ExecUpdate checked for whether ExecBRUpdateTriggers had returned a new
tuple value by seeing if the returned tuple was pointer-equal to the old
one. But the "old one" was in estate->es_junkFilter's result slot, which
would be scribbled on if we had done an EvalPlanQual update in response to
a concurrent update of the target tuple; therefore we were comparing a
dangling pointer to a live one. Given the right set of circumstances we
could get a false match, resulting in not forcing the tuple to be stored in
the slot we thought it was stored in. In the case reported by Maxim Boguk
in bug #5798, this led to "cannot extract system attribute from virtual
tuple" failures when trying to do "RETURNING ctid". I believe there is a
very-low-probability chance of more serious errors, such as generating
incorrect index entries based on the original rather than the
trigger-modified version of the row.
In HEAD, change all of ExecBRInsertTriggers, ExecIRInsertTriggers,
ExecBRUpdateTriggers, and ExecIRUpdateTriggers so that they continue to
have similar APIs. In the back branches I just changed
ExecBRUpdateTriggers, since there is no bug in the ExecBRInsertTriggers
case.
This commit provides the core code and documentation needed. A contrib
module test case will follow shortly.
Shigeru Hanada, Jan Urbanski, Heikki Linnakangas
Flattening of subquery range tables during setrefs.c could lead to the
rangetable indexes in PlanRowMark nodes not matching up with the column
names previously assigned to the corresponding resjunk ctid (resp. tableoid
or wholerow) columns. Typical symptom would be either a "cannot extract
system attribute from virtual tuple" error or an Assert failure. This
wasn't a problem before 9.0 because we didn't support FOR UPDATE below the
top query level, and so the final flattening could never renumber an RTE
that was relevant to FOR UPDATE. Fix by using a plan-tree-wide unique
number for each PlanRowMark to label the associated resjunk columns, so
that the number need not change during flattening.
Per report from David Johnston (though I'm darned if I can see how this got
past initial testing of the relevant code). Back-patch to 9.0.
This adds collation support for columns and domains, a COLLATE clause
to override it per expression, and B-tree index support.
Peter Eisentraut
reviewed by Pavel Stehule, Itagaki Takahiro, Robert Haas, Noah Misch
Until now, our Serializable mode has in fact been what's called Snapshot
Isolation, which allows some anomalies that could not occur in any
serialized ordering of the transactions. This patch fixes that using a
method called Serializable Snapshot Isolation, based on research papers by
Michael J. Cahill (see README-SSI for full references). In Serializable
Snapshot Isolation, transactions run like they do in Snapshot Isolation,
but a predicate lock manager observes the reads and writes performed and
aborts transactions if it detects that an anomaly might occur. This method
produces some false positives, ie. it sometimes aborts transactions even
though there is no anomaly.
To track reads we implement predicate locking, see storage/lmgr/predicate.c.
Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared
memory is finite, so when a transaction takes many tuple-level locks on a
page, the locks are promoted to a single page-level lock, and further to a
single relation level lock if necessary. To lock key values with no matching
tuple, a sequential scan always takes a relation-level lock, and an index
scan acquires a page-level lock that covers the search key, whether or not
there are any matching keys at the moment.
A predicate lock doesn't conflict with any regular locks or with another
predicate locks in the normal sense. They're only used by the predicate lock
manager to detect the danger of anomalies. Only serializable transactions
participate in predicate locking, so there should be no extra overhead for
for other transactions.
Predicate locks can't be released at commit, but must be remembered until
all the transactions that overlapped with it have completed. That means that
we need to remember an unbounded amount of predicate locks, so we apply a
lossy but conservative method of tracking locks for committed transactions.
If we run short of shared memory, we overflow to a new "pg_serial" SLRU
pool.
We don't currently allow Serializable transactions in Hot Standby mode.
That would be hard, because even read-only transactions can cause anomalies
that wouldn't otherwise occur.
Serializable isolation mode now means the new fully serializable level.
Repeatable Read gives you the old Snapshot Isolation level that we have
always had.
Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and
Anssi Kääriäinen
In an inherited UPDATE/DELETE, each target table has its own subplan,
because it might have a column set different from other targets. This
means that the resjunk columns we add to support EvalPlanQual might be
at different physical column numbers in each subplan. The EvalPlanQual
rewrite I did for 9.0 failed to account for this, resulting in possible
misbehavior or even crashes during concurrent updates to the same row,
as seen in a recent report from Gordon Shannon. Revise the data structure
so that we track resjunk column numbers separately for each subplan.
I also chose to move responsibility for identifying the physical column
numbers back to executor startup, instead of assuming that numbers derived
during preprocess_targetlist would stay valid throughout subsequent
massaging of the plan. That's a bit slower, so we might want to consider
undoing it someday; but it would complicate the patch considerably and
didn't seem justifiable in a bug fix that has to be back-patched to 9.0.
Foreign tables are a core component of SQL/MED. This commit does
not provide a working SQL/MED infrastructure, because foreign tables
cannot yet be queried. Support for foreign table scans will need to
be added in a future patch. However, this patch creates the necessary
system catalog structure, syntax support, and support for ancillary
operations such as COMMENT and SECURITY LABEL.
Shigeru Hanada, heavily revised by Robert Haas
This is advantageous first because it allows us to hash the smaller table
regardless of the outer-join type, and second because hash join can be more
flexible than merge join in dealing with arbitrary join quals in a FULL
join. For merge join all the join quals have to be mergejoinable, but hash
join will work so long as there's at least one hashjoinable qual --- the
others can be any condition. (This is true essentially because we don't
keep per-inner-tuple match flags in merge join, while hash join can do so.)
To do this, we need a has-it-been-matched flag for each tuple in the
hashtable, not just one for the current outer tuple. The key idea that
makes this practical is that we can store the match flag in the tuple's
infomask, since there are lots of bits there that are of no interest for a
MinimalTuple. So we aren't increasing the size of the hashtable at all for
the feature.
To write this without turning the hash code into even more of a pile of
spaghetti than it already was, I rewrote ExecHashJoin in a state-machine
style, similar to ExecMergeJoin. Other than that decision, it was pretty
straightforward.
This commit replaces pg_class.relistemp with pg_class.relpersistence;
and also modifies the RangeVar node type to carry relpersistence rather
than istemp. It also removes removes rd_istemp from RelationData and
instead performs the correct computation based on relpersistence.
For clarity, we add three new macros: RelationNeedsWAL(),
RelationUsesLocalBuffers(), and RelationUsesTempNamespace(), so that we
can clarify the purpose of each check that previous depended on
rd_istemp.
This is intended as infrastructure for the upcoming unlogged tables
patch, as well as for future possible work on global temporary tables.
This is a heavily revised version of builtin_knngist_core-0.9. The
ordering operators are no longer mixed in with actual quals, which would
have confused not only humans but significant parts of the planner.
Instead, ordering operators are carried separately throughout planning and
execution.
Since the API for ambeginscan and amrescan functions had to be changed
anyway, this commit takes the opportunity to rationalize that a bit.
RelationGetIndexScan no longer forces a premature index_rescan call;
instead, callers of index_beginscan must call index_rescan too. Aside from
making the AM-side initialization logic a bit less peculiar, this has the
advantage that we do not make a useless extra am_rescan call when there are
runtime key values. AMs formerly could not assume that the key values
passed to amrescan were actually valid; now they can.
Teodor Sigaev and Tom Lane
There were corner cases in which the planner would attempt to inline such
a function, which would result in a failure at runtime due to loss of
information about exactly what the result record type is. Fix by disabling
inlining when the function's recorded result type is RECORD. There might
be some sub-cases where inlining could still be allowed, but this is a
simple and backpatchable fix, so leave refinements for another day.
Per bug #5777 from Nate Carson.
Back-patch to all supported branches. 8.1 happens to avoid a core-dump
here, but it still does the wrong thing.
If we have Limit->Result->Sort, the Result might be projecting a tlist
that contains a set-returning function. If so, it's possible for the
SRF to sometimes return zero rows, which means we could need to fetch
more than N rows from the Sort in order to satisfy LIMIT N.
So top-N sorting cannot be used in this scenario.
Fix things so that top-N sorting can be used in child Sort nodes of a
MergeAppend node, when there is a LIMIT and no intervening joins or
grouping. Actually doing this on the executor side isn't too bad,
but it's a bit messier to get the planner to cost it properly.
Per gripe from Robert Haas.
In passing, fix an oversight in the original top-N-sorting patch:
query_planner should not assume that a LIMIT can be used to make an
explicit sort cheaper when there will be grouping or aggregation in
between. Possibly this should be back-patched, but I'm not sure the
mistake is serious enough to be a real problem in practice.
and related routines.
We already had a redundant FunctionCallInfoData struct in FuncExprState,
but were using that copy only in set-returning-function cases, to avoid
keeping function evaluation state in the expression tree for the benefit
of plpgsql's "simple expression" logic. But of course that didn't work
anyway. Given the recent fixes in plpgsql there is no need to have two
separate behaviors here. Getting rid of the local FunctionCallInfoData
structs should make things a little faster (because we don't need to do
InitFunctionCallInfoData each time), and it also makes for a noticeable
reduction in stack space consumption during recursive calls.
This patch eliminates the former need to sort the output of an Append scan
when an ordered scan of an inheritance tree is wanted. This should be
particularly useful for fast-start cases such as queries with LIMIT.
Original patch by Greg Stark, with further hacking by Hans-Jurgen Schonig,
Robert Haas, and Tom Lane.
This patch adds the SQL-standard concept of an INSTEAD OF trigger, which
is fired instead of performing a physical insert/update/delete. The
trigger function is passed the entire old and/or new rows of the view,
and must figure out what to do to the underlying tables to implement
the update. So this feature can be used to implement updatable views
using trigger programming style rather than rule hacking.
In passing, this patch corrects the names of some columns in the
information_schema.triggers view. It seems the SQL committee renamed
them somewhere between SQL:99 and SQL:2003.
Dean Rasheed, reviewed by Bernd Helmle; some additional hacking by me.
transaction snapshots, i.e. a snapshot registered at the beginning of
a transaction. Change variable naming and comments to reflect this reality
in preparation for a future, truly serializable mode, e.g.
Serializable Snapshot Isolation (SSI).
For the moment transaction snapshots are still used to implement
SERIALIZABLE, but hopefully not for too much longer. Patch by Kevin
Grittner and Dan Ports with review and some minor wording changes by me.
returning "record" actually do have the same rowtype. This is needed because
the parser can't realistically enforce that they will all have the same typmod,
as seen in a recent example from David Wheeler.
Back-patch to 8.0, which is as far back as we have the notion of RECORD
subtypes being distinguished by typmod. Wheeler's example depends on
8.4-and-up features, but I suspect there may be ways to provoke similar
failures before 8.4.
ExecModifyTable(). This avoids memory leakage when trigger functions leave
junk behind in that context (as they more or less must). Problem and solution
identified by Dean Rasheed.
I'm a bit concerned about the longevity of this solution --- once a plan can
have multiple ModifyTable nodes, we are very possibly going to have to do
something different. But it should hold up for 9.0.
unqualified names.
- Add a missing_ok parameter to get_tablespace_oid.
- Avoid duplicating get_tablespace_od guts in objectNamesToOids.
- Add a missing_ok parameter to get_database_oid.
- Replace get_roleid and get_role_checked with get_role_oid.
- Add get_namespace_oid, get_language_oid, get_am_oid.
- Refactor existing code to use new interfaces.
Thanks to KaiGai Kohei for the review.
list in ExecLockRows() forgot to allow for the possibility that some of the
rowmarks are for child tables that aren't relevant to the current row.
Per report from Kenichiro Tanaka.
a pass-by-reference datatype with a nontrivial projection step.
We were using the same memory context for the projection operation as for
the temporary context used by the hashtable routines in execGrouping.c.
However, the hashtable routines feel free to reset their temp context at
any time, which'd lead to destroying input data that was still needed.
Report and diagnosis by Tao Ma.
Back-patch to 8.1, where the problem was introduced by the changes that
allowed us to work with "virtual" tuples instead of materializing intermediate
tuple values everywhere. The earlier code looks quite similar, but it doesn't
suffer the problem because the data gets copied into another context as a
result of having to materialize ExecProject's output tuple.
Remove bespoke code in DoCopy and RI_Initial_Check, which now instead
fabricate call ExecCheckRTPerms with a manufactured RangeTblEntry.
This is intended to make it feasible for an enhanced security provider
to actually make use of ExecutorCheckPerms_hook, but also has the
advantage that RI_Initial_Check can allow use of the fast-path when
column-level but not table-level permissions are present.
KaiGai Kohei. Reviewed (in an earlier version) by Stephen Frost, and by me.
Some further changes to the comments by me.
from defining non-self-conflicting constraints.
Jeff Davis
Note: I (tgl) objected to removing this check in 9.0 on the grounds that it
was an important sanity check in new, poorly tested code. However, it should
be all right to remove it for 9.1, since we'll get field testing from the
9.0 branch.
relation using the general PARAM_EXEC executor parameter mechanism, rather
than the ad-hoc kluge of passing the outer tuple down through ExecReScan.
The previous method was hard to understand and could never be extended to
handle parameters coming from multiple join levels. This patch doesn't
change the set of possible plans nor have any significant performance effect,
but it's necessary infrastructure for future generalization of the concept
of an inner indexscan plan.
ExecReScan's second parameter is now unused, so it's removed.
This hook allows a loadable module to gain control when table permissions
are checked. It is expected to be used by an eventual SE-PostgreSQL
implementation, but there are other possible applications as well. A
sample contrib module can be found in the archives at:
http://archives.postgresql.org/pgsql-hackers/2010-05/msg01095.php
Robert Haas and Stephen Frost
is treated like end-of-input, if nulls sort last in that column and we are not
doing outer-join filling for that input. In such a case, the tuple cannot
join to anything from the other input (because we assume mergejoinable
operators are strict), and neither can any tuple following it in the sort
order. If we're not interested in doing outer-join filling we can just
pretend the tuple and its successors aren't there at all. This can save a
great deal of time in situations where there are many nulls in the join
column, as in a recent example from Scott Marlowe. Also, since the planner
tends to not count nulls in its mergejoin scan selectivity estimates, this
is an important fix to make the runtime behavior more like the estimate.
I regard this as an omission in the patch I wrote years ago to teach mergejoin
that tuples containing nulls aren't joinable, so I'm back-patching it. But
only to 8.3 --- in older versions, we didn't have a solid notion of whether
nulls sort high or low, so attempting to apply this optimization could break
things.
archival or hot standby should be WAL-logged, instead of deducing that from
other options like archive_mode. This replaces recovery_connections GUC in
the primary, where it now has no effect, but it's still used in the standby
to enable/disable hot standby.
Remove the WAL-logging of "unlogged operations", like creating an index
without WAL-logging and fsyncing it at the end. Instead, we keep a copy of
the wal_mode setting and the settings that affect how much shared memory a
hot standby server needs to track master transactions (max_connections,
max_prepared_xacts, max_locks_per_xact) in pg_control. Whenever the settings
change, at server restart, write a WAL record noting the new settings and
update pg_control. This allows us to notice the change in those settings in
the standby at the right moment, they used to be included in checkpoint
records, but that meant that a changed value was not reflected in the
standby until the first checkpoint after the change.
Bump PG_CONTROL_VERSION and XLOG_PAGE_MAGIC. Whack XLOG_PAGE_MAGIC back to
the sequence it used to follow, before hot standby and subsequent patches
changed it to 0x9003.
catalog entries via SearchSysCache and related operations. Although, at the
time that these callbacks are called by elog.c, we have not officially aborted
the current transaction, it still seems rather risky to initiate any new
catalog fetches. In all these cases the needed information is readily
available in the caller and so it's just a matter of a bit of extra notation
to pass it to the callback.
Per crash report from Dennis Koegel. I've concluded that the real fix for
his problem is to clear the error context stack at entry to proc_exit, but
it still seems like a good idea to make the callbacks a bit less fragile
for other cases.
Backpatch to 8.4. We could go further back, but the patch doesn't apply
cleanly. In the absence of proof that this fixes something and isn't just
paranoia, I'm not going to expend the effort.
Add some checks that seem logically necessary, in particular let's make
real sure that HS slave sessions cannot create temp tables. (If they did
they would think that temp tables belonging to the master's session with
the same BackendId were theirs. We *must* not allow myTempNamespace to
become set in a slave session.)
Change setval() and nextval() so that they are only allowed on temp sequences
in a read-only transaction. This seems consistent with what we allow for
table modifications in read-only transactions. Since an HS slave can't have a
temp sequence, this also provides a nicer cure for the setval PANIC reported
by Erik Rijkers.
Make the error messages more uniform, and have them mention the specific
command being complained of. This seems worth the trifling amount of extra
code, since people are likely to see such messages a lot more than before.
being assigned to, in case the expression to be assigned is a FieldStore that
would need to modify that value. The need for this was foreseen some time
ago, but not implemented then because we did not have arrays of composites.
Now we do, but the point evidently got overlooked in that patch. Net result
is that updating a field of an array element doesn't work right, as
illustrated if you try the new regression test on an unpatched backend.
Noted while experimenting with EXPLAIN VERBOSE, which has also got some issues
in this area.
Backpatch to 8.3, where arrays of composites were introduced.
The purpose of this change is to eliminate the need for every caller
of SearchSysCache, SearchSysCacheCopy, SearchSysCacheExists,
GetSysCacheOid, and SearchSysCacheList to know the maximum number
of allowable keys for a syscache entry (currently 4). This will
make it far easier to increase the maximum number of keys in a
future release should we choose to do so, and it makes the code
shorter, too.
Design and review by Tom Lane.
This patch allows the frame to start from CURRENT ROW (in either RANGE or
ROWS mode), and it also adds support for ROWS n PRECEDING and ROWS n FOLLOWING
start and end points. (RANGE value PRECEDING/FOLLOWING isn't there yet ---
the grammar works, but that's all.)
Hitoshi Harada, reviewed by Pavel Stehule
Move rd_targblock, rd_fsm_nblocks, and rd_vm_nblocks from relcache to the smgr
relation entries, so that they will get reset to InvalidBlockNumber whenever
an smgr-level flush happens. Because we now send smgr invalidation messages
immediately (not at end of transaction) when a relation truncation occurs,
this ensures that other backends will reset their values before they next
access the relation. We no longer need the unreliable assumption that a
VACUUM that's doing a truncation will hold its AccessExclusive lock until
commit --- in fact, we can intentionally release that lock as soon as we've
completed the truncation. This patch therefore reverts (most of) Alvaro's
patch of 2009-11-10, as well as my marginal hacking on it yesterday. We can
also get rid of assorted no-longer-needed relcache flushes, which are far more
expensive than an smgr flush because they kill a lot more state.
In passing this patch fixes smgr_redo's failure to perform visibility-map
truncation, and cleans up some rather dubious assumptions in freespace.c and
visibilitymap.c about when rd_fsm_nblocks and rd_vm_nblocks can be out of
date.
being called as aggregates, and to get the aggregate transition state memory
context if needed. Use it instead of poking directly into AggState and
WindowAggState in places that shouldn't know so much.
We should have done this in 8.4, probably, but better late than never.
Revised version of a patch by Hitoshi Harada.
VACUUM FULL INPLACE), along with a boatload of subsidiary code and complexity.
Per discussion, the use case for this method of vacuuming is no longer large
enough to justify maintaining it; not to mention that we don't wish to invest
the work that would be needed to make it play nicely with Hot Standby.
Aside from the code directly related to old-style VACUUM FULL, this commit
removes support for certain WAL record types that could only be generated
within VACUUM FULL, redirect-pointer removal in heap_page_prune, and
nontransactional generation of cache invalidation sinval messages (the last
being the sticking point for Hot Standby).
We still have to retain all code that copes with finding HEAP_MOVED_OFF and
HEAP_MOVED_IN flag bits on existing tuples. This can't be removed as long
as we want to support in-place update from pre-9.0 databases.
of shared or nailed system catalogs. This has two key benefits:
* The new CLUSTER-based VACUUM FULL can be applied safely to all catalogs.
* We no longer have to use an unsafe reindex-in-place approach for reindexing
shared catalogs.
CLUSTER on nailed catalogs now works too, although I left it disabled on
shared catalogs because the resulting pg_index.indisclustered update would
only be visible in one database.
Since reindexing shared system catalogs is now fully transactional and
crash-safe, the former special cases in REINDEX behavior have been removed;
shared catalogs are treated the same as non-shared.
This commit does not do anything about the recently-discussed problem of
deadlocks between VACUUM FULL/CLUSTER on a system catalog and other
concurrent queries; will address that in a separate patch. As a stopgap,
parallel_schedule has been tweaked to run vacuum.sql by itself, to avoid
such failures during the regression tests.
heap_sync() to the callers, because heap_sync() is sometimes called even
if the operation itself is WAL-logged. This eliminates the bogus unlogged
records from CLUSTER that Simon Riggs reported, patch by Fujii Masao.
We show the number of buckets, the number of batches (and also the original
number if it has changed), and the peak space used by the hash table. Minor
executor changes to track peak space used.
We need to free the OID list returned by ExecInsertIndexTuples to avoid
a query-lifespan memory leak. When many rows require rechecking, this
can be a significant leak --- it's even more than the space used for the
queued trigger events.
Dean Rasheed
