Since the dawn of time (aka Postgres95) multiple pins of the same
buffer by one backend have been optimized not to modify the shared
refcount more than once. This optimization has always used a NBuffer
sized array in each backend keeping track of a backend's pins.

That array (PrivateRefCount) was one of the biggest per-backend memory
allocations, depending on the shared_buffers setting. Besides the
waste of memory it also has proven to be a performance bottleneck when
assertions are enabled as we make sure that there's no remaining pins
left at the end of transactions. Also, on servers with lots of memory
and a correspondingly high shared_buffers setting the amount of random
memory accesses can also lead to poor cpu cache efficiency.

Because of these reasons a backend's buffers pins are now kept track
of in a small statically sized array that overflows into a hash table
when necessary. Benchmarks have shown neutral to positive performance
results with considerably lower memory usage.

Patch by me, review by Robert Haas.

Discussion: 20140321182231.GA17111@alap3.anarazel.de

Indicates if column has ever been local/non-inherited

Also update docs to mention which function are super-user-only.

Report by sys-milan@statpro.com

Backpatch through 9.4

The list of posting lists it's dealing with can contain placeholders for
deleted posting lists. The placeholders are kept around so that they can
be WAL-logged, but we must be careful to not try to access them.

This fixes bug #11280, reported by Mårten Svantesson. Backpatch to 9.4,
where the compressed data leaf page code was added.

It was changed when ALTER TABLESPACE / MOVE was added but then not
updated when that was moved back out.

Haribabu Kommi, reviewed by Muhammad Asif Naeem

Commit a16bac36 let "configure" detect
the system getaddrinfo() when building under 64-bit MinGW-w64.  However,
src/include/port/win32/sys/socket.h assumes all native Windows
configurations use our replacement.  This change placates buildfarm
member jacana until we establish a plan for getaddrinfo() on Windows.

In commit 45e02e32, we intentionally
disallowed updates on individual elements of oidvector columns.  While that
still seems like a sane idea in the abstract, we (I) forgot that citext's
"upgrade from unpackaged" script did in fact perform exactly such updates,
in order to fix the problem that citext indexes should have a collation
but would not in databases dumped or upgraded from pre-9.1 installations.

Even if we wanted to add casts to allow such updates, there's no practical
way to do so in the back branches, so the only real alternative is to make
citext's kluge even klugier.  In this patch, I cast the oidvector to text,
fix its contents with regexp_replace, and cast back to oidvector.  (Ugh!)

Since the aforementioned commit went into all active branches, we have to
fix this in all branches that contain the now-broken update script.

Per report from Eric Malm.

This reverts commit e23014f3.

As the side effect of the reverted commit, when the unit is
specified, the reloption was stored in the catalog with the unit.
This broke pg_dump (specifically, it prevented pg_dump from
outputting restorable backup regarding the reloption) and
turned the buildfarm red. Revert the commit until the fixed
version is ready.

This is useful to allow to set GUCs to values that include spaces;
something that wasn't previously possible. The primary case motivating
this is the desire to set default_transaction_isolation to 'repeatable
read' on a per connection basis, but other usecases like seach_path do
also exist.

This introduces a slight backward incompatibility: Previously a \ in
an option value would have been passed on literally, now it'll be
taken as an escape.

The relevant mailing list discussion starts with
20140204125823.GJ12016@awork2.anarazel.de.

This introduces an infrastructure which allows us to specify the units
like ms (milliseconds) in integer relation option, like GUC parameter.
Currently only autovacuum_vacuum_cost_delay reloption can accept
the units.

Reviewed by Michael Paquier

Previously, only a single-byte character was allowed as an
escape. This patch allows it to be a multi-byte character, though it
still must be a single character.

Reviewed by Heikki Linnakangas and Tom Lane.

If SELECT FOR UPDATE NOWAIT tries to lock a tuple that is concurrently
being updated, it might fail to honor its NOWAIT specification and block
instead of raising an error.

Fix by adding a no-wait flag to EvalPlanQualFetch which it can pass down
to heap_lock_tuple; also use it in EvalPlanQualFetch itself to avoid
blocking while waiting for a concurrent transaction.

Authors: Craig Ringer and Thomas Munro, tweaked by Álvaro
http://www.postgresql.org/message-id/51FB6703.9090801@2ndquadrant.com

Per Thomas Munro in the course of his SKIP LOCKED feature submission,
who also provided one of the isolation test specs.

Backpatch to 9.4, because that's as far back as it applies without
conflicts (although the bug goes all the way back).  To that branch also
backpatch Thomas Munro's new NOWAIT test cases, committed in master by
Heikki as commit 9ee16b49 .

This commit also adds the include guards to those header files.

Michael Paquier

In some cases, not all Vars were being correctly marked as having been
modified for updatable security barrier views, which resulted in invalid
plans (eg: when security barrier views were created over top of
inheiritance structures).

In passing, be sure to update both varattno and varonattno, as _equalVar
won't consider the Vars identical otherwise.  This isn't known to cause
any issues with updatable security barrier views, but was noticed as
missing while working on RLS and makes sense to get fixed.

Back-patch to 9.4 where updatable security barrier views were
introduced.

Author: Fabien COELHO <coelho@cri.ensmp.fr>

Spotted by Peter Geoghegan.

Use SECURITY_LOCAL_USERID_CHANGE while building temporary tables;
only escalate to SECURITY_RESTRICTED_OPERATION while potentially
running user-supplied code.  The more secure mode was preventing
temp table creation.  Add regression tests to cover this problem.

This fixes Bug #11208 reported by Bruno Emanuel de Andrade Silva.

Backpatch to 9.4, where the bug was introduced.

Author: David Rowley

Fabrízio de Royes Mello

That allows to run those tests against a postmaster listening on a
nonstandard port without requiring to export PGPORT in postmaster's
environment.

This still doesn't support connecting to a nondefault host without
configuring it in postmaster's environment. That's harder and less
frequently used though. So this is a useful step.

That allows parallel installcheck to succeed inside contrib/. The
output is not particularly pretty unless make's -O option to
synchronize the output is used.

There's other tests, outside contrib, that use a hardcoded,
non-unique, database name. Those prohibit paralell installcheck to be
used across more directories; but that's something for a separate
patch.

Fabrízio de Royes Mello, reviewed by Michael Paquier

Prevent automatic oid assignment when in binary upgrade mode.  Also
throw an error when contrib/pg_upgrade_support functions are called when
not in binary upgrade mode.

This prevent automatically-assigned oids from conflicting with later
pre-assigned oids coming from the old cluster.  It also makes sure oids
are preserved in call important cases.

Done for clarity

Now that 8.3 is not supported, we can use a CTE and not temp tables.
This allows for auto-oid assignment protection in a future patch.

Reverts commits 73d78e11 and
b0488e5c.  Also reverts pg_upgrade
changes.

This trims down the code, and is in preparation for hardening pg_upgrade
against auto-oid assignment.

Add to commit 73d78e11

Also adjust pg_upgrade to not use this method for optional TOAST table
creation.

Patch by Fabrízio de Royes Mello

Report by Pavel Raiskup

Also, remove OID preservation mention, mention non-text dump formats

Backpatch through 9.4

There's no point in setting up a context error callback when doing
conditional lock acquisition, because we never actually wait and so the
user wouldn't be able to see the context message anywhere.  In fact,
this is more in line with what ConditionalXactLockTableWait is doing.

Backpatch to 9.4, where this was added.

The symbol was added by 71901ab6; the original code was introduced by
6868ed74.  Development of both overlapped which is why we apparently
failed to notice.

This is a (very slight) behavior change, so I'm not backpatching this to
9.4 for now, even though the symbol does exist there.

Event triggers want to know the OID of the interesting object created,
which is the main type.  The array created as part of the operation is
just a subsidiary object which is not of much interest.

Other DDL commands are already returning the OID, which is required for
future additional event trigger work.  This is merely making these
commands in line with the rest of utility command support.