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 changes the snapshot builder so that it
remembers 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. To
avoid ABI breakage, we store the array of the initial running transactions
in the static variables InitialRunningXacts and NInitialRunningXacts,
instead of storing those in SnapBuild or ReorderBuffer.
This approach has a false positive; we could end up adding the transaction
that didn't change catalog to the snapshot since we cannot distinguish
whether the transaction has catalog changes only by checking the COMMIT
record. It doesn't have the information on which (sub) transaction has
catalog changes, and XACT_XINFO_HAS_INVALS doesn't necessarily indicate
that the transaction has catalog change. But that won't be a problem since
we use snapshot built during decoding only to read system catalogs.
On the master branch, we took a more future-proof approach by writing
catalog modifying transactions to the serialized snapshot which avoids the
above false positive. But we cannot backpatch it because of a change in
the 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
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
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
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
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.
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
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.
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
Ordinarily the functions called in this loop ought to have plenty
of CFIs themselves; but we've now seen a case where no such CFI is
reached, making the loop uninterruptible. Even though that's from
a recently-introduced bug, it seems prudent to install a CFI at
the loop level in all branches.
Per discussion of bug #17558 from Andrew Kesper (an actual fix for
that bug will follow).
Discussion: https://postgr.es/m/17558-3f6599ffcf52fd4a@postgresql.org
Remove the test case added by commit fac1b470a, which never actually
worked to expose the problem it claimed to test. Replace it with
a case that does expose the problem, and also covers the SRF-not-
at-the-top deficiency repaired in 1aa8dad41.
Richard Guo, with some editorialization by me
Discussion: https://postgr.es/m/17564-c7472c2f90ef2da3@postgresql.org
Commit fac1b470a thought we could check for set-returning functions
by testing only the top-level node in an expression tree. This is
wrong in itself, and to make matters worse it encouraged others
to make the same mistake, by exporting tlist.c's special-purpose
IS_SRF_CALL() as a widely-visible macro. I can't find any evidence
that anyone's taken the bait, but it was only a matter of time.
Use expression_returns_set() instead, and stuff the IS_SRF_CALL()
genie back in its bottle, this time with a warning label. I also
added a couple of cross-reference comments.
After a fair amount of fooling around, I've despaired of making
a robust test case that exposes the bug reliably, so no test case
here. (Note that the test case added by fac1b470a is itself
broken, in that it doesn't notice if you remove the code change.
The repro given by the bug submitter currently doesn't fail either
in v15 or HEAD, though I suspect that may indicate an unrelated bug.)
Per bug #17564 from Martijn van Oosterhout. Back-patch to v13,
as the faulty patch was.
Discussion: https://postgr.es/m/17564-c7472c2f90ef2da3@postgresql.org
The sto_using_cursor and sto_using_select tests were coded to exercise
every permutation of their test steps, but AFAICS there is no value in
exercising more than one. This matters because each permutation costs
about six seconds, thanks to the "pg_sleep(6)". Perhaps we could
reduce that, but the useless permutations seem worth getting rid of
in any case. (Note that sto_using_hash_index got it right already.)
While here, clean up some other sloppiness such as an unused table.
This doesn't make too much difference in interactive testing, since the
wasted time is typically masked by parallelization with other tests.
However, the buildfarm runs this as a serial step, which means we can
expect to shave ~40 seconds from every buildfarm run. That makes it
worth back-patching.
Discussion: https://postgr.es/m/2515192.1659454702@sss.pgh.pa.us
We've heard a couple of reports of people having trouble with
multi-gigabyte-sized query-texts files. It occurred to me that on
32-bit platforms, there could be an issue with integer overflow
of calculations associated with the total query text size.
Address that with several changes:
1. Limit pg_stat_statements.max to INT_MAX / 2 not INT_MAX.
The hashtable code will bound it to that anyway unless "long"
is 64 bits. We still need overflow guards on its use, but
this helps.
2. Add a check to prevent extending the query-texts file to
more than MaxAllocHugeSize. If it got that big, qtext_load_file
would certainly fail, so there's not much point in allowing it.
Without this, we'd need to consider whether extent, query_offset,
and related variables shouldn't be off_t not size_t.
3. Adjust the comparisons in need_gc_qtexts() to be done in 64-bit
arithmetic on all platforms. It appears possible that under duress
those multiplications could overflow 32 bits, yielding a false
conclusion that we need to garbage-collect the texts file, which
could lead to repeatedly garbage-collecting after every hash table
insertion.
Per report from Bruno da Silva. I'm not convinced that these
issues fully explain his problem; there may be some other bug that's
contributing to the query-texts file becoming so large in the first
place. But it did get that big, so #2 is a reasonable defense,
and #3 could explain the reported performance difficulties.
(See also commit 8bbe4cbd9, which addressed some related bugs.
The second Discussion: link is the thread that led up to that.)
This issue is old, and is primarily a problem for old platforms,
so back-patch.
Discussion: https://postgr.es/m/CAB+Nuk93fL1Q9eLOCotvLP07g7RAv4vbdrkm0cVQohDVMpAb9A@mail.gmail.com
Discussion: https://postgr.es/m/5601D354.5000703@BlueTreble.com
I thought commit fd96d14d9 had plugged all the holes of this sort,
but no, function RTEs could produce oversize tuples too, either
via long coldeflists or just from multiple functions in one RTE.
(I'm pretty sure the other variants of base RTEs aren't a problem,
because they ultimately refer to either a table or a sub-SELECT,
whose widths are enforced elsewhere. But we explicitly allow join
RTEs to be overwidth, as long as you don't try to form their
tuple result.)
Per further discussion of bug #17561. As before, patch all branches.
Discussion: https://postgr.es/m/17561-80350151b9ad2ad4@postgresql.org
errno was not reported correctly after attempting to clone a file,
leading to incorrect error reports. While scanning through the code, I
have not noticed any similar mistakes.
Error introduced in 3a769d8.
Author: Justin Pryzby
Discussion: https://postgr.es/m/20220731134135.GY15006@telsasoft.com
Backpatch-through: 12
The new test is from commit 9e4f914b5e.
With this setting messages have SQL error numbers included, so that
needs to be provided for in the pattern looked for.
Backpatch to all live branches like the original.
A RowExpr with more than MaxTupleAttributeNumber columns would fail at
execution anyway, since we cannot form a tuple datum with more than that
many columns. While heap_form_tuple() has a check for too many columns,
it emerges that there are some intermediate bits of code that don't
check and can be driven to failure with sufficiently many columns.
Checking this at parse time seems like the most appropriate place to
install a defense, since we already check SELECT list length there.
While at it, make the SELECT-list-length error use the same errcode
(TOO_MANY_COLUMNS) as heap_form_tuple does, rather than the generic
PROGRAM_LIMIT_EXCEEDED.
Per bug #17561 from Egor Chindyaskin. The given test case crashes
in all supported branches (and probably a lot further back),
so patch all.
Discussion: https://postgr.es/m/17561-80350151b9ad2ad4@postgresql.org
On FreeBSD, the new test fails due to a WAL file being removed before
the standby has had the chance to copy it. Fix by adding a replication
slot to prevent the removal until after the standby has connected.
Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reported-by: Matthias van de Meent <boekewurm+postgres@gmail.com>
Discussion: https://postgr.es/m/CAEze2Wj5nau_qpjbwihvmXLfkAWOZ5TKdbnqOc6nKSiRJEoPyQ@mail.gmail.com
Crash recovery on standby may encounter missing directories
when replaying database-creation WAL records. Prior to this
patch, the standby would fail to recover in such a case;
however, the directories could be legitimately missing.
Consider the following sequence of commands:
CREATE DATABASE
DROP DATABASE
DROP TABLESPACE
If, after replaying the last WAL record and removing the
tablespace directory, the standby crashes and has to replay the
create database record again, crash recovery must be able to continue.
A fix for this problem was already attempted in 49d9cfc68b, but it
was reverted because of design issues. This new version is based
on Robert Haas' proposal: any missing tablespaces are created
during recovery before reaching consistency. Tablespaces
are created as real directories, and should be deleted
by later replay. CheckRecoveryConsistency ensures
they have disappeared.
The problems detected by this new code are reported as PANIC,
except when allow_in_place_tablespaces is set to ON, in which
case they are WARNING. Apart from making tests possible, this
gives users an escape hatch in case things don't go as planned.
Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Author: Asim R Praveen <apraveen@pivotal.io>
Author: Paul Guo <paulguo@gmail.com>
Reviewed-by: Anastasia Lubennikova <lubennikovaav@gmail.com> (older versions)
Reviewed-by: Fujii Masao <masao.fujii@oss.nttdata.com> (older versions)
Reviewed-by: Michaël Paquier <michael@paquier.xyz>
Diagnosed-by: Paul Guo <paulguo@gmail.com>
Discussion: https://postgr.es/m/CAEET0ZGx9AvioViLf7nbR_8tH9-=27DN5xWJ2P9-ROH16e4JUA@mail.gmail.com
This is a backpatch to branches 10-14 of the following commits:
7170f2159f Allow "in place" tablespaces.
c6f2f01611 Fix pg_basebackup with in-place tablespaces.
f6f0db4d62 Fix pg_tablespace_location() with in-place tablespaces
7a7cd84893 doc: Remove mention to in-place tablespaces for pg_tablespace_location()
5344723755 Remove unnecessary Windows-specific basebackup code.
In-place tablespaces were introduced as a testing helper mechanism, but
they are going to be used for a bugfix in WAL replay to be backpatched
to all stable branches.
I (Álvaro) had to adjust some code to account for lack of
get_dirent_type() in branches prior to 14.
Author: Thomas Munro <thomas.munro@gmail.com>
Author: Michaël Paquier <michael@paquier.xyz>
Author: Álvaro Herrera <alvherre@alvh.no-ip.org>
Discussion: https://postgr.es/m/20220722081858.omhn2in5zt3g4nek@alvherre.pgsql
We have a few commands that "can't run in a transaction block",
meaning that if they complete their processing but then we fail
to COMMIT, we'll be left with inconsistent on-disk state.
However, the existing defenses for this are only watertight for
simple query protocol. In extended protocol, we didn't commit
until receiving a Sync message. Since the client is allowed to
issue another command instead of Sync, we're in trouble if that
command fails or is an explicit ROLLBACK. In any case, sitting
in an inconsistent state while waiting for a client message
that might not come seems pretty risky.
This case wasn't reachable via libpq before we introduced pipeline
mode, but it's always been an intended aspect of extended query
protocol, and likely there are other clients that could reach it
before.
To fix, set a flag in PreventInTransactionBlock that tells
exec_execute_message to force an immediate commit. This seems
to be the approach that does least damage to existing working
cases while still preventing the undesirable outcomes.
While here, add some documentation to protocol.sgml that explicitly
says how to use pipelining. That's latent in the existing docs if
you know what to look for, but it's better to spell it out; and it
provides a place to document this new behavior.
Per bug #17434 from Yugo Nagata. It's been wrong for ages,
so back-patch to all supported branches.
Discussion: https://postgr.es/m/17434-d9f7a064ce2a88a3@postgresql.org
We didn't explicitly say that random() uses a randomly-chosen seed
if you haven't called setseed(). Do so.
Also, remove ref/set.sgml's no-longer-accurate (and never very
relevant) statement that the seed value is multiplied by 2^31-1.
Back-patch to v12 where set.sgml's claim stopped being true.
The claim that we use a source of random bits as seed was debatable
before 4203842a1, too, so v12 seems like a good place to stop.
Per question from Carl Sopchak.
Discussion: https://postgr.es/m/f37bb937-9d99-08f0-4de7-80c91a3cfc2e@sopchak.me
Due to lack of concern for the case in the dependency code, it's
possible to drop a column of a composite type even though stored
queries have references to the dropped column via functions-in-FROM
that return the composite type. There are "soft" references,
namely FROM-clause aliases for such columns, and "hard" references,
that is actual Vars referring to them. The right fix for hard
references is to add dependencies preventing the drop; something
we've known for many years and not done (and this commit still doesn't
address it). A "soft" reference shouldn't prevent a drop though.
We've been around on this before (cf. 9b35ddce9, 2c4debbd0), but
nobody had noticed that the current behavior can result in dump/reload
failures, because ruleutils.c can print more column aliases than the
underlying composite type now has. So we need to rejigger the
column-alias-handling code to treat such columns as dropped and not
print aliases for them.
Rather than writing new code for this, I used expandRTE() which already
knows how to figure out which function result columns are dropped.
I'd initially thought maybe we could use expandRTE() in all cases, but
that fails for EXPLAIN's purposes, because the planner strips a lot of
RTE infrastructure that expandRTE() needs. So this patch just uses it
for unplanned function RTEs and otherwise does things the old way.
If there is a hard reference (Var), then removing the column alias
causes us to fail to print the Var, since there's no longer a name
to print. Failing seems less desirable than printing a made-up
name, so I made it print "?dropped?column?" instead.
Per report from Timo Stolz. Back-patch to all supported branches.
Discussion: https://postgr.es/m/5c91267e-3b6d-5795-189c-d15a55d61dbb@nullachtvierzehn.de
When a non-exclusive backup is canceled, do_pg_abort_backup() is called
and resets some variables set by pg_backup_start (pg_start_backup in v14
or before). But previously it forgot to reset the session state indicating
whether a non-exclusive backup is in progress or not in this session.
This issue could cause an assertion failure when the session running
BASE_BACKUP is terminated after it executed pg_backup_start and
pg_backup_stop (pg_stop_backup in v14 or before). Also it could cause
a segmentation fault when pg_backup_stop is called after BASE_BACKUP
in the same session is canceled.
This commit fixes the issue by making do_pg_abort_backup reset
that session state.
Back-patch to all supported branches.
Author: Fujii Masao
Reviewed-by: Kyotaro Horiguchi, Masahiko Sawada, Michael Paquier, Robert Haas
Discussion: https://postgr.es/m/3374718f-9fbf-a950-6d66-d973e027f44c@oss.nttdata.com
Multiple non-exclusive backups are able to be run conrrently in different
sessions. But, in the same session, only one non-exclusive backup can be
run at the same moment. If pg_backup_start (pg_start_backup in v14 or before)
is called in the middle of another non-exclusive backup in the same session,
an error is thrown.
However, previously, in logical replication walsender mode, even if that
walsender session had already called pg_backup_start and started
a non-exclusive backup, it could execute BASE_BACKUP command and
start another non-exclusive backup. Which caused subsequent pg_backup_stop
to throw an error because BASE_BACKUP unexpectedly reset the session state
marked by pg_backup_start.
This commit prevents BASE_BACKUP command in the middle of another
non-exclusive backup in the same session.
Back-patch to all supported branches.
Author: Fujii Masao
Reviewed-by: Kyotaro Horiguchi, Masahiko Sawada, Michael Paquier, Robert Haas
Discussion: https://postgr.es/m/3374718f-9fbf-a950-6d66-d973e027f44c@oss.nttdata.com
The patch that added regcollation doesn't seem to have been too
thorough about supporting it everywhere that other reg* types
are supported. Fix that. (The find_expr_references omission
is moderately serious, since it could result in missing expression
dependencies. The others are less exciting.)
Noted while fixing bug #17483. Back-patch to v13 where
regcollation was added.
Discussion: https://postgr.es/m/1423433.1652722406@sss.pgh.pa.us
The motivation for this is to ensure successful transmission of the
values of constants of regconfig and other reg* types. The remote
will be reading them with search_path = 'pg_catalog', so schema
qualification is necessary when referencing objects in other schemas.
Per bug #17483 from Emmanuel Quincerot. Back-patch to all supported
versions. (There's some other stuff to do here, but it's less
back-patchable.)
Discussion: https://postgr.es/m/1423433.1652722406@sss.pgh.pa.us
Commit 4518c798 blocks signals for a short region of code, but it
assumed that whatever called it had the signal mask set to UnBlockSig on
entry. That may be true today (or may even not be, in extensions in the
wild), but it would be better not to make that assumption. We should
save-and-restore the caller's signal mask.
The PG_SETMASK() portability macro couldn't be used for that, which is
why it wasn't done before. But... considering that commit a65e0864
established back in 9.6 that supported POSIX systems have sigprocmask(),
and that this is POSIX-only code, there is no reason not to use standard
sigprocmask() directly to achieve that.
Back-patch to all supported releases, like 4518c798 and 80845b7c.
Reviewed-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/CA%2BhUKGKx6Biq7_UuV0kn9DW%2B8QWcpJC1qwhizdtD9tN-fn0H0g%40mail.gmail.com
On Linux, we call posix_fallocate() on shm_open()'d memory to avoid
later potential SIGBUS (see commit 899bd785).
Based on field reports of systems stuck in an EINTR retry loop there,
there, we made it possible to break out of that loop via slightly odd
coding where the CHECK_FOR_INTERRUPTS() call was somewhat removed from
the loop (see commit 422952ee).
On further reflection, that was not a great choice for at least two
reasons:
1. If interrupts were held, the CHECK_FOR_INTERRUPTS() would do nothing
and the EINTR error would be surfaced to the user.
2. If EINTR was reported but neither QueryCancelPending nor
ProcDiePending was set, then we'd dutifully retry, but with a bit more
understanding of how posix_fallocate() works, it's now clear that you
can get into a loop that never terminates. posix_fallocate() is not a
function that can do some of the job and tell you about progress if it's
interrupted, it has to undo what it's done so far and report EINTR, and
if signals keep arriving faster than it can complete (cf recovery
conflict signals), you're stuck.
Therefore, for now, we'll simply block most signals to guarantee
progress. SIGQUIT is not blocked (see InitPostmasterChild()), because
its expected handler doesn't return, and unblockable signals like
SIGCONT are not expected to arrive at a high rate. For good measure,
we'll include the ftruncate() call in the blocked region, and add a
retry loop.
Back-patch to all supported releases.
Reported-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Reported-by: Nicola Contu <nicola.contu@gmail.com>
Reviewed-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20220701154105.jjfutmngoedgiad3%40alvherre.pgsql