/*
* Mask a btree page before performing consistency checks on it.
*/
void
btree_mask(char *pagedata, BlockNumber blkno)
{
Page page = (Page) pagedata;
BTPageOpaque maskopaq;
mask_page_lsn_and_checksum(page);
mask_page_hint_bits(page);
mask_unused_space(page);
maskopaq = (BTPageOpaque) PageGetSpecialPointer(page);
if (P_ISDELETED(maskopaq))
{
/*
* Mask page content on a DELETED page since it will be re-initialized
* during replay. See btree_xlog_unlink_page() for details.
*/
mask_page_content(page);
}
else if (P_ISLEAF(maskopaq))
{
/*
* In btree leaf pages, it is possible to modify the LP_FLAGS without
* emitting any WAL record. Hence, mask the line pointer flags. See
* _bt_killitems(), _bt_check_unique() for details.
*/
mask_lp_flags(page);
}
/*
* BTP_HAS_GARBAGE is just an un-logged hint bit. So, mask it. See
* _bt_killitems(), _bt_check_unique() for details.
*/
maskopaq->btpo_flags &= ~BTP_HAS_GARBAGE;
/*
* During replay of a btree page split, we don't set the BTP_SPLIT_END
* flag of the right sibling and initialize the cycle_id to 0 for the same
* page. See btree_xlog_split() for details.
*/
maskopaq->btpo_flags &= ~BTP_SPLIT_END;
maskopaq->btpo_cycleid = 0;
}
/*
* Add an item to a page being built.
*
* The main difference between this routine and a bare PageAddItem call
* is that this code knows that the leftmost data item on a non-leaf
* btree page doesn't need to have a key. Therefore, it strips such
* items down to just the item header.
*
* This is almost like nbtinsert.c's _bt_pgaddtup(), but we can't use
* that because it assumes that P_RIGHTMOST() will return the correct
* answer for the page. Here, we don't know yet if the page will be
* rightmost. Offset P_FIRSTKEY is always the first data key.
*/
static void
_bt_sortaddtup(Page page,
Size itemsize,
IndexTuple itup,
OffsetNumber itup_off)
{
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
IndexTupleData trunctuple;
if (!P_ISLEAF(opaque) && itup_off == P_FIRSTKEY)
{
trunctuple = *itup;
trunctuple.t_info = sizeof(IndexTupleData);
itup = &trunctuple;
itemsize = sizeof(IndexTupleData);
}
if (PageAddItem(page, (Item) itup, itemsize, itup_off,
false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to the index page");
}
/*
* pgstat_btree_page -- check tuples in a btree page
*/
static void
pgstat_btree_page(pgstattuple_type *stat, Relation rel, BlockNumber blkno,
BufferAccessStrategy bstrategy)
{
Buffer buf;
Page page;
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL, bstrategy);
LockBuffer(buf, BT_READ);
page = BufferGetPage(buf);
/* Page is valid, see what to do with it */
if (PageIsNew(page))
{
/* fully empty page */
stat->free_space += BLCKSZ;
}
else
{
BTPageOpaque opaque;
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (opaque->btpo_flags & (BTP_DELETED | BTP_HALF_DEAD))
{
/* recyclable page */
stat->free_space += BLCKSZ;
}
else if (P_ISLEAF(opaque))
{
pgstat_index_page(stat, page, P_FIRSTDATAKEY(opaque),
PageGetMaxOffsetNumber(page));
}
else
{
/* root or node */
}
}
_bt_relbuf(rel, buf);
}
/*
* _bt_search() -- Search the tree for a particular scankey,
* or more precisely for the first leaf page it could be on.
*
* Return value is a stack of parent-page pointers. *bufP is set to the
* address of the leaf-page buffer, which is read-locked and pinned.
* No locks are held on the parent pages, however!
*
* NOTE that the returned buffer is read-locked regardless of the access
* parameter. However, access = BT_WRITE will allow an empty root page
* to be created and returned. When access = BT_READ, an empty index
* will result in *bufP being set to InvalidBuffer.
*/
BTStack
_bt_search(Relation rel, int keysz, ScanKey scankey,
Buffer *bufP, int access)
{
BTStack stack_in = NULL;
/* Get the root page to start with */
*bufP = _bt_getroot(rel, access);
/* If index is empty and access = BT_READ, no root page is created. */
if (!BufferIsValid(*bufP))
return (BTStack) NULL;
/* Loop iterates once per level descended in the tree */
for (;;)
{
Page page;
BTPageOpaque opaque;
OffsetNumber offnum;
ItemId itemid;
BTItem btitem;
IndexTuple itup;
BlockNumber blkno;
BlockNumber par_blkno;
BTStack new_stack;
/*
* Race -- the page we just grabbed may have split since we read
* its pointer in the parent (or metapage). If it has, we may
* need to move right to its new sibling. Do that.
*/
*bufP = _bt_moveright(rel, *bufP, keysz, scankey, BT_READ);
/* if this is a leaf page, we're done */
page = BufferGetPage(*bufP);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (P_ISLEAF(opaque))
break;
/*
* Find the appropriate item on the internal page, and get the
* child page that it points to.
*/
offnum = _bt_binsrch(rel, *bufP, keysz, scankey);
itemid = PageGetItemId(page, offnum);
btitem = (BTItem) PageGetItem(page, itemid);
itup = &(btitem->bti_itup);
blkno = ItemPointerGetBlockNumber(&(itup->t_tid));
par_blkno = BufferGetBlockNumber(*bufP);
/*
* We need to save the location of the index entry we chose in the
* parent page on a stack. In case we split the tree, we'll use
* the stack to work back up to the parent page. We also save the
* actual downlink (TID) to uniquely identify the index entry, in
* case it moves right while we're working lower in the tree. See
* the paper by Lehman and Yao for how this is detected and
* handled. (We use the child link to disambiguate duplicate keys
* in the index -- Lehman and Yao disallow duplicate keys.)
*/
new_stack = (BTStack) palloc(sizeof(BTStackData));
new_stack->bts_blkno = par_blkno;
new_stack->bts_offset = offnum;
memcpy(&new_stack->bts_btitem, btitem, sizeof(BTItemData));
new_stack->bts_parent = stack_in;
/* drop the read lock on the parent page, acquire one on the child */
_bt_relbuf(rel, *bufP);
*bufP = _bt_getbuf(rel, blkno, BT_READ);
/* okay, all set to move down a level */
stack_in = new_stack;
}
return stack_in;
}
/*----------
* _bt_compare() -- Compare scankey to a particular tuple on the page.
*
* keysz: number of key conditions to be checked (might be less than the
* total length of the scan key!)
* page/offnum: location of btree item to be compared to.
*
* This routine returns:
* <0 if scankey < tuple at offnum;
* 0 if scankey == tuple at offnum;
* >0 if scankey > tuple at offnum.
* NULLs in the keys are treated as sortable values. Therefore
* "equality" does not necessarily mean that the item should be
* returned to the caller as a matching key!
*
* CRUCIAL NOTE: on a non-leaf page, the first data key is assumed to be
* "minus infinity": this routine will always claim it is less than the
* scankey. The actual key value stored (if any, which there probably isn't)
* does not matter. This convention allows us to implement the Lehman and
* Yao convention that the first down-link pointer is before the first key.
* See backend/access/nbtree/README for details.
*----------
*/
int32
_bt_compare(Relation rel,
int keysz,
ScanKey scankey,
Page page,
OffsetNumber offnum)
{
TupleDesc itupdesc = RelationGetDescr(rel);
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
BTItem btitem;
IndexTuple itup;
int i;
/*
* Force result ">" if target item is first data item on an internal
* page --- see NOTE above.
*/
if (!P_ISLEAF(opaque) && offnum == P_FIRSTDATAKEY(opaque))
return 1;
btitem = (BTItem) PageGetItem(page, PageGetItemId(page, offnum));
itup = &(btitem->bti_itup);
/*
* The scan key is set up with the attribute number associated with
* each term in the key. It is important that, if the index is
* multi-key, the scan contain the first k key attributes, and that
* they be in order. If you think about how multi-key ordering works,
* you'll understand why this is.
*
* We don't test for violation of this condition here, however. The
* initial setup for the index scan had better have gotten it right
* (see _bt_first).
*/
for (i = 0; i < keysz; i++)
{
ScanKey entry = &scankey[i];
Datum datum;
bool isNull;
int32 result;
datum = index_getattr(itup, entry->sk_attno, itupdesc, &isNull);
/* see comments about NULLs handling in btbuild */
if (entry->sk_flags & SK_ISNULL) /* key is NULL */
{
if (isNull)
result = 0; /* NULL "=" NULL */
else
result = 1; /* NULL ">" NOT_NULL */
}
else if (isNull) /* key is NOT_NULL and item is NULL */
{
result = -1; /* NOT_NULL "<" NULL */
}
else
{
result = DatumGetInt32(FunctionCall2(&entry->sk_func,
entry->sk_argument,
datum));
}
/* if the keys are unequal, return the difference */
if (result != 0)
return result;
}
/* if we get here, the keys are equal */
return 0;
}
uint32 check_index_page(Relation rel, PageHeader header, char *buffer, int block) {
uint32 nerrs = 0;
BTPageOpaque opaque = NULL;
/* check basic page header */
nerrs += check_page_header(header, block);
/* (block==0) means it's a meta-page, otherwise it's a regular index-page */
if (block == BTREE_METAPAGE) {
BTMetaPageData * mpdata = BTPageGetMeta(buffer);
ereport(DEBUG2, (errmsg("[%d] is a meta-page [magic=%d, version=%d]", block, mpdata->btm_magic, mpdata->btm_version)));
if (mpdata->btm_magic != BTREE_MAGIC) {
ereport(WARNING,(errmsg("[%d] metapage contains invalid magic number %d (should be %d)", block, mpdata->btm_magic, BTREE_MAGIC)));
nerrs++;
}
if (mpdata->btm_version != BTREE_VERSION) {
ereport(WARNING,(errmsg("[%d] metapage contains invalid version %d (should be %d)", block, mpdata->btm_version, BTREE_VERSION)));
nerrs++;
}
/* FIXME Check that the btm_root/btm_fastroot is between 1 and number of index blocks */
/* FIXME Check that the btm_level/btm_fastlevel is equal to the level fo the root block */
} else {
opaque = (BTPageOpaque)(buffer + header->pd_special);
/* check there's enough space for index-relevant data */
if (header->pd_special > BLCKSZ - sizeof(BTPageOpaque)) {
ereport(WARNING,
(errmsg("[%d] there's not enough special space for index data (%d > %d)",
block,
(int) sizeof(BTPageOpaque),
BLCKSZ - header->pd_special)));
nerrs++;
}
/*
* if the page is a leaf page, then level needs to be 0. Otherwise,
* it should be > 0. Deleted pages don't have a level, the level
* field is interleaved with an xid.
*/
if (!P_ISDELETED(opaque))
{
if (P_ISLEAF(opaque))
{
if (opaque-> btpo.level != 0) {
ereport(WARNING,
(errmsg("[%d] is leaf page, but level %d is not zero",
block, opaque->btpo.level)));
nerrs++;
}
}
else
{
if (opaque-> btpo.level == 0) {
ereport(WARNING,
(errmsg("[%d] is a non-leaf page, but level is zero",
block)));
nerrs++;
}
}
}
}
return nerrs;
}
/*
* For a newly inserted heap tid, check if an entry with this tid
* already exists in a unique index. If it does, abort the inserting
* transaction.
*/
static void
_bt_validate_tid(Relation irel, ItemPointer h_tid)
{
MIRROREDLOCK_BUFMGR_DECLARE;
BlockNumber blkno;
BlockNumber num_pages;
Buffer buf;
Page page;
BTPageOpaque opaque;
IndexTuple itup;
OffsetNumber maxoff,
minoff,
offnum;
elog(DEBUG1, "validating tid (%d,%d) for index (%s)",
ItemPointerGetBlockNumber(h_tid), ItemPointerGetOffsetNumber(h_tid),
RelationGetRelationName(irel));
blkno = BTREE_METAPAGE + 1;
num_pages = RelationGetNumberOfBlocks(irel);
MIRROREDLOCK_BUFMGR_LOCK;
for (; blkno < num_pages; blkno++)
{
buf = ReadBuffer(irel, blkno);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (!PageIsNew(page))
_bt_checkpage(irel, buf);
if (P_ISLEAF(opaque))
{
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offnum));
if (ItemPointerEquals(&itup->t_tid, h_tid))
{
Form_pg_attribute key_att = RelationGetDescr(irel)->attrs[0];
Oid key = InvalidOid;
bool isnull;
if (key_att->atttypid == OIDOID)
{
key = DatumGetInt32(
index_getattr(itup, 1, RelationGetDescr(irel), &isnull));
elog(ERROR, "found tid (%d,%d), %s (%d) already in index (%s)",
ItemPointerGetBlockNumber(h_tid), ItemPointerGetOffsetNumber(h_tid),
NameStr(key_att->attname), key, RelationGetRelationName(irel));
}
else
{
elog(ERROR, "found tid (%d,%d) already in index (%s)",
ItemPointerGetBlockNumber(h_tid), ItemPointerGetOffsetNumber(h_tid),
RelationGetRelationName(irel));
}
}
}
}
ReleaseBuffer(buf);
}
MIRROREDLOCK_BUFMGR_UNLOCK;
}
/*
* btvacuumpage --- VACUUM one page
*
* This processes a single page for btvacuumscan(). In some cases we
* must go back and re-examine previously-scanned pages; this routine
* recurses when necessary to handle that case.
*
* blkno is the page to process. orig_blkno is the highest block number
* reached by the outer btvacuumscan loop (the same as blkno, unless we
* are recursing to re-examine a previous page).
*/
static void
btvacuumpage(BTVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)
{
MIRROREDLOCK_BUFMGR_DECLARE;
IndexVacuumInfo *info = vstate->info;
IndexBulkDeleteResult *stats = vstate->stats;
IndexBulkDeleteCallback callback = vstate->callback;
void *callback_state = vstate->callback_state;
Relation rel = info->index;
bool delete_now;
BlockNumber recurse_to;
Buffer buf;
Page page;
BTPageOpaque opaque;
restart:
delete_now = false;
recurse_to = P_NONE;
/* call vacuum_delay_point while not holding any buffer lock */
vacuum_delay_point();
/*
* We can't use _bt_getbuf() here because it always applies
* _bt_checkpage(), which will barf on an all-zero page. We want to
* recycle all-zero pages, not fail. Also, we want to use a nondefault
* buffer access strategy.
*/
// -------- MirroredLock ----------
MIRROREDLOCK_BUFMGR_LOCK;
buf = ReadBufferWithStrategy(rel, blkno, info->strategy);
LockBuffer(buf, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (!PageIsNew(page))
_bt_checkpage(rel, buf);
/*
* If we are recursing, the only case we want to do anything with is a
* live leaf page having the current vacuum cycle ID. Any other state
* implies we already saw the page (eg, deleted it as being empty). In
* particular, we don't want to risk adding it to freePages twice.
*/
if (blkno != orig_blkno)
{
if (_bt_page_recyclable(page) ||
P_IGNORE(opaque) ||
!P_ISLEAF(opaque) ||
opaque->btpo_cycleid != vstate->cycleid)
{
_bt_relbuf(rel, buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
// -------- MirroredLock ----------
return;
}
}
/* Page is valid, see what to do with it */
if (_bt_page_recyclable(page))
{
/* Okay to recycle this page */
if (vstate->nFreePages < vstate->maxFreePages)
vstate->freePages[vstate->nFreePages++] = blkno;
vstate->totFreePages++;
stats->pages_deleted++;
}
else if (P_ISDELETED(opaque))
{
/* Already deleted, but can't recycle yet */
stats->pages_deleted++;
}
else if (P_ISHALFDEAD(opaque))
{
/* Half-dead, try to delete */
delete_now = true;
}
else if (P_ISLEAF(opaque))
{
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable;
OffsetNumber offnum,
minoff,
maxoff;
/*
* Trade in the initial read lock for a super-exclusive write lock on
* this page. We must get such a lock on every leaf page over the
* course of the vacuum scan, whether or not it actually contains any
* deletable tuples --- see nbtree/README.
*/
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBufferForCleanup(buf);
/*
* Check whether we need to recurse back to earlier pages. What we
* are concerned about is a page split that happened since we started
* the vacuum scan. If the split moved some tuples to a lower page
* then we might have missed 'em. If so, set up for tail recursion.
* (Must do this before possibly clearing btpo_cycleid below!)
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid &&
!(opaque->btpo_flags & BTP_SPLIT_END) &&
!P_RIGHTMOST(opaque) &&
opaque->btpo_next < orig_blkno)
recurse_to = opaque->btpo_next;
/*
* Scan over all items to see which ones need deleted according to the
* callback function.
*/
ndeletable = 0;
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(page);
if (callback)
{
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
IndexTuple itup;
ItemPointer htup;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offnum));
htup = &(itup->t_tid);
if (callback(htup, callback_state))
deletable[ndeletable++] = offnum;
}
}
/*
* Apply any needed deletes. We issue just one _bt_delitems() call
* per page, so as to minimize WAL traffic.
*/
if (ndeletable > 0)
{
_bt_delitems(rel, buf, deletable, ndeletable, true);
stats->tuples_removed += ndeletable;
/* must recompute maxoff */
maxoff = PageGetMaxOffsetNumber(page);
}
else
{
/*
* If the page has been split during this vacuum cycle, it seems
* worth expending a write to clear btpo_cycleid even if we don't
* have any deletions to do. (If we do, _bt_delitems takes care
* of this.) This ensures we won't process the page again.
*
* We treat this like a hint-bit update because there's no need to
* WAL-log it.
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid)
{
opaque->btpo_cycleid = 0;
SetBufferCommitInfoNeedsSave(buf);
}
}
/*
* If it's now empty, try to delete; else count the live tuples. We
* don't delete when recursing, though, to avoid putting entries into
* freePages out-of-order (doesn't seem worth any extra code to handle
* the case).
*/
if (minoff > maxoff)
delete_now = (blkno == orig_blkno);
else
stats->num_index_tuples += maxoff - minoff + 1;
}
if (delete_now)
{
MemoryContext oldcontext;
int ndel;
/* Run pagedel in a temp context to avoid memory leakage */
MemoryContextReset(vstate->pagedelcontext);
oldcontext = MemoryContextSwitchTo(vstate->pagedelcontext);
ndel = _bt_pagedel(rel, buf, NULL, info->vacuum_full);
/* count only this page, else may double-count parent */
if (ndel)
stats->pages_deleted++;
/*
* During VACUUM FULL it's okay to recycle deleted pages immediately,
* since there can be no other transactions scanning the index. Note
* that we will only recycle the current page and not any parent pages
* that _bt_pagedel might have recursed to; this seems reasonable in
* the name of simplicity. (Trying to do otherwise would mean we'd
* have to sort the list of recyclable pages we're building.)
*/
if (ndel && info->vacuum_full)
{
if (vstate->nFreePages < vstate->maxFreePages)
vstate->freePages[vstate->nFreePages++] = blkno;
vstate->totFreePages++;
}
MemoryContextSwitchTo(oldcontext);
/* pagedel released buffer, so we shouldn't */
}
else
_bt_relbuf(rel, buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
// -------- MirroredLock ----------
/*
* This is really tail recursion, but if the compiler is too stupid to
* optimize it as such, we'd eat an uncomfortably large amount of stack
* space per recursion level (due to the deletable[] array). A failure is
* improbable since the number of levels isn't likely to be large ... but
* just in case, let's hand-optimize into a loop.
*/
if (recurse_to != P_NONE)
{
blkno = recurse_to;
goto restart;
}
}
/*----------
* _bt_compare() -- Compare scankey to a particular tuple on the page.
*
* The passed scankey must be an insertion-type scankey (see nbtree/README),
* but it can omit the rightmost column(s) of the index.
*
* keysz: number of key conditions to be checked (might be less than the
* number of index columns!)
* page/offnum: location of btree item to be compared to.
*
* This routine returns:
* <0 if scankey < tuple at offnum;
* 0 if scankey == tuple at offnum;
* >0 if scankey > tuple at offnum.
* NULLs in the keys are treated as sortable values. Therefore
* "equality" does not necessarily mean that the item should be
* returned to the caller as a matching key!
*
* CRUCIAL NOTE: on a non-leaf page, the first data key is assumed to be
* "minus infinity": this routine will always claim it is less than the
* scankey. The actual key value stored (if any, which there probably isn't)
* does not matter. This convention allows us to implement the Lehman and
* Yao convention that the first down-link pointer is before the first key.
* See backend/access/nbtree/README for details.
*----------
*/
int32
_bt_compare(Relation rel,
int keysz,
ScanKey scankey,
Page page,
OffsetNumber offnum)
{
TupleDesc itupdesc = RelationGetDescr(rel);
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
IndexTuple itup;
int i;
/*
* Force result ">" if target item is first data item on an internal page
* --- see NOTE above.
*/
if (!P_ISLEAF(opaque) && offnum == P_FIRSTDATAKEY(opaque))
return 1;
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
/*
* The scan key is set up with the attribute number associated with each
* term in the key. It is important that, if the index is multi-key, the
* scan contain the first k key attributes, and that they be in order. If
* you think about how multi-key ordering works, you'll understand why
* this is.
*
* We don't test for violation of this condition here, however. The
* initial setup for the index scan had better have gotten it right (see
* _bt_first).
*/
for (i = 1; i <= keysz; i++)
{
Datum datum;
bool isNull;
int32 result;
datum = index_getattr(itup, scankey->sk_attno, itupdesc, &isNull);
/* see comments about NULLs handling in btbuild */
if (scankey->sk_flags & SK_ISNULL) /* key is NULL */
{
if (isNull)
result = 0; /* NULL "=" NULL */
else if (scankey->sk_flags & SK_BT_NULLS_FIRST)
result = -1; /* NULL "<" NOT_NULL */
else
result = 1; /* NULL ">" NOT_NULL */
}
else if (isNull) /* key is NOT_NULL and item is NULL */
{
if (scankey->sk_flags & SK_BT_NULLS_FIRST)
result = 1; /* NOT_NULL ">" NULL */
else
result = -1; /* NOT_NULL "<" NULL */
}
else
{
/*
* The sk_func needs to be passed the index value as left arg and
* the sk_argument as right arg (they might be of different
* types). Since it is convenient for callers to think of
* _bt_compare as comparing the scankey to the index item, we have
* to flip the sign of the comparison result. (Unless it's a DESC
* column, in which case we *don't* flip the sign.)
*/
result = DatumGetInt32(FunctionCall2Coll(&scankey->sk_func,
scankey->sk_collation,
datum,
scankey->sk_argument));
if (!(scankey->sk_flags & SK_BT_DESC))
result = -result;
}
/* if the keys are unequal, return the difference */
if (result != 0)
return result;
scankey++;
}
/* if we get here, the keys are equal */
return 0;
}
/* -------------------------------------------------
* GetBTPageStatistics()
*
* Collect statistics of single b-tree page
* -------------------------------------------------
*/
static void
GetBTPageStatistics(block_t blkno, buf_id_t buffer, BTPageStat *stat)
{
page_p page = BUF_PAGE(buffer);
struct page_hdr * phdr = (struct page_hdr *) page;
item_id_t maxoff = PAGE_MAX_ITEM_ID(page);
struct bt_page_opaque * opaque = (struct bt_page_opaque *) PAGE_SPECIAL_PTR(page);
int item_size = 0;
int off;
stat->blkno = blkno;
stat->max_avail = BLK_SZ - (BLK_SZ - phdr->pd_special + PAGE_HDR_SZ);
stat->dead_items = stat->live_items = 0;
stat->page_size = PAGE_SZ(page);
/* page type (flags) */
if (P_ISDELETED(opaque))
{
stat->type = 'd';
stat->btpo.xact = opaque->btpo.xact;
return;
}
else if (P_IGNORE(opaque))
stat->type = 'e';
else if (P_ISLEAF(opaque))
stat->type = 'l';
else if (P_ISROOT(opaque))
stat->type = 'r';
else
stat->type = 'i';
/* btpage opaque data */
stat->btpo_prev = opaque->btpo_prev;
stat->btpo_next = opaque->btpo_next;
stat->btpo.level = opaque->btpo.level;
stat->btpo_flags = opaque->btpo_flags;
stat->btpo_cycleid = opaque->btpo_cycleid;
/* count live and dead tuples, and free space */
for (off = FIRST_ITEM_ID; off <= maxoff; off++)
{
struct index_tuple * itup;
struct item_id * id = PAGE_ITEM_ID(page, off);
itup = (struct index_tuple *) PAGE_GET_ITEM(page, id);
item_size += INDEX_TUPLE_SZ(itup);
if (!ITEMID_DEAD(id))
stat->live_items++;
else
stat->dead_items++;
}
stat->free_size = page_free_space(page);
if ((stat->live_items + stat->dead_items) > 0)
stat->avg_item_size = item_size / (stat->live_items + stat->dead_items);
else
stat->avg_item_size = 0;
}
/* -------------------------------------------------
* GetBTPageStatistics()
*
* Collect statistics of single b-tree page
* -------------------------------------------------
*/
static void
GetBTPageStatistics(BlockNumber blkno, Buffer buffer, BTPageStat *stat)
{
Page page = BufferGetPage(buffer);
PageHeader phdr = (PageHeader) page;
OffsetNumber maxoff = PageGetMaxOffsetNumber(page);
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
int item_size = 0;
int off;
stat->blkno = blkno;
stat->max_avail = BLCKSZ - (BLCKSZ - phdr->pd_special + SizeOfPageHeaderData);
stat->dead_items = stat->live_items = 0;
stat->page_size = PageGetPageSize(page);
/* page type (flags) */
if (P_ISDELETED(opaque))
{
stat->type = 'd';
stat->btpo.xact = opaque->btpo.xact;
return;
}
else if (P_IGNORE(opaque))
stat->type = 'e';
else if (P_ISLEAF(opaque))
stat->type = 'l';
else if (P_ISROOT(opaque))
stat->type = 'r';
else
stat->type = 'i';
/* btpage opaque data */
stat->btpo_prev = opaque->btpo_prev;
stat->btpo_next = opaque->btpo_next;
stat->btpo.level = opaque->btpo.level;
stat->btpo_flags = opaque->btpo_flags;
stat->btpo_cycleid = opaque->btpo_cycleid;
/* count live and dead tuples, and free space */
for (off = FirstOffsetNumber; off <= maxoff; off++)
{
IndexTuple itup;
ItemId id = PageGetItemId(page, off);
itup = (IndexTuple) PageGetItem(page, id);
item_size += IndexTupleSize(itup);
if (!ItemIdIsDead(id))
stat->live_items++;
else
stat->dead_items++;
}
stat->free_size = PageGetFreeSpace(page);
if ((stat->live_items + stat->dead_items) > 0)
stat->avg_item_size = item_size / (stat->live_items + stat->dead_items);
else
stat->avg_item_size = 0;
}
Datum
readindex(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
readindexinfo *info;
Relation irel = NULL;
Relation hrel = NULL;
MIRROREDLOCK_BUFMGR_DECLARE;
if (SRF_IS_FIRSTCALL())
{
Oid irelid = PG_GETARG_OID(0);
TupleDesc tupdesc;
MemoryContext oldcontext;
AttrNumber outattnum;
TupleDesc itupdesc;
int i;
AttrNumber attno;
irel = index_open(irelid, AccessShareLock);
itupdesc = RelationGetDescr(irel);
outattnum = FIXED_COLUMN + itupdesc->natts;
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
tupdesc = CreateTemplateTupleDesc(outattnum, false);
attno = 1;
TupleDescInitEntry(tupdesc, attno++, "ictid", TIDOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "hctid", TIDOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "aotid", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "istatus", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "hstatus", TEXTOID, -1, 0);
for (i = 0; i < itupdesc->natts; i++)
{
Form_pg_attribute attr = itupdesc->attrs[i];
TupleDescInitEntry(tupdesc, attno++, NameStr(attr->attname), attr->atttypid, attr->atttypmod, 0);
}
funcctx->tuple_desc = BlessTupleDesc(tupdesc);
info = (readindexinfo *) palloc(sizeof(readindexinfo));
funcctx->user_fctx = (void *) info;
info->outattnum = outattnum;
info->ireloid = irelid;
hrel = relation_open(irel->rd_index->indrelid, AccessShareLock);
if (hrel->rd_rel != NULL &&
(hrel->rd_rel->relstorage == 'a' ||
hrel->rd_rel->relstorage == 'c'))
{
relation_close(hrel, AccessShareLock);
hrel = NULL;
info->hreloid = InvalidOid;
}
else
info->hreloid = irel->rd_index->indrelid;
info->num_pages = RelationGetNumberOfBlocks(irel);
info->blkno = BTREE_METAPAGE + 1;
info->page = NULL;
MemoryContextSwitchTo(oldcontext);
}
funcctx = SRF_PERCALL_SETUP();
info = (readindexinfo *) funcctx->user_fctx;
/*
* Open the relations (on first call, we did that above already).
* We unfortunately have to look up the relcache entry on every call,
* because if we store it in the cross-call context, we won't get a
* chance to release it if the function isn't run to completion,
* e.g. because of a LIMIT clause. We only lock the relation on the
* first call, and keep the lock until completion, however.
*/
if (!irel)
irel = index_open(info->ireloid, NoLock);
if (!hrel && info->hreloid != InvalidOid)
hrel = heap_open(info->hreloid, NoLock);
while (info->blkno < info->num_pages)
{
Datum values[255];
bool nulls[255];
ItemPointerData itid;
HeapTuple tuple;
Datum result;
if (info->page == NULL)
{
Buffer buf;
/*
* Make copy of the page, because we cannot hold a buffer pin
* across calls (we wouldn't have a chance to release it, if the
* function isn't run to completion.)
*/
info->page = palloc(BLCKSZ);
MIRROREDLOCK_BUFMGR_LOCK;
buf = ReadBuffer(irel, info->blkno);
memcpy(info->page, BufferGetPage(buf), BLCKSZ);
ReleaseBuffer(buf);
MIRROREDLOCK_BUFMGR_UNLOCK;
info->opaque = (BTPageOpaque) PageGetSpecialPointer(info->page);
info->minoff = P_FIRSTDATAKEY(info->opaque);
info->maxoff = PageGetMaxOffsetNumber(info->page);
info->offnum = info->minoff;
}
if (!P_ISLEAF(info->opaque) || info->offnum > info->maxoff)
{
pfree(info->page);
info->page = NULL;
info->blkno++;
continue;
}
MemSet(nulls, false, info->outattnum * sizeof(bool));
ItemPointerSet(&itid, info->blkno, info->offnum);
values[0] = ItemPointerGetDatum(&itid);
readindextuple(info, irel, hrel, values, nulls);
info->offnum = OffsetNumberNext(info->offnum);
tuple = heap_form_tuple(funcctx->tuple_desc, values, nulls);
result = HeapTupleGetDatum(tuple);
if (hrel != NULL)
heap_close(hrel, NoLock);
index_close(irel, NoLock);
SRF_RETURN_NEXT(funcctx, result);
}
if (hrel != NULL)
heap_close(hrel, AccessShareLock);
index_close(irel, AccessShareLock);
SRF_RETURN_DONE(funcctx);
}
/* ------------------------------------------------------
* pgstatindex()
*
* Usage: SELECT * FROM pgstatindex('t1_pkey');
* ------------------------------------------------------
*/
Datum
pgstatindex(PG_FUNCTION_ARGS)
{
text *relname = PG_GETARG_TEXT_P(0);
Relation rel;
RangeVar *relrv;
Datum result;
BlockNumber nblocks;
BlockNumber blkno;
BTIndexStat indexStat;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use pgstattuple functions"))));
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = relation_openrv(relrv, AccessShareLock);
if (!IS_INDEX(rel) || !IS_BTREE(rel))
elog(ERROR, "relation \"%s\" is not a btree index",
RelationGetRelationName(rel));
/*
* Reject attempts to read non-local temporary relations; we would be
* likely to get wrong data since we have no visibility into the owning
* session's local buffers.
*/
if (RELATION_IS_OTHER_TEMP(rel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
/*
* Read metapage
*/
{
Buffer buffer = ReadBuffer(rel, 0);
Page page = BufferGetPage(buffer);
BTMetaPageData *metad = BTPageGetMeta(page);
indexStat.version = metad->btm_version;
indexStat.level = metad->btm_level;
indexStat.root_blkno = metad->btm_root;
ReleaseBuffer(buffer);
}
/* -- init counters -- */
indexStat.root_pages = 0;
indexStat.internal_pages = 0;
indexStat.leaf_pages = 0;
indexStat.empty_pages = 0;
indexStat.deleted_pages = 0;
indexStat.max_avail = 0;
indexStat.free_space = 0;
indexStat.fragments = 0;
/*
* Scan all blocks except the metapage
*/
nblocks = RelationGetNumberOfBlocks(rel);
for (blkno = 1; blkno < nblocks; blkno++)
{
Buffer buffer;
Page page;
BTPageOpaque opaque;
/* Read and lock buffer */
buffer = ReadBuffer(rel, blkno);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
page = BufferGetPage(buffer);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
/* Determine page type, and update totals */
if (P_ISLEAF(opaque))
{
int max_avail;
max_avail = BLCKSZ - (BLCKSZ - ((PageHeader) page)->pd_special + SizeOfPageHeaderData);
indexStat.max_avail += max_avail;
indexStat.free_space += PageGetFreeSpace(page);
indexStat.leaf_pages++;
/*
* If the next leaf is on an earlier block, it means a
* fragmentation.
*/
if (opaque->btpo_next != P_NONE && opaque->btpo_next < blkno)
indexStat.fragments++;
}
else if (P_ISDELETED(opaque))
indexStat.deleted_pages++;
else if (P_IGNORE(opaque))
indexStat.empty_pages++;
else if (P_ISROOT(opaque))
indexStat.root_pages++;
else
indexStat.internal_pages++;
/* Unlock and release buffer */
LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
ReleaseBuffer(buffer);
}
relation_close(rel, AccessShareLock);
/*----------------------------
* Build a result tuple
*----------------------------
*/
{
TupleDesc tupleDesc;
int j;
char *values[10];
HeapTuple tuple;
/* 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");
j = 0;
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", indexStat.version);
values[j] = palloc(32);
snprintf(values[j++], 32, "%d", indexStat.level);
values[j] = palloc(32);
snprintf(values[j++], 32, INT64_FORMAT,
(indexStat.root_pages +
indexStat.leaf_pages +
indexStat.internal_pages +
indexStat.deleted_pages +
indexStat.empty_pages) * BLCKSZ);
values[j] = palloc(32);
snprintf(values[j++], 32, "%u", indexStat.root_blkno);
values[j] = palloc(32);
snprintf(values[j++], 32, INT64_FORMAT, indexStat.internal_pages);
values[j] = palloc(32);
snprintf(values[j++], 32, INT64_FORMAT, indexStat.leaf_pages);
values[j] = palloc(32);
snprintf(values[j++], 32, INT64_FORMAT, indexStat.empty_pages);
values[j] = palloc(32);
snprintf(values[j++], 32, INT64_FORMAT, indexStat.deleted_pages);
values[j] = palloc(32);
if (indexStat.max_avail > 0)
snprintf(values[j++], 32, "%.2f",
100.0 - (double) indexStat.free_space / (double) indexStat.max_avail * 100.0);
else
snprintf(values[j++], 32, "NaN");
values[j] = palloc(32);
if (indexStat.leaf_pages > 0)
snprintf(values[j++], 32, "%.2f",
(double) indexStat.fragments / (double) indexStat.leaf_pages * 100.0);
else
snprintf(values[j++], 32, "NaN");
tuple = BuildTupleFromCStrings(TupleDescGetAttInMetadata(tupleDesc),
values);
result = HeapTupleGetDatum(tuple);
}
PG_RETURN_DATUM(result);
}
Datum
readindex(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
readindexinfo *info;
MIRROREDLOCK_BUFMGR_DECLARE;
if (SRF_IS_FIRSTCALL())
{
Oid irelid = PG_GETARG_OID(0);
TupleDesc tupdesc;
MemoryContext oldcontext;
AttrNumber outattnum;
Relation irel;
TupleDesc itupdesc;
int i;
AttrNumber attno;
irel = index_open(irelid, AccessShareLock);
itupdesc = RelationGetDescr(irel);
outattnum = FIXED_COLUMN + itupdesc->natts;
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
tupdesc = CreateTemplateTupleDesc(outattnum, false);
attno = 1;
TupleDescInitEntry(tupdesc, attno++, "ictid", TIDOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "hctid", TIDOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "aotid", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "istatus", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, attno++, "hstatus", TEXTOID, -1, 0);
for (i = 0; i < itupdesc->natts; i++)
{
Form_pg_attribute attr = itupdesc->attrs[i];
TupleDescInitEntry(tupdesc, attno++, NameStr(attr->attname), attr->atttypid, attr->atttypmod, 0);
}
funcctx->tuple_desc = BlessTupleDesc(tupdesc);
info = (readindexinfo *) palloc(sizeof(readindexinfo));
funcctx->user_fctx = (void *) info;
info->outattnum = outattnum;
info->irel = irel;
info->hrel = relation_open(irel->rd_index->indrelid, AccessShareLock);
if (info->hrel->rd_rel != NULL &&
(info->hrel->rd_rel->relstorage == 'a' ||
info->hrel->rd_rel->relstorage == 'c'))
{
relation_close(info->hrel, AccessShareLock);
info->hrel = NULL;
}
info->num_pages = RelationGetNumberOfBlocks(irel);
info->blkno = BTREE_METAPAGE + 1;
info->page = NULL;
MemoryContextSwitchTo(oldcontext);
}
funcctx = SRF_PERCALL_SETUP();
info = (readindexinfo *) funcctx->user_fctx;
while (info->blkno < info->num_pages)
{
Datum values[255];
bool nulls[255];
ItemPointerData itid;
HeapTuple tuple;
Datum result;
if (info->page == NULL)
{
MIRROREDLOCK_BUFMGR_LOCK;
info->buf = ReadBuffer(info->irel, info->blkno);
info->page = BufferGetPage(info->buf);
info->opaque = (BTPageOpaque) PageGetSpecialPointer(info->page);
info->minoff = P_FIRSTDATAKEY(info->opaque);
info->maxoff = PageGetMaxOffsetNumber(info->page);
info->offnum = info->minoff;
MIRROREDLOCK_BUFMGR_UNLOCK;
}
if (!P_ISLEAF(info->opaque) || info->offnum > info->maxoff)
{
ReleaseBuffer(info->buf);
info->page = NULL;
info->blkno++;
continue;
}
MemSet(nulls, false, info->outattnum * sizeof(bool));
ItemPointerSet(&itid, info->blkno, info->offnum);
values[0] = ItemPointerGetDatum(&itid);
readindextuple(info, values, nulls);
info->offnum = OffsetNumberNext(info->offnum);
tuple = heap_form_tuple(funcctx->tuple_desc, values, nulls);
result = HeapTupleGetDatum(tuple);
SRF_RETURN_NEXT(funcctx, result);
}
if (info->hrel != NULL)
relation_close(info->hrel, AccessShareLock);
index_close(info->irel, AccessShareLock);
SRF_RETURN_DONE(funcctx);
}
/*
* _bt_endpoint() -- Find the first or last page in the index, and scan
* from there to the first key satisfying all the quals.
*
* This is used by _bt_first() to set up a scan when we've determined
* that the scan must start at the beginning or end of the index (for
* a forward or backward scan respectively). Exit conditions are the
* same as for _bt_first().
*/
static bool
_bt_endpoint(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
BTScanOpaque so = (BTScanOpaque) scan->opaque;
Buffer buf;
Page page;
BTPageOpaque opaque;
OffsetNumber start;
BTScanPosItem *currItem;
/*
* Scan down to the leftmost or rightmost leaf page. This is a simplified
* version of _bt_search(). We don't maintain a stack since we know we
* won't need it.
*/
buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir));
if (!BufferIsValid(buf))
{
/*
* Empty index. Lock the whole relation, as nothing finer to lock
* exists.
*/
PredicateLockRelation(rel, scan->xs_snapshot);
so->currPos.buf = InvalidBuffer;
return false;
}
PredicateLockPage(rel, BufferGetBlockNumber(buf), scan->xs_snapshot);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
Assert(P_ISLEAF(opaque));
if (ScanDirectionIsForward(dir))
{
/* There could be dead pages to the left, so not this: */
/* Assert(P_LEFTMOST(opaque)); */
start = P_FIRSTDATAKEY(opaque);
}
else if (ScanDirectionIsBackward(dir))
{
Assert(P_RIGHTMOST(opaque));
start = PageGetMaxOffsetNumber(page);
}
else
{
elog(ERROR, "invalid scan direction: %d", (int) dir);
start = 0; /* keep compiler quiet */
}
/* remember which buffer we have pinned */
so->currPos.buf = buf;
/* initialize moreLeft/moreRight appropriately for scan direction */
if (ScanDirectionIsForward(dir))
{
so->currPos.moreLeft = false;
so->currPos.moreRight = true;
}
else
{
so->currPos.moreLeft = true;
so->currPos.moreRight = false;
}
so->numKilled = 0; /* just paranoia */
so->markItemIndex = -1; /* ditto */
/*
* Now load data from the first page of the scan.
*/
if (!_bt_readpage(scan, dir, start))
{
/*
* There's no actually-matching data on this page. Try to advance to
* the next page. Return false if there's no matching data at all.
*/
if (!_bt_steppage(scan, dir))
return false;
}
/* Drop the lock, but not pin, on the current page */
LockBuffer(so->currPos.buf, BUFFER_LOCK_UNLOCK);
/* OK, itemIndex says what to return */
currItem = &so->currPos.items[so->currPos.itemIndex];
scan->xs_ctup.t_self = currItem->heapTid;
if (scan->xs_want_itup)
scan->xs_itup = (IndexTuple) (so->currTuples + currItem->tupleOffset);
return true;
}
/**
* @brief Read the left-most leaf page by walking down on index tree structure
* from root node.
*
* Process flow
* -# Open index file and read meta page
* -# Get block number of root page
* -# Read "fast root" page
* -# Read left child page until reaching left-most leaf page
*
* After calling this function, the members of BTReader are the following:
* - smgr : Smgr relation of the existing index file.
* - blkno : block number of left-most leaf page. If there is no leaf page,
* InvalidBlockNumber is set.
* - offnum : InvalidOffsetNumber is set.
* - page : Left-most leaf page, or undefined if no leaf page.
*
* @param reader [in/out] B-Tree index reader
* @return true iff there are some tuples
*/
static bool
BTReaderInit(BTReader *reader, Relation rel)
{
BTPageOpaque metaopaque;
BTMetaPageData *metad;
BTPageOpaque opaque;
BlockNumber blkno;
/*
* HACK: We cannot use smgropen because smgrs returned from it
* will be closed automatically when we assign a new file node.
*
* XXX: It might be better to open the previous relfilenode with
* smgropen *after* RelationSetNewRelfilenode.
*/
memset(&reader->smgr, 0, sizeof(reader->smgr));
#if PG_VERSION_NUM >= 90100
reader->smgr.smgr_rnode.node = rel->rd_node;
reader->smgr.smgr_rnode.backend =
rel->rd_backend == MyBackendId ? MyBackendId : InvalidBackendId;
#else
reader->smgr.smgr_rnode = rel->rd_node;
#endif
reader->smgr.smgr_which = 0; /* md.c */
reader->blkno = InvalidBlockNumber;
reader->offnum = InvalidOffsetNumber;
reader->page = palloc(BLCKSZ);
/*
* Read meta page and check sanity of it.
*
* XXX: It might be better to do REINDEX against corrupted indexes
* instead of raising errors because we've spent long time for data
* loading...
*/
BTReaderReadPage(reader, BTREE_METAPAGE);
metaopaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
metad = BTPageGetMeta(reader->page);
if (!(metaopaque->btpo_flags & BTP_META) ||
metad->btm_magic != BTREE_MAGIC)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" is not a reader",
RelationGetRelationName(rel))));
if (metad->btm_version != BTREE_VERSION)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("version mismatch in index \"%s\": file version %d,"
" code version %d",
RelationGetRelationName(rel),
metad->btm_version, BTREE_VERSION)));
if (metad->btm_root == P_NONE)
{
/* No root page; We ignore the index in the subsequent build. */
reader->blkno = InvalidBlockNumber;
return false;
}
/* Go to the fast root page. */
blkno = metad->btm_fastroot;
BTReaderReadPage(reader, blkno);
opaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
/* Walk down to the left-most leaf page */
while (!P_ISLEAF(opaque))
{
ItemId firstid;
IndexTuple itup;
/* Get the block number of the left child */
firstid = PageGetItemId(reader->page, P_FIRSTDATAKEY(opaque));
itup = (IndexTuple) PageGetItem(reader->page, firstid);
blkno = ItemPointerGetBlockNumber(&(itup->t_tid));
/* Go down to children */
for (;;)
{
BTReaderReadPage(reader, blkno);
opaque = (BTPageOpaque) PageGetSpecialPointer(reader->page);
if (!P_IGNORE(opaque))
break;
if (P_RIGHTMOST(opaque))
{
/* We reach end of the index without any valid leaves. */
reader->blkno = InvalidBlockNumber;
return false;
}
blkno = opaque->btpo_next;
}
}
return true;
}
/*
* _bt_binsrch() -- Do a binary search for a key on a particular page.
*
* The passed scankey must be an insertion-type scankey (see nbtree/README),
* but it can omit the rightmost column(s) of the index.
*
* When nextkey is false (the usual case), we are looking for the first
* item >= scankey. When nextkey is true, we are looking for the first
* item strictly greater than scankey.
*
* On a leaf page, _bt_binsrch() returns the OffsetNumber of the first
* key >= given scankey, or > scankey if nextkey is true. (NOTE: in
* particular, this means it is possible to return a value 1 greater than the
* number of keys on the page, if the scankey is > all keys on the page.)
*
* On an internal (non-leaf) page, _bt_binsrch() returns the OffsetNumber
* of the last key < given scankey, or last key <= given scankey if nextkey
* is true. (Since _bt_compare treats the first data key of such a page as
* minus infinity, there will be at least one key < scankey, so the result
* always points at one of the keys on the page.) This key indicates the
* right place to descend to be sure we find all leaf keys >= given scankey
* (or leaf keys > given scankey when nextkey is true).
*
* This procedure is not responsible for walking right, it just examines
* the given page. _bt_binsrch() has no lock or refcount side effects
* on the buffer.
*/
OffsetNumber
_bt_binsrch(Relation rel,
Buffer buf,
int keysz,
ScanKey scankey,
bool nextkey)
{
Page page;
BTPageOpaque opaque;
OffsetNumber low,
high;
int32 result,
cmpval;
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
low = P_FIRSTDATAKEY(opaque);
high = PageGetMaxOffsetNumber(page);
/*
* If there are no keys on the page, return the first available slot. Note
* this covers two cases: the page is really empty (no keys), or it
* contains only a high key. The latter case is possible after vacuuming.
* This can never happen on an internal page, however, since they are
* never empty (an internal page must have children).
*/
if (high < low)
return low;
/*
* Binary search to find the first key on the page >= scan key, or first
* key > scankey when nextkey is true.
*
* For nextkey=false (cmpval=1), the loop invariant is: all slots before
* 'low' are < scan key, all slots at or after 'high' are >= scan key.
*
* For nextkey=true (cmpval=0), the loop invariant is: all slots before
* 'low' are <= scan key, all slots at or after 'high' are > scan key.
*
* We can fall out when high == low.
*/
high++; /* establish the loop invariant for high */
cmpval = nextkey ? 0 : 1; /* select comparison value */
while (high > low)
{
OffsetNumber mid = low + ((high - low) / 2);
/* We have low <= mid < high, so mid points at a real slot */
result = _bt_compare(rel, keysz, scankey, page, mid);
if (result >= cmpval)
low = mid + 1;
else
high = mid;
}
/*
* At this point we have high == low, but be careful: they could point
* past the last slot on the page.
*
* On a leaf page, we always return the first key >= scan key (resp. >
* scan key), which could be the last slot + 1.
*/
if (P_ISLEAF(opaque))
return low;
/*
* On a non-leaf page, return the last key < scan key (resp. <= scan key).
* There must be one if _bt_compare() is playing by the rules.
*/
Assert(low > P_FIRSTDATAKEY(opaque));
return OffsetNumberPrev(low);
}
static Datum
pgstatindex_impl(Relation rel, FunctionCallInfo fcinfo)
{
Datum result;
BlockNumber nblocks;
BlockNumber blkno;
BTIndexStat indexStat;
BufferAccessStrategy bstrategy = GetAccessStrategy(BAS_BULKREAD);
if (!IS_INDEX(rel) || !IS_BTREE(rel))
elog(ERROR, "relation \"%s\" is not a btree index",
RelationGetRelationName(rel));
/*
* Reject attempts to read non-local temporary relations; we would be
* likely to get wrong data since we have no visibility into the owning
* session's local buffers.
*/
if (RELATION_IS_OTHER_TEMP(rel))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary tables of other sessions")));
/*
* Read metapage
*/
{
Buffer buffer = ReadBufferExtended(rel, MAIN_FORKNUM, 0, RBM_NORMAL, bstrategy);
Page page = BufferGetPage(buffer);
BTMetaPageData *metad = BTPageGetMeta(page);
indexStat.version = metad->btm_version;
indexStat.level = metad->btm_level;
indexStat.root_blkno = metad->btm_root;
ReleaseBuffer(buffer);
}
/* -- init counters -- */
indexStat.internal_pages = 0;
indexStat.leaf_pages = 0;
indexStat.empty_pages = 0;
indexStat.deleted_pages = 0;
indexStat.max_avail = 0;
indexStat.free_space = 0;
indexStat.fragments = 0;
/*
* Scan all blocks except the metapage
*/
nblocks = RelationGetNumberOfBlocks(rel);
for (blkno = 1; blkno < nblocks; blkno++)
{
Buffer buffer;
Page page;
BTPageOpaque opaque;
CHECK_FOR_INTERRUPTS();
/* Read and lock buffer */
buffer = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL, bstrategy);
LockBuffer(buffer, BUFFER_LOCK_SHARE);
page = BufferGetPage(buffer);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
/* Determine page type, and update totals */
if (P_ISDELETED(opaque))
indexStat.deleted_pages++;
else if (P_IGNORE(opaque))
indexStat.empty_pages++; /* this is the "half dead" state */
else if (P_ISLEAF(opaque))
{
int max_avail;
max_avail = BLCKSZ - (BLCKSZ - ((PageHeader) page)->pd_special + SizeOfPageHeaderData);
indexStat.max_avail += max_avail;
indexStat.free_space += PageGetFreeSpace(page);
indexStat.leaf_pages++;
/*
* If the next leaf is on an earlier block, it means a
* fragmentation.
*/
if (opaque->btpo_next != P_NONE && opaque->btpo_next < blkno)
indexStat.fragments++;
}
else
indexStat.internal_pages++;
/* Unlock and release buffer */
LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
ReleaseBuffer(buffer);
}
relation_close(rel, AccessShareLock);
/*----------------------------
* Build a result tuple
*----------------------------
*/
{
TupleDesc tupleDesc;
int j;
char *values[10];
HeapTuple tuple;
/* 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");
j = 0;
values[j++] = psprintf("%d", indexStat.version);
values[j++] = psprintf("%d", indexStat.level);
values[j++] = psprintf(INT64_FORMAT,
(1 + /* include the metapage in index_size */
indexStat.leaf_pages +
indexStat.internal_pages +
indexStat.deleted_pages +
indexStat.empty_pages) * BLCKSZ);
values[j++] = psprintf("%u", indexStat.root_blkno);
values[j++] = psprintf(INT64_FORMAT, indexStat.internal_pages);
values[j++] = psprintf(INT64_FORMAT, indexStat.leaf_pages);
values[j++] = psprintf(INT64_FORMAT, indexStat.empty_pages);
values[j++] = psprintf(INT64_FORMAT, indexStat.deleted_pages);
if (indexStat.max_avail > 0)
values[j++] = psprintf("%.2f",
100.0 - (double) indexStat.free_space / (double) indexStat.max_avail * 100.0);
else
values[j++] = pstrdup("NaN");
if (indexStat.leaf_pages > 0)
values[j++] = psprintf("%.2f",
(double) indexStat.fragments / (double) indexStat.leaf_pages * 100.0);
else
values[j++] = pstrdup("NaN");
tuple = BuildTupleFromCStrings(TupleDescGetAttInMetadata(tupleDesc),
values);
result = HeapTupleGetDatum(tuple);
}
return result;
}
/*
* _bt_search() -- Search the tree for a particular scankey,
* or more precisely for the first leaf page it could be on.
*
* The passed scankey must be an insertion-type scankey (see nbtree/README),
* but it can omit the rightmost column(s) of the index.
*
* When nextkey is false (the usual case), we are looking for the first
* item >= scankey. When nextkey is true, we are looking for the first
* item strictly greater than scankey.
*
* Return value is a stack of parent-page pointers. *bufP is set to the
* address of the leaf-page buffer, which is read-locked and pinned.
* No locks are held on the parent pages, however!
*
* NOTE that the returned buffer is read-locked regardless of the access
* parameter. However, access = BT_WRITE will allow an empty root page
* to be created and returned. When access = BT_READ, an empty index
* will result in *bufP being set to InvalidBuffer. Also, in BT_WRITE mode,
* any incomplete splits encountered during the search will be finished.
*/
BTStack
_bt_search(Relation rel, int keysz, ScanKey scankey, bool nextkey,
Buffer *bufP, int access)
{
BTStack stack_in = NULL;
/* Get the root page to start with */
*bufP = _bt_getroot(rel, access);
/* If index is empty and access = BT_READ, no root page is created. */
if (!BufferIsValid(*bufP))
return (BTStack) NULL;
/* Loop iterates once per level descended in the tree */
for (;;)
{
Page page;
BTPageOpaque opaque;
OffsetNumber offnum;
ItemId itemid;
IndexTuple itup;
BlockNumber blkno;
BlockNumber par_blkno;
BTStack new_stack;
/*
* Race -- the page we just grabbed may have split since we read its
* pointer in the parent (or metapage). If it has, we may need to
* move right to its new sibling. Do that.
*
* In write-mode, allow _bt_moveright to finish any incomplete splits
* along the way. Strictly speaking, we'd only need to finish an
* incomplete split on the leaf page we're about to insert to, not on
* any of the upper levels (they is taken care of in _bt_getstackbuf,
* if the leaf page is split and we insert to the parent page). But
* this is a good opportunity to finish splits of internal pages too.
*/
*bufP = _bt_moveright(rel, *bufP, keysz, scankey, nextkey,
(access == BT_WRITE), stack_in,
BT_READ);
/* if this is a leaf page, we're done */
page = BufferGetPage(*bufP);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (P_ISLEAF(opaque))
break;
/*
* Find the appropriate item on the internal page, and get the child
* page that it points to.
*/
offnum = _bt_binsrch(rel, *bufP, keysz, scankey, nextkey);
itemid = PageGetItemId(page, offnum);
itup = (IndexTuple) PageGetItem(page, itemid);
blkno = ItemPointerGetBlockNumber(&(itup->t_tid));
par_blkno = BufferGetBlockNumber(*bufP);
/*
* We need to save the location of the index entry we chose in the
* parent page on a stack. In case we split the tree, we'll use the
* stack to work back up to the parent page. We also save the actual
* downlink (TID) to uniquely identify the index entry, in case it
* moves right while we're working lower in the tree. See the paper
* by Lehman and Yao for how this is detected and handled. (We use the
* child link to disambiguate duplicate keys in the index -- Lehman
* and Yao disallow duplicate keys.)
*/
new_stack = (BTStack) palloc(sizeof(BTStackData));
new_stack->bts_blkno = par_blkno;
new_stack->bts_offset = offnum;
memcpy(&new_stack->bts_btentry, itup, sizeof(IndexTupleData));
new_stack->bts_parent = stack_in;
/* drop the read lock on the parent page, acquire one on the child */
*bufP = _bt_relandgetbuf(rel, *bufP, blkno, BT_READ);
/* okay, all set to move down a level */
stack_in = new_stack;
}
return stack_in;
}
/*
* _bt_endpoint() -- Find the first or last key in the index.
*
* This is used by _bt_first() to set up a scan when we've determined
* that the scan must start at the beginning or end of the index (for
* a forward or backward scan respectively).
*/
static bool
_bt_endpoint(IndexScanDesc scan, ScanDirection dir)
{
Relation rel;
Buffer buf;
Page page;
BTPageOpaque opaque;
ItemPointer current;
OffsetNumber maxoff;
OffsetNumber start;
BlockNumber blkno;
BTItem btitem;
IndexTuple itup;
BTScanOpaque so;
bool res;
bool continuescan;
rel = scan->indexRelation;
current = &(scan->currentItemData);
so = (BTScanOpaque) scan->opaque;
/*
* Scan down to the leftmost or rightmost leaf page. This is a
* simplified version of _bt_search(). We don't maintain a stack
* since we know we won't need it.
*/
buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir));
if (!BufferIsValid(buf))
{
/* empty index... */
ItemPointerSetInvalid(current);
so->btso_curbuf = InvalidBuffer;
return false;
}
blkno = BufferGetBlockNumber(buf);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
Assert(P_ISLEAF(opaque));
maxoff = PageGetMaxOffsetNumber(page);
if (ScanDirectionIsForward(dir))
{
/* There could be dead pages to the left, so not this: */
/* Assert(P_LEFTMOST(opaque)); */
start = P_FIRSTDATAKEY(opaque);
}
else if (ScanDirectionIsBackward(dir))
{
Assert(P_RIGHTMOST(opaque));
start = PageGetMaxOffsetNumber(page);
if (start < P_FIRSTDATAKEY(opaque)) /* watch out for empty
* page */
start = P_FIRSTDATAKEY(opaque);
}
else
{
elog(ERROR, "invalid scan direction: %d", (int) dir);
start = 0; /* keep compiler quiet */
}
ItemPointerSet(current, blkno, start);
/* remember which buffer we have pinned */
so->btso_curbuf = buf;
/*
* Left/rightmost page could be empty due to deletions, if so step
* till we find a nonempty page.
*/
if (start > maxoff)
{
if (!_bt_step(scan, &buf, dir))
return false;
start = ItemPointerGetOffsetNumber(current);
page = BufferGetPage(buf);
}
btitem = (BTItem) PageGetItem(page, PageGetItemId(page, start));
itup = &(btitem->bti_itup);
/* see if we picked a winner */
if (_bt_checkkeys(scan, itup, dir, &continuescan))
{
/* yes, return it */
scan->xs_ctup.t_self = itup->t_tid;
res = true;
}
else if (continuescan)
{
/* no, but there might be another one that is */
res = _bt_next(scan, dir);
}
else
{
/* no tuples in the index match this scan key */
ItemPointerSetInvalid(current);
so->btso_curbuf = InvalidBuffer;
_bt_relbuf(rel, buf);
res = false;
}
return res;
}
/*
* Post vacuum, iterate over all entries in index, check if the h_tid
* of each entry exists and is not dead. For specific system tables,
* also ensure that the key in index entry matches the corresponding
* attribute in the heap tuple.
*/
void
_bt_validate_vacuum(Relation irel, Relation hrel, TransactionId oldest_xmin)
{
MIRROREDLOCK_BUFMGR_DECLARE;
BlockNumber blkno;
BlockNumber num_pages;
Buffer ibuf = InvalidBuffer;
Buffer hbuf = InvalidBuffer;
Page ipage;
BTPageOpaque opaque;
IndexTuple itup;
HeapTupleData htup;
OffsetNumber maxoff,
minoff,
offnum;
Oid ioid,
hoid;
bool isnull;
blkno = BTREE_METAPAGE + 1;
num_pages = RelationGetNumberOfBlocks(irel);
elog(LOG, "btvalidatevacuum: index %s, heap %s",
RelationGetRelationName(irel), RelationGetRelationName(hrel));
MIRROREDLOCK_BUFMGR_LOCK;
for (; blkno < num_pages; blkno++)
{
ibuf = ReadBuffer(irel, blkno);
ipage = BufferGetPage(ibuf);
opaque = (BTPageOpaque) PageGetSpecialPointer(ipage);
if (!PageIsNew(ipage))
_bt_checkpage(irel, ibuf);
if (P_ISLEAF(opaque))
{
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(ipage);
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
itup = (IndexTuple) PageGetItem(ipage,
PageGetItemId(ipage, offnum));
ItemPointerCopy(&itup->t_tid, &htup.t_self);
/*
* TODO: construct a tid bitmap based on index tids
* and fetch heap tids in order afterwards. That will
* also allow validating if a heap tid appears twice
* in a unique index.
*/
if (!heap_release_fetch(hrel, SnapshotAny, &htup,
&hbuf, true, NULL))
{
elog(ERROR, "btvalidatevacuum: tid (%d,%d) from index %s "
"not found in heap %s",
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel),
RelationGetRelationName(hrel));
}
switch (HeapTupleSatisfiesVacuum(hrel, htup.t_data, oldest_xmin, hbuf))
{
case HEAPTUPLE_RECENTLY_DEAD:
case HEAPTUPLE_LIVE:
case HEAPTUPLE_INSERT_IN_PROGRESS:
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* these tuples are considered alive by vacuum */
break;
case HEAPTUPLE_DEAD:
elog(ERROR, "btvalidatevacuum: vacuum did not remove "
"dead tuple (%d,%d) from heap %s and index %s",
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(hrel),
RelationGetRelationName(irel));
break;
default:
elog(ERROR, "btvalidatevacuum: invalid visibility");
break;
}
switch(RelationGetRelid(irel))
{
case DatabaseOidIndexId:
case TypeOidIndexId:
case ClassOidIndexId:
case ConstraintOidIndexId:
hoid = HeapTupleGetOid(&htup);
ioid = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
if (hoid != ioid)
{
elog(ERROR,
"btvalidatevacuum: index oid(%d) != heap oid(%d)"
" tuple (%d,%d) index %s", ioid, hoid,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
break;
case GpRelationNodeOidIndexId:
hoid = heap_getattr(&htup, 1, RelationGetDescr(hrel), &isnull);
ioid = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
if (hoid != ioid)
{
elog(ERROR,
"btvalidatevacuum: index oid(%d) != heap oid(%d)"
" tuple (%d,%d) index %s", ioid, hoid,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
int4 hsegno = heap_getattr(&htup, 2, RelationGetDescr(hrel), &isnull);
int4 isegno = index_getattr(itup, 2, RelationGetDescr(irel), &isnull);
if (isegno != hsegno)
{
elog(ERROR,
"btvalidatevacuum: index segno(%d) != heap segno(%d)"
" tuple (%d,%d) index %s", isegno, hsegno,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
break;
default:
break;
}
if (RelationGetNamespace(irel) == PG_AOSEGMENT_NAMESPACE)
{
int4 isegno = index_getattr(itup, 1, RelationGetDescr(irel), &isnull);
int4 hsegno = heap_getattr(&htup, 1, RelationGetDescr(hrel), &isnull);
if (isegno != hsegno)
{
elog(ERROR,
"btvalidatevacuum: index segno(%d) != heap segno(%d)"
" tuple (%d,%d) index %s", isegno, hsegno,
ItemPointerGetBlockNumber(&itup->t_tid),
ItemPointerGetOffsetNumber(&itup->t_tid),
RelationGetRelationName(irel));
}
}
}
}
if (BufferIsValid(ibuf))
ReleaseBuffer(ibuf);
}
if (BufferIsValid(hbuf))
ReleaseBuffer(hbuf);
MIRROREDLOCK_BUFMGR_UNLOCK;
}
/*
* _bt_binsrch() -- Do a binary search for a key on a particular page.
*
* The scankey we get has the compare function stored in the procedure
* entry of each data struct. We invoke this regproc to do the
* comparison for every key in the scankey.
*
* On a leaf page, _bt_binsrch() returns the OffsetNumber of the first
* key >= given scankey. (NOTE: in particular, this means it is possible
* to return a value 1 greater than the number of keys on the page,
* if the scankey is > all keys on the page.)
*
* On an internal (non-leaf) page, _bt_binsrch() returns the OffsetNumber
* of the last key < given scankey. (Since _bt_compare treats the first
* data key of such a page as minus infinity, there will be at least one
* key < scankey, so the result always points at one of the keys on the
* page.) This key indicates the right place to descend to be sure we
* find all leaf keys >= given scankey.
*
* This procedure is not responsible for walking right, it just examines
* the given page. _bt_binsrch() has no lock or refcount side effects
* on the buffer.
*/
OffsetNumber
_bt_binsrch(Relation rel,
Buffer buf,
int keysz,
ScanKey scankey)
{
TupleDesc itupdesc;
Page page;
BTPageOpaque opaque;
OffsetNumber low,
high;
int32 result;
itupdesc = RelationGetDescr(rel);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
low = P_FIRSTDATAKEY(opaque);
high = PageGetMaxOffsetNumber(page);
/*
* If there are no keys on the page, return the first available slot.
* Note this covers two cases: the page is really empty (no keys), or
* it contains only a high key. The latter case is possible after
* vacuuming. This can never happen on an internal page, however,
* since they are never empty (an internal page must have children).
*/
if (high < low)
return low;
/*
* Binary search to find the first key on the page >= scan key. Loop
* invariant: all slots before 'low' are < scan key, all slots at or
* after 'high' are >= scan key. We can fall out when high == low.
*/
high++; /* establish the loop invariant for high */
while (high > low)
{
OffsetNumber mid = low + ((high - low) / 2);
/* We have low <= mid < high, so mid points at a real slot */
result = _bt_compare(rel, keysz, scankey, page, mid);
if (result > 0)
low = mid + 1;
else
high = mid;
}
/*
* At this point we have high == low, but be careful: they could point
* past the last slot on the page.
*
* On a leaf page, we always return the first key >= scan key (which
* could be the last slot + 1).
*/
if (P_ISLEAF(opaque))
return low;
/*
* On a non-leaf page, return the last key < scan key. There must be
* one if _bt_compare() is playing by the rules.
*/
Assert(low > P_FIRSTDATAKEY(opaque));
return OffsetNumberPrev(low);
}
/*
* btvacuumpage --- VACUUM one page
*
* This processes a single page for btvacuumscan(). In some cases we
* must go back and re-examine previously-scanned pages; this routine
* recurses when necessary to handle that case.
*
* blkno is the page to process. orig_blkno is the highest block number
* reached by the outer btvacuumscan loop (the same as blkno, unless we
* are recursing to re-examine a previous page).
*/
static void
btvacuumpage(BTVacState *vstate, BlockNumber blkno, BlockNumber orig_blkno)
{
IndexVacuumInfo *info = vstate->info;
IndexBulkDeleteResult *stats = vstate->stats;
IndexBulkDeleteCallback callback = vstate->callback;
void *callback_state = vstate->callback_state;
Relation rel = info->index;
bool delete_now;
BlockNumber recurse_to;
Buffer buf;
Page page;
BTPageOpaque opaque = NULL;
restart:
delete_now = false;
recurse_to = P_NONE;
/* call vacuum_delay_point while not holding any buffer lock */
vacuum_delay_point();
/*
* We can't use _bt_getbuf() here because it always applies
* _bt_checkpage(), which will barf on an all-zero page. We want to
* recycle all-zero pages, not fail. Also, we want to use a nondefault
* buffer access strategy.
*/
buf = ReadBufferExtended(rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
info->strategy);
LockBuffer(buf, BT_READ);
page = BufferGetPage(buf);
if (!PageIsNew(page))
{
_bt_checkpage(rel, buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
}
/*
* If we are recursing, the only case we want to do anything with is a
* live leaf page having the current vacuum cycle ID. Any other state
* implies we already saw the page (eg, deleted it as being empty).
*/
if (blkno != orig_blkno)
{
if (_bt_page_recyclable(page) ||
P_IGNORE(opaque) ||
!P_ISLEAF(opaque) ||
opaque->btpo_cycleid != vstate->cycleid)
{
_bt_relbuf(rel, buf);
return;
}
}
/* Page is valid, see what to do with it */
if (_bt_page_recyclable(page))
{
/* Okay to recycle this page */
RecordFreeIndexPage(rel, blkno);
vstate->totFreePages++;
stats->pages_deleted++;
}
else if (P_ISDELETED(opaque))
{
/* Already deleted, but can't recycle yet */
stats->pages_deleted++;
}
else if (P_ISHALFDEAD(opaque))
{
/* Half-dead, try to delete */
delete_now = true;
}
else if (P_ISLEAF(opaque))
{
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable;
OffsetNumber offnum,
minoff,
maxoff;
/*
* Trade in the initial read lock for a super-exclusive write lock on
* this page. We must get such a lock on every leaf page over the
* course of the vacuum scan, whether or not it actually contains any
* deletable tuples --- see nbtree/README.
*/
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBufferForCleanup(buf);
/*
* Remember highest leaf page number we've taken cleanup lock on; see
* notes in btvacuumscan
*/
if (blkno > vstate->lastBlockLocked)
vstate->lastBlockLocked = blkno;
/*
* Check whether we need to recurse back to earlier pages. What we
* are concerned about is a page split that happened since we started
* the vacuum scan. If the split moved some tuples to a lower page
* then we might have missed 'em. If so, set up for tail recursion.
* (Must do this before possibly clearing btpo_cycleid below!)
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid &&
!(opaque->btpo_flags & BTP_SPLIT_END) &&
!P_RIGHTMOST(opaque) &&
opaque->btpo_next < orig_blkno)
recurse_to = opaque->btpo_next;
/*
* Scan over all items to see which ones need deleted according to the
* callback function.
*/
ndeletable = 0;
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(page);
if (callback)
{
for (offnum = minoff;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
IndexTuple itup;
ItemPointer htup;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offnum));
htup = &(itup->t_tid);
/*
* During Hot Standby we currently assume that
* XLOG_BTREE_VACUUM records do not produce conflicts. That is
* only true as long as the callback function depends only
* upon whether the index tuple refers to heap tuples removed
* in the initial heap scan. When vacuum starts it derives a
* value of OldestXmin. Backends taking later snapshots could
* have a RecentGlobalXmin with a later xid than the vacuum's
* OldestXmin, so it is possible that row versions deleted
* after OldestXmin could be marked as killed by other
* backends. The callback function *could* look at the index
* tuple state in isolation and decide to delete the index
* tuple, though currently it does not. If it ever did, we
* would need to reconsider whether XLOG_BTREE_VACUUM records
* should cause conflicts. If they did cause conflicts they
* would be fairly harsh conflicts, since we haven't yet
* worked out a way to pass a useful value for
* latestRemovedXid on the XLOG_BTREE_VACUUM records. This
* applies to *any* type of index that marks index tuples as
* killed.
*/
if (callback(htup, callback_state))
deletable[ndeletable++] = offnum;
}
}
/*
* Apply any needed deletes. We issue just one _bt_delitems_vacuum()
* call per page, so as to minimize WAL traffic.
*/
if (ndeletable > 0)
{
/*
* Notice that the issued XLOG_BTREE_VACUUM WAL record includes
* all information to the replay code to allow it to get a cleanup
* lock on all pages between the previous lastBlockVacuumed and
* this page. This ensures that WAL replay locks all leaf pages at
* some point, which is important should non-MVCC scans be
* requested. This is currently unused on standby, but we record
* it anyway, so that the WAL contains the required information.
*
* Since we can visit leaf pages out-of-order when recursing,
* replay might end up locking such pages an extra time, but it
* doesn't seem worth the amount of bookkeeping it'd take to avoid
* that.
*/
_bt_delitems_vacuum(rel, buf, deletable, ndeletable,
vstate->lastBlockVacuumed);
/*
* Remember highest leaf page number we've issued a
* XLOG_BTREE_VACUUM WAL record for.
*/
if (blkno > vstate->lastBlockVacuumed)
vstate->lastBlockVacuumed = blkno;
stats->tuples_removed += ndeletable;
/* must recompute maxoff */
maxoff = PageGetMaxOffsetNumber(page);
}
else
{
/*
* If the page has been split during this vacuum cycle, it seems
* worth expending a write to clear btpo_cycleid even if we don't
* have any deletions to do. (If we do, _bt_delitems_vacuum takes
* care of this.) This ensures we won't process the page again.
*
* We treat this like a hint-bit update because there's no need to
* WAL-log it.
*/
if (vstate->cycleid != 0 &&
opaque->btpo_cycleid == vstate->cycleid)
{
opaque->btpo_cycleid = 0;
MarkBufferDirtyHint(buf, true);
}
}
/*
* If it's now empty, try to delete; else count the live tuples. We
* don't delete when recursing, though, to avoid putting entries into
* freePages out-of-order (doesn't seem worth any extra code to handle
* the case).
*/
if (minoff > maxoff)
delete_now = (blkno == orig_blkno);
else
stats->num_index_tuples += maxoff - minoff + 1;
}
if (delete_now)
{
MemoryContext oldcontext;
int ndel;
/* Run pagedel in a temp context to avoid memory leakage */
MemoryContextReset(vstate->pagedelcontext);
oldcontext = MemoryContextSwitchTo(vstate->pagedelcontext);
ndel = _bt_pagedel(rel, buf);
/* count only this page, else may double-count parent */
if (ndel)
stats->pages_deleted++;
MemoryContextSwitchTo(oldcontext);
/* pagedel released buffer, so we shouldn't */
}
else
_bt_relbuf(rel, buf);
/*
* This is really tail recursion, but if the compiler is too stupid to
* optimize it as such, we'd eat an uncomfortably large amount of stack
* space per recursion level (due to the deletable[] array). A failure is
* improbable since the number of levels isn't likely to be large ... but
* just in case, let's hand-optimize into a loop.
*/
if (recurse_to != P_NONE)
{
blkno = recurse_to;
goto restart;
}
}
/* checks the individual attributes of the tuple */
uint32 check_index_tuple_attributes(Relation rel, PageHeader header, int block, int i, char *buffer) {
IndexTuple tuple;
uint32 nerrs = 0;
int j, off;
bits8 * bitmap;
BTPageOpaque opaque;
ereport(DEBUG2,(errmsg("[%d:%d] checking attributes for the tuple", block, i)));
/* get the index tuple and info about the page */
tuple = (IndexTuple)(buffer + header->pd_linp[i].lp_off);
opaque = (BTPageOpaque)(buffer + header->pd_special);
/* current attribute offset - always starts at (buffer + off) */
off = header->pd_linp[i].lp_off + IndexInfoFindDataOffset(tuple->t_info);
ereport(DEBUG3,(errmsg("[%d:%d] tuple has %d attributes", block, (i+1),
RelationGetNumberOfAttributes(rel))));
bitmap = (bits8*)(buffer + header->pd_linp[i].lp_off + sizeof(IndexTupleData));
/* TODO This is mostly copy'n'paste from check_heap_tuple_attributes,
so maybe it could be refactored to share the code. */
/* For left-most tuples on non-leaf pages, there are no data actually
(see src/backend/access/nbtree/README, last paragraph in section "Notes
About Data Representation")
Use P_LEFTMOST/P_ISLEAF to identify such cases (for the leftmost item only)
and set len = 0.
*/
if (P_LEFTMOST(opaque) && (! P_ISLEAF(opaque)) && (i == 0)) {
ereport(DEBUG3, (errmsg("[%d:%d] leftmost tuple on non-leaf block => no data, skipping", block, i)));
return nerrs;
}
/* check all the index attributes */
for (j = 0; j < rel->rd_att->natts; j++) {
/* default length of the attribute */
int len = rel->rd_att->attrs[j]->attlen;
/* copy from src/backend/commands/analyze.c */
bool is_varlena = (!rel->rd_att->attrs[j]->attbyval && len == -1);
bool is_varwidth = (!rel->rd_att->attrs[j]->attbyval && len < 0); /* thus it's "len = -2" */
/* if the attribute is marked as NULL (in the tuple header), skip to the next attribute */
if (IndexTupleHasNulls(tuple) && att_isnull(j, bitmap)) {
ereport(DEBUG3, (errmsg("[%d:%d] attribute '%s' is NULL (skipping)", block, (i+1), rel->rd_att->attrs[j]->attname.data)));
continue;
}
/* fix the alignment (see src/include/access/tupmacs.h) */
off = att_align_pointer(off, rel->rd_att->attrs[j]->attalign, rel->rd_att->attrs[j]->attlen, buffer+off);
if (is_varlena) {
/*
other interesting macros (see postgres.h) - should do something about those ...
VARATT_IS_COMPRESSED(PTR) VARATT_IS_4B_C(PTR)
VARATT_IS_EXTERNAL(PTR) VARATT_IS_1B_E(PTR)
VARATT_IS_SHORT(PTR) VARATT_IS_1B(PTR)
VARATT_IS_EXTENDED(PTR) (!VARATT_IS_4B_U(PTR))
*/
len = VARSIZE_ANY(buffer + off);
if (len < 0) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' has negative length < 0 (%d)", block, (i+1), rel->rd_att->attrs[j]->attname.data, len)));
++nerrs;
break;
}
if (VARATT_IS_COMPRESSED(buffer + off)) {
/* the raw length should be less than 1G (and positive) */
if ((VARRAWSIZE_4B_C(buffer + off) < 0) || (VARRAWSIZE_4B_C(buffer + off) > 1024*1024)) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' has invalid length %d (should be between 0 and 1G)", block, (i+1), rel->rd_att->attrs[j]->attname.data, VARRAWSIZE_4B_C(buffer + off))));
++nerrs;
/* no break here, this does not break the page structure - we may check the other attributes */
}
}
/* FIXME Check if the varlena value may be detoasted. */
} else if (is_varwidth) {
/* get the C-string length (at most to the end of tuple), +1 as it does not include '\0' at the end */
/* if the string is not properly terminated, then this returns 'remaining space + 1' so it's detected */
len = strnlen(buffer + off, header->pd_linp[i].lp_off + len + header->pd_linp[i].lp_len - off) + 1;
}
/* Check if the length makes sense (is not negative and does not overflow
* the tuple end, stop validating the other rows (we don't know where to
* continue anyway). */
if (off + len > (header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' (off=%d len=%d) overflows tuple end (off=%d, len=%d)", block, (i+1), rel->rd_att->attrs[j]->attname.data, off, len, header->pd_linp[i].lp_off, header->pd_linp[i].lp_len)));
++nerrs;
break;
}
/* skip to the next attribute */
off += len;
ereport(DEBUG3,(errmsg("[%d:%d] attribute '%s' len=%d", block, (i+1), rel->rd_att->attrs[j]->attname.data, len)));
}
ereport(DEBUG3,(errmsg("[%d:%d] last attribute ends at %d, tuple ends at %d", block, (i+1), off, header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)));
/* after the last attribute, the offset should be exactly the same as the end of the tuple */
if (MAXALIGN(off) != header->pd_linp[i].lp_off + header->pd_linp[i].lp_len) {
ereport(WARNING, (errmsg("[%d:%d] the last attribute ends at %d but the tuple ends at %d", block, (i+1), off, header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)));
++nerrs;
}
return nerrs;
}
