLogRedoAction action;
HashPageOpaque pageopaque;
xldata = (xl_hash_vacuum_one_page *) XLogRecGetData(record);
/*
* If we have any conflict processing to do, it must happen before we
* update the page.
*
* Hash index records that are marked as LP_DEAD and being removed during
* hash index tuple insertion can conflict with standby queries. You might
* think that vacuum records would conflict as well, but we've handled
* that already. XLOG_HEAP2_CLEANUP_INFO records provide the highest xid
* cleaned by the vacuum of the heap and so we can resolve any conflicts
* just once when that arrives. After that we know that no conflicts
* exist from individual hash index vacuum records on that index.
*/
if (InHotStandby)
{
TransactionId latestRemovedXid =
hash_xlog_vacuum_get_latestRemovedXid(record);
RelFileNode rnode;
XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(latestRemovedXid, rnode);
}
action = XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &buffer);
if (action == BLK_NEEDS_REDO)
{
char *ptr;
Size len;
ptr = XLogRecGetBlockData(record, 0, &len);
page = (Page) BufferGetPage(buffer);
if (len > 0)
{
OffsetNumber *unused;
OffsetNumber *unend;
unused = (OffsetNumber *) ptr;
unend = (OffsetNumber *) ((char *) ptr + len);
if ((unend - unused) > 0)
PageIndexMultiDelete(page, unused, unend - unused);
}
/*
* Mark the page as not containing any LP_DEAD items. See comments
* in _hash_vacuum_one_page() for details.
*/
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
pageopaque->hasho_flag &= ~LH_PAGE_HAS_DEAD_TUPLES;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
if (XLogReadBufferForRedo(record, 1, &metabuf) == BLK_NEEDS_REDO)
{
Page metapage;
HashMetaPage metap;
metapage = BufferGetPage(metabuf);
metap = HashPageGetMeta(metapage);
metap->hashm_ntuples -= xldata->ntuples;
PageSetLSN(metapage, lsn);
MarkBufferDirty(metabuf);
}
if (BufferIsValid(metabuf))
UnlockReleaseBuffer(metabuf);
}
