/*
* _hash_get_oldblock_from_newbucket() -- get the block number of a bucket
* from which current (new) bucket is being split.
*/
BlockNumber
_hash_get_oldblock_from_newbucket(Relation rel, Bucket new_bucket)
{
Bucket old_bucket;
uint32 mask;
Buffer metabuf;
HashMetaPage metap;
BlockNumber blkno;
/*
* To get the old bucket from the current bucket, we need a mask to modulo
* into lower half of table. This mask is stored in meta page as
* hashm_lowmask, but here we can't rely on the same, because we need a
* value of lowmask that was prevalent at the time when bucket split was
* started. Masking the most significant bit of new bucket would give us
* old bucket.
*/
mask = (((uint32) 1) << (fls(new_bucket) - 1)) - 1;
old_bucket = new_bucket & mask;
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
blkno = BUCKET_TO_BLKNO(metap, old_bucket);
_hash_relbuf(rel, metabuf);
return blkno;
}
/*
* replay a hash index insert without split
*/
static void
hash_xlog_insert(XLogReaderState *record)
{
HashMetaPage metap;
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_insert *xlrec = (xl_hash_insert *) XLogRecGetData(record);
Buffer buffer;
Page page;
if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
Size datalen;
char *datapos = XLogRecGetBlockData(record, 0, &datalen);
page = BufferGetPage(buffer);
if (PageAddItem(page, (Item) datapos, datalen, xlrec->offnum,
false, false) == InvalidOffsetNumber)
elog(PANIC, "hash_xlog_insert: failed to add item");
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
/*
* Note: in normal operation, we'd update the metapage while still
* holding lock on the page we inserted into. But during replay it's
* not necessary to hold that lock, since no other index updates can
* be happening concurrently.
*/
page = BufferGetPage(buffer);
metap = HashPageGetMeta(page);
metap->hashm_ntuples += 1;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
}
/*
* replay a hash index bitmap page
*/
static void
hash_xlog_init_bitmap_page(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
Buffer bitmapbuf;
Buffer metabuf;
Page page;
HashMetaPage metap;
uint32 num_buckets;
xl_hash_init_bitmap_page *xlrec = (xl_hash_init_bitmap_page *) XLogRecGetData(record);
/*
* Initialize bitmap page
*/
bitmapbuf = XLogInitBufferForRedo(record, 0);
_hash_initbitmapbuffer(bitmapbuf, xlrec->bmsize, true);
PageSetLSN(BufferGetPage(bitmapbuf), lsn);
MarkBufferDirty(bitmapbuf);
UnlockReleaseBuffer(bitmapbuf);
/* add the new bitmap page to the metapage's list of bitmaps */
if (XLogReadBufferForRedo(record, 1, &metabuf) == BLK_NEEDS_REDO)
{
/*
* Note: in normal operation, we'd update the metapage while still
* holding lock on the bitmap page. But during replay it's not
* necessary to hold that lock, since nobody can see it yet; the
* creating transaction hasn't yet committed.
*/
page = BufferGetPage(metabuf);
metap = HashPageGetMeta(page);
num_buckets = metap->hashm_maxbucket + 1;
metap->hashm_mapp[metap->hashm_nmaps] = num_buckets + 1;
metap->hashm_nmaps++;
PageSetLSN(page, lsn);
MarkBufferDirty(metabuf);
}
if (BufferIsValid(metabuf))
UnlockReleaseBuffer(metabuf);
}
/*
* _hash_get_newblock_from_oldbucket() -- get the block number of a bucket
* that will be generated after split from old bucket.
*
* This is used to find the new bucket from old bucket based on current table
* half. It is mainly required to finish the incomplete splits where we are
* sure that not more than one bucket could have split in progress from old
* bucket.
*/
BlockNumber
_hash_get_newblock_from_oldbucket(Relation rel, Bucket old_bucket)
{
Bucket new_bucket;
Buffer metabuf;
HashMetaPage metap;
BlockNumber blkno;
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
new_bucket = _hash_get_newbucket_from_oldbucket(rel, old_bucket,
metap->hashm_lowmask,
metap->hashm_maxbucket);
blkno = BUCKET_TO_BLKNO(metap, new_bucket);
_hash_relbuf(rel, metabuf);
return blkno;
}
/*
* _hash_getcachedmetap() -- Returns cached metapage data.
*
* If metabuf is not InvalidBuffer, caller must hold a pin, but no lock, on
* the metapage. If not set, we'll set it before returning if we have to
* refresh the cache, and return with a pin but no lock on it; caller is
* responsible for releasing the pin.
*
* We refresh the cache if it's not initialized yet or force_refresh is true.
*/
HashMetaPage
_hash_getcachedmetap(Relation rel, Buffer *metabuf, bool force_refresh)
{
Page page;
Assert(metabuf);
if (force_refresh || rel->rd_amcache == NULL)
{
char *cache = NULL;
/*
* It's important that we don't set rd_amcache to an invalid value.
* Either MemoryContextAlloc or _hash_getbuf could fail, so don't
* install a pointer to the newly-allocated storage in the actual
* relcache entry until both have succeeeded.
*/
if (rel->rd_amcache == NULL)
cache = MemoryContextAlloc(rel->rd_indexcxt,
sizeof(HashMetaPageData));
/* Read the metapage. */
if (BufferIsValid(*metabuf))
LockBuffer(*metabuf, BUFFER_LOCK_SHARE);
else
*metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ,
LH_META_PAGE);
page = BufferGetPage(*metabuf);
/* Populate the cache. */
if (rel->rd_amcache == NULL)
rel->rd_amcache = cache;
memcpy(rel->rd_amcache, HashPageGetMeta(page),
sizeof(HashMetaPageData));
/* Release metapage lock, but keep the pin. */
LockBuffer(*metabuf, BUFFER_LOCK_UNLOCK);
}
return (HashMetaPage) rel->rd_amcache;
}
/*
* replay for update meta page
*/
static void
hash_xlog_update_meta_page(XLogReaderState *record)
{
HashMetaPage metap;
XLogRecPtr lsn = record->EndRecPtr;
xl_hash_update_meta_page *xldata = (xl_hash_update_meta_page *) XLogRecGetData(record);
Buffer metabuf;
Page page;
if (XLogReadBufferForRedo(record, 0, &metabuf) == BLK_NEEDS_REDO)
{
page = BufferGetPage(metabuf);
metap = HashPageGetMeta(page);
metap->hashm_ntuples = xldata->ntuples;
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);
}
/*
* 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;
}
/*
* 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;
}
/* ------------------------------------------------
* hash_bitmap_info()
*
* Get bitmap information for a particular overflow page
*
* Usage: SELECT * FROM hash_bitmap_info('con_hash_index'::regclass, 5);
* ------------------------------------------------
*/
Datum
hash_bitmap_info(PG_FUNCTION_ARGS)
{
Oid indexRelid = PG_GETARG_OID(0);
uint64 ovflblkno = PG_GETARG_INT64(1);
HashMetaPage metap;
Buffer metabuf,
mapbuf;
BlockNumber bitmapblkno;
Page mappage;
bool bit = false;
TupleDesc tupleDesc;
Relation indexRel;
uint32 ovflbitno;
int32 bitmappage,
bitmapbit;
HeapTuple tuple;
int i,
j;
Datum values[3];
bool nulls[3];
uint32 *freep;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use raw page functions"))));
indexRel = index_open(indexRelid, AccessShareLock);
if (!IS_HASH(indexRel))
elog(ERROR, "relation \"%s\" is not a hash index",
RelationGetRelationName(indexRel));
if (RELATION_IS_OTHER_TEMP(indexRel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
if (ovflblkno >= RelationGetNumberOfBlocks(indexRel))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("block number " UINT64_FORMAT " is out of range for relation \"%s\"",
ovflblkno, RelationGetRelationName(indexRel))));
/* Read the metapage so we can determine which bitmap page to use */
metabuf = _hash_getbuf(indexRel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* Reject attempt to read the bit for a metapage or bitmap page; this is
* only meaningful for overflow pages.
*/
if (ovflblkno == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid overflow block number %u",
(BlockNumber) ovflblkno)));
for (i = 0; i < metap->hashm_nmaps; i++)
if (metap->hashm_mapp[i] == ovflblkno)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid overflow block number %u",
(BlockNumber) ovflblkno)));
/*
* Identify overflow bit number. This will error out for primary bucket
* pages, and we've already rejected the metapage and bitmap pages above.
*/
ovflbitno = _hash_ovflblkno_to_bitno(metap, (BlockNumber) ovflblkno);
bitmappage = ovflbitno >> BMPG_SHIFT(metap);
bitmapbit = ovflbitno & BMPG_MASK(metap);
if (bitmappage >= metap->hashm_nmaps)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid overflow block number %u",
(BlockNumber) ovflblkno)));
bitmapblkno = metap->hashm_mapp[bitmappage];
_hash_relbuf(indexRel, metabuf);
/* Check the status of bitmap bit for overflow page */
mapbuf = _hash_getbuf(indexRel, bitmapblkno, HASH_READ, LH_BITMAP_PAGE);
mappage = BufferGetPage(mapbuf);
freep = HashPageGetBitmap(mappage);
bit = ISSET(freep, bitmapbit) != 0;
_hash_relbuf(indexRel, mapbuf);
index_close(indexRel, AccessShareLock);
/* 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");
tupleDesc = BlessTupleDesc(tupleDesc);
MemSet(nulls, 0, sizeof(nulls));
j = 0;
values[j++] = Int64GetDatum((int64) bitmapblkno);
values[j++] = Int32GetDatum(bitmapbit);
values[j++] = BoolGetDatum(bit);
tuple = heap_form_tuple(tupleDesc, values, nulls);
PG_RETURN_DATUM(HeapTupleGetDatum(tuple));
}
/*
* _hash_metapinit() -- Initialize the metadata page of a hash index,
* the initial buckets, and the initial bitmap page.
*
* The initial number of buckets is dependent on num_tuples, an estimate
* of the number of tuples to be loaded into the index initially. The
* chosen number of buckets is returned.
*
* 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.
*/
uint32
_hash_metapinit(Relation rel, double num_tuples, ForkNumber forkNum)
{
HashMetaPage metap;
HashPageOpaque pageopaque;
Buffer metabuf;
Buffer buf;
Page pg;
int32 data_width;
int32 item_width;
int32 ffactor;
double dnumbuckets;
uint32 num_buckets;
uint32 log2_num_buckets;
uint32 i;
/* safety check */
if (RelationGetNumberOfBlocksInFork(rel, forkNum) != 0)
elog(ERROR, "cannot initialize non-empty hash index \"%s\"",
RelationGetRelationName(rel));
/*
* Determine the target fill factor (in tuples per bucket) for this index.
* The idea is to make the fill factor correspond to pages about as full
* as the user-settable fillfactor parameter says. We can compute it
* exactly since the index datatype (i.e. uint32 hash key) is fixed-width.
*/
data_width = sizeof(uint32);
item_width = MAXALIGN(sizeof(IndexTupleData)) + MAXALIGN(data_width) +
sizeof(ItemIdData); /* include the line pointer */
ffactor = RelationGetTargetPageUsage(rel, HASH_DEFAULT_FILLFACTOR) / item_width;
/* keep to a sane range */
if (ffactor < 10)
ffactor = 10;
/*
* Choose the number of initial bucket pages to match the fill factor
* given the estimated number of tuples. We round up the result to the
* next power of 2, however, and always force at least 2 bucket pages. The
* upper limit is determined by considerations explained in
* _hash_expandtable().
*/
dnumbuckets = num_tuples / ffactor;
if (dnumbuckets <= 2.0)
num_buckets = 2;
else if (dnumbuckets >= (double) 0x40000000)
num_buckets = 0x40000000;
else
num_buckets = ((uint32) 1) << _hash_log2((uint32) dnumbuckets);
log2_num_buckets = _hash_log2(num_buckets);
Assert(num_buckets == (((uint32) 1) << log2_num_buckets));
Assert(log2_num_buckets < HASH_MAX_SPLITPOINTS);
/*
* We initialize the metapage, the first N bucket pages, and the first
* bitmap page in sequence, using _hash_getnewbuf to cause smgrextend()
* calls to occur. This ensures that the smgr level has the right idea of
* the physical index length.
*/
metabuf = _hash_getnewbuf(rel, HASH_METAPAGE, forkNum);
pg = BufferGetPage(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_page_id = HASHO_PAGE_ID;
metap = HashPageGetMeta(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 = HashGetMaxBitmapSize(pg);
/* 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)
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));
/*
* Label the index with its primary hash support function's OID. This is
* pretty useless for normal operation (in fact, hashm_procid is not used
* anywhere), but it might be handy for forensic purposes so we keep it.
*/
metap->hashm_procid = index_getprocid(rel, 1, HASHPROC);
/*
* We initialize the index with N buckets, 0 .. N-1, occupying physical
* blocks 1 to N. The first freespace bitmap page is in block N+1. Since
* N is a power of 2, we can set the masks this way:
*/
metap->hashm_maxbucket = metap->hashm_lowmask = num_buckets - 1;
metap->hashm_highmask = (num_buckets << 1) - 1;
MemSet(metap->hashm_spares, 0, sizeof(metap->hashm_spares));
MemSet(metap->hashm_mapp, 0, sizeof(metap->hashm_mapp));
/* Set up mapping for one spare page after the initial splitpoints */
metap->hashm_spares[log2_num_buckets] = 1;
metap->hashm_ovflpoint = log2_num_buckets;
metap->hashm_firstfree = 0;
/*
* Release buffer lock on the metapage while we initialize buckets.
* Otherwise, we'll be in interrupt holdoff and the CHECK_FOR_INTERRUPTS
* won't accomplish anything. It's a bad idea to hold buffer locks for
* long intervals in any case, since that can block the bgwriter.
*/
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/*
* Initialize the first N buckets
*/
for (i = 0; i < num_buckets; i++)
{
/* Allow interrupts, in case N is huge */
CHECK_FOR_INTERRUPTS();
buf = _hash_getnewbuf(rel, BUCKET_TO_BLKNO(metap, i), forkNum);
pg = BufferGetPage(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_page_id = HASHO_PAGE_ID;
_hash_wrtbuf(rel, buf);
}
/* Now reacquire buffer lock on metapage */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
/*
* Initialize first bitmap page
*/
_hash_initbitmap(rel, metap, num_buckets + 1, forkNum);
/* all done */
_hash_wrtbuf(rel, metabuf);
return num_buckets;
}
/*
* _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.
*
* Add the tuples (itups) to wbuf in this function. We could do that in the
* caller as well, but the advantage of doing it here is we can easily write
* the WAL for XLOG_HASH_SQUEEZE_PAGE operation. Addition of tuples and
* removal of overflow page has to done as an atomic operation, otherwise
* during replay on standby users might find duplicate records.
*
* 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 page, that's next to
* ovflbuf in the bucket chain. We don't acquire the lock on page that's
* prior to ovflbuf in chain if it is same as wbuf because the caller already
* has a lock on same.
*/
BlockNumber
_hash_freeovflpage(Relation rel, Buffer bucketbuf, Buffer ovflbuf,
Buffer wbuf, IndexTuple *itups, OffsetNumber *itup_offsets,
Size *tups_size, uint16 nitups,
BufferAccessStrategy bstrategy)
{
HashMetaPage metap;
Buffer metabuf;
Buffer mapbuf;
BlockNumber ovflblkno;
BlockNumber prevblkno;
BlockNumber blkno;
BlockNumber nextblkno;
BlockNumber writeblkno;
HashPageOpaque ovflopaque;
Page ovflpage;
Page mappage;
uint32 *freep;
uint32 ovflbitno;
int32 bitmappage,
bitmapbit;
Bucket bucket PG_USED_FOR_ASSERTS_ONLY;
Buffer prevbuf = InvalidBuffer;
Buffer nextbuf = InvalidBuffer;
bool update_metap = false;
/* 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;
writeblkno = BufferGetBlockNumber(wbuf);
bucket = ovflopaque->hasho_bucket;
/*
* 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. Concurrency issues are avoided by using lock chaining as
* described atop hashbucketcleanup.
*/
if (BlockNumberIsValid(prevblkno))
{
if (prevblkno == writeblkno)
prevbuf = wbuf;
else
prevbuf = _hash_getbuf_with_strategy(rel,
prevblkno,
HASH_WRITE,
LH_BUCKET_PAGE | LH_OVERFLOW_PAGE,
bstrategy);
}
if (BlockNumberIsValid(nextblkno))
nextbuf = _hash_getbuf_with_strategy(rel,
nextblkno,
HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
/* 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 = _hash_ovflblkno_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 */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
/* read the bitmap page to clear the bitmap bit */
mapbuf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BITMAP_PAGE);
mappage = BufferGetPage(mapbuf);
freep = HashPageGetBitmap(mappage);
Assert(ISSET(freep, bitmapbit));
/* Get write-lock on metapage to update firstfree */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
/* This operation needs to log multiple tuples, prepare WAL for that */
if (RelationNeedsWAL(rel))
XLogEnsureRecordSpace(HASH_XLOG_FREE_OVFL_BUFS, 4 + nitups);
START_CRIT_SECTION();
/*
* we have to insert tuples 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 them).
*/
if (nitups > 0)
{
_hash_pgaddmultitup(rel, wbuf, itups, itup_offsets, nitups);
MarkBufferDirty(wbuf);
}
/*
* Reinitialize the freed overflow page. Just zeroing the page won't
* work, because WAL replay routines expect pages to be initialized. See
* explanation of RBM_NORMAL mode atop XLogReadBufferExtended. We are
* careful to make the special space valid here so that tools like
* pageinspect won't get confused.
*/
_hash_pageinit(ovflpage, BufferGetPageSize(ovflbuf));
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = InvalidBlockNumber;
ovflopaque->hasho_nextblkno = InvalidBlockNumber;
ovflopaque->hasho_bucket = -1;
ovflopaque->hasho_flag = LH_UNUSED_PAGE;
ovflopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(ovflbuf);
if (BufferIsValid(prevbuf))
{
Page prevpage = BufferGetPage(prevbuf);
HashPageOpaque prevopaque = (HashPageOpaque) PageGetSpecialPointer(prevpage);
Assert(prevopaque->hasho_bucket == bucket);
prevopaque->hasho_nextblkno = nextblkno;
MarkBufferDirty(prevbuf);
}
if (BufferIsValid(nextbuf))
{
Page nextpage = BufferGetPage(nextbuf);
HashPageOpaque nextopaque = (HashPageOpaque) PageGetSpecialPointer(nextpage);
Assert(nextopaque->hasho_bucket == bucket);
nextopaque->hasho_prevblkno = prevblkno;
MarkBufferDirty(nextbuf);
}
/* Clear the bitmap bit to indicate that this overflow page is free */
CLRBIT(freep, bitmapbit);
MarkBufferDirty(mapbuf);
/* if this is now the first free page, update hashm_firstfree */
if (ovflbitno < metap->hashm_firstfree)
{
metap->hashm_firstfree = ovflbitno;
update_metap = true;
MarkBufferDirty(metabuf);
}
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
xl_hash_squeeze_page xlrec;
XLogRecPtr recptr;
int i;
xlrec.prevblkno = prevblkno;
xlrec.nextblkno = nextblkno;
xlrec.ntups = nitups;
xlrec.is_prim_bucket_same_wrt = (wbuf == bucketbuf);
xlrec.is_prev_bucket_same_wrt = (wbuf == prevbuf);
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashSqueezePage);
/*
* bucket buffer needs to be registered to ensure that we can acquire
* a cleanup lock on it during replay.
*/
if (!xlrec.is_prim_bucket_same_wrt)
XLogRegisterBuffer(0, bucketbuf, REGBUF_STANDARD | REGBUF_NO_IMAGE);
XLogRegisterBuffer(1, wbuf, REGBUF_STANDARD);
if (xlrec.ntups > 0)
{
XLogRegisterBufData(1, (char *) itup_offsets,
nitups * sizeof(OffsetNumber));
for (i = 0; i < nitups; i++)
XLogRegisterBufData(1, (char *) itups[i], tups_size[i]);
}
XLogRegisterBuffer(2, ovflbuf, REGBUF_STANDARD);
/*
* If prevpage and the writepage (block in which we are moving tuples
* from overflow) are same, then no need to separately register
* prevpage. During replay, we can directly update the nextblock in
* writepage.
*/
if (BufferIsValid(prevbuf) && !xlrec.is_prev_bucket_same_wrt)
XLogRegisterBuffer(3, prevbuf, REGBUF_STANDARD);
if (BufferIsValid(nextbuf))
XLogRegisterBuffer(4, nextbuf, REGBUF_STANDARD);
XLogRegisterBuffer(5, mapbuf, REGBUF_STANDARD);
XLogRegisterBufData(5, (char *) &bitmapbit, sizeof(uint32));
if (update_metap)
{
XLogRegisterBuffer(6, metabuf, REGBUF_STANDARD);
XLogRegisterBufData(6, (char *) &metap->hashm_firstfree, sizeof(uint32));
}
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_SQUEEZE_PAGE);
PageSetLSN(BufferGetPage(wbuf), recptr);
PageSetLSN(BufferGetPage(ovflbuf), recptr);
if (BufferIsValid(prevbuf) && !xlrec.is_prev_bucket_same_wrt)
PageSetLSN(BufferGetPage(prevbuf), recptr);
if (BufferIsValid(nextbuf))
PageSetLSN(BufferGetPage(nextbuf), recptr);
PageSetLSN(BufferGetPage(mapbuf), recptr);
if (update_metap)
PageSetLSN(BufferGetPage(metabuf), recptr);
}
END_CRIT_SECTION();
/* release previous bucket if it is not same as write bucket */
if (BufferIsValid(prevbuf) && prevblkno != writeblkno)
_hash_relbuf(rel, prevbuf);
if (BufferIsValid(ovflbuf))
_hash_relbuf(rel, ovflbuf);
if (BufferIsValid(nextbuf))
_hash_relbuf(rel, nextbuf);
_hash_relbuf(rel, mapbuf);
_hash_relbuf(rel, metabuf);
return nextblkno;
}
/*
* _hash_addovflpage
*
* Add an overflow page to the bucket whose last page is pointed to by 'buf'.
*
* On entry, the caller must hold a pin but no lock on 'buf'. The pin is
* dropped before exiting (we assume the caller is not interested in 'buf'
* anymore) if not asked to retain. The pin will be retained only for the
* primary bucket. The returned overflow page will be pinned and
* write-locked; it is guaranteed to be empty.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* That buffer is returned in the same state.
*
* NB: since this could be executed concurrently by multiple processes,
* one should not assume that the returned overflow page will be the
* immediate successor of the originally passed 'buf'. Additional overflow
* pages might have been added to the bucket chain in between.
*/
Buffer
_hash_addovflpage(Relation rel, Buffer metabuf, Buffer buf, bool retain_pin)
{
Buffer ovflbuf;
Page page;
Page ovflpage;
HashPageOpaque pageopaque;
HashPageOpaque ovflopaque;
HashMetaPage metap;
Buffer mapbuf = InvalidBuffer;
Buffer newmapbuf = InvalidBuffer;
BlockNumber blkno;
uint32 orig_firstfree;
uint32 splitnum;
uint32 *freep = NULL;
uint32 max_ovflpg;
uint32 bit;
uint32 bitmap_page_bit;
uint32 first_page;
uint32 last_bit;
uint32 last_page;
uint32 i,
j;
bool page_found = false;
/*
* Write-lock the tail page. Here, we need to maintain locking order such
* that, first acquire the lock on tail page of bucket, then on meta page
* to find and lock the bitmap page and if it is found, then lock on meta
* page is released, then finally acquire the lock on new overflow buffer.
* We need this locking order to avoid deadlock with backends that are
* doing inserts.
*
* Note: We could have avoided locking many buffers here if we made two
* WAL records for acquiring an overflow page (one to allocate an overflow
* page and another to add it to overflow bucket chain). However, doing
* so can leak an overflow page, if the system crashes after allocation.
* Needless to say, it is better to have a single record from a
* performance point of view as well.
*/
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
/* probably redundant... */
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
/* loop to find current tail page, in case someone else inserted too */
for (;;)
{
BlockNumber nextblkno;
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
nextblkno = pageopaque->hasho_nextblkno;
if (!BlockNumberIsValid(nextblkno))
break;
/* we assume we do not need to write the unmodified page */
if (retain_pin)
{
/* pin will be retained only for the primary bucket page */
Assert((pageopaque->hasho_flag & LH_PAGE_TYPE) == LH_BUCKET_PAGE);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
}
else
_hash_relbuf(rel, buf);
retain_pin = false;
buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
}
/* Get exclusive lock on the meta page */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
_hash_checkpage(rel, metabuf, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/* start search at hashm_firstfree */
orig_firstfree = metap->hashm_firstfree;
first_page = orig_firstfree >> BMPG_SHIFT(metap);
bit = orig_firstfree & BMPG_MASK(metap);
i = first_page;
j = bit / BITS_PER_MAP;
bit &= ~(BITS_PER_MAP - 1);
/* outer loop iterates once per bitmap page */
for (;;)
{
BlockNumber mapblkno;
Page mappage;
uint32 last_inpage;
/* want to end search with the last existing overflow page */
splitnum = metap->hashm_ovflpoint;
max_ovflpg = metap->hashm_spares[splitnum] - 1;
last_page = max_ovflpg >> BMPG_SHIFT(metap);
last_bit = max_ovflpg & BMPG_MASK(metap);
if (i > last_page)
break;
Assert(i < metap->hashm_nmaps);
mapblkno = metap->hashm_mapp[i];
if (i == last_page)
last_inpage = last_bit;
else
last_inpage = BMPGSZ_BIT(metap) - 1;
/* Release exclusive lock on metapage while reading bitmap page */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
mapbuf = _hash_getbuf(rel, mapblkno, HASH_WRITE, LH_BITMAP_PAGE);
mappage = BufferGetPage(mapbuf);
freep = HashPageGetBitmap(mappage);
for (; bit <= last_inpage; j++, bit += BITS_PER_MAP)
{
if (freep[j] != ALL_SET)
{
page_found = true;
/* Reacquire exclusive lock on the meta page */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
/* convert bit to bit number within page */
bit += _hash_firstfreebit(freep[j]);
bitmap_page_bit = bit;
/* convert bit to absolute bit number */
bit += (i << BMPG_SHIFT(metap));
/* Calculate address of the recycled overflow page */
blkno = bitno_to_blkno(metap, bit);
/* Fetch and init the recycled page */
ovflbuf = _hash_getinitbuf(rel, blkno);
goto found;
}
}
/* No free space here, try to advance to next map page */
_hash_relbuf(rel, mapbuf);
mapbuf = InvalidBuffer;
i++;
j = 0; /* scan from start of next map page */
bit = 0;
/* Reacquire exclusive lock on the meta page */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
}
/*
* No free pages --- have to extend the relation to add an overflow page.
* First, check to see if we have to add a new bitmap page too.
*/
if (last_bit == (uint32) (BMPGSZ_BIT(metap) - 1))
{
/*
* We create the new bitmap page with all pages marked "in use".
* Actually two pages in the new bitmap's range will exist
* immediately: the bitmap page itself, and the following page which
* is the one we return to the caller. Both of these are correctly
* marked "in use". Subsequent pages do not exist yet, but it is
* convenient to pre-mark them as "in use" too.
*/
bit = metap->hashm_spares[splitnum];
/* metapage already has a write lock */
if (metap->hashm_nmaps >= HASH_MAX_BITMAPS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("out of overflow pages in hash index \"%s\"",
RelationGetRelationName(rel))));
newmapbuf = _hash_getnewbuf(rel, bitno_to_blkno(metap, bit), MAIN_FORKNUM);
}
else
{
/*
* Nothing to do here; since the page will be past the last used page,
* we know its bitmap bit was preinitialized to "in use".
*/
}
/* Calculate address of the new overflow page */
bit = BufferIsValid(newmapbuf) ?
metap->hashm_spares[splitnum] + 1 : metap->hashm_spares[splitnum];
blkno = bitno_to_blkno(metap, bit);
/*
* Fetch the page with _hash_getnewbuf to ensure smgr's idea of the
* relation length stays in sync with ours. XXX It's annoying to do this
* with metapage write lock held; would be better to use a lock that
* doesn't block incoming searches.
*
* It is okay to hold two buffer locks here (one on tail page of bucket
* and other on new overflow page) since there cannot be anyone else
* contending for access to ovflbuf.
*/
ovflbuf = _hash_getnewbuf(rel, blkno, MAIN_FORKNUM);
found:
/*
* Do the update. No ereport(ERROR) until changes are logged. We want to
* log the changes for bitmap page and overflow page together to avoid
* loss of pages in case the new page is added.
*/
START_CRIT_SECTION();
if (page_found)
{
Assert(BufferIsValid(mapbuf));
/* mark page "in use" in the bitmap */
SETBIT(freep, bitmap_page_bit);
MarkBufferDirty(mapbuf);
}
else
{
/* update the count to indicate new overflow page is added */
metap->hashm_spares[splitnum]++;
if (BufferIsValid(newmapbuf))
{
_hash_initbitmapbuffer(newmapbuf, metap->hashm_bmsize, false);
MarkBufferDirty(newmapbuf);
/* add the new bitmap page to the metapage's list of bitmaps */
metap->hashm_mapp[metap->hashm_nmaps] = BufferGetBlockNumber(newmapbuf);
metap->hashm_nmaps++;
metap->hashm_spares[splitnum]++;
}
MarkBufferDirty(metabuf);
/*
* for new overflow page, we don't need to explicitly set the bit in
* bitmap page, as by default that will be set to "in use".
*/
}
/*
* Adjust hashm_firstfree to avoid redundant searches. But don't risk
* changing it if someone moved it while we were searching bitmap pages.
*/
if (metap->hashm_firstfree == orig_firstfree)
{
metap->hashm_firstfree = bit + 1;
MarkBufferDirty(metabuf);
}
/* initialize new overflow page */
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = BufferGetBlockNumber(buf);
ovflopaque->hasho_nextblkno = InvalidBlockNumber;
ovflopaque->hasho_bucket = pageopaque->hasho_bucket;
ovflopaque->hasho_flag = LH_OVERFLOW_PAGE;
ovflopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(ovflbuf);
/* logically chain overflow page to previous page */
pageopaque->hasho_nextblkno = BufferGetBlockNumber(ovflbuf);
MarkBufferDirty(buf);
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
XLogRecPtr recptr;
xl_hash_add_ovfl_page xlrec;
xlrec.bmpage_found = page_found;
xlrec.bmsize = metap->hashm_bmsize;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashAddOvflPage);
XLogRegisterBuffer(0, ovflbuf, REGBUF_WILL_INIT);
XLogRegisterBufData(0, (char *) &pageopaque->hasho_bucket, sizeof(Bucket));
XLogRegisterBuffer(1, buf, REGBUF_STANDARD);
if (BufferIsValid(mapbuf))
{
XLogRegisterBuffer(2, mapbuf, REGBUF_STANDARD);
XLogRegisterBufData(2, (char *) &bitmap_page_bit, sizeof(uint32));
}
if (BufferIsValid(newmapbuf))
XLogRegisterBuffer(3, newmapbuf, REGBUF_WILL_INIT);
XLogRegisterBuffer(4, metabuf, REGBUF_STANDARD);
XLogRegisterBufData(4, (char *) &metap->hashm_firstfree, sizeof(uint32));
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_ADD_OVFL_PAGE);
PageSetLSN(BufferGetPage(ovflbuf), recptr);
PageSetLSN(BufferGetPage(buf), recptr);
if (BufferIsValid(mapbuf))
PageSetLSN(BufferGetPage(mapbuf), recptr);
if (BufferIsValid(newmapbuf))
PageSetLSN(BufferGetPage(newmapbuf), recptr);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
END_CRIT_SECTION();
if (retain_pin)
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
else
_hash_relbuf(rel, buf);
if (BufferIsValid(mapbuf))
_hash_relbuf(rel, mapbuf);
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
if (BufferIsValid(newmapbuf))
_hash_relbuf(rel, newmapbuf);
return ovflbuf;
}
/*
* _hash_doinsert() -- Handle insertion of a single index tuple.
*
* This routine is called by the public interface routines, hashbuild
* and hashinsert. By here, itup is completely filled in.
*/
void
_hash_doinsert(Relation rel, IndexTuple itup)
{
Buffer buf;
Buffer metabuf;
HashMetaPage metap;
BlockNumber blkno;
Page page;
HashPageOpaque pageopaque;
Size itemsz;
bool do_expand;
uint32 hashkey;
Bucket bucket;
/*
* Get the hash key for the item (it's stored in the index tuple itself).
*/
hashkey = _hash_get_indextuple_hashkey(itup);
/* compute item size too */
itemsz = IndexTupleDSize(*itup);
itemsz = MAXALIGN(itemsz); /* be safe, PageAddItem will do this but we
* need to be consistent */
/*
* 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));
/*
* Check whether the item can fit on a hash page at all. (Eventually, we
* ought to try to apply TOAST methods if not.) Note that at this point,
* itemsz doesn't include the ItemId.
*
* XXX this is useless code if we are only storing hash keys.
*/
if (itemsz > HashMaxItemSize((Page) metap))
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %lu exceeds hash maximum %lu",
(unsigned long) itemsz,
(unsigned long) HashMaxItemSize((Page) metap)),
errhint("Values larger than a buffer page cannot be indexed.")));
/*
* 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);
/* release lock on metapage, but keep pin since we'll need it again */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/*
* Acquire share lock on target bucket; then we can release split lock.
*/
_hash_getlock(rel, blkno, HASH_SHARE);
_hash_droplock(rel, 0, HASH_SHARE);
/* Fetch the primary bucket page for the bucket */
buf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BUCKET_PAGE);
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
Assert(pageopaque->hasho_bucket == bucket);
/* Do the insertion */
while (PageGetFreeSpace(page) < itemsz)
{
/*
* no space on this page; check for an overflow page
*/
BlockNumber nextblkno = pageopaque->hasho_nextblkno;
if (BlockNumberIsValid(nextblkno))
{
/*
* ovfl page exists; go get it. if it doesn't have room, we'll
* find out next pass through the loop test above.
*/
_hash_relbuf(rel, buf);
buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
page = BufferGetPage(buf);
}
else
{
/*
* we're at the end of the bucket chain and we haven't found a
* page with enough room. allocate a new overflow page.
*/
/* release our write lock without modifying buffer */
_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);
/* chain to a new overflow page */
buf = _hash_addovflpage(rel, metabuf, buf);
page = BufferGetPage(buf);
/* should fit now, given test above */
Assert(PageGetFreeSpace(page) >= itemsz);
}
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
Assert(pageopaque->hasho_flag == LH_OVERFLOW_PAGE);
Assert(pageopaque->hasho_bucket == bucket);
}
/* found page with enough space, so add the item here */
(void) _hash_pgaddtup(rel, buf, itemsz, itup);
/* write and release the modified page */
_hash_wrtbuf(rel, buf);
/* We can drop the bucket lock now */
_hash_droplock(rel, blkno, HASH_SHARE);
/*
* Write-lock the metapage so we can increment the tuple count. After
* incrementing it, check to see if it's time for a split.
*/
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
metap->hashm_ntuples += 1;
/* Make sure this stays in sync with _hash_expandtable() */
do_expand = metap->hashm_ntuples >
(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1);
/* Write out the metapage and drop lock, but keep pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/* Attempt to split if a split is needed */
if (do_expand)
_hash_expandtable(rel, metabuf);
/* Finally drop our pin on the metapage */
_hash_dropbuf(rel, metabuf);
}
/*
* Attempt to expand the hash table by creating one new bucket.
*
* This will silently do nothing if we don't get cleanup lock on old or
* new bucket.
*
* Complete the pending splits and remove the tuples from old bucket,
* if there are any left over from the previous split.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state.
*/
void
_hash_expandtable(Relation rel, Buffer metabuf)
{
HashMetaPage metap;
Bucket old_bucket;
Bucket new_bucket;
uint32 spare_ndx;
BlockNumber start_oblkno;
BlockNumber start_nblkno;
Buffer buf_nblkno;
Buffer buf_oblkno;
Page opage;
Page npage;
HashPageOpaque oopaque;
HashPageOpaque nopaque;
uint32 maxbucket;
uint32 highmask;
uint32 lowmask;
bool metap_update_masks = false;
bool metap_update_splitpoint = false;
restart_expand:
/*
* Write-lock the meta page. It used to be necessary to acquire a
* heavyweight lock to begin a split, but that is no longer required.
*/
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
_hash_checkpage(rel, metabuf, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* Check to see if split is still needed; someone else might have already
* done one while we waited for the lock.
*
* Make sure this stays in sync with _hash_doinsert()
*/
if (metap->hashm_ntuples <=
(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1))
goto fail;
/*
* Can't split anymore if maxbucket has reached its maximum possible
* value.
*
* Ideally we'd allow bucket numbers up to UINT_MAX-1 (no higher because
* the calculation maxbucket+1 mustn't overflow). Currently we restrict
* to half that because of overflow looping in _hash_log2() and
* insufficient space in hashm_spares[]. It's moot anyway because an
* index with 2^32 buckets would certainly overflow BlockNumber and hence
* _hash_alloc_buckets() would fail, but if we supported buckets smaller
* than a disk block then this would be an independent constraint.
*
* If you change this, see also the maximum initial number of buckets in
* _hash_init().
*/
if (metap->hashm_maxbucket >= (uint32) 0x7FFFFFFE)
goto fail;
/*
* Determine which bucket is to be split, and attempt to take cleanup lock
* on the old bucket. If we can't get the lock, give up.
*
* The cleanup lock protects us not only against other backends, but
* against our own backend as well.
*
* The cleanup lock is mainly to protect the split from concurrent
* inserts. See src/backend/access/hash/README, Lock Definitions for
* further details. Due to this locking restriction, if there is any
* pending scan, the split will give up which is not good, but harmless.
*/
new_bucket = metap->hashm_maxbucket + 1;
old_bucket = (new_bucket & metap->hashm_lowmask);
start_oblkno = BUCKET_TO_BLKNO(metap, old_bucket);
buf_oblkno = _hash_getbuf_with_condlock_cleanup(rel, start_oblkno, LH_BUCKET_PAGE);
if (!buf_oblkno)
goto fail;
opage = BufferGetPage(buf_oblkno);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
/*
* We want to finish the split from a bucket as there is no apparent
* benefit by not doing so and it will make the code complicated to finish
* the split that involves multiple buckets considering the case where new
* split also fails. We don't need to consider the new bucket for
* completing the split here as it is not possible that a re-split of new
* bucket starts when there is still a pending split from old bucket.
*/
if (H_BUCKET_BEING_SPLIT(oopaque))
{
/*
* Copy bucket mapping info now; refer the comment in code below where
* we copy this information before calling _hash_splitbucket to see
* why this is okay.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
/*
* Release the lock on metapage and old_bucket, before completing the
* split.
*/
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
LockBuffer(buf_oblkno, BUFFER_LOCK_UNLOCK);
_hash_finish_split(rel, metabuf, buf_oblkno, old_bucket, maxbucket,
highmask, lowmask);
/* release the pin on old buffer and retry for expand. */
_hash_dropbuf(rel, buf_oblkno);
goto restart_expand;
}
/*
* Clean the tuples remained from the previous split. This operation
* requires cleanup lock and we already have one on the old bucket, so
* let's do it. We also don't want to allow further splits from the bucket
* till the garbage of previous split is cleaned. This has two
* advantages; first, it helps in avoiding the bloat due to garbage and
* second is, during cleanup of bucket, we are always sure that the
* garbage tuples belong to most recently split bucket. On the contrary,
* if we allow cleanup of bucket after meta page is updated to indicate
* the new split and before the actual split, the cleanup operation won't
* be able to decide whether the tuple has been moved to the newly created
* bucket and ended up deleting such tuples.
*/
if (H_NEEDS_SPLIT_CLEANUP(oopaque))
{
/*
* Copy bucket mapping info now; refer to the comment in code below
* where we copy this information before calling _hash_splitbucket to
* see why this is okay.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
/* Release the metapage lock. */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
hashbucketcleanup(rel, old_bucket, buf_oblkno, start_oblkno, NULL,
maxbucket, highmask, lowmask, NULL, NULL, true,
NULL, NULL);
_hash_dropbuf(rel, buf_oblkno);
goto restart_expand;
}
/*
* There shouldn't be any active scan on new bucket.
*
* Note: it is safe to compute the new bucket's blkno here, even though we
* may still need to update the BUCKET_TO_BLKNO mapping. This is because
* the current value of hashm_spares[hashm_ovflpoint] correctly shows
* where we are going to put a new splitpoint's worth of buckets.
*/
start_nblkno = BUCKET_TO_BLKNO(metap, new_bucket);
/*
* If the split point is increasing we need to allocate a new batch of
* bucket pages.
*/
spare_ndx = _hash_spareindex(new_bucket + 1);
if (spare_ndx > metap->hashm_ovflpoint)
{
uint32 buckets_to_add;
Assert(spare_ndx == metap->hashm_ovflpoint + 1);
/*
* We treat allocation of buckets as a separate WAL-logged action.
* Even if we fail after this operation, won't leak bucket pages;
* rather, the next split will consume this space. In any case, even
* without failure we don't use all the space in one split operation.
*/
buckets_to_add = _hash_get_totalbuckets(spare_ndx) - new_bucket;
if (!_hash_alloc_buckets(rel, start_nblkno, buckets_to_add))
{
/* can't split due to BlockNumber overflow */
_hash_relbuf(rel, buf_oblkno);
goto fail;
}
}
/*
* Physically allocate the new bucket's primary page. We want to do this
* before changing the metapage's mapping info, in case we can't get the
* disk space. Ideally, we don't need to check for cleanup lock on new
* bucket as no other backend could find this bucket unless meta page is
* updated. However, it is good to be consistent with old bucket locking.
*/
buf_nblkno = _hash_getnewbuf(rel, start_nblkno, MAIN_FORKNUM);
if (!IsBufferCleanupOK(buf_nblkno))
{
_hash_relbuf(rel, buf_oblkno);
_hash_relbuf(rel, buf_nblkno);
goto fail;
}
/*
* Since we are scribbling on the pages in the shared buffers, establish a
* critical section. Any failure in this next code leaves us with a big
* problem: the metapage is effectively corrupt but could get written back
* to disk.
*/
START_CRIT_SECTION();
/*
* Okay to proceed with split. Update the metapage bucket mapping info.
*/
metap->hashm_maxbucket = new_bucket;
if (new_bucket > metap->hashm_highmask)
{
/* Starting a new doubling */
metap->hashm_lowmask = metap->hashm_highmask;
metap->hashm_highmask = new_bucket | metap->hashm_lowmask;
metap_update_masks = true;
}
/*
* If the split point is increasing we need to adjust the hashm_spares[]
* array and hashm_ovflpoint so that future overflow pages will be created
* beyond this new batch of bucket pages.
*/
if (spare_ndx > metap->hashm_ovflpoint)
{
metap->hashm_spares[spare_ndx] = metap->hashm_spares[metap->hashm_ovflpoint];
metap->hashm_ovflpoint = spare_ndx;
metap_update_splitpoint = true;
}
MarkBufferDirty(metabuf);
/*
* Copy bucket mapping info now; this saves re-accessing the meta page
* inside _hash_splitbucket's inner loop. Note that once we drop the
* split lock, other splits could begin, so these values might be out of
* date before _hash_splitbucket finishes. That's okay, since all it
* needs is to tell which of these two buckets to map hashkeys into.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
opage = BufferGetPage(buf_oblkno);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
/*
* Mark the old bucket to indicate that split is in progress. (At
* operation end, we will clear the split-in-progress flag.) Also, for a
* primary bucket page, hasho_prevblkno stores the number of buckets that
* existed as of the last split, so we must update that value here.
*/
oopaque->hasho_flag |= LH_BUCKET_BEING_SPLIT;
oopaque->hasho_prevblkno = maxbucket;
MarkBufferDirty(buf_oblkno);
npage = BufferGetPage(buf_nblkno);
/*
* initialize the new bucket's primary page and mark it to indicate that
* split is in progress.
*/
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
nopaque->hasho_prevblkno = maxbucket;
nopaque->hasho_nextblkno = InvalidBlockNumber;
nopaque->hasho_bucket = new_bucket;
nopaque->hasho_flag = LH_BUCKET_PAGE | LH_BUCKET_BEING_POPULATED;
nopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(buf_nblkno);
/* XLOG stuff */
if (RelationNeedsWAL(rel))
{
xl_hash_split_allocate_page xlrec;
XLogRecPtr recptr;
xlrec.new_bucket = maxbucket;
xlrec.old_bucket_flag = oopaque->hasho_flag;
xlrec.new_bucket_flag = nopaque->hasho_flag;
xlrec.flags = 0;
XLogBeginInsert();
XLogRegisterBuffer(0, buf_oblkno, REGBUF_STANDARD);
XLogRegisterBuffer(1, buf_nblkno, REGBUF_WILL_INIT);
XLogRegisterBuffer(2, metabuf, REGBUF_STANDARD);
if (metap_update_masks)
{
xlrec.flags |= XLH_SPLIT_META_UPDATE_MASKS;
XLogRegisterBufData(2, (char *) &metap->hashm_lowmask, sizeof(uint32));
XLogRegisterBufData(2, (char *) &metap->hashm_highmask, sizeof(uint32));
}
if (metap_update_splitpoint)
{
xlrec.flags |= XLH_SPLIT_META_UPDATE_SPLITPOINT;
XLogRegisterBufData(2, (char *) &metap->hashm_ovflpoint,
sizeof(uint32));
XLogRegisterBufData(2,
(char *) &metap->hashm_spares[metap->hashm_ovflpoint],
sizeof(uint32));
}
XLogRegisterData((char *) &xlrec, SizeOfHashSplitAllocPage);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_SPLIT_ALLOCATE_PAGE);
PageSetLSN(BufferGetPage(buf_oblkno), recptr);
PageSetLSN(BufferGetPage(buf_nblkno), recptr);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
END_CRIT_SECTION();
/* drop lock, but keep pin */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
/* Relocate records to the new bucket */
_hash_splitbucket(rel, metabuf,
old_bucket, new_bucket,
buf_oblkno, buf_nblkno, NULL,
maxbucket, highmask, lowmask);
/* all done, now release the pins on primary buckets. */
_hash_dropbuf(rel, buf_oblkno);
_hash_dropbuf(rel, buf_nblkno);
return;
/* Here if decide not to split or fail to acquire old bucket lock */
fail:
/* We didn't write the metapage, so just drop lock */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
}
/*
* _hash_init_metabuffer() -- Initialize the metadata page of a hash index.
*/
void
_hash_init_metabuffer(Buffer buf, double num_tuples, RegProcedure procid,
uint16 ffactor, bool initpage)
{
HashMetaPage metap;
HashPageOpaque pageopaque;
Page page;
double dnumbuckets;
uint32 num_buckets;
uint32 spare_index;
uint32 i;
/*
* Choose the number of initial bucket pages to match the fill factor
* given the estimated number of tuples. We round up the result to the
* total number of buckets which has to be allocated before using its
* _hashm_spare element. However always force at least 2 bucket pages. The
* upper limit is determined by considerations explained in
* _hash_expandtable().
*/
dnumbuckets = num_tuples / ffactor;
if (dnumbuckets <= 2.0)
num_buckets = 2;
else if (dnumbuckets >= (double) 0x40000000)
num_buckets = 0x40000000;
else
num_buckets = _hash_get_totalbuckets(_hash_spareindex(dnumbuckets));
spare_index = _hash_spareindex(num_buckets);
Assert(spare_index < HASH_MAX_SPLITPOINTS);
page = BufferGetPage(buf);
if (initpage)
_hash_pageinit(page, BufferGetPageSize(buf));
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
pageopaque->hasho_prevblkno = InvalidBlockNumber;
pageopaque->hasho_nextblkno = InvalidBlockNumber;
pageopaque->hasho_bucket = -1;
pageopaque->hasho_flag = LH_META_PAGE;
pageopaque->hasho_page_id = HASHO_PAGE_ID;
metap = HashPageGetMeta(page);
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 = HashGetMaxBitmapSize(page);
/* 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)
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));
/*
* Label the index with its primary hash support function's OID. This is
* pretty useless for normal operation (in fact, hashm_procid is not used
* anywhere), but it might be handy for forensic purposes so we keep it.
*/
metap->hashm_procid = procid;
/*
* We initialize the index with N buckets, 0 .. N-1, occupying physical
* blocks 1 to N. The first freespace bitmap page is in block N+1.
*/
metap->hashm_maxbucket = num_buckets - 1;
/*
* Set highmask as next immediate ((2 ^ x) - 1), which should be
* sufficient to cover num_buckets.
*/
metap->hashm_highmask = (1 << (_hash_log2(num_buckets + 1))) - 1;
metap->hashm_lowmask = (metap->hashm_highmask >> 1);
MemSet(metap->hashm_spares, 0, sizeof(metap->hashm_spares));
MemSet(metap->hashm_mapp, 0, sizeof(metap->hashm_mapp));
/* Set up mapping for one spare page after the initial splitpoints */
metap->hashm_spares[spare_index] = 1;
metap->hashm_ovflpoint = spare_index;
metap->hashm_firstfree = 0;
/*
* Set pd_lower just past the end of the metadata. This is to log full
* page image of metapage in xloginsert.c.
*/
((PageHeader) page)->pd_lower =
((char *) metap + sizeof(HashMetaPageData)) - (char *) page;
}
/*
* _hash_init() -- Initialize the metadata page of a hash index,
* the initial buckets, and the initial bitmap page.
*
* The initial number of buckets is dependent on num_tuples, an estimate
* of the number of tuples to be loaded into the index initially. The
* chosen number of buckets is returned.
*
* 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.
*/
uint32
_hash_init(Relation rel, double num_tuples, ForkNumber forkNum)
{
Buffer metabuf;
Buffer buf;
Buffer bitmapbuf;
Page pg;
HashMetaPage metap;
RegProcedure procid;
int32 data_width;
int32 item_width;
int32 ffactor;
uint32 num_buckets;
uint32 i;
bool use_wal;
/* safety check */
if (RelationGetNumberOfBlocksInFork(rel, forkNum) != 0)
elog(ERROR, "cannot initialize non-empty hash index \"%s\"",
RelationGetRelationName(rel));
/*
* WAL log creation of pages if the relation is persistent, or this is the
* init fork. Init forks for unlogged relations always need to be WAL
* logged.
*/
use_wal = RelationNeedsWAL(rel) || forkNum == INIT_FORKNUM;
/*
* Determine the target fill factor (in tuples per bucket) for this index.
* The idea is to make the fill factor correspond to pages about as full
* as the user-settable fillfactor parameter says. We can compute it
* exactly since the index datatype (i.e. uint32 hash key) is fixed-width.
*/
data_width = sizeof(uint32);
item_width = MAXALIGN(sizeof(IndexTupleData)) + MAXALIGN(data_width) +
sizeof(ItemIdData); /* include the line pointer */
ffactor = RelationGetTargetPageUsage(rel, HASH_DEFAULT_FILLFACTOR) / item_width;
/* keep to a sane range */
if (ffactor < 10)
ffactor = 10;
procid = index_getprocid(rel, 1, HASHSTANDARD_PROC);
/*
* We initialize the metapage, the first N bucket pages, and the first
* bitmap page in sequence, using _hash_getnewbuf to cause smgrextend()
* calls to occur. This ensures that the smgr level has the right idea of
* the physical index length.
*
* Critical section not required, because on error the creation of the
* whole relation will be rolled back.
*/
metabuf = _hash_getnewbuf(rel, HASH_METAPAGE, forkNum);
_hash_init_metabuffer(metabuf, num_tuples, procid, ffactor, false);
MarkBufferDirty(metabuf);
pg = BufferGetPage(metabuf);
metap = HashPageGetMeta(pg);
/* XLOG stuff */
if (use_wal)
{
xl_hash_init_meta_page xlrec;
XLogRecPtr recptr;
xlrec.num_tuples = num_tuples;
xlrec.procid = metap->hashm_procid;
xlrec.ffactor = metap->hashm_ffactor;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashInitMetaPage);
XLogRegisterBuffer(0, metabuf, REGBUF_WILL_INIT);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_INIT_META_PAGE);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
num_buckets = metap->hashm_maxbucket + 1;
/*
* Release buffer lock on the metapage while we initialize buckets.
* Otherwise, we'll be in interrupt holdoff and the CHECK_FOR_INTERRUPTS
* won't accomplish anything. It's a bad idea to hold buffer locks for
* long intervals in any case, since that can block the bgwriter.
*/
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
/*
* Initialize and WAL Log the first N buckets
*/
for (i = 0; i < num_buckets; i++)
{
BlockNumber blkno;
/* Allow interrupts, in case N is huge */
CHECK_FOR_INTERRUPTS();
blkno = BUCKET_TO_BLKNO(metap, i);
buf = _hash_getnewbuf(rel, blkno, forkNum);
_hash_initbuf(buf, metap->hashm_maxbucket, i, LH_BUCKET_PAGE, false);
MarkBufferDirty(buf);
if (use_wal)
log_newpage(&rel->rd_node,
forkNum,
blkno,
BufferGetPage(buf),
true);
_hash_relbuf(rel, buf);
}
/* Now reacquire buffer lock on metapage */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
/*
* Initialize bitmap page
*/
bitmapbuf = _hash_getnewbuf(rel, num_buckets + 1, forkNum);
_hash_initbitmapbuffer(bitmapbuf, metap->hashm_bmsize, false);
MarkBufferDirty(bitmapbuf);
/* add the new bitmap page to the metapage's list of bitmaps */
/* metapage already has a write lock */
if (metap->hashm_nmaps >= HASH_MAX_BITMAPS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("out of overflow pages in hash index \"%s\"",
RelationGetRelationName(rel))));
metap->hashm_mapp[metap->hashm_nmaps] = num_buckets + 1;
metap->hashm_nmaps++;
MarkBufferDirty(metabuf);
/* XLOG stuff */
if (use_wal)
{
xl_hash_init_bitmap_page xlrec;
XLogRecPtr recptr;
xlrec.bmsize = metap->hashm_bmsize;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashInitBitmapPage);
XLogRegisterBuffer(0, bitmapbuf, REGBUF_WILL_INIT);
/*
* This is safe only because nobody else can be modifying the index at
* this stage; it's only visible to the transaction that is creating
* it.
*/
XLogRegisterBuffer(1, metabuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_INIT_BITMAP_PAGE);
PageSetLSN(BufferGetPage(bitmapbuf), recptr);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
/* all done */
_hash_relbuf(rel, bitmapbuf);
_hash_relbuf(rel, metabuf);
return num_buckets;
}
/*
* 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);
}
}
/*
* Attempt to expand the hash table by creating one new bucket.
*
* This will silently do nothing if it cannot get the needed locks.
*
* The caller should hold no locks on the hash index.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state.
*/
void
_hash_expandtable(Relation rel, Buffer metabuf)
{
HashMetaPage metap;
Bucket old_bucket;
Bucket new_bucket;
uint32 spare_ndx;
BlockNumber start_oblkno;
BlockNumber start_nblkno;
uint32 maxbucket;
uint32 highmask;
uint32 lowmask;
/*
* Obtain the page-zero lock to assert the right to begin a split (see
* README).
*
* Note: deadlock should be impossible here. Our own backend could only be
* holding bucket sharelocks due to stopped indexscans; those will not
* block other holders of the page-zero lock, who are only interested in
* acquiring bucket sharelocks themselves. Exclusive bucket locks are
* only taken here and in hashbulkdelete, and neither of these operations
* needs any additional locks to complete. (If, due to some flaw in this
* reasoning, we manage to deadlock anyway, it's okay to error out; the
* index will be left in a consistent state.)
*/
_hash_getlock(rel, 0, HASH_EXCLUSIVE);
/* Write-lock the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
_hash_checkpage(rel, metabuf, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* Check to see if split is still needed; someone else might have already
* done one while we waited for the lock.
*
* Make sure this stays in sync with _hash_doinsert()
*/
if (metap->hashm_ntuples <=
(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1))
goto fail;
/*
* Can't split anymore if maxbucket has reached its maximum possible
* value.
*
* Ideally we'd allow bucket numbers up to UINT_MAX-1 (no higher because
* the calculation maxbucket+1 mustn't overflow). Currently we restrict
* to half that because of overflow looping in _hash_log2() and
* insufficient space in hashm_spares[]. It's moot anyway because an
* index with 2^32 buckets would certainly overflow BlockNumber and hence
* _hash_alloc_buckets() would fail, but if we supported buckets smaller
* than a disk block then this would be an independent constraint.
*
* If you change this, see also the maximum initial number of buckets in
* _hash_metapinit().
*/
if (metap->hashm_maxbucket >= (uint32) 0x7FFFFFFE)
goto fail;
/*
* Determine which bucket is to be split, and attempt to lock the old
* bucket. If we can't get the lock, give up.
*
* The lock protects us against other backends, but not against our own
* backend. Must check for active scans separately.
*/
new_bucket = metap->hashm_maxbucket + 1;
old_bucket = (new_bucket & metap->hashm_lowmask);
start_oblkno = BUCKET_TO_BLKNO(metap, old_bucket);
if (_hash_has_active_scan(rel, old_bucket))
goto fail;
if (!_hash_try_getlock(rel, start_oblkno, HASH_EXCLUSIVE))
goto fail;
/*
* Likewise lock the new bucket (should never fail).
*
* Note: it is safe to compute the new bucket's blkno here, even though we
* may still need to update the BUCKET_TO_BLKNO mapping. This is because
* the current value of hashm_spares[hashm_ovflpoint] correctly shows
* where we are going to put a new splitpoint's worth of buckets.
*/
start_nblkno = BUCKET_TO_BLKNO(metap, new_bucket);
if (_hash_has_active_scan(rel, new_bucket))
elog(ERROR, "scan in progress on supposedly new bucket");
if (!_hash_try_getlock(rel, start_nblkno, HASH_EXCLUSIVE))
elog(ERROR, "could not get lock on supposedly new bucket");
/*
* If the split point is increasing (hashm_maxbucket's log base 2
* increases), we need to allocate a new batch of bucket pages.
*/
spare_ndx = _hash_log2(new_bucket + 1);
if (spare_ndx > metap->hashm_ovflpoint)
{
Assert(spare_ndx == metap->hashm_ovflpoint + 1);
/*
* The number of buckets in the new splitpoint is equal to the total
* number already in existence, i.e. new_bucket. Currently this maps
* one-to-one to blocks required, but someday we may need a more
* complicated calculation here.
*/
if (!_hash_alloc_buckets(rel, start_nblkno, new_bucket))
{
/* can't split due to BlockNumber overflow */
_hash_droplock(rel, start_oblkno, HASH_EXCLUSIVE);
_hash_droplock(rel, start_nblkno, HASH_EXCLUSIVE);
goto fail;
}
}
/*
* Okay to proceed with split. Update the metapage bucket mapping info.
*
* Since we are scribbling on the metapage data right in the shared
* buffer, any failure in this next little bit leaves us with a big
* problem: the metapage is effectively corrupt but could get written back
* to disk. We don't really expect any failure, but just to be sure,
* establish a critical section.
*/
START_CRIT_SECTION();
metap->hashm_maxbucket = new_bucket;
if (new_bucket > metap->hashm_highmask)
{
/* Starting a new doubling */
metap->hashm_lowmask = metap->hashm_highmask;
metap->hashm_highmask = new_bucket | metap->hashm_lowmask;
}
/*
* If the split point is increasing (hashm_maxbucket's log base 2
* increases), we need to adjust the hashm_spares[] array and
* hashm_ovflpoint so that future overflow pages will be created beyond
* this new batch of bucket pages.
*/
if (spare_ndx > metap->hashm_ovflpoint)
{
metap->hashm_spares[spare_ndx] = metap->hashm_spares[metap->hashm_ovflpoint];
metap->hashm_ovflpoint = spare_ndx;
}
/* Done mucking with metapage */
END_CRIT_SECTION();
/*
* Copy bucket mapping info now; this saves re-accessing the meta page
* inside _hash_splitbucket's inner loop. Note that once we drop the
* split lock, other splits could begin, so these values might be out of
* date before _hash_splitbucket finishes. That's okay, since all it
* needs is to tell which of these two buckets to map hashkeys into.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
/* Write out the metapage and drop lock, but keep pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/* Release split lock; okay for other splits to occur now */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
/* Relocate records to the new bucket */
_hash_splitbucket(rel, metabuf, old_bucket, new_bucket,
start_oblkno, start_nblkno,
maxbucket, highmask, lowmask);
/* Release bucket locks, allowing others to access them */
_hash_droplock(rel, start_oblkno, HASH_EXCLUSIVE);
_hash_droplock(rel, start_nblkno, HASH_EXCLUSIVE);
return;
/* Here if decide not to split or fail to acquire old bucket lock */
fail:
/* We didn't write the metapage, so just drop lock */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/* Release split lock */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
}
/*
* 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;
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;
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(&local_metapage, 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;
bucket_buf = buf;
hashbucketcleanup(rel, cur_bucket, bucket_buf, blkno, info->strategy,
local_metapage.hashm_maxbucket,
local_metapage.hashm_highmask,
local_metapage.hashm_lowmask, &tuples_removed,
&num_index_tuples, split_cleanup,
callback, callback_state);
_hash_dropbuf(rel, bucket_buf);
/* 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 */
return stats;
}
/* ------------------------------------------------
* hash_metapage_info()
*
* Get the meta-page information for a hash index
*
* Usage: SELECT * FROM hash_metapage_info(get_raw_page('con_hash_index', 0))
* ------------------------------------------------
*/
Datum
hash_metapage_info(PG_FUNCTION_ARGS)
{
bytea *raw_page = PG_GETARG_BYTEA_P(0);
Page page;
HashMetaPageData *metad;
TupleDesc tupleDesc;
HeapTuple tuple;
int i,
j;
Datum values[16];
bool nulls[16];
Datum spares[HASH_MAX_SPLITPOINTS];
Datum mapp[HASH_MAX_BITMAPS];
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use raw page functions"))));
page = verify_hash_page(raw_page, LH_META_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");
tupleDesc = BlessTupleDesc(tupleDesc);
metad = HashPageGetMeta(page);
MemSet(nulls, 0, sizeof(nulls));
j = 0;
values[j++] = Int64GetDatum((int64) metad->hashm_magic);
values[j++] = Int64GetDatum((int64) metad->hashm_version);
values[j++] = Float8GetDatum(metad->hashm_ntuples);
values[j++] = Int32GetDatum((int32) metad->hashm_ffactor);
values[j++] = Int32GetDatum((int32) metad->hashm_bsize);
values[j++] = Int32GetDatum((int32) metad->hashm_bmsize);
values[j++] = Int32GetDatum((int32) metad->hashm_bmshift);
values[j++] = Int64GetDatum((int64) metad->hashm_maxbucket);
values[j++] = Int64GetDatum((int64) metad->hashm_highmask);
values[j++] = Int64GetDatum((int64) metad->hashm_lowmask);
values[j++] = Int64GetDatum((int64) metad->hashm_ovflpoint);
values[j++] = Int64GetDatum((int64) metad->hashm_firstfree);
values[j++] = Int64GetDatum((int64) metad->hashm_nmaps);
values[j++] = ObjectIdGetDatum((Oid) metad->hashm_procid);
for (i = 0; i < HASH_MAX_SPLITPOINTS; i++)
spares[i] = Int64GetDatum((int64) metad->hashm_spares[i]);
values[j++] = PointerGetDatum(construct_array(spares,
HASH_MAX_SPLITPOINTS,
INT8OID,
8, FLOAT8PASSBYVAL, 'd'));
for (i = 0; i < HASH_MAX_BITMAPS; i++)
mapp[i] = Int64GetDatum((int64) metad->hashm_mapp[i]);
values[j++] = PointerGetDatum(construct_array(mapp,
HASH_MAX_BITMAPS,
INT8OID,
8, FLOAT8PASSBYVAL, 'd'));
tuple = heap_form_tuple(tupleDesc, values, nulls);
PG_RETURN_DATUM(HeapTupleGetDatum(tuple));
}
/*
* _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_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;
}
/*
* _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;
}
/* ------------------------------------------------------
* pgstathashindex()
*
* Usage: SELECT * FROM pgstathashindex('hashindex');
* ------------------------------------------------------
*/
Datum
pgstathashindex(PG_FUNCTION_ARGS)
{
Oid relid = PG_GETARG_OID(0);
BlockNumber nblocks;
BlockNumber blkno;
Relation rel;
HashIndexStat stats;
BufferAccessStrategy bstrategy;
HeapTuple tuple;
TupleDesc tupleDesc;
Datum values[8];
bool nulls[8];
Buffer metabuf;
HashMetaPage metap;
float8 free_percent;
uint64 total_space;
rel = index_open(relid, AccessShareLock);
/* index_open() checks that it's an index */
if (!IS_HASH(rel))
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("relation \"%s\" is not a HASH 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 indexes of other sessions")));
/* Get the information we need from the metapage. */
memset(&stats, 0, sizeof(stats));
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
stats.version = metap->hashm_version;
stats.space_per_page = metap->hashm_bsize;
_hash_relbuf(rel, metabuf);
/* Get the current relation length */
nblocks = RelationGetNumberOfBlocks(rel);
/* prepare access strategy for this index */
bstrategy = GetAccessStrategy(BAS_BULKREAD);
/* Start from blkno 1 as 0th block is metapage */
for (blkno = 1; blkno < nblocks; blkno++)
{
Buffer buf;
Page page;
CHECK_FOR_INTERRUPTS();
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
bstrategy);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = (Page) BufferGetPage(buf);
if (PageIsNew(page))
stats.unused_pages++;
else 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))));
else
{
HashPageOpaque opaque;
int pagetype;
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
pagetype = opaque->hasho_flag & LH_PAGE_TYPE;
if (pagetype == LH_BUCKET_PAGE)
{
stats.bucket_pages++;
GetHashPageStats(page, &stats);
}
else if (pagetype == LH_OVERFLOW_PAGE)
{
stats.overflow_pages++;
GetHashPageStats(page, &stats);
}
else if (pagetype == LH_BITMAP_PAGE)
stats.bitmap_pages++;
else if (pagetype == LH_UNUSED_PAGE)
stats.unused_pages++;
else
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("unexpected page type 0x%04X in HASH index \"%s\" block %u",
opaque->hasho_flag, RelationGetRelationName(rel),
BufferGetBlockNumber(buf))));
}
UnlockReleaseBuffer(buf);
}
/* Done accessing the index */
index_close(rel, AccessShareLock);
/* Count unused pages as free space. */
stats.free_space += stats.unused_pages * stats.space_per_page;
/*
* Total space available for tuples excludes the metapage and the bitmap
* pages.
*/
total_space = (nblocks - (stats.bitmap_pages + 1)) * stats.space_per_page;
if (total_space == 0)
free_percent = 0.0;
else
free_percent = 100.0 * stats.free_space / total_space;
/*
* 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");
tupleDesc = BlessTupleDesc(tupleDesc);
/*
* Build and return the tuple
*/
MemSet(nulls, 0, sizeof(nulls));
values[0] = Int32GetDatum(stats.version);
values[1] = Int64GetDatum((int64) stats.bucket_pages);
values[2] = Int64GetDatum((int64) stats.overflow_pages);
values[3] = Int64GetDatum((int64) stats.bitmap_pages);
values[4] = Int64GetDatum((int64) stats.unused_pages);
values[5] = Int64GetDatum(stats.live_items);
values[6] = Int64GetDatum(stats.dead_items);
values[7] = Float8GetDatum(free_percent);
tuple = heap_form_tuple(tupleDesc, values, nulls);
PG_RETURN_DATUM(HeapTupleGetDatum(tuple));
}
/*
* _hash_addovflpage
*
* Add an overflow page to the bucket whose last page is pointed to by 'buf'.
*
* On entry, the caller must hold a pin but no lock on 'buf'. The pin is
* dropped before exiting (we assume the caller is not interested in 'buf'
* anymore). The returned overflow page will be pinned and write-locked;
* it is guaranteed to be empty.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* That buffer is returned in the same state.
*
* The caller must hold at least share lock on the bucket, to ensure that
* no one else tries to compact the bucket meanwhile. This guarantees that
* 'buf' won't stop being part of the bucket while it's unlocked.
*
* NB: since this could be executed concurrently by multiple processes,
* one should not assume that the returned overflow page will be the
* immediate successor of the originally passed 'buf'. Additional overflow
* pages might have been added to the bucket chain in between.
*/
Buffer
_hash_addovflpage(Relation rel, Buffer metabuf, Buffer buf)
{
Buffer ovflbuf;
Page page;
Page ovflpage;
HashPageOpaque pageopaque;
HashPageOpaque ovflopaque;
/*CS3223*/ /*declare variables*/
HashMetaPage metap;
Bucket bucket;
int i;
int index;
int bitIndexInElement;
uint32 ovflElement;
uint32 *tempPointer;
/* allocate and lock an empty overflow page */
ovflbuf = _hash_getovflpage(rel, metabuf);
/*CS3223*/
metap = HashPageGetMeta(BufferGetPage(metabuf));
/* find bucket number of primary page */
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
bucket = pageopaque -> hasho_bucket;
/*
* Write-lock the tail page. It is okay to hold two buffer locks here
* since there cannot be anyone else contending for access to ovflbuf.
*/
_hash_chgbufaccess(rel, buf, HASH_NOLOCK, HASH_WRITE);
/* probably redundant... */
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
/* loop to find current tail page, in case someone else inserted too */
//for (;;)
/*CS3223*/
while (i>=0)
{
BlockNumber nextblkno;
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
nextblkno = pageopaque->hasho_nextblkno;
if (!BlockNumberIsValid(nextblkno))
break;
/* we assume we do not need to write the unmodified page */
_hash_relbuf(rel, buf);
buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
/*CS3223*/
i++;
}
/* now that we have correct backlink, initialize new overflow page */
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = BufferGetBlockNumber(buf);
ovflopaque->hasho_nextblkno = InvalidBlockNumber;
ovflopaque->hasho_bucket = pageopaque->hasho_bucket;
ovflopaque->hasho_flag = LH_OVERFLOW_PAGE;
ovflopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(ovflbuf);
/* logically chain overflow page to previous page */
pageopaque->hasho_nextblkno = BufferGetBlockNumber(ovflbuf);
_hash_wrtbuf(rel, buf);
/*CS3223*/
/* if length of the bucket chain is 1, only one ovflpage was added, which means the bucket was not split before */
if (i == 1) {
index = bucket / 32;
bitIndexInElement = bucket % 32;
ovflElement = (uint32)metap->ovflBkts[index];
ovflElement = ovflElement | (1 << bitIndexInElement);
tempPointer = &(metap->ovflBkts[index]);
*tempPointer = ovflElement;
}
return ovflbuf;
}
/*
* _hash_getovflpage()
*
* Find an available overflow page and return it. The returned buffer
* is pinned and write-locked, and has had _hash_pageinit() applied,
* but it is caller's responsibility to fill the special space.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* That buffer is left in the same state at exit.
*/
static Buffer
_hash_getovflpage(Relation rel, Buffer metabuf)
{
HashMetaPage metap;
Buffer mapbuf = 0;
Buffer newbuf;
BlockNumber blkno;
uint32 orig_firstfree;
uint32 splitnum;
uint32 *freep = NULL;
uint32 max_ovflpg;
uint32 bit;
uint32 first_page;
uint32 last_bit;
uint32 last_page;
uint32 i,
j;
/* Get exclusive lock on the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
_hash_checkpage(rel, metabuf, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/* start search at hashm_firstfree */
orig_firstfree = metap->hashm_firstfree;
first_page = orig_firstfree >> BMPG_SHIFT(metap);
bit = orig_firstfree & BMPG_MASK(metap);
i = first_page;
j = bit / BITS_PER_MAP;
bit &= ~(BITS_PER_MAP - 1);
/* outer loop iterates once per bitmap page */
for (;;)
{
BlockNumber mapblkno;
Page mappage;
uint32 last_inpage;
/* want to end search with the last existing overflow page */
splitnum = metap->hashm_ovflpoint;
max_ovflpg = metap->hashm_spares[splitnum] - 1;
last_page = max_ovflpg >> BMPG_SHIFT(metap);
last_bit = max_ovflpg & BMPG_MASK(metap);
if (i > last_page)
break;
Assert(i < metap->hashm_nmaps);
mapblkno = metap->hashm_mapp[i];
if (i == last_page)
last_inpage = last_bit;
else
last_inpage = BMPGSZ_BIT(metap) - 1;
/* Release exclusive lock on metapage while reading bitmap page */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
mapbuf = _hash_getbuf(rel, mapblkno, HASH_WRITE, LH_BITMAP_PAGE);
mappage = BufferGetPage(mapbuf);
freep = HashPageGetBitmap(mappage);
for (; bit <= last_inpage; j++, bit += BITS_PER_MAP)
{
if (freep[j] != ALL_SET)
goto found;
}
/* No free space here, try to advance to next map page */
_hash_relbuf(rel, mapbuf);
i++;
j = 0; /* scan from start of next map page */
bit = 0;
/* Reacquire exclusive lock on the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
}
/*
* No free pages --- have to extend the relation to add an overflow page.
* First, check to see if we have to add a new bitmap page too.
*/
if (last_bit == (uint32) (BMPGSZ_BIT(metap) - 1))
{
/*
* We create the new bitmap page with all pages marked "in use".
* Actually two pages in the new bitmap's range will exist
* immediately: the bitmap page itself, and the following page which
* is the one we return to the caller. Both of these are correctly
* marked "in use". Subsequent pages do not exist yet, but it is
* convenient to pre-mark them as "in use" too.
*/
bit = metap->hashm_spares[splitnum];
_hash_initbitmap(rel, metap, bitno_to_blkno(metap, bit), MAIN_FORKNUM);
metap->hashm_spares[splitnum]++;
}
else
{
/*
* Nothing to do here; since the page will be past the last used page,
* we know its bitmap bit was preinitialized to "in use".
*/
}
/* Calculate address of the new overflow page */
bit = metap->hashm_spares[splitnum];
blkno = bitno_to_blkno(metap, bit);
/*
* Fetch the page with _hash_getnewbuf to ensure smgr's idea of the
* relation length stays in sync with ours. XXX It's annoying to do this
* with metapage write lock held; would be better to use a lock that
* doesn't block incoming searches.
*/
newbuf = _hash_getnewbuf(rel, blkno, MAIN_FORKNUM);
metap->hashm_spares[splitnum]++;
/*
* Adjust hashm_firstfree to avoid redundant searches. But don't risk
* changing it if someone moved it while we were searching bitmap pages.
*/
if (metap->hashm_firstfree == orig_firstfree)
metap->hashm_firstfree = bit + 1;
/* Write updated metapage and release lock, but not pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
return newbuf;
found:
/* convert bit to bit number within page */
bit += _hash_firstfreebit(freep[j]);
/* mark page "in use" in the bitmap */
SETBIT(freep, bit);
_hash_wrtbuf(rel, mapbuf);
/* Reacquire exclusive lock on the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
/* convert bit to absolute bit number */
bit += (i << BMPG_SHIFT(metap));
/* Calculate address of the recycled overflow page */
blkno = bitno_to_blkno(metap, bit);
/*
* Adjust hashm_firstfree to avoid redundant searches. But don't risk
* changing it if someone moved it while we were searching bitmap pages.
*/
if (metap->hashm_firstfree == orig_firstfree)
{
metap->hashm_firstfree = bit + 1;
/* Write updated metapage and release lock, but not pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
}
else
{
/* We didn't change the metapage, so no need to write */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
}
/* Fetch, init, and return the recycled page */
return _hash_getinitbuf(rel, blkno);
}
/*
* 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);
}
/*
* 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.
*/
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;
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 */
return stats;
}
/*
* 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);
}