/*
* _bt_walk_left() -- step left one page, if possible
*
* The given buffer must be pinned and read-locked. This will be dropped
* before stepping left. On return, we have pin and read lock on the
* returned page, instead.
*
* Returns InvalidBuffer if there is no page to the left (no lock is held
* in that case).
*
* When working on a non-leaf level, it is possible for the returned page
* to be half-dead; the caller should check that condition and step left
* again if it's important.
*/
static Buffer
_bt_walk_left(Relation rel, Buffer buf)
{
Page page;
BTPageOpaque opaque;
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
for (;;)
{
BlockNumber obknum;
BlockNumber lblkno;
BlockNumber blkno;
int tries;
/* if we're at end of tree, release buf and return failure */
if (P_LEFTMOST(opaque))
{
_bt_relbuf(rel, buf);
break;
}
/* remember original page we are stepping left from */
obknum = BufferGetBlockNumber(buf);
/* step left */
blkno = lblkno = opaque->btpo_prev;
_bt_relbuf(rel, buf);
/* check for interrupts while we're not holding any buffer lock */
CHECK_FOR_INTERRUPTS();
buf = _bt_getbuf(rel, blkno, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
/*
* If this isn't the page we want, walk right till we find what we
* want --- but go no more than four hops (an arbitrary limit). If we
* don't find the correct page by then, the most likely bet is that
* the original page got deleted and isn't in the sibling chain at all
* anymore, not that its left sibling got split more than four times.
*
* Note that it is correct to test P_ISDELETED not P_IGNORE here,
* because half-dead pages are still in the sibling chain. Caller
* must reject half-dead pages if wanted.
*/
tries = 0;
for (;;)
{
if (!P_ISDELETED(opaque) && opaque->btpo_next == obknum)
{
/* Found desired page, return it */
return buf;
}
if (P_RIGHTMOST(opaque) || ++tries > 4)
break;
blkno = opaque->btpo_next;
buf = _bt_relandgetbuf(rel, buf, blkno, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
}
/* Return to the original page to see what's up */
buf = _bt_relandgetbuf(rel, buf, obknum, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (P_ISDELETED(opaque))
{
/*
* It was deleted. Move right to first nondeleted page (there
* must be one); that is the page that has acquired the deleted
* one's keyspace, so stepping left from it will take us where we
* want to be.
*/
for (;;)
{
if (P_RIGHTMOST(opaque))
elog(ERROR, "fell off the end of index \"%s\"",
RelationGetRelationName(rel));
blkno = opaque->btpo_next;
buf = _bt_relandgetbuf(rel, buf, blkno, BT_READ);
page = BufferGetPage(buf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (!P_ISDELETED(opaque))
break;
}
/*
* Now return to top of loop, resetting obknum to point to this
* nondeleted page, and try again.
*/
}
else
{
/*
* It wasn't deleted; the explanation had better be that the page
* to the left got split or deleted. Without this check, we'd go
* into an infinite loop if there's anything wrong.
*/
if (opaque->btpo_prev == lblkno)
elog(ERROR, "could not find left sibling of block %u in index \"%s\"",
obknum, RelationGetRelationName(rel));
/* Okay to try again with new lblkno value */
}
}
return InvalidBuffer;
}
/*
* _bt_getroot() -- Get the root page of the btree.
*
* Since the root page can move around the btree file, we have to read
* its location from the metadata page, and then read the root page
* itself. If no root page exists yet, we have to create one. The
* standard class of race conditions exists here; I think I covered
* them all in the Hopi Indian rain dance of lock requests below.
*
* The access type parameter (BT_READ or BT_WRITE) controls whether
* a new root page will be created or not. If access = BT_READ,
* and no root page exists, we just return InvalidBuffer. For
* BT_WRITE, we try to create the root page if it doesn't exist.
* NOTE that the returned root page will have only a read lock set
* on it even if access = BT_WRITE!
*
* The returned page is not necessarily the true root --- it could be
* a "fast root" (a page that is alone in its level due to deletions).
* Also, if the root page is split while we are "in flight" to it,
* what we will return is the old root, which is now just the leftmost
* page on a probably-not-very-wide level. For most purposes this is
* as good as or better than the true root, so we do not bother to
* insist on finding the true root. We do, however, guarantee to
* return a live (not deleted or half-dead) page.
*
* On successful return, the root page is pinned and read-locked.
* The metadata page is not locked or pinned on exit.
*/
Buffer
_bt_getroot(Relation rel, int access)
{
Buffer metabuf;
Page metapg;
BTPageOpaque metaopaque;
Buffer rootbuf;
Page rootpage;
BTPageOpaque rootopaque;
BlockNumber rootblkno;
uint32 rootlevel;
BTMetaPageData *metad;
MIRROREDLOCK_BUFMGR_MUST_ALREADY_BE_HELD;
/*
* Try to use previously-cached metapage data to find the root. This
* normally saves one buffer access per index search, which is a very
* helpful savings in bufmgr traffic and hence contention.
*/
if (rel->rd_amcache != NULL)
{
metad = (BTMetaPageData *) rel->rd_amcache;
/* We shouldn't have cached it if any of these fail */
Assert(metad->btm_magic == BTREE_MAGIC);
Assert(metad->btm_version == BTREE_VERSION);
Assert(metad->btm_root != P_NONE);
rootblkno = metad->btm_fastroot;
Assert(rootblkno != P_NONE);
rootlevel = metad->btm_fastlevel;
rootbuf = _bt_getbuf(rel, rootblkno, BT_READ);
rootpage = BufferGetPage(rootbuf);
rootopaque = (BTPageOpaque) PageGetSpecialPointer(rootpage);
/*
* Since the cache might be stale, we check the page more carefully
* here than normal. We *must* check that it's not deleted. If it's
* not alone on its level, then we reject too --- this may be overly
* paranoid but better safe than sorry. Note we don't check P_ISROOT,
* because that's not set in a "fast root".
*/
if (!P_IGNORE(rootopaque) &&
rootopaque->btpo.level == rootlevel &&
P_LEFTMOST(rootopaque) &&
P_RIGHTMOST(rootopaque))
{
/* OK, accept cached page as the root */
return rootbuf;
}
_bt_relbuf(rel, rootbuf);
/* Cache is stale, throw it away */
if (rel->rd_amcache)
pfree(rel->rd_amcache);
rel->rd_amcache = NULL;
}
metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metaopaque = (BTPageOpaque) PageGetSpecialPointer(metapg);
metad = BTPageGetMeta(metapg);
/* sanity-check the metapage */
if (!(metaopaque->btpo_flags & BTP_META) ||
metad->btm_magic != BTREE_MAGIC)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
errmsg("index \"%s\" is not a btree",
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 no root page initialized yet, do it */
if (metad->btm_root == P_NONE)
{
/* If access = BT_READ, caller doesn't want us to create root yet */
if (access == BT_READ)
{
_bt_relbuf(rel, metabuf);
return InvalidBuffer;
}
// Fetch gp_persistent_relation_node information that will be added to XLOG record.
RelationFetchGpRelationNodeForXLog(rel);
/* trade in our read lock for a write lock */
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
LockBuffer(metabuf, BT_WRITE);
/*
* Race condition: if someone else initialized the metadata between
* the time we released the read lock and acquired the write lock, we
* must avoid doing it again.
*/
if (metad->btm_root != P_NONE)
{
/*
* Metadata initialized by someone else. In order to guarantee no
* deadlocks, we have to release the metadata page and start all
* over again. (Is that really true? But it's hardly worth trying
* to optimize this case.)
*/
_bt_relbuf(rel, metabuf);
return _bt_getroot(rel, access);
}
/*
* Get, initialize, write, and leave a lock of the appropriate type on
* the new root page. Since this is the first page in the tree, it's
* a leaf as well as the root.
*/
rootbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
rootblkno = BufferGetBlockNumber(rootbuf);
rootpage = BufferGetPage(rootbuf);
rootopaque = (BTPageOpaque) PageGetSpecialPointer(rootpage);
rootopaque->btpo_prev = rootopaque->btpo_next = P_NONE;
rootopaque->btpo_flags = (BTP_LEAF | BTP_ROOT);
rootopaque->btpo.level = 0;
rootopaque->btpo_cycleid = 0;
/* NO ELOG(ERROR) till meta is updated */
START_CRIT_SECTION();
metad->btm_root = rootblkno;
metad->btm_level = 0;
metad->btm_fastroot = rootblkno;
metad->btm_fastlevel = 0;
MarkBufferDirty(rootbuf);
MarkBufferDirty(metabuf);
/* XLOG stuff */
if (!rel->rd_istemp)
{
xl_btree_newroot xlrec;
XLogRecPtr recptr;
XLogRecData rdata;
xl_btreenode_set(&(xlrec.btreenode), rel);
xlrec.rootblk = rootblkno;
xlrec.level = 0;
rdata.data = (char *) &xlrec;
rdata.len = SizeOfBtreeNewroot;
rdata.buffer = InvalidBuffer;
rdata.next = NULL;
recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT, &rdata);
PageSetLSN(rootpage, recptr);
PageSetTLI(rootpage, ThisTimeLineID);
PageSetLSN(metapg, recptr);
PageSetTLI(metapg, ThisTimeLineID);
}
END_CRIT_SECTION();
/*
* Send out relcache inval for metapage change (probably unnecessary
* here, but let's be safe).
*/
CacheInvalidateRelcache(rel);
/*
* swap root write lock for read lock. There is no danger of anyone
* else accessing the new root page while it's unlocked, since no one
* else knows where it is yet.
*/
LockBuffer(rootbuf, BUFFER_LOCK_UNLOCK);
LockBuffer(rootbuf, BT_READ);
/* okay, metadata is correct, release lock on it */
_bt_relbuf(rel, metabuf);
}
else
{
rootblkno = metad->btm_fastroot;
Assert(rootblkno != P_NONE);
rootlevel = metad->btm_fastlevel;
/*
* Cache the metapage data for next time
*/
rel->rd_amcache = MemoryContextAlloc(rel->rd_indexcxt,
sizeof(BTMetaPageData));
memcpy(rel->rd_amcache, metad, sizeof(BTMetaPageData));
/*
* We are done with the metapage; arrange to release it via first
* _bt_relandgetbuf call
*/
rootbuf = metabuf;
for (;;)
{
rootbuf = _bt_relandgetbuf(rel, rootbuf, rootblkno, BT_READ);
rootpage = BufferGetPage(rootbuf);
rootopaque = (BTPageOpaque) PageGetSpecialPointer(rootpage);
if (!P_IGNORE(rootopaque))
break;
/* it's dead, Jim. step right one page */
if (P_RIGHTMOST(rootopaque))
elog(ERROR, "no live root page found in index \"%s\"",
RelationGetRelationName(rel));
rootblkno = rootopaque->btpo_next;
}
/* Note: can't check btpo.level on deleted pages */
if (rootopaque->btpo.level != rootlevel)
elog(ERROR, "root page %u of index \"%s\" has level %u, expected %u",
rootblkno, RelationGetRelationName(rel),
rootopaque->btpo.level, rootlevel);
}
/*
* By here, we have a pin and read lock on the root page, and no lock set
* on the metadata page. Return the root page's buffer.
*/
return rootbuf;
}
/* 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;
}
