Currently, test_slru's shmem_startup_hook unconditionally generates
new LWLock tranche IDs.  This is fine on non-EXEC_BACKEND builds,
where only the postmaster executes this hook, but on EXEC_BACKEND
builds, every backend executes it, too.  To fix, only generate the
tranche IDs in the postmaster process by checking the
IsUnderPostmaster variable.

This is arguably a bug fix and could be back-patched, but since the
damage is limited to some extra unused tranche IDs in a test
module, I'm not going to bother.

Reported-by: Sami Imseih <samimseih@gmail.com>
Reviewed-by: Sami Imseih <samimseih@gmail.com>
Discussion: https://postgr.es/m/CAA5RZ0vaAuonaf12CeDddQJu5xKL%2B6xVyS%2B_q1%2BcH%3D33JXV82w%40mail.gmail.com

This section claims that each backend executes the
shmem_startup_hook shortly after attaching to shared memory, which
is true for EXEC_BACKEND builds, but not for others.  This commit
adds this important detail.

Oversight in commit 964152c.

Reported-by: Sami Imseih <samimseih@gmail.com>
Reviewed-by: Sami Imseih <samimseih@gmail.com>
Discussion: https://postgr.es/m/CAA5RZ0vEGT1eigGbVt604LkXP6mUPMwPMxQoRCbFny44w%2B9EUQ%40mail.gmail.com
Backpatch-through: 17

The startup process does not process shared invalidation messages, only
sending them, and never calls AtEOXact_SMgr() which clean up any
unpinned SMgrRelations.  Hence, it is never able to free SMgrRelations
on a periodic basis, bloating its hashtable over time.

Like the checkpointer and the bgwriter, this commit takes a conservative
approach by freeing periodically SMgrRelations when replaying a
checkpoint record, either online or shutdown, so as the startup process
has a way to perform a periodic cleanup.

Issue caused by 21d9c3e, so backpatch down to v17.

Author: Jingtang Zhang <mrdrivingduck@gmail.com>
Reviewed-by: Yuhang Qiu <iamqyh@gmail.com>
Discussion: https://postgr.es/m/28C687D4-F335-417E-B06C-6612A0BD5A10@gmail.com
Backpatch-through: 17

All the calls replaced by this commit use 4-byte integers for their
variables used in input of my_log2().  Hence, the limit against
too-large inputs does not really apply.  Thresholds are also applied, as
of:
- In nodeAgg.c, the number of partitions is limited by
HASHAGG_MAX_PARTITIONS.
- In nodeHash.c, ExecChooseHashTableSize() caps its maximum number of
buckets based on HashJoinTuple and palloc() allocation limit.
- In worker.c, the number of subxacts tracked by ApplySubXactData uses
uint32, making pg_ceil_log2_64() safe to use directly.

Several approaches have been discussed, like an integration with
thresholds in pg_bitutils.h, but it was found confusing.  This uses
Dean's idea, which gives a simpler result than what I came up with to be
able to remove dynahash.h.  dynahash.h will be removed in a follow-up
commit, removing some duplication with the ceil log2 routines.

Reviewed-by: Peter Eisentraut <peter@eisentraut.org>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Discussion: https://postgr.es/m/CAEZATCUJPQD_7sC-wErak2CQGNa6bj2hY-mr8wsBki=kX7f2_A@mail.gmail.com

All the callers of my_log2() are now limited inside dynahash.c, so let's
remove this header.  The same capability is provided by pg_bitutils.h
already.

Discussion: https://postgr.es/m/CAEZATCUJPQD_7sC-wErak2CQGNa6bj2hY-mr8wsBki=kX7f2_A@mail.gmail.com

In CREATE TABLE ... LIKE, any check constraints copied from the source
table should be set to valid if they are ENFORCED (the default).

Bug introduced in commit ca87c41.

Author: jian he <jian.universality@gmail.com>
Discussion: https://www.postgresql.org/message-id/CACJufxH%3D%2Bod8Wy0P4L3_GpapNwLUP3oAes5UFRJ7yTxrM_M5kg%40mail.gmail.com

autoconf builds have compiled this file with -ftree-vectorize since
commit 8870917, but meson builds seem to have missed the memo.

Reviewed-by: Jeff Davis <pgsql@j-davis.com>
Discussion: https://postgr.es/m/aL85CeasM51-0D1h%40nathan
Backpatch-through: 16

If the hash functions used for hashing tuples leaked any memory,
we failed to clean that up, resulting in query-lifespan memory
leakage in queries using hashed subplans.  One way that could
happen is if the values being hashed require de-toasting, since
most of our hash functions don't trouble to clean up de-toasted
inputs.

Prior to commit bf6c614, this leakage was largely masked
because TupleHashTableMatch would reset hashtable->tempcxt
(via execTuplesMatch).  But it doesn't do that anymore, and
that's not really the right place for this anyway: doing it
there could reset the tempcxt many times per hash lookup,
or not at all.  Instead put reset calls into ExecHashSubPlan
and buildSubPlanHash.  Along the way to that, rearrange
ExecHashSubPlan so that there's just one place to call
MemoryContextReset instead of several.

This amounts to accepting the de-facto API spec that the caller
of the TupleHashTable routines is responsible for resetting the
tempcxt adequately often.  Although the other callers seem to
get this right, it was not documented anywhere, so add a comment
about it.

Bug: #19040
Reported-by: Haiyang Li <mohen.lhy@alibaba-inc.com>
Author: Haiyang Li <mohen.lhy@alibaba-inc.com>
Reviewed-by: Fei Changhong <feichanghong@qq.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/19040-c9b6073ef814f48c@postgresql.org
Backpatch-through: 13

Instead of building a separate memory context that's used just
for running hash functions, make the hash functions run in the
per-tuple context of the node's innerecontext.  This saves a
little space at runtime, and it avoids needing to reset two
contexts instead of one inside buildSubPlanHash's main loop.

This largely reverts commit 133924e.  That's safe to do now
because bf6c614 decoupled the evaluation context used by
TupleHashTableMatch from that used for hash function evaluation,
so that there's no longer a risk of resetting the innerecontext
too soon.

Per discussion of bug #19040, although this is not directly
a fix for that.

Author: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Haiyang Li <mohen.lhy@alibaba-inc.com>
Reviewed-by: Fei Changhong <feichanghong@qq.com>
Discussion: https://postgr.es/m/19040-c9b6073ef814f48c@postgresql.org

If sizeof(Pointer) is 4 then sizeof(SortItem) will be 12, so that
if data->numrows is odd then we placed the values array at a location
that's not a multiple of 8.  That was fine when sizeof(Datum) was also
4, but in the wake of commit 2a600a9 it makes some alignment-picky
machines unhappy.  (You need a 32-bit machine that nonetheless expects
8-byte alignment of 8-byte quantities, which is an odd-seeming
combination but it does exist outside the Intel universe.)

To fix, MAXALIGN the space allocated to the SortItem array.
In passing, let's make the "len" variable be Size not int,
just for paranoia's sake.

This code was arguably not too safe even before 2a600a9, but at
present I don't see a strong argument for back-patching.

Reported-by: Tomas Vondra <tomas@vondra.me>
Author: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/87036018-8d70-40ad-a0ac-192b07bd7b04@vondra.me

GucSource_Names was documented as being in guc.c, but since 0a20ff5
it is located in guc_tables.c.  The reference to the location of
GucSource_Names is important, as GucSource needs to be kept in sync with
GucSource_Names.

Author: David G. Johnston <david.g.johnston@gmail.com>
Discussion: https://postgr.es/m/CAKFQuwYPgAHWPYjPzK7iXzhSZ6MKR8w20_Nz7ZXpOvx=kZbs7A@mail.gmail.com
Backpatch-through: 16

hash_xlog.h included descriptions for the blocks used in WAL records
that were was not completely consistent with how the records are
generated, with one block missing for SQUEEZE_PAGE, and inconsistent
descriptions used for block 0 in VACUUM_ONE_PAGE and MOVE_PAGE_CONTENTS.

This information was incorrect since c11453c, cross-checking the
logic for the record generation.

Author: Kirill Reshke <reshkekirill@gmail.com>
Reviewed-by: Andrey Borodin <x4mmm@yandex-team.ru>
Discussion: https://postgr.es/m/CALdSSPj1j=a1d1hVA3oabRFz0hSU3KKrYtZPijw4UPUM7LY9zw@mail.gmail.com
Backpatch-through: 13

Mention that the new variants of random(min, max) are affected by
setseed(), like the original functions.

Reported-by: Marcos Pegoraro <marcos@f10.com.br>
Discussion: https://postgr.es/m/CAB-JLwb1=drA3Le6uZXDBi_tCpeS1qm6XQU7dKwac_x91Z4qDg@mail.gmail.com

Incorrect indentation introduced by commit faf071b.

The test assumes that a backend will execute COMMIT PREPARED on the
publisher and hit the injection point commit-after-delay-checkpoint within
the commit critical section. This should cause the apply worker on the
subscriber to wait for the transaction to complete.

However, the test does not guarantee that the injection point is actually
triggered, creating a race condition where the apply worker may proceed
prematurely during COMMIT PREPARED.

This commit resolves the issue by explicitly waiting for the injection
point to be hit before continuing with the test, ensuring consistent and
reliable behavior.

Author: Zhijie Hou <houzj.fnst@fujitsu.com>
Reviewed-by: shveta malik <shveta.malik@gmail.com>
Discussion: https://postgr.es/m/TY4PR01MB1690751D1CA8C128B0770EC6F9409A@TY4PR01MB16907.jpnprd01.prod.outlook.com

The Sun Studio compiler complains about an empty declaration here.

Note for future historians:  This does not mean that this compiler is
still of current interest for anyone using PostgreSQL.  But we can let
this small fix be its parting gift.

Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://www.postgresql.org/message-id/flat/a0f817ee-fb86-483a-8a14-b6f7f5991b6e%40eisentraut.org

Since commit f5e0186 (Raise C requirement to C11), we effectively
require at least GCC version 4.7, so checks for older versions can be
removed.

Reviewed-by: Andres Freund <andres@anarazel.de>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://www.postgresql.org/message-id/flat/a0f817ee-fb86-483a-8a14-b6f7f5991b6e%40eisentraut.org