Previously, we relied on HEAP2_NEW_CID records and XACT_INVALIDATION
records to know if the transaction has modified the catalog, and that
information is not serialized to snapshot. Therefore, after the restart,
if the logical decoding decodes only the commit record of the transaction
that has actually modified a catalog, we will miss adding its XID to the
snapshot. Thus, we will end up looking at catalogs with the wrong
snapshot.
To fix this problem, this change adds the list of transaction IDs and
sub-transaction IDs, that have modified catalogs and are running during
snapshot serialization, to the serialized snapshot. After restart or
otherwise, when we restore from such a serialized snapshot, the
corresponding list is restored in memory. Now, when decoding a COMMIT
record, we check both the list and the ReorderBuffer to see if the
transaction has modified catalogs.
Since this adds additional information to the serialized snapshot, we
cannot backpatch it. For back branches, we took another approach.
We remember the last-running-xacts list of the decoded RUNNING_XACTS
record after restoring the previously serialized snapshot. Then, we mark
the transaction as containing catalog changes if it's in the list of
initial running transactions and its commit record has
XACT_XINFO_HAS_INVALS. This doesn't require any file format changes but
the transaction will end up being added to the snapshot even if it has
only relcache invalidations. But that won't be a problem since we use
snapshot built during decoding only to read system catalogs.
This commit bumps SNAPBUILD_VERSION because of a change in SnapBuild.
Reported-by: Mike Oh
Author: Masahiko Sawada
Reviewed-by: Amit Kapila, Shi yu, Takamichi Osumi, Kyotaro Horiguchi, Bertrand Drouvot, Ahsan Hadi
Backpatch-through: 10
Discussion: https://postgr.es/m/81D0D8B0-E7C4-4999-B616-1E5004DBDCD2%40amazon.com
As reported by Yura Sokolov, scanning the snapshot->xip array has
noticeable impact on scalability when there are a large number of
concurrent writers. Use the optimized (on x86-64) routine from b6ef16756
to speed up searches through the [sub]xip arrays. One benchmark showed
a 5% increase in transaction throughput with 128 concurrent writers,
and a 50% increase in a pathological case of 1024 writers. While a hash
table would have scaled even better, it was ultimately rejected because
of concerns around code complexity and memory allocation. Credit to Andres
Freund for the idea to optimize linear search using SIMD instructions.
Nathan Bossart
Reviewed by: Andres Freund, John Naylor, Bharath Rupireddy, Masahiko Sawada
Discussion: https://postgr.es/m/20220713170950.GA3116318%40nathanxps13
fmgr_sql must make expanded-datum arguments read-only, because
it's possible that the function body will pass the argument to
more than one callee function. If one of those functions takes
the datum's R/W property as license to scribble on it, then later
callees will see an unexpected value, leading to wrong answers.
From a performance standpoint, it'd be nice to skip this in the
common case that the argument value is passed to only one callee.
However, detecting that seems fairly hard, and certainly not
something that I care to attempt in a back-patched bug fix.
Per report from Adam Mackler. This has been broken since we
invented expanded datums, so back-patch to all supported branches.
Discussion: https://postgr.es/m/WScDU5qfoZ7PB2gXwNqwGGgDPmWzz08VdydcPFLhOwUKZcdWbblbo-0Lku-qhuEiZoXJ82jpiQU4hOjOcrevYEDeoAvz6nR0IU4IHhXnaCA=@mackler.email
Discussion: https://postgr.es/m/187436.1660143060@sss.pgh.pa.us
The messages were using 'two-phase transaction' at some places and
'prepared transaction' at other places. Make them consistently use
'prepared transaction'.
Reported-by: Ekaterina Kiryanova
Author: Peter Smith
Reviewed by: Amit Kapila
Backpatch-through: 15
Discussion: https://postgr.es/m/745414e7-efb2-a6ae-5b83-fcbdf35aabc8@postgrespro.ru
Use SSE2 intrinsics to speed up the search, where available. Otherwise,
use a simple 'for' loop. The motivation to add this now is to speed up
XidInMVCCSnapshot(), which is the reason only unsigned 32-bit integer
arrays are optimized. Other types are left for future work, as is the
extension of this technique to non-x86 platforms.
Nathan Bossart
Reviewed by: Andres Freund, Bharath Rupireddy, Masahiko Sawada
Discussion: https://postgr.es/m/20220713170950.GA3116318%40nathanxps13
This commit addresses a few things around GUCs:
- The TCP-related parameters (the four tcp_keepalives_* and
client_connection_check_interval are listed in postgresql.conf.sample in
a subsection called "TCP settings" of "CONNECTIONS AND AUTHENTICATION",
but they did not have their own group name in guc.c.
- enable_group_by_reordering, stats_fetch_consistency and
recovery_prefetch had an inconsistent description, missing a dot at the
end.
- In postgresql.conf.sample, "Process title" should not have a section
of its own, but it should be a subsection of "REPORTING AND LOGGING".
This impacts the contents of pg_settings, which could be seen as a
compatibility break, so no backpatch is done. This is similar to the
cleanup done in a55a984.
Author: Shinya Kato
Discussion: https://postgr.es/m/5e0c9c608624eafbba910c344282cb14@oss.nttdata.com
Commit 623cc673 removed gettimeofday(), and commits 24c3ce8f and
495ed0ef removed support for very old Windows releases with low accuracy
timers, but references to those things were left behind in comments.
Reported-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/295419.1659918447%40sss.pgh.pa.us
Commit 964d01ae9 was a few bricks shy of a load here: the script
checked whether gen_node_support.pl itself had been updated since it
was last run, but not whether any of its input files had been updated.
Fix that. While here, scrape the list of input files from the
Makefiles rather than having a duplicate copy, as we do for most
other lists of source files.
In passing, improve gen_node_support.pl's error report for an
incorrect file list.
Per gripe from Amit Kapila.
Discussion: https://postgr.es/m/CAA4eK1KQk4vP-3mTAz26h-PRUZaGu8Fc=q-ZKSajsAthH0A15w@mail.gmail.com
Per buildfarm, the output order of \dx+ isn't consistent across
locales. Apply NO_LOCALE to force C locale. There might be a
more localized way, but I'm not seeing it offhand, and anyway
there is nothing in this test module that particularly cares
about locales.
Security: CVE-2022-2625
Previously, if an extension script did CREATE OR REPLACE and there was
an existing object not belonging to the extension, it would overwrite
the object and adopt it into the extension. This is problematic, first
because the overwrite is probably unintentional, and second because we
didn't change the object's ownership. Thus a hostile user could create
an object in advance of an expected CREATE EXTENSION command, and would
then have ownership rights on an extension object, which could be
modified for trojan-horse-type attacks.
Hence, forbid CREATE OR REPLACE of an existing object unless it already
belongs to the extension. (Note that we've always forbidden replacing
an object that belongs to some other extension; only the behavior for
previously-free-standing objects changes here.)
For the same reason, also fail CREATE IF NOT EXISTS when there is
an existing object that doesn't belong to the extension.
Our thanks to Sven Klemm for reporting this problem.
Security: CVE-2022-2625
At some point in the past some headers (net/if.h on some BSDs in 2009,
netinet/tcp.h on IRIX in 2000, sys/ucred.h in 2013 on then older openbsd),
only compiled if other heades were included first, complicating configure
tests. More recent tests indicate that that's not required anymore.
Reviewed-By: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/20220807012914.ydz73yte6j3coulo@awork3.anarazel.de
Previously we fell back to __FUNCTION__ and then NULL. As __func__ is in C99
that shouldn't be necessary anymore.
Solution.pm defined HAVE_FUNCNAME__FUNCTION instead of
HAVE_FUNCNAME__FUNC (originating in 4164e6636e), as at some point in the past
MSVC only supported __FUNCTION__. Our minimum version supports __func__.
Reviewed-By: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/20220807012914.ydz73yte6j3coulo@awork3.anarazel.de
strtof() is in C99 and all targeted systems have it. We can remove the
configure probe and some dead code, but we still need replacement code
for a couple of systems that have known buggy implementations selected
via platform template.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/152683.1659830125%40sss.pgh.pa.us
Previously we bothered to forward-declare struct timezone, following man
pages on typical systems, but POSIX actually says the argument (which we
ignore anyway) is void *. Drop a line.
While here, add an assertion that nobody actually uses the tzp argument.
Previously we did extra work to select between Windows APIs needed on
older releases, but now we can just use the higher resolution function
directly.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/CA%2BhUKGKwRpvGfcfq2qNVAQS2Wg1B9eA9QRhAmVSyJt1zsCN2sQ%40mail.gmail.com
Buildfarm member jacana (MinGW) has been complaining that
get_iso_localename is defined but not used. This is evidently
fallout from the recent removal of VS2013 support in pg_locale.c.
Rearrange the #ifs so that get_iso_localename and its subroutine
search_locale_enum won't get built on MinGW.
I also noticed that a comment in get_iso_localename cross-
referenced a comment in IsoLocaleName that isn't there anymore.
Put back what I think is the referenced material.
RelationCopyStorageUsingBuffer thought it could skip copying
empty pages, but of course that does not work at all, because
subsequent blocks will be out of place.
Also fix it to acquire share lock on the source buffer. It *might*
be safe to not do that, but it's not very certain, and I don't think
this code deserves any benefit of the doubt.
Dilip Kumar, per complaint from me
Discussion: https://postgr.es/m/3679800.1659654066@sss.pgh.pa.us
There's no known supported system needing 1 argument gettimeofday()
support. The test for it was added a long time ago (92c6bf9775). Remove.
Until now we tested whether a gettimeofday() fallback is needed when
targetting windows. Which lead to the odd result that HAVE_GETTIMEOFDAY only
being defined when targetting MinGW (which has gettimeofday() since at least
2007). As the fallback is specific to msvc, remove the configure code and
rename src/port/gettimeofday.c to src/port/win32gettimeofday.c.
While at it, also remove the definition of struct timezone, a forward
declaration of the struct is sufficient.
Reviewed-By: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-By: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/20220806000311.ywx65iuchvj4qn2k@awork3.anarazel.de
Commit 59be1c942a already tried to make
src/test/recovery/t/033_replay_tsp_drops more reliable, but it wasn't
enough. Try to improve on that by making this use of a replication slot
to be more like others. Also, don't drop the slot.
Make a few other stylistic changes while at it. It's still quite slow,
which is another thing that we need to fix in this script.
Backpatch to all supported branches.
Discussion: https://postgr.es/m/349302.1659191875@sss.pgh.pa.us
Now that lstat() reports junction points with S_IFLNK/S_ISLINK(), and
unlink() can unlink them, there is no need for conditional code for
Windows in a few places. That was expressed by testing for WIN32 or
S_ISLNK, which we can now constant-fold.
The coding around pgwin32_is_junction() was a bit suspect anyway, as we
never checked for errors, and we also know that errors can be spuriously
reported because of transient sharing violations on this OS. The
lstat()-based code has handling for that.
This also reverts 4fc6b6ee on master only. That was done because
lstat() didn't previously work for symlinks (junction points), but now
it does.
Tested-by: Andrew Dunstan <andrew@dunslane.net>
Discussion: https://postgr.es/m/CA%2BhUKGLfOOeyZpm5ByVcAt7x5Pn-%3DxGRNCvgiUPVVzjFLtnY0w%40mail.gmail.com
Junction points will be reported with S_ISLNK(x.st_mode), simulating
POSIX lstat(). stat() will follow pseudo-symlinks, like in POSIX (but
only one level before giving up, unlike in POSIX).
This completes a TODO left by commit bed90759fc.
Tested-by: Andrew Dunstan <andrew@dunslane.net> (earlier version)
Discussion: https://postgr.es/m/CA%2BhUKGLfOOeyZpm5ByVcAt7x5Pn-%3DxGRNCvgiUPVVzjFLtnY0w%40mail.gmail.com
strtoll was backfilled with either __strtoll or strtoq on systems without
strtoll. The last such system on the buildfarm was an ancient HP-UX animal. We
don't support HP-UX anymore, so remove.
On other systems strtoll was present, but did not have a declaration. The last
known instance on the buildfarm was running an ancient OSX and shut down in
2019.
Author: Andres Freund <andres@anarazel.de>
Reviewed-by: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/20220804013546.h65najrzig764jar@awork3.anarazel.de
nbtree deduplication passes add tuples from the original/target page to
a temp page, merging as necessary. The temp page is copied back to the
target permanent page in the critical section. This is similar to the
approach taken by nbtree page splits.
Adjust comments that referred to updating the original page in-place as
tuples were merged. These were left over from earlier versions of the
deduplication patch that didn't yet use a temp page.
On closer inspection, mcv.c isn't as broken for ScalarArrayOpExpr
as I thought. The Var-on-right issue is real enough, but actually
it does cope fine with a NULL array constant --- I was misled by
an XXX comment suggesting it didn't. Undo that part of the code
change, and replace the XXX comment with something less misleading.
Since v14, the extended stats machinery will try to estimate for
otherwise-unsupported boolean expressions if they match an expression
available from an extended stats object. mcv.c did not get the memo
about this, and would spit up with "unknown clause type". Fortunately
the case is easy to handle, since we can expect the expression yields
boolean.
While here, replace some not-terribly-on-point assertions with
simpler runtime tests for lookup failure. That seems appropriate
so that we get an elog not a crash if we somehow get to the new
it-should-be-a-bool-expression code with a subexpression that
doesn't match any stats column.
Per report from Danny Shemesh. Thanks to Justin Pryzby for
preliminary investigation.
Discussion: https://postgr.es/m/CAFZC=QqD6=27wQPOW1pbRa98KPyuyn+7cL_Ay_Ck-roZV84vHg@mail.gmail.com
statext_is_compatible_clause_internal() checked that the arguments
of a ScalarArrayOpExpr are one Var and one Const, but it would allow
cases where the Const was on the left. Subsequent uses of the clause
are not expecting that and would suffer assertion failures or core
dumps. mcv.c also had not bothered to cope with the case of a NULL
array constant, which seems really unacceptably sloppy of somebody.
(Although our tools failed us there too, since AFAIK neither Coverity
nor any compiler warned of the obvious use-of-uninitialized-variable
condition.) It seems best to handle that by having
statext_is_compatible_clause_internal() reject it.
Noted while fixing bug #17570. Back-patch to v13 where the
extended stats code grew some awareness of ScalarArrayOpExpr.
This makes it more convenient for git config to contain the
blame.ignoreRevsFile setting; otherwise current git versions complain if
the file is not present.
I constructed the file for each branch by scraping the file in branch
master for commits that appear in that branch. Because a few additional
pgindent commits have been added to the list in master since the list
was first created, this also propagates those to branches 14 and 15
where the file already existed. Also, some entries appear to have been
made using author-date rather than committer-date in the format string,
so some timestamps are changed. Also remove bogus whitespace in the
suggested `git log` format string.
Backpatch to all supported branches.
Discussion: https://postgr.es/m/20220711163138.o72evdeus5f5yy5z@alvherre.pgsql
Commit a4d75c86b improved the extended-stats logic to allow extended
stats to be collected on expressions not just bare Vars. To apply
such stats, we first verify that the user has permissions to read all
columns used in the stats. (If not, the query will likely fail at
runtime, but the planner ought not do so.) That had to get extended
to check permissions of columns appearing within such expressions,
but the code for that was completely wrong: it applied pull_varattnos
to the wrong pointer, leading to "unrecognized node type" failures.
Furthermore, although you couldn't get to this because of that bug,
it failed to account for the attnum offset applied by pull_varattnos.
This escaped recognition so far because the code in question is not
reached when the user has whole-table SELECT privilege (which is the
common case), and because only subexpressions not specially handled
by statext_is_compatible_clause_internal() are at risk.
I think a large part of the reason for this bug is under-documentation
of what statext_is_compatible_clause() is doing and what its arguments
are, so do some work on the comments to try to improve that.
Per bug #17570 from Alexander Kozhemyakin. Patch by Richard Guo;
comments and other cosmetic improvements by me. (Thanks also to
Japin Li for diagnosis.) Back-patch to v14 where the bug came in.
Discussion: https://postgr.es/m/17570-f2f2e0f4bccf0965@postgresql.org
Having additional triggers in a test table made the ORDER BY clauses in
old queries underspecified. Add another column there for stability.
Per sporadic buildfarm pink.
When inserting a view referencing a foreign table that has WITH CHECK
OPTION constraints, in single-insert mode postgres_fdw retrieves the
data that was actually inserted on the remote side so that the WITH
CHECK OPTION constraints are enforced with the data locally, but in
batch-insert mode it cannot currently retrieve the data (except for the
row first inserted through the view), resulting in enforcing the WITH
CHECK OPTION constraints with the data passed from the core (except for
the first-inserted row), which led to incorrect results when inserting
into a view referencing a foreign table in which a remote BEFORE ROW
INSERT trigger changes the rows inserted through the view so that they
violate the view's WITH CHECK OPTION constraint. Also, the query
inserting into the view caused an assertion failure in assert-enabled
builds.
Fix these by disabling batch insertion when inserting into such a view.
Back-patch to v14 where batch insertion was added.
Discussion: https://postgr.es/m/CAPmGK17LpbTZs4m4a_6THP54UBeK9fHvX8aVVA%2BC6yEZDZwQcg%40mail.gmail.com
Using ATSimpleRecursion() in ATPrepCmd() to do so as bbb927b4db did is
not correct, because ATPrepCmd() can't distinguish between triggers that
may be cloned and those that may not, so would wrongly try to recurse
for the latter category of triggers.
So this commit restores the code in EnableDisableTrigger() that
86f575948c had added to do the recursion, which would do it only for
triggers that may be cloned, that is, row-level triggers. This also
changes tablecmds.c such that ATExecCmd() is able to pass the value of
ONLY flag down to EnableDisableTrigger() using its new 'recurse'
parameter.
This also fixes what seems like an oversight of 86f575948c that the
recursion to partition triggers would only occur if EnableDisableTrigger()
had actually changed the trigger. It is more apt to recurse to inspect
partition triggers even if the parent's trigger didn't need to be
changed: only then can we be certain that all descendants share the same
state afterwards.
Backpatch all the way back to 11, like bbb927b4db. Care is taken not
to break ABI compatibility (and that no catversion bump is needed.)
Co-authored-by: Amit Langote <amitlangote09@gmail.com>
Reviewed-by: Dmitry Koval <d.koval@postgrespro.ru>
Discussion: https://postgr.es/m/CA+HiwqG-cZT3XzGAnEgZQLoQbyfJApVwOTQaCaas1mhpf+4V5A@mail.gmail.com
clock_gettime() is in SUSv2 and all targeted Unix systems have it.
Remove a chunk of fallback code for old Unix is no longer reachable on
modern systems, and untested as of the retirement of build farm animal
prairiedog.
There is no need to retain a HAVE_CLOCK_GETTIME macro here, because it
is already used in a context with Unix and Windows code paths.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA+hUKGJ3LHeP9w5Fgzdr4G8AnEtJ=z=p6hGDEm4qYGEUX5B6fQ@mail.gmail.com
Commit dd299df818, which made heap TID a tiebreaker nbtree index
column, introduced new rules on page space management to make suffix
truncation safe for v4+ indexes. New pivot tuples (generated by suffix
truncation during leaf page splits) sometimes require dedicated extra
space at the end of a new leaf page high key/pivot to store a heap TID
using a special representation (a representation only used in pivot
tuples).
The definition of "1/3 of a page" was reduced by a single MAXALIGN()
quantum for v4 indexes to make sure that the final enlarged pivot tuple
always fit, even with a split point whose firstright tuple happened to
already be at the "1/3 of a page" limit (limit for non-pivot tuples).
Internal pages (which only contain pivot tuples) stuck with the original
"1/3 of a page" definition. This scheme made it impossible for any page
split to fail to free enough space for its newitem, which is never okay.
The macro that determines whether non-pivot tuples exceed their "1/3 of
a leaf page" restriction was structured as if space was needed for all
three tuples during a leaf page split (the new pivot plus two very large
adjoining non-pivots that are separated by the split). This was subtly
wrong, in that it accidentally relied on implementation details that
could (at least in theory) change in the future.
To fix, make the macro subtract a single MAXALIGN() quantum, once. The
macro evaluates to exactly the same value as before in practice. But it
no longer depends on the current layout of nbtree's special area struct.
No backpatch, since this isn't a live bug.
Author: Peter Geoghegan <pg@bowt.ie>
Reported-By: Robert Haas <robertmhaas@gmail.com>
Diagnosed-By: Robert Haas <robertmhaas@gmail.com>
Discussion: https://postgr.es/m/CA+Tgmoa7UBxivM7f6Ocx_qbq4=ky3uXc+WZNOBcVX+kvJvWOEA@mail.gmail.com
Amit Kapila