39dcfda2d fixed an incorrect reuse of ResultRelInfos for AFTER triggers
when the ResultRelInfo needed to have a different ri_RootResultRelInfo.
That caused an issue in logical replication apply workers as
finish_edata() neglects to call ExecCloseResultRelations() and instead
relies on ExecCleanupTupleRouting() to close relations opened during
partitioning's tuple routing. Since 39dcfda2d, because we may have done
some additional table_opens() calls due to having to create an additional
ResultRelInfo because of requirements to have a different
ri_RootResultRelInfo, we should now be explicitly closing any relations
opened on ResultRelInfos in EState's es_trig_target_relations.
Since finish_edate() seems to want to avoid calling
ExecCloseResultRelations(), add a new external function named
ExecCloseTrigTargetRelations().
Reported-by: Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com>
Author: Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com>
Author: David Rowley <dgrowleyml@gmail.com>
Reviewed-by: Zhijie Hou (Fujitsu) <houzj.fnst@fujitsu.com>
Discussion: https://postgr.es/m/OS9PR01MB121491E7E05950D108AF9A6D8F5A72@OS9PR01MB12149.jpnprd01.prod.outlook.com
Backpatch-through: 15
This reverts commit a0b6ef29a5, along with
its follow-up 2e123e3c2b ("Silence compiler
warning from older compilers"), which only adjusted code introduced by
the former.
The change failed with an empty table and an invalid default, and the
best way to deal with that will involve an addition to the TAM API, so
it's not ready for relese 19 now.
Discussion: https://postgr.es/m/7033D663-DDB4-4B35-922C-F33DE53B1502@gmail.com
Make sure that function declarations use names that exactly match the
corresponding names from function definitions in a few places. Most of
these inconsistencies were introduced during Postgres 19 development.
This commit was written with help from clang-tidy, by mechanically
applying the same rules as similar clean-up commits (the earliest such
commit was commit 035ce1fe).
This moves the implementation of ExecProcNodeInstr, the ExecProcNode variant
that gets used when instrumentation is on, to be defined in instrument.c
instead of execProcNode.c, and marks functions it uses as inline.
This allows compilers to generate an optimized implementation, and shows a 4
to 12% reduction in instrumentation overhead for queries that move lots of
rows.
Author: Lukas Fittl <lukas@fittl.com>
Suggested-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CAP53PkzdBK8VJ1fS4AZ481LgMN8f9mJiC39ZRHqkFUSYq6KWmg@mail.gmail.com
In a plain join, we can just summarily discard an input tuple
with null join key(s), since it cannot match anything from
the other side of the join (assuming a strict join operator).
However, if the tuple comes from the outer side of an outer join
then we have to emit it with null-extension of the other side.
Up to now, hash joins did that by inserting the tuple into the hash
table as though it were a normal tuple. This is unnecessarily
inefficient though, since the required processing is far simpler than
for a potentially-matchable tuple. Worse, if there are a lot of such
tuples they will bloat the hash bucket they go into, possibly causing
useless repeated attempts to split that bucket or increase the number
of batches. We have a report of a large join vainly creating many
thousands of batches when faced with such input.
This patch improves the situation by keeping such tuples out of the
hash table altogether, instead pushing them into a separate tuplestore
from which we return them later. (One might consider trying to return
them immediately; but that would require substantial refactoring, and
it doesn't work anyway for cases where we rescan an unmodified hash
table.) This works even in parallel hash joins, because whichever
worker reads a null-keyed tuple can just return it; there's no need
for consultation with other workers. Thus the tuplestores are local
storage even in a parallel join.
A pre-existing buglet that I noticed while analyzing the code's
behavior is that ExecHashRemoveNextSkewBucket fails to decrement
hashtable->skewTuples for tuples moved into the main hash table
from the skew hash table. This invalidates ExecHashTableInsert's
calculation of the number of main-hash-table tuples, though probably
not by a lot since we expect the skew table to be small relative
to the main one. Nonetheless, let's fix that too while we're here.
Bug: #18909
Reported-by: Sergey Koposov <Sergey.Koposov@ed.ac.uk>
Author: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Chao Li <li.evan.chao@gmail.com>
Discussion: https://postgr.es/m/3061845.1746486714@sss.pgh.pa.us
Remove a bunch of #include lines from execnodes.h. Most of these
requier suitable typedefs to be added, so that it still compiles
standalone. In one case, the fix is to move a struct definition to the
one .c file where it is needed.
Also some light clean up in plannodes.h and genam.h, though not as
extensive as in execnodes.h.
Author: Álvaro Herrera <alvherre@kurilemu.de>
Author: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/202603131240.ihwqdxnj7w2o@alvherre.pgsql
This commit includes various optimizations to improve the performance of
tuple deformation.
We now precalculate CompactAttribute's attcacheoff, which allows us to
remove the code from the deform routines which was setting the
attcacheoff. Setting the attcacheoff is now handled by
TupleDescFinalize(), which must be called before the TupleDesc is used for
anything. Having TupleDescFinalize() means we can store the first
attribute in the TupleDesc which does not have an offset cached. That
allows us to add a dedicated deforming loop to deform all attributes up
to the final one with an attcacheoff set, or up to the first NULL
attribute, whichever comes first.
Here we also improve tuple deformation performance of tuples with NULLs.
Previously, if the HEAP_HASNULL bit was set in the tuple's t_infomask,
deforming would, one-by-one, check each and every bit in the NULL bitmap
to see if it was zero. Now, we process the NULL bitmap 1 byte at a time
rather than 1 bit at a time to find the attnum with the first NULL. We
can now deform the tuple without checking for NULLs up to just before that
attribute.
We also record the maximum attribute number which is guaranteed to exist
in the tuple, that is, has a NOT NULL constraint and isn't an
atthasmissing attribute. When deforming only attributes prior to the
guaranteed attnum, we've no need to access the tuple's natt count. As an
additional optimization, we only count fixed-width columns when
calculating the maximum guaranteed column, as this eliminates the need to
emit code to fetch byref types in the deformation loop for guaranteed
attributes.
Some locations in the code deform tuples that have yet to go through NOT
NULL constraint validation. We're unable to perform the guaranteed
attribute optimization when that's the case. This optimization is opt-in
via the TupleTableSlot using the TTS_FLAG_OBEYS_NOT_NULL_CONSTRAINTS
flag.
This commit also adds a more efficient way of populating the isnull
array by using a bit-wise SWAR trick which performs multiplication on the
inverse of the tuple's bitmap byte and masking out all but the lower bit
of each of the boolean's byte. This results in much more optimal code
when compared to determining the NULLness via att_isnull(). 8 isnull
elements are processed at once using this method, which means we need to
round the tts_isnull array size up to the next 8 bytes. The palloc code
does this anyway, but the round-up needed to be formalized so as not to
overwrite the sentinel byte in MEMORY_CONTEXT_CHECKING builds. Doing
this also allows the NULL-checking deforming loop to more efficiently
check the isnull array, rather than doing the bit-wise processing for each
attribute that att_isnull() does.
The level of performance improvement from these changes seems to vary
depending on the CPU architecture. Apple's M chips seem particularly
fond of the changes, with some of the tested deform-heavy queries going
over twice as fast as before. With x86-64, the speedups aren't quite as
large. With tables containing only a small number of columns, the
speedups will be less.
Author: David Rowley <dgrowleyml@gmail.com>
Reviewed-by: Chao Li <li.evan.chao@gmail.com>
Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by: John Naylor <johncnaylorls@gmail.com>
Reviewed-by: Amit Langote <amitlangote09@gmail.com>
Reviewed-by: Zsolt Parragi <zsolt.parragi@percona.com>
Reviewed-by: Álvaro Herrera <alvherre@kurilemu.de>
Reviewed-by: Junwang Zhao <zhjwpku@gmail.com>
Discussion: https://postgr.es/m/CAApHDvpoFjaj3%2Bw_jD5uPnGazaw41A71tVJokLDJg2zfcigpMQ%40mail.gmail.com
Previously, ALTER TABLE ADD COLUMN always forced a table rewrite when
the column type was a domain with constraints (CHECK or NOT NULL), even
if the default value satisfied those constraints. This was because
contain_volatile_functions() considers CoerceToDomain immutable, so
the code conservatively assumed any constrained domain might fail.
Improve this by using soft error handling (ErrorSaveContext) to evaluate
the CoerceToDomain expression at ALTER TABLE time. If the default value
passes the domain's constraints, the value is stored as a "missing"
attribute default and no table rewrite is needed. If the constraint
check fails, we fall back to a table rewrite, preserving the historical
behavior that constraint violations are only raised when the table
actually contains rows.
Domains with volatile constraint expressions always require a table
rewrite since the constraint result could differ per evaluation and
cannot be cached.
Author: Jian He <jian.universality@gmail.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Andrew Dunstan <andrew@dunslane.net>
Reviewed-by: Viktor Holmberg <viktor.holmberg@aiven.io>
Discussion: https://postgr.es/m/CACJufxE_+iZBR1i49k_AHigppPwLTJi6km8NOsC7FWvKdEmmXg@mail.gmail.com
We've been nibbling away at removing uses of "long" for a long time,
since its width is platform-dependent. Here's one more: change the
remaining "long" fields in Plan nodes to Cardinality, since the three
surviving examples all represent group-count estimates. The upstream
planner code was converted to Cardinality some time ago; for example
the corresponding fields in Path nodes are type Cardinality, as are
the arguments of the make_foo_path functions. Downstream in the
executor, it turns out that these all feed to the table-size argument
of BuildTupleHashTable. Change that to "double" as well, and fix it
so that it safely clamps out-of-range values to the uint32 limit of
simplehash.h, as was not being done before.
Essentially, this is removing all the artificial datatype-dependent
limitations on these values from upstream processing, and applying
just one clamp at the moment where we're forced to do so by the
datatype choices of simplehash.h.
Also, remove BuildTupleHashTable's misguided attempt to enforce
work_mem/hash_mem_limit. It doesn't have enough information
(particularly not the expected tuple width) to do that accurately,
and it has no real business second-guessing the caller's choice.
For all these plan types, it's really the planner's responsibility
to not choose a hashed implementation if the hashtable is expected
to exceed hash_mem_limit. The previous patch improved the
accuracy of those estimates, and even if BuildTupleHashTable had
more information it should arrive at the same conclusions.
Reported-by: Jeff Janes <jeff.janes@gmail.com>
Author: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/CAMkU=1zia0JfW_QR8L5xA2vpa0oqVuiapm78h=WpNsHH13_9uw@mail.gmail.com
For several types of plan nodes that use TupleHashTables, the
planner estimated the expected size of the table as basically
numEntries * (MAXALIGN(dataWidth) + MAXALIGN(SizeofHeapTupleHeader)).
This is pretty far off, especially for small data widths, because
it doesn't account for the overhead of the simplehash.h hash table
nor for any per-tuple "additional space" the plan node may request.
Jeff Janes noted a case where the estimate was off by about a factor
of three, even though the obvious hazards such as inaccurate estimates
of numEntries or dataWidth didn't apply.
To improve matters, create functions provided by the relevant executor
modules that can estimate the required sizes with reasonable accuracy.
(We're still not accounting for effects like allocator padding, but
this at least gets the first-order effects correct.)
I added functions that can estimate the tuple table sizes for
nodeSetOp and nodeSubplan; these rely on an estimator for
TupleHashTables in general, and that in turn relies on one for
simplehash.h hash tables. That feels like kind of a lot of mechanism,
but if we take any short-cuts we're violating modularity boundaries.
The other places that use TupleHashTables are nodeAgg, which took
pains to get its numbers right already, and nodeRecursiveunion.
I did not try to improve the situation for nodeRecursiveunion because
there's nothing to improve: we are not making an estimate of the hash
table size, and it wouldn't help us to do so because we have no
non-hashed alternative implementation. On top of that, our estimate
of the number of entries to be hashed in that module is so suspect
that we'd likely often choose the wrong implementation if we did have
two ways to do it.
Reported-by: Jeff Janes <jeff.janes@gmail.com>
Author: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/CAMkU=1zia0JfW_QR8L5xA2vpa0oqVuiapm78h=WpNsHH13_9uw@mail.gmail.com
For all extant uses of TupleHashTables, execGrouping.c itself does
nothing with the "tablecxt" except to allocate new hash entries in it,
and the callers do nothing with it except to reset the whole context.
So this is an ideal use-case for a BumpContext, and the hash tables
are frequently big enough for the savings to be significant.
(Commit cc721c459 already taught nodeAgg.c this idea, but neglected
the other callers of BuildTupleHashTable.)
While at it, let's clean up some ill-advised leftovers from rebasing
TupleHashTables on simplehash.h:
* Many comments and variable names were based on the idea that the
tablecxt holds the whole TupleHashTable, whereas now it only holds the
hashed tuples (plus any caller-defined "additional storage"). Rename
to names like tuplescxt and tuplesContext, and adjust the comments.
Also adjust the memory context names to be like "<Foo> hashed tuples".
* Make ResetTupleHashTable() reset the tuplescxt rather than relying
on the caller to do so; that was fairly bizarre and seems like a
recipe for leaks. This is less efficient in the case where nodeAgg.c
uses the same tuplescxt for several different hashtables, but only
microscopically so because mcxt.c will short-circuit the extra resets
via its isReset flag. I judge the extra safety and intellectual
cleanliness well worth those few cycles.
* Remove the long-obsolete "allow_jit" check added by ac88807f9;
instead, just Assert that metacxt and tuplescxt are different.
We need that anyway for this definition of ResetTupleHashTable() to
be safe.
There is a side issue of the extent to which this change invalidates
the planner's estimates of hashtable memory consumption. However,
those estimates are already pretty bad, so improving them seems like
it can be a separate project. This change is useful to do first to
establish consistent executor behavior that the planner can expect.
A loose end not addressed here is that the "entrysize" calculation
in BuildTupleHashTable seems wrong: "sizeof(TupleHashEntryData) +
additionalsize" corresponds neither to the size of the simplehash
entries nor to the total space needed per tuple. It's questionable
why BuildTupleHashTable is second-guessing its caller's nbuckets
choice at all, since the original source of the number should have had
more information. But that all seems wrapped up with the planner's
estimation logic, so let's leave it for the planned followup patch.
Reported-by: Jeff Janes <jeff.janes@gmail.com>
Reported-by: David Rowley <dgrowleyml@gmail.com>
Author: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: David Rowley <dgrowleyml@gmail.com>
Discussion: https://postgr.es/m/CAMkU=1zia0JfW_QR8L5xA2vpa0oqVuiapm78h=WpNsHH13_9uw@mail.gmail.com
Discussion: https://postgr.es/m/2268409.1761512111@sss.pgh.pa.us
This patch adds support for a new SQL command:
ALTER SUBSCRIPTION ... REFRESH SEQUENCES
This command updates the sequence entries present in the
pg_subscription_rel catalog table with the INIT state to trigger
resynchronization.
In addition to the new command, the following subscription commands have
been enhanced to automatically refresh sequence mappings:
ALTER SUBSCRIPTION ... REFRESH PUBLICATION
ALTER SUBSCRIPTION ... ADD PUBLICATION
ALTER SUBSCRIPTION ... DROP PUBLICATION
ALTER SUBSCRIPTION ... SET PUBLICATION
These commands will perform the following actions:
Add newly published sequences that are not yet part of the subscription.
Remove sequences that are no longer included in the publication.
This ensures that sequence replication remains aligned with the current
state of the publication on the publisher side.
Note that the actual synchronization of sequence data/values will be
handled in a subsequent patch that introduces a dedicated sequence sync
worker.
Author: Vignesh C <vignesh21@gmail.com>
Reviewed-by: Amit Kapila <amit.kapila16@gmail.com>
Reviewed-by: shveta malik <shveta.malik@gmail.com>
Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com>
Reviewed-by: Hayato Kuroda <kuroda.hayato@fujitsu.com>
Reviewed-by: Dilip Kumar <dilipbalaut@gmail.com>
Reviewed-by: Peter Smith <smithpb2250@gmail.com>
Reviewed-by: Nisha Moond <nisha.moond412@gmail.com>
Reviewed-by: Shlok Kyal <shlok.kyal.oss@gmail.com>
Reviewed-by: Chao Li <li.evan.chao@gmail.com>
Reviewed-by: Hou Zhijie <houzj.fnst@fujitsu.com>
Discussion: https://postgr.es/m/CAA4eK1LC+KJiAkSrpE_NwvNdidw9F2os7GERUeSxSKv71gXysQ@mail.gmail.com
If an INSERT has an ON CONFLICT DO UPDATE clause, the executor must
check that the target relation supports UPDATE as well as INSERT. In
particular, it must check that the target relation has a REPLICA
IDENTITY if it publishes updates. Formerly, it was not doing this
check, which could lead to silently breaking replication.
Fix by adding such a check to CheckValidResultRel(), which requires
adding a new onConflictAction argument. In back-branches, preserve ABI
compatibility by introducing a wrapper function with the original
signature.
Author: Zhijie Hou <houzj.fnst@fujitsu.com>
Reviewed-by: Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Tested-by: Chao Li <li.evan.chao@gmail.com>
Discussion: https://postgr.es/m/OS3PR01MB57180C87E43A679A730482DF94B62@OS3PR01MB5718.jpnprd01.prod.outlook.com
Backpatch-through: 13
Commit e2d4ef8de8 (the fix for CVE-2017-7484) added security checks
to the selectivity estimation functions to prevent them from running
user-supplied operators on data obtained from pg_statistic if the user
lacks privileges to select from the underlying table. In cases
involving inheritance/partitioning, those checks were originally
performed against the child RTE (which for plain inheritance might
actually refer to the parent table). Commit 553d2ec271 then extended
that to also check the parent RTE, allowing access if the user had
permissions on either the parent or the child. It turns out, however,
that doing any checks using the child RTE is incorrect, since
securityQuals is set to NULL when creating an RTE for an inheritance
child (whether it refers to the parent table or the child table), and
therefore such checks do not correctly account for any RLS policies or
security barrier views. Therefore, do the security checks using only
the parent RTE. This is consistent with how RLS policies are applied,
and the executor's ACL checks, both of which use only the parent
table's permissions/policies. Similar checks are performed in the
extended stats code, so update that in the same way, centralizing all
the checks in a new function.
In addition, note that these checks by themselves are insufficient to
ensure that the user has access to the table's data because, in a
query that goes via a view, they only check that the view owner has
permissions on the underlying table, not that the current user has
permissions on the view itself. In the selectivity estimation
functions, there is no easy way to navigate from underlying tables to
views, so add permissions checks for all views mentioned in the query
to the planner startup code. If the user lacks permissions on a view,
a permissions error will now be reported at planner-startup, and the
selectivity estimation functions will not be run.
Checking view permissions at planner-startup in this way is a little
ugly, since the same checks will be repeated at executor-startup.
Longer-term, it might be better to move all the permissions checks
from the executor to the planner so that permissions errors can be
reported sooner, instead of creating a plan that won't ever be run.
However, such a change seems too far-reaching to be back-patched.
Back-patch to all supported versions. In v13, there is the added
complication that UPDATEs and DELETEs on inherited target tables are
planned using inheritance_planner(), which plans each inheritance
child table separately, so that the selectivity estimation functions
do not know that they are dealing with a child table accessed via its
parent. Handle that by checking access permissions on the top parent
table at planner-startup, in the same way as we do for views. Any
securityQuals on the top parent table are moved down to the child
tables by inheritance_planner(), so they continue to be checked by the
selectivity estimation functions.
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Noah Misch <noah@leadboat.com>
Backpatch-through: 13
Security: CVE-2025-8713
This enhancement builds upon the infrastructure introduced in commit
228c370868, which enables the preservation of deleted tuples and their
origin information on the subscriber. This capability is crucial for
handling concurrent transactions replicated from remote nodes.
The update introduces support for detecting update_deleted conflicts
during the application of update operations on the subscriber. When an
update operation fails to locate the target row-typically because it has
been concurrently deleted-we perform an additional table scan. This scan
uses the SnapshotAny mechanism and we do this additional scan only when
the retain_dead_tuples option is enabled for the relevant subscription.
The goal of this scan is to locate the most recently deleted tuple-matching
the old column values from the remote update-that has not yet been removed
by VACUUM and is still visible according to our slot (i.e., its deletion
is not older than conflict-detection-slot's xmin). If such a tuple is
found, the system reports an update_deleted conflict, including the origin
and transaction details responsible for the deletion.
This provides a groundwork for more robust and accurate conflict
resolution process, preventing unexpected behavior by correctly
identifying cases where a remote update clashes with a deletion from
another origin.
Author: Zhijie Hou <houzj.fnst@fujitsu.com>
Reviewed-by: shveta malik <shveta.malik@gmail.com>
Reviewed-by: Nisha Moond <nisha.moond412@gmail.com>
Reviewed-by: Dilip Kumar <dilipbalaut@gmail.com>
Reviewed-by: Hayato Kuroda <kuroda.hayato@fujitsu.com>
Reviewed-by: Amit Kapila <amit.kapila16@gmail.com>
Discussion: https://postgr.es/m/OS0PR01MB5716BE80DAEB0EE2A6A5D1F5949D2@OS0PR01MB5716.jpnprd01.prod.outlook.com
As pointed out by Tom Lane, the patch introduced fragile and invasive
design around plan invalidation handling when locking of prunable
partitions was deferred from plancache.c to the executor. In
particular, it violated assumptions about CachedPlan immutability and
altered executor APIs in ways that are difficult to justify given the
added complexity and overhead.
This also removes the firstResultRels field added to PlannedStmt in
commit 28317de72, which was intended to support deferred locking of
certain ModifyTable result relations.
Reported-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/605328.1747710381@sss.pgh.pa.us
Commit cbc127917e introduced tracking of unpruned relids to avoid
processing pruned relations, and changed ExecInitModifyTable() to
initialize only unpruned result relations. As a result, MERGE
statements that prune all target partitions can now lead to crashes
or incorrect behavior during execution.
The crash occurs because some executor code paths rely on
ModifyTableState.resultRelInfo[0] being present and initialized,
even when no result relations remain after pruning. For example,
ExecMerge() and ExecMergeNotMatched() use the first resultRelInfo
to determine the appropriate action. Similarly,
ExecInitPartitionInfo() assumes that at least one result relation
exists.
To preserve these assumptions, ExecInitModifyTable() now includes the
first result relation in the initialized result relation list if all
result relations for that ModifyTable were pruned. To enable that,
ExecDoInitialPruning() ensures the first relation is locked if it was
pruned and locking is necessary.
To support this exception to the pruning logic, PlannedStmt now
includes a list of RT indexes identifying the first result relation
of each ModifyTable node in the plan. This allows
ExecDoInitialPruning() to check whether each such relation was
pruned and, if so, lock it if necessary.
Bug: #18830
Reported-by: Robins Tharakan <tharakan@gmail.com>
Diagnozed-by: Tender Wang <tndrwang@gmail.com>
Diagnozed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Co-authored-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Reviewed-by: Tender Wang <tndrwang@gmail.com>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Discussion: https://postgr.es/m/18830-1f31ea1dc930d444%40postgresql.org
Before executing a cached generic plan, AcquireExecutorLocks() in
plancache.c locks all relations in a plan's range table to ensure the
plan is safe for execution. However, this locks runtime-prunable
relations that will later be pruned during "initial" runtime pruning,
introducing unnecessary overhead.
This commit defers locking for such relations to executor startup and
ensures that if the CachedPlan is invalidated due to concurrent DDL
during this window, replanning is triggered. Deferring these locks
avoids unnecessary locking overhead for pruned partitions, resulting
in significant speedup, particularly when many partitions are pruned
during initial runtime pruning.
* Changes to locking when executing generic plans:
AcquireExecutorLocks() now locks only unprunable relations, that is,
those found in PlannedStmt.unprunableRelids (introduced in commit
cbc127917e), to avoid locking runtime-prunable partitions
unnecessarily. The remaining locks are taken by
ExecDoInitialPruning(), which acquires them only for partitions that
survive pruning.
This deferral does not affect the locks required for permission
checking in InitPlan(), which takes place before initial pruning.
ExecCheckPermissions() now includes an Assert to verify that all
relations undergoing permission checks, none of which can be in the
set of runtime-prunable relations, are properly locked.
* Plan invalidation handling:
Deferring locks introduces a window where prunable relations may be
altered by concurrent DDL, invalidating the plan. A new function,
ExecutorStartCachedPlan(), wraps ExecutorStart() to detect and handle
invalidation caused by deferred locking. If invalidation occurs,
ExecutorStartCachedPlan() updates CachedPlan using the new
UpdateCachedPlan() function and retries execution with the updated
plan. To ensure all code paths that may be affected by this handle
invalidation properly, all callers of ExecutorStart that may execute a
PlannedStmt from a CachedPlan have been updated to use
ExecutorStartCachedPlan() instead.
UpdateCachedPlan() replaces stale plans in CachedPlan.stmt_list. A new
CachedPlan.stmt_context, created as a child of CachedPlan.context,
allows freeing old PlannedStmts while preserving the CachedPlan
structure and its statement list. This ensures that loops over
statements in upstream callers of ExecutorStartCachedPlan() remain
intact.
ExecutorStart() and ExecutorStart_hook implementations now return a
boolean value indicating whether plan initialization succeeded with a
valid PlanState tree in QueryDesc.planstate, or false otherwise, in
which case QueryDesc.planstate is NULL. Hook implementations are
required to call standard_ExecutorStart() at the beginning, and if it
returns false, they should do the same without proceeding.
* Testing:
To verify these changes, the delay_execution module tests scenarios
where cached plans become invalid due to changes in prunable relations
after deferred locks.
* Note to extension authors:
ExecutorStart_hook implementations must verify plan validity after
calling standard_ExecutorStart(), as explained earlier. For example:
if (prev_ExecutorStart)
plan_valid = prev_ExecutorStart(queryDesc, eflags);
else
plan_valid = standard_ExecutorStart(queryDesc, eflags);
if (!plan_valid)
return false;
<extension-code>
return true;
Extensions accessing child relations, especially prunable partitions,
via ExecGetRangeTableRelation() must now ensure their RT indexes are
present in es_unpruned_relids (introduced in commit cbc127917e), or
they will encounter an error. This is a strict requirement after this
change, as only relations in that set are locked.
The idea of deferring some locks to executor startup, allowing locks
for prunable partitions to be skipped, was first proposed by Tom Lane.
Reviewed-by: Robert Haas <robertmhaas@gmail.com> (earlier versions)
Reviewed-by: David Rowley <dgrowleyml@gmail.com> (earlier versions)
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us> (earlier versions)
Reviewed-by: Tomas Vondra <tomas@vondra.me>
Reviewed-by: Junwang Zhao <zhjwpku@gmail.com>
Discussion: https://postgr.es/m/CA+HiwqFGkMSge6TgC9KQzde0ohpAycLQuV7ooitEEpbKB0O_mg@mail.gmail.com
This commit introduces changes to track unpruned relations explicitly,
making it possible for top-level plan nodes, such as ModifyTable and
LockRows, to avoid processing partitions pruned during initial
pruning. Scan-level nodes, such as Append and MergeAppend, already
avoid the unnecessary processing by accessing partition pruning
results directly via part_prune_index. In contrast, top-level nodes
cannot access pruning results directly and need to determine which
partitions remain unpruned.
To address this, this commit introduces a new bitmapset field,
es_unpruned_relids, which the executor uses to track the set of
unpruned relations. This field is referenced during plan
initialization to skip initializing certain nodes for pruned
partitions. It is initialized with PlannedStmt.unprunableRelids,
a new field that the planner populates with RT indexes of relations
that cannot be pruned during runtime pruning. These include relations
not subject to partition pruning and those required for execution
regardless of pruning.
PlannedStmt.unprunableRelids is computed during set_plan_refs() by
removing the RT indexes of runtime-prunable relations, identified
from PartitionPruneInfos, from the full set of relation RT indexes.
ExecDoInitialPruning() then updates es_unpruned_relids by adding
partitions that survive initial pruning.
To support this, PartitionedRelPruneInfo and PartitionedRelPruningData
now include a leafpart_rti_map[] array that maps partition indexes to
their corresponding RT indexes. The former is used in set_plan_refs()
when constructing unprunableRelids, while the latter is used in
ExecDoInitialPruning() to convert partition indexes returned by
get_matching_partitions() into RT indexes, which are then added to
es_unpruned_relids.
These changes make it possible for ModifyTable and LockRows nodes to
process only relations that remain unpruned after initial pruning.
ExecInitModifyTable() trims lists, such as resultRelations,
withCheckOptionLists, returningLists, and updateColnosLists, to
consider only unpruned partitions. It also creates ResultRelInfo
structs only for these partitions. Similarly, child RowMarks for
pruned relations are skipped.
By avoiding unnecessary initialization of structures for pruned
partitions, these changes improve the performance of updates and
deletes on partitioned tables during initial runtime pruning.
Due to ExecInitModifyTable() changes as described above, EXPLAIN on a
plan for UPDATE and DELETE that uses runtime initial pruning no longer
lists partitions pruned during initial pruning.
Reviewed-by: Robert Haas <robertmhaas@gmail.com> (earlier versions)
Reviewed-by: Tomas Vondra <tomas@vondra.me>
Discussion: https://postgr.es/m/CA+HiwqFGkMSge6TgC9KQzde0ohpAycLQuV7ooitEEpbKB0O_mg@mail.gmail.com
This turns GistTranslateCompareType() into a callback function of the
gist index AM instead of a standalone function. The existing callers
are changed to use IndexAmTranslateCompareType(). This then makes
that code not hardcoded toward gist.
This means in particular that the temporal keys code is now
independent of gist. Also, this generalizes commit 74edabce7a, so
other index access methods other than the previously hardcoded ones
could now work as REPLICA IDENTITY in a logical replication
subscriber.
Author: Mark Dilger <mark.dilger@enterprisedb.com>
Co-authored-by: Peter Eisentraut <peter@eisentraut.org>
Discussion: https://www.postgresql.org/message-id/flat/E72EAA49-354D-4C2E-8EB9-255197F55330@enterprisedb.com
This allows the RETURNING list of INSERT/UPDATE/DELETE/MERGE queries
to explicitly return old and new values by using the special aliases
"old" and "new", which are automatically added to the query (if not
already defined) while parsing its RETURNING list, allowing things
like:
RETURNING old.colname, new.colname, ...
RETURNING old.*, new.*
Additionally, a new syntax is supported, allowing the names "old" and
"new" to be changed to user-supplied alias names, e.g.:
RETURNING WITH (OLD AS o, NEW AS n) o.colname, n.colname, ...
This is useful when the names "old" and "new" are already defined,
such as inside trigger functions, allowing backwards compatibility to
be maintained -- the interpretation of any existing queries that
happen to already refer to relations called "old" or "new", or use
those as aliases for other relations, is not changed.
For an INSERT, old values will generally be NULL, and for a DELETE,
new values will generally be NULL, but that may change for an INSERT
with an ON CONFLICT ... DO UPDATE clause, or if a query rewrite rule
changes the command type. Therefore, we put no restrictions on the use
of old and new in any DML queries.
Dean Rasheed, reviewed by Jian He and Jeff Davis.
Discussion: https://postgr.es/m/CAEZATCWx0J0-v=Qjc6gXzR=KtsdvAE7Ow=D=mu50AgOe+pvisQ@mail.gmail.com
Previously, the caller needed to allocate the memory and the
TupleHashTable would store a pointer to it. That wastes space for the
palloc overhead as well as the size of the pointer itself.
Now, the TupleHashTable relies on the caller to correctly specify the
additionalsize, and allocates that amount of space. The caller can
then request a pointer into that space.
Discussion: https://postgr.es/m/b9cbf0219a9859dc8d240311643ff4362fd9602c.camel@j-davis.com
Reviewed-by: Heikki Linnakangas
It was reasonable to preserve the old API of BuildTupleHashTable()
in the back branches, but in HEAD we should actively discourage use
of that version. There are no remaining callers in core, so just
get rid of it. Then rename BuildTupleHashTableExt() back to
BuildTupleHashTable().
While at it, fix up the miserably-poorly-maintained header comment
for BuildTupleHashTable[Ext]. It looks like more than one patch in
this area has had the opinion that updating comments is beneath them.
Discussion: https://postgr.es/m/538343.1734646986@sss.pgh.pa.us
The original design for set operations involved appending the two
input relations into one and adding a flag column that allows
distinguishing which side each row came from. Then the SetOp node
pries them apart again based on the flag. This is bizarre. The
only apparent reason to do it is that when sorting, we'd only need
one Sort node not two. But since sorting is at least O(N log N),
sorting all the data is actually worse than sorting each side
separately --- plus, we have no chance of taking advantage of
presorted input. On top of that, adding the flag column frequently
requires an additional projection step that adds cycles, and then
the Append node isn't free either. Let's get rid of all of that
and make the SetOp node have two separate children, using the
existing outerPlan/innerPlan infrastructure.
This initial patch re-implements nodeSetop.c and does a bare minimum
of work on the planner side to generate correctly-shaped plans.
In particular, I've tried not to change the cost estimates here,
so that the visible changes in the regression test results will only
involve removal of useless projection steps and not any changes in
whether to use sorted vs hashed mode.
For SORTED mode, we combine successive identical tuples from each
input into groups, and then merge-join the groups. The tuple
comparisons now use SortSupport instead of simple equality, but
the group-formation part should involve roughly the same number of
tuple comparisons as before. The cross-comparisons between left and
right groups probably add to that, but I'm not sure to quantify how
many more comparisons we might need.
For HASHED mode, nodeSetop's logic is almost the same as before,
just refactored into two separate loops instead of one loop that
has an assumption that it will see all the left-hand inputs first.
In both modes, I added early-exit logic to not bother reading the
right-hand relation if the left-hand input is empty, since neither
INTERSECT nor EXCEPT modes can produce any output if the left input
is empty. This could have been done before in the hashed mode, but
not in sorted mode. Sorted mode can also stop as soon as it exhausts
the left input; any remaining right-hand tuples cannot have matches.
Also, this patch adds some infrastructure for detecting whether
child plan nodes all output the same type of tuple table slot.
If they do, the hash table logic can use slightly more efficient
code based on assuming that that's the input slot type it will see.
We'll make use of that infrastructure in other plan node types later.
Patch by me; thanks to Richard Guo and David Rowley for review.
Discussion: https://postgr.es/m/1850138.1731549611@sss.pgh.pa.us
0f5738202 adjusted the execGrouping.c code so it made use of ExprStates to
generate hash values. That commit made a wrong assumption that the slot
type to pass to ExecBuildHash32FromAttrs() is always &TTSOpsMinimalTuple.
That's not the case as the slot type depends on the slot type passed to
LookupTupleHashEntry(), which for nodeRecursiveunion.c, could be any of
the current slot types.
Here we fix this by adding a new parameter to BuildTupleHashTableExt()
to allow the slot type to be passed in. In the case of nodeSubplan.c
and nodeAgg.c the slot type is always &TTSOpsVirtual, so for both of
those cases, it's beneficial to pass the known slot type as that allows
ExecBuildHash32FromAttrs() to skip adding the tuple deform step to the
resulting ExprState. Another possible fix would have been to have
ExecBuildHash32FromAttrs() set "fetch.kind" to NULL so that
ExecComputeSlotInfo() always determines the EEOP_INNER_FETCHSOME is
required, however, that option isn't favorable as slows down aggregation
and hashed subplan evaluation due to the extra (needless) deform step.
Thanks to Nathan Bossart for bisecting to find the offending commit
based on Paul's report.
Reported-by: Paul Ramsey <pramsey@cleverelephant.ca>
Discussion: https://postgr.es/m/99F064C1-B3EB-4BE7-97D2-D2A0AA487A71@cleverelephant.ca
This speeds up obtaining hash values for GROUP BY and hashed SubPlans by
using the ExprState support for hashing, thus allowing JIT compilation for
obtaining hash values for these operations.
This, even without JIT compilation, has been shown to improve Hash
Aggregate performance in some cases by around 15% and hashed NOT IN
queries in one case by over 30%, however, real-world cases are likely to
see smaller gains as the test cases used were purposefully designed to
have high hashing overheads by keeping the hash table small to prevent
additional memory overheads that would be a factor when working with large
hash tables.
In passing, fix a hypothetical bug in ExecBuildHash32Expr() so that the
initial value is stored directly in the ExprState's result field if
there are no expressions to hash. None of the current users of this
function use an initial value, so the bug is only hypothetical.
Reviewed-by: Andrei Lepikhov <lepihov@gmail.com>
Discussion: https://postgr.es/m/CAApHDvpYSO3kc9UryMevWqthTBrxgfd9djiAjKHMPUSQeX9vdQ@mail.gmail.com
get_equal_strategy_number_for_am() gets the equal strategy number for
an AM. This currently only supports btree and hash. In the more
general case, this also depends on the operator class (see for example
GistTranslateStratnum()). To support that, replace this function with
get_equal_strategy_number() that takes an opclass and derives it from
there. (This function already existed before as a static function, so
the signature is kept for simplicity.)
This patch is only a refactoring, it doesn't add support for other
index AMs such as gist. This will be done separately.
Reviewed-by: Paul Jungwirth <pj@illuminatedcomputing.com>
Reviewed-by: vignesh C <vignesh21@gmail.com>
Discussion: https://www.postgresql.org/message-id/flat/CA+renyUApHgSZF9-nd-a0+OPGharLQLO=mDHcY4_qQ0+noCUVg@mail.gmail.com
Our parallel-mode code only works when we are executing a query
in full, so ExecutePlan must disable parallel mode when it is
asked to do partial execution. The previous logic for this
involved passing down a flag (variously named execute_once or
run_once) from callers of ExecutorRun or PortalRun. This is
overcomplicated, and unsurprisingly some of the callers didn't
get it right, since it requires keeping state that not all of
them have handy; not to mention that the requirements for it were
undocumented. That led to assertion failures in some corner
cases. The only state we really need for this is the existing
QueryDesc.already_executed flag, so let's just put all the
responsibility in ExecutePlan. (It could have been done in
ExecutorRun too, leading to a slightly shorter patch -- but if
there's ever more than one caller of ExecutePlan, it seems better
to have this logic in the subroutine than the callers.)
This makes those ExecutorRun/PortalRun parameters unnecessary.
In master it seems okay to just remove them, returning the
API for those functions to what it was before parallelism.
Such an API break is clearly not okay in stable branches,
but for them we can just leave the parameters in place after
documenting that they do nothing.
Per report from Yugo Nagata, who also reviewed and tested
this patch. Back-patch to all supported branches.
Discussion: https://postgr.es/m/20241206062549.710dc01cf91224809dd6c0e1@sraoss.co.jp
This patch provides the additional logging information in the following
conflict scenarios while applying changes:
insert_exists: Inserting a row that violates a NOT DEFERRABLE unique constraint.
update_differ: Updating a row that was previously modified by another origin.
update_exists: The updated row value violates a NOT DEFERRABLE unique constraint.
update_missing: The tuple to be updated is missing.
delete_differ: Deleting a row that was previously modified by another origin.
delete_missing: The tuple to be deleted is missing.
For insert_exists and update_exists conflicts, the log can include the origin
and commit timestamp details of the conflicting key with track_commit_timestamp
enabled.
update_differ and delete_differ conflicts can only be detected when
track_commit_timestamp is enabled on the subscriber.
We do not offer additional logging for exclusion constraint violations because
these constraints can specify rules that are more complex than simple equality
checks. Resolving such conflicts won't be straightforward. This area can be
further enhanced if required.
Author: Hou Zhijie
Reviewed-by: Shveta Malik, Amit Kapila, Nisha Moond, Hayato Kuroda, Dilip Kumar
Discussion: https://postgr.es/m/OS0PR01MB5716352552DFADB8E9AD1D8994C92@OS0PR01MB5716.jpnprd01.prod.outlook.com
Here we add ExprState support for obtaining a 32-bit hash value from a
list of expressions. This allows both faster hashing and also JIT
compilation of these expressions. This is especially useful when hash
joins have multiple join keys as the previous code called ExecEvalExpr on
each hash join key individually and that was inefficient as tuple
deformation would have only taken into account one key at a time, which
could lead to walking the tuple once for each join key. With the new
code, we'll determine the maximum attribute required and deform the tuple
to that point only once.
Some performance tests done with this change have shown up to a 20%
performance increase of a query containing a Hash Join without JIT
compilation and up to a 26% performance increase when JIT is enabled and
optimization and inlining were performed by the JIT compiler. The
performance increase with 1 join column was less with a 14% increase
with and without JIT. This test was done using a fairly small hash
table and a large number of hash probes. The increase will likely be
less with large tables, especially ones larger than L3 cache as memory
pressure is more likely to be the limiting factor there.
This commit only addresses Hash Joins, but lays expression evaluation
and JIT compilation infrastructure for other hashing needs such as Hash
Aggregate.
Author: David Rowley
Reviewed-by: Alexey Dvoichenkov <alexey@hyperplane.net>
Reviewed-by: Tels <nospam-pg-abuse@bloodgate.com>
Discussion: https://postgr.es/m/CAApHDvoexAxgQFNQD_GRkr2O_eJUD1-wUGm%3Dm0L%2BGc%3DT%3DkEa4g%40mail.gmail.com
This allows the target relation of MERGE to be an auto-updatable or
trigger-updatable view, and includes support for WITH CHECK OPTION,
security barrier views, and security invoker views.
A trigger-updatable view must have INSTEAD OF triggers for every type
of action (INSERT, UPDATE, and DELETE) mentioned in the MERGE command.
An auto-updatable view must not have any INSTEAD OF triggers. Mixing
auto-update and trigger-update actions (i.e., having a partial set of
INSTEAD OF triggers) is not supported.
Rule-updatable views are also not supported, since there is no
rewriter support for non-SELECT rules with MERGE operations.
Dean Rasheed, reviewed by Jian He and Alvaro Herrera.
Discussion: https://postgr.es/m/CAEZATCVcB1g0nmxuEc-A+gGB0HnfcGQNGYH7gS=7rq0u0zOBXA@mail.gmail.com
This commit removes unnecessary ExecExprFreeContext() calls in
ExecEnd* routines because the actual cleanup is managed by
FreeExecutorState(). With no callers remaining for
ExecExprFreeContext(), this commit also removes the function.
This commit also drops redundant ExecClearTuple() calls, because
ExecResetTupleTable() in ExecEndPlan() already takes care of
resetting and dropping all TupleTableSlots initialized with
ExecInitScanTupleSlot() and ExecInitExtraTupleSlot().
After these modifications, the ExecEnd*() routines for ValuesScan,
NamedTuplestoreScan, and WorkTableScan became redundant. So, this
commit removes them.
Reviewed-by: Robert Haas
Discussion: https://postgr.es/m/CA+HiwqFGkMSge6TgC9KQzde0ohpAycLQuV7ooitEEpbKB0O_mg@mail.gmail.com
Commit 89e46da5e5 allowed using BTREE indexes that are neither
PRIMARY KEY nor REPLICA IDENTITY on the subscriber during apply of
update/delete. This patch extends that functionality to also allow HASH
indexes.
We explored supporting other index access methods as well but they don't
have a fixed strategy for equality operation which is required by the
current infrastructure in logical replication to scan the indexes.
Author: Kuroda Hayato
Reviewed-by: Peter Smith, Onder Kalaci, Amit Kapila
Discussion: https://postgr.es/m/TYAPR01MB58669D7414E59664E17A5827F522A@TYAPR01MB5866.jpnprd01.prod.outlook.com
