Here we convert CompactAttribute.attalign from a char, which is directly
derived from pg_attribute.attalign into a uint8, which stores the number
of bytes to align the column's value by in the tuple.
This allows tuple deformation and tuple size calculations to move away
from using the inefficient att_align_nominal() macro, which manually
checks each TYPALIGN_* char to translate that into the alignment bytes
for the given type. Effectively, this commit changes those to TYPEALIGN
calls, which are branchless and only perform some simple arithmetic with
some bit-twiddling.
The removed branches were often mispredicted by CPUs, especially so in
real-world tables which often contain a mishmash of different types
with different alignment requirements.
Author: David Rowley
Reviewed-by: Andres Freund, Victor Yegorov
Discussion: https://postgr.es/m/CAApHDvrBztXP3yx=NKNmo3xwFAFhEdyPnvrDg3=M0RhDs+4vYw@mail.gmail.com
The new compact_attrs array stores a few select fields from
FormData_pg_attribute in a more compact way, using only 16 bytes per
column instead of the 104 bytes that FormData_pg_attribute uses. Using
CompactAttribute allows performance-critical operations such as tuple
deformation to be performed without looking at the FormData_pg_attribute
element in TupleDesc which means fewer cacheline accesses.
For some workloads, tuple deformation can be the most CPU intensive part
of processing the query. Some testing with 16 columns on a table
where the first column is variable length showed around a 10% increase in
transactions per second for an OLAP type query performing aggregation on
the 16th column. However, in certain cases, the increases were much
higher, up to ~25% on one AMD Zen4 machine.
This also makes pg_attribute.attcacheoff redundant. A follow-on commit
will remove it, thus shrinking the FormData_pg_attribute struct by 4
bytes.
Author: David Rowley
Reviewed-by: Andres Freund, Victor Yegorov
Discussion: https://postgr.es/m/CAApHDvrBztXP3yx=NKNmo3xwFAFhEdyPnvrDg3=M0RhDs+4vYw@mail.gmail.com
Remove the 'whenTaken' and 'lsn' fields from SnapshotData. After the
removal of the "snapshot too old" feature, they were never set to a
non-zero value.
This largely reverts commit 3e2f3c2e42, which added the
OldestActiveSnapshot tracking, and the init_toast_snapshot()
function. That was only required for setting the 'whenTaken' and 'lsn'
fields. SnapshotToast is now a constant again, like SnapshotSelf and
SnapshotAny. I kept a thin get_toast_snapshot() wrapper around
SnapshotToast though, to check that you have a registered or active
snapshot. That's still a useful sanity check.
Reviewed-by: Nathan Bossart, Andres Freund, Tom Lane
Discussion: https://www.postgresql.org/message-id/cd4b4f8c-e63a-41c0-95f6-6e6cd9b83f6d@iki.fi
TAP tests can write
$node->init(no_data_checksums => 1);
to initialize a cluster explicitly without checksums. Currently, this
is the default, but this change allows running all tests with
checksums enabled, like
PG_TEST_INITDB_EXTRA_OPTS=--data-checksums meson test ...
And this also prepares the tests for when we switch the default to
checksums enabled.
The pg_checksums tests need to disable checksums so it can test its
own functionality of enabling checksums. The amcheck/pg_amcheck tests
need to disable checksums because they manually introduce corruption
that they want to detect, but with checksums enabled, the checksum
verification will fail before they even get to their work.
Author: Greg Sabino Mullane <greg@turnstep.com>
Reviewed-by: Nathan Bossart <nathandbossart@gmail.com>
Reviewed-by: Peter Eisentraut <peter@eisentraut.org>
Discussion: https://www.postgresql.org/message-id/flat/CAKAnmmKwiMHik5AHmBEdf5vqzbOBbcwEPHo4-PioWeAbzwcTOQ@mail.gmail.com
Currently, when amcheck validates a unique constraint, it visits the heap for
each index tuple. This commit implements skipping keys, which have only one
non-dedeuplicated index tuple (quite common case for unique indexes). That
gives substantial economy on index checking time.
Reported-by: Noah Misch
Discussion: https://postgr.es/m/20240325020323.fd.nmisch%40google.com
Author: Alexander Korotkov, Pavel Borisov
This commit removes unused variables and routines from some perl code
that have accumulated across the years. This touches the following
areas:
- Wait event generation script.
- AdjustUpgrade.pm.
- TAP perl code
Author: Alexander Lakhin
Reviewed-by: Dagfinn Ilmari Mannsåker
Discussion: https://postgr.es/m/70b340bc-244a-589d-ef8b-d8aebb707a84@gmail.com
* Don't forget to pfree() the right page when it's to be ignored.
* Report error on unexpected non-leaf right page even if this page is not
to be ignored. This restores the logic which was unintendedly changed
in 97e5b0026f.
Reported-by: Pavel Borisov
This is a very unlikely condition during checking a B-tree unique constraint,
meaning that the index structure is violated badly, and we shouldn't continue
checking to avoid endless loops, etc. So it's worth immediately throwing an
error.
Reported-by: Peter Geoghegan
Discussion: https://postgr.es/m/CAH2-Wzk%2B2116uOXdOViA27SHcr31WKPgmjsxXLBs_aTxAeThzg%40mail.gmail.com
Author: Pavel Borisov
5ae2087202 implemented a cross-page unique constraint check by loading
the right sibling to the BtreeCheckState.target variable. This is wrong,
because bt_target_page_check() shouldn't change the target page. Also,
BtreeCheckState.target shouldn't be changed alone without
BtreeCheckState.targetblock.
However, the above didn't cause any visible bugs for the following reasons.
1. We do a cross-page unique constraint check only for leaf index pages.
2. The only way target page get accessed after a cross-page unique constraint
check is loading of the lowkey.
3. The only place lowkey is used is bt_child_highkey_check(), and that applies
only to non-leafs.
The reasons above don't diminish the fact that changing BtreeCheckState.target
for a cross-page unique constraint check is wrong. This commit changes this
check to temporarily store the right sibling to the local variable.
Reported-by: Peter Geoghegan
Discussion: https://postgr.es/m/CAH2-Wzk%2B2116uOXdOViA27SHcr31WKPgmjsxXLBs_aTxAeThzg%40mail.gmail.com
Author: Pavel Borisov
This commit introduces a new data structure BtreeLastVisibleEntry comprising
information about the last visible heap entry with the current value of key.
Usage of this data structure allows us to avoid passing all this information
as individual function arguments.
Reported-by: Alexander Korotkov
Discussion: https://www.postgresql.org/message-id/CAPpHfdsVbB9ToriaB1UHuOKwjKxiZmTFQcEF%3DjuzzC_nby31uA%40mail.gmail.com
Author: Pavel Borisov, Alexander Korotkov
This fixes various typos, duplicated words, and tiny bits of whitespace
mainly in code comments but also in docs.
Author: Daniel Gustafsson <daniel@yesql.se>
Author: Heikki Linnakangas <hlinnaka@iki.fi>
Author: Alexander Lakhin <exclusion@gmail.com>
Author: David Rowley <dgrowleyml@gmail.com>
Author: Nazir Bilal Yavuz <byavuz81@gmail.com>
Discussion: https://postgr.es/m/3F577953-A29E-4722-98AD-2DA9EFF2CBB8@yesql.se
It might happen that the varlena value wasn't compressed by index_form_tuple()
due to current storage parameters. If compression is currently enabled, we
need to compress such values to match index tuple coming from the heap.
Backpatch to all supported versions.
Discussion: https://postgr.es/m/flat/7bdbe559-d61a-4ae4-a6e1-48abdf3024cc%40postgrespro.ru
Author: Andrey Borodin
Reviewed-by: Alexander Lakhin, Michael Zhilin, Jian He, Alexander Korotkov
Backpatch-through: 12
In the heap, tuples may contain short varlena datum with both 1B header and 4B
headers. But the corresponding index tuple should always have such varlena's
with 1B headers. So, for fingerprinting, we need to convert.
Backpatch to all supported versions.
Discussion: https://postgr.es/m/flat/7bdbe559-d61a-4ae4-a6e1-48abdf3024cc%40postgrespro.ru
Author: Michael Zhilin
Reviewed-by: Alexander Lakhin, Andrey Borodin, Jian He, Alexander Korotkov
Backpatch-through: 12
Use GUC_ACTION_SAVE rather than GUC_ACTION_SET, necessary for working
with parallel query.
Now that the call requires more arguments, wrap the call in a new
function to avoid code duplication and offer a place for a comment.
Discussion: https://postgr.es/m/E1rhJpO-0027Wf-9L@gemulon.postgresql.org
There are a lot of Perl scripts in the tree, mostly code generation
and TAP tests. Occasionally, these scripts produce warnings. These
are probably always mistakes on the developer side (true positives).
Typical examples are warnings from genbki.pl or related when you make
a mess in the catalog files during development, or warnings from tests
when they massage a config file that looks different on different
hosts, or mistakes during merges (e.g., duplicate subroutine
definitions), or just mistakes that weren't noticed because there is a
lot of output in a verbose build.
This changes all warnings into fatal errors, by replacing
use warnings;
by
use warnings FATAL => 'all';
in all Perl files.
Discussion: https://www.postgresql.org/message-id/flat/06f899fd-1826-05ab-42d6-adeb1fd5e200%40eisentraut.org
Teach _bt_binsrch (and related helper routines like _bt_search and
_bt_compare) about the initial positioning requirements of backward
scans. Routines like _bt_binsrch already know all about "nextkey"
searches, so it seems natural to teach them about "goback"/backward
searches, too. These concepts are closely related, and are much easier
to understand when discussed together.
Now that certain implementation details are hidden from _bt_first, it's
straightforward to add a new optimization: backward scans using the <
strategy now avoid extra leaf page accesses in certain "boundary cases".
Consider the following example, which uses the tenk1 table (and its
tenk1_hundred index) from the standard regression tests:
SELECT * FROM tenk1 WHERE hundred < 12 ORDER BY hundred DESC LIMIT 1;
Before this commit, nbtree would scan two leaf pages, even though it was
only really necessary to scan one leaf page. We'll now descend straight
to the leaf page containing a (12, -inf) high key instead. The scan
will locate matching non-pivot tuples with "hundred" values starting
from the value 11. The scan won't waste a page access on the right
sibling leaf page, which cannot possibly contain any matching tuples.
You can think of the optimization added by this commit as disabling an
optimization (the _bt_compare "!pivotsearch" behavior that was added to
Postgres 12 in commit dd299df8) for a small subset of cases where it was
always counterproductive.
Equivalently, you can think of the new optimization as extending the
"pivotsearch" behavior that page deletion by VACUUM has long required
(since the aforementioned Postgres 12 commit went in) to other, similar
cases. Obviously, this isn't strictly necessary for these new cases
(unlike VACUUM, _bt_first is prepared to move the scan to the left once
on the leaf level), but the underlying principle is the same.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Matthias van de Meent <boekewurm+postgres@gmail.com>
Discussion: https://postgr.es/m/CAH2-Wz=XPzM8HzaLPq278Vms420mVSHfgs9wi5tjFKHcapZCEw@mail.gmail.com
This prevents false-positive reports about "the first child of leftmost
target page is not leftmost of its level", "block %u is not leftmost"
and "left link/right link pair". They appeared if amcheck ran before
VACUUM cleaned things, after a cluster exited recovery between the
first-stage and second-stage WAL records of a deletion. Back-patch to
v11 (all supported versions).
Reviewed by Peter Geoghegan.
Discussion: https://postgr.es/m/20231005025232.c7.nmisch@google.com
Add the 'checkunique' argument to bt_index_check() and bt_index_parent_check().
When the flag is specified the procedures will check the unique constraint
violation for unique indexes. Only one heap entry for all equal keys in
the index should be visible (including posting list entries). Report an error
otherwise.
pg_amcheck called with the --checkunique option will do the same check for all
the indexes it checks.
Author: Anastasia Lubennikova <lubennikovaav@gmail.com>
Author: Pavel Borisov <pashkin.elfe@gmail.com>
Author: Maxim Orlov <orlovmg@gmail.com>
Reviewed-by: Mark Dilger <mark.dilger@enterprisedb.com>
Reviewed-by: Zhihong Yu <zyu@yugabyte.com>
Reviewed-by: Peter Geoghegan <pg@bowt.ie>
Reviewed-by: Aleksander Alekseev <aleksander@timescale.com>
Discussion: https://postgr.es/m/CALT9ZEHRn5xAM5boga0qnrCmPV52bScEK2QnQ1HmUZDD301JEg%40mail.gmail.com
Since C99, there can be a trailing comma after the last value in an
enum definition. A lot of new code has been introducing this style on
the fly. Some new patches are now taking an inconsistent approach to
this. Some add the last comma on the fly if they add a new last
value, some are trying to preserve the existing style in each place,
some are even dropping the last comma if there was one. We could
nudge this all in a consistent direction if we just add the trailing
commas everywhere once.
I omitted a few places where there was a fixed "last" value that will
always stay last. I also skipped the header files of libpq and ecpg,
in case people want to use those with older compilers. There were
also a small number of cases where the enum type wasn't used anywhere
(but the enum values were), which ended up confusing pgindent a bit,
so I left those alone.
Discussion: https://www.postgresql.org/message-id/flat/386f8c45-c8ac-4681-8add-e3b0852c1620%40eisentraut.org
Under interval_ops, some equal values are distinguishable. One such
pair is '24:00:00' and '1 day'. With that being so, btequalimage()
breaches the documented contract for the "equalimage" btree support
function. This can cause incorrect results from index-only scans.
Users should REINDEX any btree indexes having interval-type columns.
After updating, pg_amcheck will report an error for almost all such
indexes. This fix makes interval_ops simply omit the support function,
like numeric_ops does. Back-pack to v13, where btequalimage() first
appeared. In back branches, for the benefit of old catalog content,
btequalimage() code will return false for type "interval". Going
forward, back-branch initdb will include the catalog change.
Reviewed by Peter Geoghegan.
Discussion: https://postgr.es/m/20231011013317.22.nmisch@google.com
Split nbtree's _bt_getbuf function is two: code that read locks or write
locks existing pages remains in _bt_getbuf, while code that deals with
allocating new pages is moved to a new, dedicated function called
_bt_allocbuf. This simplifies most _bt_getbuf callers, since it is no
longer necessary for them to pass a heaprel argument. Many of the
changes to nbtree from commit 61b313e4 can be reverted. This minimizes
the divergence between HEAD/PostgreSQL 16 and earlier release branches.
_bt_allocbuf replaces the previous nbtree idiom of passing P_NEW to
_bt_getbuf. There are only 3 affected call sites, all of which continue
to pass a heaprel for recovery conflict purposes. Note that nbtree's
use of P_NEW was superficial; nbtree never actually relied on the P_NEW
code paths in bufmgr.c, so this change is strictly mechanical.
GiST already took the same approach; it has a dedicated function for
allocating new pages called gistNewBuffer(). That factor allowed commit
61b313e4 to make much more targeted changes to GiST.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Heikki Linnakangas <hlinnaka@iki.fi>
Discussion: https://postgr.es/m/CAH2-Wz=8Z9qY58bjm_7TAHgtW6RzZ5Ke62q5emdCEy9BAzwhmg@mail.gmail.com
While executing maintenance operations (ANALYZE, CLUSTER, REFRESH
MATERIALIZED VIEW, REINDEX, or VACUUM), set search_path to
'pg_catalog, pg_temp' to prevent inconsistent behavior.
Functions that are used for functional indexes, in index expressions,
or in materialized views and depend on a different search path must be
declared with CREATE FUNCTION ... SET search_path='...'.
This change addresses a security risk introduced in commit 60684dd834,
where a role with MAINTAIN privileges on a table may be able to
escalate privileges to the table owner. That commit is not yet part of
any release, so no need to backpatch.
Discussion: https://postgr.es/m/e44327179e5c9015c8dda67351c04da552066017.camel%40j-davis.com
Reviewed-by: Greg Stark
Reviewed-by: Nathan Bossart
Run pgindent, pgperltidy, and reformat-dat-files.
This set of diffs is a bit larger than typical. We've updated to
pg_bsd_indent 2.1.2, which properly indents variable declarations that
have multi-line initialization expressions (the continuation lines are
now indented one tab stop). We've also updated to perltidy version
20230309 and changed some of its settings, which reduces its desire to
add whitespace to lines to make assignments etc. line up. Going
forward, that should make for fewer random-seeming changes to existing
code.
Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
a9c70b46 added the statistics view pg_stat_io which contained columns
"io_context" and "io_object". Given that the columns are in the
pg_stat_io view, the "io" prefix is somewhat redundant, so remove it.
The code variables referring to these fields are kept unchanged so as
they can keep their context about I/O.
Bump catalog version.
Author: Melanie Plageman
Reviewed-by: Kyotaro Horiguchi, Fabrízio de Royes Mello
Discussion: https://postgr.es/m/CAAKRu_aAQoJWrvT2BYYQvJChFKra_O-5ra3jhzKJZqWsTR1CPQ@mail.gmail.com
This fixes many spelling mistakes in comments, but a few references to
invalid parameter names, function names and option names too in comments
and also some in string constants
Also, fix an #undef that was undefining the incorrect definition
Author: Alexander Lakhin
Reviewed-by: Justin Pryzby
Discussion: https://postgr.es/m/d5f68d19-c0fc-91a9-118d-7c6a5a3f5fad@gmail.com
This breaks out the background and interactive psql functionality into a
new class, PostgreSQL::Test::BackgroundPsql. Sessions are still initiated
via PostgreSQL::Test::Cluster, but once started they can be manipulated by
the new helper functions which intend to make querying easier. A sample
session for a command which can be expected to finish at a later time can
be seen below.
my $session = $node->background_psql('postgres');
$bsession->query_until(qr/start/, q(
\echo start
CREATE INDEX CONCURRENTLY idx ON t(a);
));
$bsession->quit;
Patch by Andres Freund with some additional hacking by me.
Author: Andres Freund <andres@anarazel.de>
Reviewed-by: Andrew Dunstan <andrew@dunslane.net>
Discussion: https://postgr.es/m/20230130194350.zj5v467x4jgqt3d6@awork3.anarazel.de
This is done in preparation for logical decoding on standby, which needs to
include whether visibility affecting WAL records are about a (user) catalog
table. Which is only known for the table, not the indexes.
It's also nice to be able to pass the heap relation to GlobalVisTestFor() in
vacuumRedirectAndPlaceholder().
Author: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
Discussion: https://postgr.es/m/21b700c3-eecf-2e05-a699-f8c78dd31ec7@gmail.com
Commit e88754a196 caused that case
to be reported as corruption, but Peter Geoghegan pointed out that
it can legitimately happen in the case of a speculative insertion
that aborts, so we'd better not flag it as corruption after all.
Back-patch to v14, like the commit that introduced the issue.
Discussion: http://postgr.es/m/CAH2-WzmEabzcPTxSY-NXKH6Qt3FkAPYHGQSe2PtvGgj17ZQkCw@mail.gmail.com
Commit bbc1376b39 checked that if
a redirected line pointer pointed to a tuple, the tuple should be
marked both HEAP_ONLY_TUPLE and HEAP_UPDATED. But Andres Freund
pointed out that *any* tuple that is marked HEAP_ONLY_TUPLE should
be marked HEAP_UPDATED, not just one that is the target of a
redirected line pointer. Do that instead.
To see why this is better, consider a redirect line pointer A
which points to a heap-only tuple B which points (via CTID)
to another heap-only tuple C. With the old code, we'd complain
if B was not marked HEAP_UPDATED, but with this change, we'll
complain if either B or C is not marked HEAP_UPDATED.
(Note that, with or without this commit, if either B or C were
not marked HEAP_ONLY_TUPLE, we would also complain about that.)
Discussion: http://postgr.es/m/CA%2BTgmobLypZx%3DcOH%2ByY1GZmCruaoucHm77A6y_-Bo%3Dh-_3H28g%40mail.gmail.com
Commit bbc1376b39 added a new lp_valid[]
array which records whether or not a line pointer was thought to be
valid, but entries could sometimes get set to true in cases where that
wasn't actually safe. Fix that.
Suppose A is a redirect line pointer and B is the other line pointer
to which it points. The old code could mishandle this situation in a
couple of different ways. First, if B was unused, we'd complain about
corruption but still set lp_valid[A] = true, causing later code
to try to access the B as if it were pointing to a tuple. Second,
if B was dead, we wouldn't complain about corruption at all, which is
an oversight, and would also set lp_valid[A] = true, which would
again confuse later code. Fix all that.
In the case where B is a redirect, the old code was correct, but
refactor things a bit anyway so that all of these cases are handled
more symmetrically. Also add an Assert() and some comments.
Andres Freund and Robert Haas
Discussion: http://postgr.es/m/20230323172607.y3lejpntjnuis5vv%40awork3.anarazel.de
In such cases, get_xid_status() doesn't set its output parameter (the
third argument), so we shouldn't fall through to code which will test
the value of that parameter. There are five existing calls to
get_xid_status(), three of which seem to already handle this case
properly. This commit tries to fix the other two.
If we're checking xmin and find that it is invalid (i.e. 0) just
report that as corruption, similar to what's already done in the
three cases that seem correct. If we're checking xmax and find
that's invalid, that's fine: it just means that the tuple hasn't
been updated or deleted.
Thanks to Andres Freund and valgrind for finding this problem, and
also to Andres for having a look at the patch. This bug seems to go
all the way back to where verify_heapam was first introduced, but
wasn't detected until recently, possibly because of the new test cases
added for update chain verification. Back-patch to v14, where this
code showed up.
Discussion: http://postgr.es/m/CA+TgmoZAYzQZqyUparXy_ks3OEOfLD9-bEXt8N-2tS1qghX9gQ@mail.gmail.com
We shouldn't set successor[whatever] to an offset number that is less
than FirstOffsetNumber or more than maxoff. We already avoided that
for redirects, but not for CTID links. Allowing bad offset numbers
into the successor[] array causes core dumps.
We shouldn't use HeapTupleHeaderIsHotUpdated() because it checks
stuff other than the status of the infomask2 bit HEAP_HOT_UPDATED.
We only care about the status of that bit, not the other stuff
that HeapTupleHeaderIsHotUpdated() checks. This mistake can cause
verify_heapam() to report corruption when none is present.
The first hunk of this patch was written by me. The other two were
written by Andres Freund. This could probably do with more review
before commit, but I'd like to try to get the buildfarm green again
sooner rather than later.
Discussion: http://postgr.es/m/20230322204552.s6cv3ybqkklhhybb@awork3.anarazel.de
Prior to this commit, we only consider each tuple or line pointer
on the page in isolation, but now we can do some validation of a line
pointer against its successor. For example, a redirect line pointer
shouldn't point to another redirect line pointer, and if a tuple
is HOT-updated, the result should be a heap-only tuple.
Himanshu Upadhyaya and Robert Haas, reviewed by Aleksander Alekseev,
Andres Freund, and Peter Geoghegan.
64bit xids can't represent xids before epoch 0 (see also be504a3e97). When
FullTransactionIdFromXidAndCtx() was passed such an xid, it'd create a 64bit
xid far into the future. Noticed while adding assertions in the course of
investigating be504a3e97, as amcheck's test create such xids.
To fix the issue, just return FirstNormalFullTransactionId in this case. A
freshly initdb'd cluster already has a newer horizon. The most minimal version
of this would make the messages for some detected corruptions differently
inaccurate. To make those cases accurate, switch
FullTransactionIdFromXidAndCtx() to use the 32bit modulo difference between
xid and nextxid to compute the 64bit xid, yielding sensible "in the future" /
"in the past" answers.
Reviewed-by: Mark Dilger <mark.dilger@enterprisedb.com>
Discussion: https://postgr.es/m/20230108002923.cyoser3ttmt63bfn@awork3.anarazel.de
Backpatch: 14-, where heapam verification was introduced
The initialization order in update_cached_xid_range() was wrong, calling
FullTransactionIdFromXidAndCtx() before setting
->next_xid. FullTransactionIdFromXidAndCtx() uses ->next_xid.
In most situations this will not cause visible issues, because the next call
to update_cached_xid_range() will use a less wrong ->next_xid. It's rare that
xids advance fast enough for this to be a problem.
Found while adding more asserts to the 64bit xid infrastructure.
Reviewed-by: Mark Dilger <mark.dilger@enterprisedb.com>
Discussion: https://postgr.es/m/20230108002923.cyoser3ttmt63bfn@awork3.anarazel.de
Backpatch: 14-, where heapam verification was introduced
John Naylor