XLogPageRead() checks immediately for an invalid WAL record header on a
standby, to be able to handle the case of continuation records that need
to be read across two different sources. As written, the check was too
generic, applying to any target LSN. Based on an analysis by Kyotaro
Horiguchi, what really matters is to make sure that the page header is
checked when attempting to read a LSN at the boundary of a segment, to
handle the case of a continuation record that spawns across multiple
pages when dealing with multiple segments, as WAL receivers are spawned
they request WAL from the beginning of a segment. This fix has been
proposed by Kyotaro Horiguchi.
This could cause standbys to loop infinitely when dealing with a
continuation record during a timeline jump, in the case where the
contents of the record in the follow-up page are invalid.
Some regression tests are added to check such scenarios, able to
reproduce the original problem. In the test, the contents of a
continuation record are overwritten with junk zeros on its follow-up
page, and replayed on standbys. This is inspired by 039_end_of_wal.pl,
and is enough to show how standbys should react on promotion by not
being stuck. Without the fix, the test would fail with a timeout. The
test to reproduce the problem has been written by Alexander Kukushkin.
The original check has been introduced in 0668719801, for a similar
problem.
Author: Kyotaro Horiguchi, Alexander Kukushkin
Reviewed-by: Michael Paquier
Discussion: https://postgr.es/m/CAFh8B=mozC+e1wGJq0H=0O65goZju+6ab5AU7DEWCSUA2OtwDg@mail.gmail.com
Backpatch-through: 13
This commit reverts 8f67f994e8 (down to v13) and c3de0f9eed (down to
v17), as these are proving to not be completely correct regarding two
aspects:
- In v17 and newer branches, c3de0f9eed38's check for epoch handling is
incorrect, and does not correctly handle frozen epochs. A logic closer
to widen_snapshot_xid() should be used. The 2PC code should try to
integrate deeper with FullTransactionIds, 5a1dfde833 being not enough.
- In v13 and newer branches, 8f67f994e8 is a workaround for the real
issue, which is that we should not attempt CLOG lookups without reaching
consistency. This exists since 728bd991c3, and this is reachable with
ProcessTwoPhaseBuffer() called by restoreTwoPhaseData() at the beginning
of recovery.
Per discussion with Noah Misch.
Discussion: https://postgr.es/m/20250116010051.f3.nmisch@google.com
Backpatch-through: 13
We should run the expression subtrees of PartitionedRelPruneInfo
structs through fix_scan_expr. Failure to do so means that
AlternativeSubPlans within those expressions won't be cleaned up
properly, resulting in "unrecognized node type" errors since v14.
It seems fairly likely that at least some of the other steps done
by fix_scan_expr are important here as well, resulting in as-yet-
undetected bugs. Therefore, I've chosen to back-patch this to
all supported branches including v13, even though the known
symptom doesn't manifest in v13.
Per bug #18778 from Alexander Lakhin.
Discussion: https://postgr.es/m/18778-24cd399df6c806af@postgresql.org
If a new catalog tuple is inserted that belongs to a catcache list
entry, and cache invalidation happens while the list entry is being
built, the list entry might miss the newly inserted tuple.
To fix, change the way we detect concurrent invalidations while a
catcache entry is being built. Keep a stack of entries that are being
built, and apply cache invalidation to those entries in addition to
the real catcache entries. This is similar to the in-progress list in
relcache.c.
Back-patch to all supported versions.
Reviewed-by: Noah Misch
Discussion: https://www.postgresql.org/message-id/2234dc98-06fe-42ed-b5db-ac17384dc880@iki.fi
This function expects an "int64" as result and stores the number of
pages to add to the index scan bitmap as an "int", multiplying its final
result by 10. For a relation large enough, this can theoretically
overflow if counting more than (INT32_MAX / 10) pages, knowing that the
number of pages is upper-bounded by MaxBlockNumber.
To avoid the overflow, this commit redefines "totalpages", used to
calculate the result, to be an "int64" rather than an "int".
Reported-by: Evgeniy Gorbanyov
Author: James Hunter
Discussion: https://www.postgresql.org/message-id/07704817-6fa0-460c-b1cf-cd18f7647041@basealt.ru
Backpatch-through: 13
Commit 27a1f8d108 missed updating the max HBA option count to
account for the new option added. Fix by bumping the counter
and adjust the relevant comment to match. Backpatch down to
all supported branches like the erroneous commit.
Reported-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/286764.1736697356@sss.pgh.pa.us
Backpatch-through: v13
When deparsing an XMLTABLE() expression, XML namespace names were not
quoted. However, since they are parsed as ColLabel tokens, some names
require double quotes to ensure that they are properly interpreted.
Fix by using quote_identifier() in the deparsing code.
Back-patch to all supported versions.
Dean Rasheed, reviewed by Tom Lane.
Discussion: https://postgr.es/m/CAEZATCXTpAS%3DncfLNTZ7YS6O5puHeLg_SUYAit%2Bcs7wsrd9Msg%40mail.gmail.com
Commit 66aaabe7 (branches 13 - 17 only) was not acceptable to the Oracle
Developer Studio compiler on build farm animal wrasse. It accidentally
used a C++ style return statement to wrap a void function. None of the
usual compilers complained, but it is right, that is not allowed in C.
Fix.
Reported-by: Michael Paquier <michael@paquier.xyz>
Discussion: https://postgr.es/m/Z33vgfVgvOnbFLN9%40paquier.xyz
It's possible that external code is calling smgrtruncate(). Any
external callers might like to consider the recent changes to
RelationTruncate(), but commit 38c579b0 should not have changed the
function prototype in the back-branches, per ABI stability policy.
Restore smgrtruncate()'s traditional argument list in the back-branches,
but make it a wrapper for a new function smgrtruncate2(). The three
callers in core can use smgrtruncate2() directly. In master (18-to-be),
smgrtruncate2() is effectively renamed to smgrtruncate(), so this wart
is cleaned up.
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Discussion: https://postgr.es/m/CA%2BhUKG%2BThae6x6%2BjmQiuALQBT2Ae1ChjMh1%3DkMvJ8y_SBJZrvA%40mail.gmail.com
pgoutput caches the attribute map of a relation, that is free()'d only
when validating a RelationSyncEntry. However, this code path is not
taken when calling any of the SQL functions able to do some logical
decoding, like pg_logical_slot_{get,peek}_changes(), leaking some memory
into CacheMemoryContext on repeated calls.
To address this, a relation's attribute map is allocated in
PGOutputData's cachectx, free()'d at the end of the execution of these
SQL functions when logical decoding ends. This is available down to 15.
v13 and v14 have a similar leak, which will be dealt with later.
Reported-by: Masahiko Sawada
Author: Vignesh C
Reviewed-by: Hou Zhijie
Discussion: https://postgr.es/m/CAD21AoDkAhQVSukOfH3_reuF-j4EU0-HxMqU3dU+bSTxsqT14Q@mail.gmail.com
Discussion: https://postgr.es/m/CALDaNm1hewNAsZ_e6FF52a=9drmkRJxtEPrzCB6-9mkJyeBBqA@mail.gmail.com
Backpatch-through: 15
Before 728bd991c3, that has improved the support for 2PC files during
recovery, the initial logic scanning files in pg_twophase was done so as
files in the future of the transaction ID horizon were checked first,
followed by a check if a transaction ID is aborted or committed which
could involve a pg_xact lookup. After this commit, these checks have
been done in reverse order.
Files detected as in the future do not have a state that can be checked
in pg_xact, hence this caused recovery to fail abruptly should an
orphaned 2PC file in the future of the transaction ID horizon exist in
pg_twophase at the beginning of recovery.
A test is added to check for this scenario, using an empty 2PC with a
transaction ID large enough to be in the future when running the test.
This test is added in 16 and older versions for now. 17 and newer
versions are impacted by a second bug caused by the addition of the
epoch in the 2PC file names. An equivalent test will be added in these
branches in a follow-up commit, once the second set of issues reported
are fixed.
Author: Vitaly Davydov, Michael Paquier
Discussion: https://postgr.es/m/11e597-676ab680-8d-374f23c0@145466129
Backpatch-through: 13
Cause parallel workers to not check datallowconn, rolcanlogin, and
ACL_CONNECT privileges. The leader already checked these things
(except for rolcanlogin which might have been checked for a different
role). Re-checking can accomplish little except to induce unexpected
failures in applications that might not even be aware that their query
has been parallelized. We already had the principle that parallel
workers rely on their leader to pass a valid set of authorization
information, so this change just extends that a bit further.
Also, modify the ReservedConnections, datconnlimit and rolconnlimit
logic so that these limits are only enforced against regular backends,
and only regular backends are counted while checking if the limits
were already reached. Previously, background processes that had an
assigned database or role were subject to these limits (with rather
random exclusions for autovac workers and walsenders), and the set of
existing processes that counted against each limit was quite haphazard
as well. The point of these limits, AFAICS, is to ensure the
availability of PGPROC slots for regular backends. Since all other
types of processes have their own separate pools of PGPROC slots, it
makes no sense either to enforce these limits against them or to count
them while enforcing the limit.
While edge-case failures of these sorts have been possible for a
long time, the problem got a good deal worse with commit 5a2fed911
(CVE-2024-10978), which caused parallel workers to make some of these
checks using the leader's current role where before we had used its
AuthenticatedUserId, thus allowing parallel queries to fail after
SET ROLE. The previous behavior was fairly accidental and I have
no desire to return to it.
This patch includes reverting 73c9f91a1, which was an emergency hack
to suppress these same checks in some cases. It wasn't complete,
as shown by a recent bug report from Laurenz Albe. We can also revert
fd4d93d26 and 492217301, which hacked around the same problems in one
regression test.
In passing, remove the special case for autovac workers in
CheckMyDatabase; it seems cleaner to have AutoVacWorkerMain pass
the INIT_PG_OVERRIDE_ALLOW_CONNS flag, now that that does what's
needed.
Like 5a2fed911, back-patch to supported branches (which sadly no
longer includes v12).
Discussion: https://postgr.es/m/1808397.1735156190@sss.pgh.pa.us
Commit aac2c9b4fd mandated such locking
and attempted to fulfill that mandate, but it missed REASSIGN OWNED.
Hence, it remained possible to lose VACUUM's inplace update of
datfrozenxid if a REASSIGN OWNED processed that database at the same
time. This didn't affect the other inplace-updated catalog, pg_class.
For pg_class, REASSIGN OWNED calls ATExecChangeOwner() instead of the
generic AlterObjectOwner_internal(), and ATExecChangeOwner() fulfills
the locking mandate.
Like in GRANT, implement this by following the locking protocol for any
catalog subject to the generic AlterObjectOwner_internal(). It would
suffice to do this for IsInplaceUpdateOid() catalogs only. Back-patch
to v13 (all supported versions).
Kirill Reshke. Reported by Alexander Kukushkin.
Discussion: https://postgr.es/m/CAFh8B=mpKjAy4Cuun-HP-f_vRzh2HSvYFG3rhVfYbfEBUhBAGg@mail.gmail.com
The pgoutput module caches publication names in a list and frees it upon
invalidation. However, the code forgot to free the actual publication
names within the list elements, as publication names are pstrdup()'d in
GetPublication(). This would cause memory to leak in
CacheMemoryContext, bloating it over time as this context is not
cleaned.
This is a problem for WAL senders running for a long time, as an
accumulation of invalidation requests would bloat its cache memory
usage. A second case, where this leak is easier to see, involves a
backend calling SQL functions like pg_logical_slot_{get,peek}_changes()
which create a new decoding context with each execution. More
publications create more bloat.
To address this, this commit adds a new memory context within the
logical decoding context and resets it each time the publication names
cache is invalidated, based on a suggestion from Amit Kapila. This
ensures that the lifespan of the publication names aligns with that of
the logical decoding context.
Contrary to the HEAD-only commit f0c569d715 that has changed
PGOutputData to track this new child memory context, the context is
tracked with a static variable whose state is reset with a MemoryContext
reset callback attached to PGOutputData->context, so as ABI
compatibility is preserved in stable branches. This approach is based
on an suggestion from Amit Kapila.
Analyzed-by: Michael Paquier, Jeff Davis
Author: Masahiko Sawada
Reviewed-by: Amit Kapila, Michael Paquier, Euler Taveira, Hou Zhijie
Discussion: https://postgr.es/m/Z0khf9EVMVLOc_YY@paquier.xyz
Backpatch-through: 13
GetSnapshotData() set TransactionXmin = MyProc->xmin, but when
SnapshotResetXmin() advanced MyProc->xmin, it did not advance
TransactionXmin correspondingly. That meant that TransactionXmin could
be older than MyProc->xmin, and XIDs between than TransactionXmin and
the real MyProc->xmin could be vacuumed away. One known consequence is
in pg_subtrans lookups: we might try to look up the status of an XID
that was already truncated away.
Back-patch to all supported versions.
Reviewed-by: Andres Freund
Discussion: https://www.postgresql.org/message-id/d27a046d-a1e4-47d1-a95c-fbabe41debb4@iki.fi
RelationTruncate() does three things, while holding an
AccessExclusiveLock and preventing checkpoints:
1. Logs the truncation.
2. Drops buffers, even if they're dirty.
3. Truncates some number of files.
Step 2 could previously be canceled if it had to wait for I/O, and step
3 could and still can fail in file APIs. All orderings of these
operations have data corruption hazards if interrupted, so we can't give
up until the whole operation is done. When dirty pages were discarded
but the corresponding blocks were left on disk due to ERROR, old page
versions could come back from disk, reviving deleted data (see
pgsql-bugs #18146 and several like it). When primary and standby were
allowed to disagree on relation size, standbys could panic (see
pgsql-bugs #18426) or revive data unknown to visibility management on
the primary (theorized).
Changes:
* WAL is now unconditionally flushed first
* smgrtruncate() is now called in a critical section, preventing
interrupts and causing PANIC on file API failure
* smgrtruncate() has a new parameter for existing fork sizes,
because it can't call smgrnblocks() itself inside a critical section
The changes apply to RelationTruncate(), smgr_redo() and
pg_truncate_visibility_map(). That last is also brought up to date with
other evolutions of the truncation protocol.
The VACUUM FileTruncate() failure mode had been discussed in older
reports than the ones referenced below, with independent analysis from
many people, but earlier theories on how to fix it were too complicated
to back-patch. The more recently invented cancellation bug was
diagnosed by Alexander Lakhin. Other corruption scenarios were spotted
by me while iterating on this patch and earlier commit 75818b3a.
Back-patch to all supported releases.
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Reviewed-by: Robert Haas <robertmhaas@gmail.com>
Reported-by: rootcause000@gmail.com
Reported-by: Alexander Lakhin <exclusion@gmail.com>
Discussion: https://postgr.es/m/18146-04e908c662113ad5%40postgresql.org
Discussion: https://postgr.es/m/18426-2d18da6586f152d6%40postgresql.org
durable_rename_excl() attempts to avoid overwriting any existing files
by using link() and unlink(), and it falls back to rename() on some
platforms (aka WIN32), which offers no such overwrite protection. Most
callers use durable_rename_excl() just in case there is an existing
file, but in practice there shouldn't be one (see below for more
details).
Furthermore, failures during durable_rename_excl() can result in
multiple hard links to the same file. As per Nathan's tests, it is
possible to end up with two links to the same file in pg_wal after a
crash just before unlink() during WAL recycling. Specifically, the test
produced links to the same file for the current WAL file and the next
one because the half-recycled WAL file was re-recycled upon restarting,
leading to WAL corruption.
This change replaces all the calls of durable_rename_excl() to
durable_rename(). This removes the protection against accidentally
overwriting an existing file, but some platforms are already living
without it and ordinarily there shouldn't be one. The function itself
is left around in case any extensions are using it. It will be removed
on HEAD via a follow-up commit.
Here is a summary of the existing callers of durable_rename_excl() (see
second discussion link at the bottom), replaced by this commit. First,
basic_archive used it to avoid overwriting an archive concurrently
created by another server, but as mentioned above, it will still
overwrite files on some platforms. Second, xlog.c uses it to recycle
past WAL segments, where an overwrite should not happen (origin of the
change at f0e37a8) because there are protections about the WAL segment
to select when recycling an entry. The third and last area is related
to the write of timeline history files. writeTimeLineHistory() will
write a new timeline history file at the end of recovery on promotion,
so there should be no such files for the same timeline.
What remains is writeTimeLineHistoryFile(), that can be used in parallel
by a WAL receiver and the startup process, and some digging of the
buildfarm shows that EEXIST from a WAL receiver can happen with an error
of "could not link file \"pg_wal/xlogtemp.NN\" to \"pg_wal/MM.history\",
which would cause an automatic restart of the WAL receiver as it is
promoted to FATAL, hence this should improve the stability of the WAL
receiver as rename() would overwrite an existing TLI history file
already fetched by the startup process at recovery.
This is the second time this change is attempted, ccfbd92 being the
first one, but this time no assertions are added for the case of a TLI
history file written concurrently by the WAL receiver or the startup
process because we can expect one to exist (some of the TAP tests are
able to trigger with a proper timing).
This commit has been originally applied on v16~ as of dac1ff3090, and
we have received more reports of this issue, where clusters can become
corrupted at replay in older stable branches with multiple links
pointing to the same physical WAL segment file. This backpatch
addresses the problem for the v13~v15 range.
Author: Nathan Bossart
Reviewed-by: Robert Haas, Kyotaro Horiguchi, Michael Paquier
Discussion: https://postgr.es/m/20220407182954.GA1231544@nathanxps13
Discussion: https://postgr.es/m/Ym6GZbqQdlalSKSG@paquier.xyz
Discussion: https://postgr.es/m/CAJhEC04tBkYPF4q2uS_rCytauvNEVqdBAzasBEokfceFhF=KDQ@mail.gmail.com
If the non-recursive part of a recursive CTE ended up using
TTSOpsBufferHeapTuple as the table slot type, then a duplicate value
could cause an Assert failure in CheckOpSlotCompatibility() when
checking the hash table for the duplicate value. The expected slot type
for the deform step was TTSOpsMinimalTuple so the Assert failed when the
TTSOpsBufferHeapTuple slot was used.
This is a long-standing bug which we likely didn't notice because it
seems much more likely that the non-recursive term would have required
projection and used a TTSOpsVirtual slot, which CheckOpSlotCompatibility
is ok with.
There doesn't seem to be any harm done here other than the Assert
failure. Both TTSOpsMinimalTuple and TTSOpsBufferHeapTuple slot types
require tuple deformation, so the EEOP_*_FETCHSOME ExprState step would
have properly existed in the ExprState.
The solution is to pass NULL for the ExecBuildGroupingEqual's 'lops'
parameter. This means the ExprState's EEOP_*_FETCHSOME step won't
expect a fixed slot type. This makes CheckOpSlotCompatibility() happy as
no checking is performed when the ExprEvalStep is not expecting a fixed
slot type.
Reported-by: Richard Guo
Reviewed-by: Tom Lane
Discussion: https://postgr.es/m/CAMbWs4-8U9q2LAtf8+ghV11zeUReA3AmrYkxzBEv0vKnDxwkKA@mail.gmail.com
Backpatch-through: 13, all supported versions
If a dshash table grows very large (e.g., the dshash table for
cumulative statistics when there are millions of tables), resizing
it may fail with an error like:
ERROR: invalid DSA memory alloc request size 1073741824
To fix, permit dshash resizing to allocate more than 1 GB by
providing the DSA_ALLOC_HUGE flag.
Reported-by: Andreas Scherbaum
Author: Matthias van de Meent
Reviewed-by: Cédric Villemain, Michael Paquier, Andres Freund
Discussion: https://postgr.es/m/80a12d59-0d5e-4c54-866c-e69cd6536471%40pgug.de
Backpatch-through: 13
Since commit 97550c0711, these failed with
"PANIC: proc_exit() called in child process" due to uninitialized or
stale MyProcPid. That was reachable if close() failed in
ClosePostmasterPorts() or setlocale(category, "C") failed, both
unlikely. Back-patch to v13 (all supported versions).
Discussion: https://postgr.es/m/20241208034614.45.nmisch@google.com
Our parallel-mode code only works when we are executing a query
in full, so ExecutePlan must disable parallel mode when it is
asked to do partial execution. The previous logic for this
involved passing down a flag (variously named execute_once or
run_once) from callers of ExecutorRun or PortalRun. This is
overcomplicated, and unsurprisingly some of the callers didn't
get it right, since it requires keeping state that not all of
them have handy; not to mention that the requirements for it were
undocumented. That led to assertion failures in some corner
cases. The only state we really need for this is the existing
QueryDesc.already_executed flag, so let's just put all the
responsibility in ExecutePlan. (It could have been done in
ExecutorRun too, leading to a slightly shorter patch -- but if
there's ever more than one caller of ExecutePlan, it seems better
to have this logic in the subroutine than the callers.)
This makes those ExecutorRun/PortalRun parameters unnecessary.
In master it seems okay to just remove them, returning the
API for those functions to what it was before parallelism.
Such an API break is clearly not okay in stable branches,
but for them we can just leave the parameters in place after
documenting that they do nothing.
Per report from Yugo Nagata, who also reviewed and tested
this patch. Back-patch to all supported branches.
Discussion: https://postgr.es/m/20241206062549.710dc01cf91224809dd6c0e1@sraoss.co.jp
pgstat_write_statsfile() discards any entries marked as dropped from
being written to the stats file at shutdown, and also included an
assertion based on the same condition.
The intention of the assertion is to track that no pgstats entries
should be left around as terminating backends should drop any entries
they still hold references on before the stats file is written by the
checkpointer, and it not worth taking down the server in this case if
there is a bug making that possible.
Let's improve the comment of this area to document clearly what's
intended.
Based on a discussion with Bertrand Drouvot and Anton A. Melnikov.
Author: Bertrand Drouvot
Discussion: https://postgr.es/m/a13e8cdf-b97a-4ecb-8f42-aaa367974e29@postgrespro.ru
Backpatch-through: 15
818119afcc has introduced the "generation" concept in pgstats entries,
incremented a counter when a pgstats entry is reinitialized, but it did
not count on the fact that backends still holding local references to
such entries need to be refreshed if the cache age is outdated. The
previous logic only updated local references when an entry was dropped,
but it needs also to consider entries that are reinitialized.
This matters for replication slot stats (as well as custom pgstats kinds
in 18~), where concurrent drops and creates of a slot could cause
incorrect stats to be locally referenced. This would lead to an
assertion failure at shutdown when writing out the stats file, as the
backend holding an outdated local reference would not be able to drop
during its shutdown sequence the stats entry that should be dropped, as
the last process holding a reference to the stats entry. The
checkpointer was then complaining about such an entry late in the
shutdown sequence, after the shutdown checkpoint is finished with the
control file updated, causing the stats file to not be generated. In
non-assert builds, the entry would just be skipped with the stats file
written.
Note that only logical replication slots use statistics.
A test case based on TAP is added to test_decoding, where a persistent
connection peeking at a slot's data is kept with concurrent drops and
creates of the same slot. This is based on the isolation test case that
Anton has sent. As it requires a node shutdown with a check to make
sure that the stats file is written with this specific sequence of
events, TAP is used instead.
Reported-by: Anton A. Melnikov
Reviewed-by: Bertrand Drouvot
Discussion: https://postgr.es/m/56bf8ff9-dd8c-47b2-872a-748ede82af99@postgrespro.ru
Backpatch-through: 15
When short-circuiting WindowAgg node evaluation on the top-level
WindowAgg node using quals on monotonic window functions, because the
WindowAgg run condition can mean there's no need to evaluate subsequent
window function results in the same partition once the run condition
becomes false, it was possible that the executor would use stale results
from the previous invocation of the window function in some cases.
A fix for this was partially done by a5832722, but that commit only
fixed the issue for non-top-level WindowAgg nodes. I mistakenly thought
that the top-level WindowAgg didn't have this issue, but Jayesh's example
case clearly shows that's incorrect. At the time, I also thought that
this only affected 32-bit systems as all window functions which then
supported run conditions returned BIGINT, however, that's wrong as
ExecProject is still called and that could cause evaluation of any other
window function belonging to the same WindowAgg node, one of which may
return a byref type.
The only queries affected by this are WindowAggs with a "Run Condition"
which contains at least one window function with a byref result type,
such as lead() or lag() on a byref column. The window clause must also
contain a PARTITION BY clause (without a PARTITION BY, execution of the
WindowAgg stops immediately when the run condition becomes false and
there's no risk of using the stale results).
Reported-by: Jayesh Dehankar
Discussion: https://postgr.es/m/193261e2c4d.3dd3cd7c1842.871636075166132237@zohocorp.com
Backpatch-through: 15, where WindowAgg run conditions were added
Usually an entry in pg_amop or pg_amproc does not need a dependency on
its amoplefttype/amoprighttype/amproclefttype/amprocrighttype types,
because there is an indirect dependency via the argument types of its
referenced operator or procedure, or via the opclass it belongs to.
However, for some support procedures in some index AMs, the argument
types of the support procedure might not mention the column data type
at all. Also, the amop/amproc entry might be treated as "loose" in
the opfamily, in which case it lacks a dependency on any particular
opclass; or it might be a cross-type entry having a reference to a
datatype that is not its opclass' opcintype.
The upshot of all this is that there are cases where a datatype can
be dropped while leaving behind amop/amproc entries that mention it,
because there is no path in pg_depend showing that those entries
depend on that type. Such entries are harmless in normal activity,
because they won't get used, but they cause problems for maintenance
actions such as dropping the operator family. They also cause pg_dump
to produce bogus output. The previous commit put a band-aid on the
DROP OPERATOR FAMILY failure, but a real fix is needed.
To fix, add pg_depend entries showing that a pg_amop/pg_amproc entry
depends on its lefttype/righttype. To avoid bloating pg_depend too
much, skip this if the referenced operator or function has that type
as an input type. (I did not bother with considering the possible
indirect dependency via the opclass' opcintype; at least in the
reported case, that wouldn't help anyway.)
Probably, the reason this has escaped notice for so long is that
add-on datatypes and relevant opclasses/opfamilies are usually
packaged as extensions nowadays, so that there's no way to drop
a type without dropping the referencing opclasses/opfamilies too.
Still, in the absence of pg_depend entries there's nothing that
constrains DROP EXTENSION to drop the opfamily entries before the
datatype, so it seems possible for a DROP failure to occur anyway.
The specific case that was reported doesn't fail in v13, because
v13 prefers to attach the support procedure to the opclass not the
opfamily. But it's surely possible to construct other edge cases
that do fail in v13, so patch that too.
Per report from Yoran Heling. Back-patch to all supported branches.
Discussion: https://postgr.es/m/Z1MVCOh1hprjK5Sf@gmai021
Yoran Heling reported a case where a data type could be dropped
while references to its OID remain behind in pg_amproc. This
causes getObjectDescription to fail, which blocks dropping the
operator family (since our DROP code likes to construct descriptions
of everything it's dropping). The proper fix for this requires
adding more pg_depend entries. But to allow DROP to go through with
already-corrupt catalogs, tweak getObjectDescription to print "???"
for the type instead of failing when it processes such an entry.
I changed the logic for pg_amop similarly, for consistency,
although it is not known that the problem can manifest in pg_amop.
Per report from Yoran Heling. Back-patch to all supported
branches (although the problem may be unreachable in v13).
Discussion: https://postgr.es/m/Z1MVCOh1hprjK5Sf@gmai021
These format codes produce or consume strings of digits, so they
should be labeled with is_digit = true, but they were not.
This has effect in only one place, where is_next_separator()
is checked to see if the preceding format code should slurp up
all the available digits. Thus, with a format such as '...SSFF3'
with remaining input '12345', the 'SS' code would consume all
five digits (and then complain about seconds being out of range)
when it should eat only two digits.
Per report from Nick Davies. This bug goes back to d589f9446
where the FFn codes were introduced, so back-patch to v13.
Discussion: https://postgr.es/m/AM8PR08MB6356AC979252CFEA78B56678B6312@AM8PR08MB6356.eurprd08.prod.outlook.com
Previously, it set only DELAY_CHKPT_COMPLETE. That was important,
because it meant that if the XLOG_SMGR_TRUNCATE record preceded a
XLOG_CHECKPOINT_ONLINE record in the WAL, then the truncation would also
happen on disk before the XLOG_CHECKPOINT_ONLINE record was
written.
However, it didn't guarantee that the sync request for the truncation
was processed before the XLOG_CHECKPOINT_ONLINE record was written. By
setting DELAY_CHKPT_START, we guarantee that if an XLOG_SMGR_TRUNCATE
record is written to WAL before the redo pointer of a concurrent
checkpoint, the sync request queued by that operation must be processed
by that checkpoint, rather than being left for the following one.
This is a refinement of commit 412ad7a556. Back-patch to all supported
releases, like that commit.
Author: Robert Haas <robertmhaas@gmail.com>
Reported-by: Thomas Munro <thomas.munro@gmail.com>
Discussion: https://postgr.es/m/CA%2BhUKG%2B-2rjGZC2kwqr2NMLBcEBp4uf59QT1advbWYF_uc%2B0Aw%40mail.gmail.com
Commit 352f6f2d used %d instead of %lu to format DWORD (unsigned long)
with psprintf(). The _WIN32_WINNT value recently changed for MinGW in
REL_15_STABLE (commit d700e8d7), so the code was suddenly being
compiled, with warnings from gcc.
The warnings were already fixed in 16+ by commits 495ed0ef and a9bc04b2
after the _WIN32_WINNT value was increase there. 14 and 13 didn't warn
because they still use a lower value for MinGW, and supported versions
of Visual Studio should compile the code in all live branches but don't
check our format string.
The change doesn't affect the result: sizeof(int) == sizeof(long) on
this platform, and the values are computed with expressions that cannot
exceed INT_MAX so were never interpreted as negative.
Back-patch the formatting change from those commits into 13-15. This
should turn CI's 15 branch green again and stop fairywren from warning
about that on 15.
Reported-by: Andres Freund <andres@anarazel.de>
Reported-by: Peter Eisentraut <peter@eisentraut.org>
Discussion: https://postgr.es/m/t2vjrcb3bloxf5qqvxjst6r7lvrefqyecxgt2koy5ho5b5glr2%40yuupmm6whgob
When using a pipeline, a transaction starts from the first command and
is committed with a Sync message or when the pipeline ends.
Functions like IsInTransactionBlock() or PreventInTransactionBlock()
were already able to understand a pipeline as being in a transaction
block, but it was not the case of CheckTransactionBlock(). This
function is called for example to generate a WARNING for SET LOCAL,
complaining that it is used outside of a transaction block.
The current state of the code caused multiple problems, like:
- SET LOCAL executed at any stage of a pipeline issued a WARNING, even
if the command was at least second in line where the pipeline is in a
transaction state.
- LOCK TABLE failed when invoked at any step of a pipeline, even if it
should be able to work within a transaction block.
The pipeline protocol assumes that the first command of a pipeline is
not part of a transaction block, and that any follow-up commands is
considered as within a transaction block.
This commit changes the backend so as an implicit transaction block is
started each time the first Execute message of a pipeline has finished
processing, with this implicit transaction block ended once a sync is
processed. The checks based on XACT_FLAGS_PIPELINING in the routines
checking if we are in a transaction block are not necessary: it is
enough to rely on the existing ones.
Some tests are added to pgbench, that can be backpatched down to v17
when \syncpipeline is involved and down to v14 where \startpipeline and
\endpipeline are available. This is unfortunately limited regarding the
error patterns that can be checked, but it provides coverage for various
pipeline combinations to check if these succeed or fail. These tests
are able to capture the case of SET LOCAL's WARNING. The author has
proposed a different feature to improve the coverage by adding similar
meta-commands to psql where error messages could be checked, something
more useful for the cases where commands cannot be used in transaction
blocks, like REINDEX CONCURRENTLY or VACUUM. This is considered as
future work for v18~.
Author: Anthonin Bonnefoy
Reviewed-by: Jelte Fennema-Nio, Michael Paquier
Discussion: https://postgr.es/m/CAO6_XqrWO8uNBQrSu5r6jh+vTGi5Oiyk4y8yXDORdE2jbzw8xw@mail.gmail.com
Backpatch-through: 13
If passed a read-write expanded object pointer, the EEOP_NULLIF
code would hand that same pointer to the equality function
and then (unless equality was reported) also return the same
pointer as its value. This is no good, because a function that
receives a read-write expanded object pointer is fully entitled
to scribble on or even delete the object, thus corrupting the
NULLIF output. (This problem is likely unobservable with the
equality functions provided in core Postgres, but it's easy to
demonstrate with one coded in plpgsql.)
To fix, make sure the pointer passed to the equality function
is read-only. We can still return the original read-write
pointer as the NULLIF result, allowing optimization of later
operations.
Per bug #18722 from Alexander Lakhin. This has been wrong
since we invented expanded objects, so back-patch to all
supported branches.
Discussion: https://postgr.es/m/18722-fd9e645448cc78b4@postgresql.org
This WARNING appeared because SearchSysCacheLocked1() read
cc_relisshared before catcache initialization, when the field is false
unconditionally. On the basis of reading false there, it constructed a
locktag as though pg_tablespace weren't relisshared. Only shared
catalogs could be affected, and only GRANT TABLESPACE was affected in
practice. SearchSysCacheLocked1() callers use one other shared-relation
syscache, DATABASEOID. DATABASEOID is initialized by the end of
CheckMyDatabase(), making the problem unreachable for pg_database.
Back-patch to v13 (all supported versions). This has no known impact
before v16, where ExecGrant_common() first appeared. Earlier branches
avoid trouble by having a separate ExecGrant_Tablespace() that doesn't
use LOCKTAG_TUPLE. However, leaving this unfixed in v15 could ensnare a
future back-patch of a SearchSysCacheLocked1() call.
Reported by Aya Iwata.
Discussion: https://postgr.es/m/OS7PR01MB11964507B5548245A7EE54E70EA212@OS7PR01MB11964.jpnprd01.prod.outlook.com
RelationSyncCache, the hash table in charge of tracking the relation
schemas sent through pgoutput, was forgetting to free the TupleDesc
associated to the two slots used to store the new and old tuples,
causing some memory to be leaked each time a relation is invalidated
when the slots of an existing relation entry are cleaned up.
This is rather hard to notice as the bloat is pretty minimal, but a
long-running WAL sender would be in trouble over time depending on the
workload. sysbench has proved to be pretty good at showing the problem,
coupled with some memory monitoring of the WAL sender.
Issue introduced in 52e4f0cd47, that has added row filters for tables
logically replicated.
Author: Boyu Yang
Reviewed-by: Michael Paquier, Hou Zhijie
Discussion: https://postgr.es/m/DM3PR84MB3442E14B340E553313B5C816E3252@DM3PR84MB3442.NAMPRD84.PROD.OUTLOOK.COM
Backpatch-through: 15
Ordinarily transformSetOperationTree will collect all UNION/
INTERSECT/EXCEPT steps into the setOperations tree of the topmost
Query, so that leaf queries do not contain any setOperations.
However, it cannot thus flatten a subquery that also contains
WITH, ORDER BY, FOR UPDATE, or LIMIT. I (tgl) forgot that in
commit 07b4c48b6 and wrote an assertion in rule deparsing that
a leaf's setOperations would always be empty.
If it were nonempty then we would want to parenthesize the subquery
to ensure that the output represents the setop nesting correctly
(e.g. UNION below INTERSECT had better get parenthesized). So
rather than just removing the faulty Assert, let's change it into
an additional case to check to decide whether to add parens. We
don't expect that the additional case will ever fire, but it's
cheap insurance.
Man Zeng and Tom Lane
Discussion: https://postgr.es/m/tencent_7ABF9B1F23B0C77606FC5FE3@qq.com
In the dim past we figured it was okay to ignore collations
when combining UNION set-operation nodes into a single N-way
UNION operation. I believe that was fine at the time, but
it stopped being fine when we added nondeterministic collations:
the semantics of distinct-ness are affected by those. v17 made
it even less fine by allowing per-child sorting operations to
be merged via MergeAppend, although I think we accidentally
avoided any live bug from that.
Add a check that collations match before deciding that two
UNION nodes are equivalent. I also failed to resist the
temptation to comment plan_union_children() a little better.
Back-patch to all supported branches (v13 now), since they
all have nondeterministic collations.
Discussion: https://postgr.es/m/3605568.1731970579@sss.pgh.pa.us
fixempties() counts the number of in-arcs in the regex NFA and then
allocates an array of that many arc pointers. If the NFA contains no
arcs, this amounts to malloc(0) for which some platforms return NULL.
The code mistakenly treats that as indicating out-of-memory. Thus,
we can get a bogus "out of memory" failure for some unsatisfiable
regexes.
This happens only in v15 and earlier, since bea3d7e38 switched to
using palloc() rather than bare malloc(). And at least of the
platforms in the buildfarm, only AIX seems to return NULL. So the
impact is pretty narrow. But I don't especially want to ship code
that is failing its own regression tests, so let's fix this for
this week's releases.
A quick code survey says that there is only the one place in
src/backend/regex/ that is at risk of doing malloc(0), so we'll just
band-aid that place. A more future-proof fix could be to install a
malloc() wrapper similar to pg_malloc(). But this code seems unlikely
to change much more in the affected branches, so that's probably
overkill.
The only known test case for this involves a complemented character
class in a bracket expression, for example [^\s\S], and we didn't
support that in v13. So it may be that the problem is unreachable
in v13. But I'm not 100% sure of that, so patch v13 too.
Discussion: https://postgr.es/m/661fd81b-f069-8f30-1a41-e195c57449b4@gmail.com
After commit 5a2fed911a, the catalog state
resulting from these commands ceased to affect sessions. Restore the
longstanding behavior, which is like beginning the session with a SET
ROLE command. If cherry-picking the CVE-2024-10978 fixes, default to
including this, too. (This fixes an unintended side effect of fixing
CVE-2024-10978.) Back-patch to v12, like that commit. The release team
decided to include v12, despite the original intent to halt v12 commits
earlier this week.
Tom Lane and Noah Misch. Reported by Etienne LAFARGE.
Discussion: https://postgr.es/m/CADOZwSb0UsEr4_UTFXC5k7=fyyK8uKXekucd+-uuGjJsGBfxgw@mail.gmail.com
Previously LogicalIncreaseRestartDecodingForSlot() accidentally
accepted any LSN as the candidate_lsn and candidate_valid after the
restart_lsn of the replication slot was updated, so it potentially
caused the restart_lsn to move backwards.
A scenario where this could happen in logical replication is: after a
logical replication restart, based on previous candidate_lsn and
candidate_valid values in memory, the restart_lsn advances upon
receiving a subscriber acknowledgment. Then, logical decoding restarts
from an older point, setting candidate_lsn and candidate_valid based
on an old RUNNING_XACTS record. Subsequent subscriber acknowledgments
then update the restart_lsn to an LSN older than the current value.
In the reported case, after WAL files were removed by a checkpoint,
the retreated restart_lsn prevented logical replication from
restarting due to missing WAL segments.
This change essentially modifies the 'if' condition to 'else if'
condition within the function. The previous code had an asymmetry in
this regard compared to LogicalIncreaseXminForSlot(), which does
almost the same thing for different fields.
The WAL removal issue was reported by Hubert Depesz Lubaczewski.
Backpatch to all supported versions, since the bug exists since 9.4
where logical decoding was introduced.
Reviewed-by: Tomas Vondra, Ashutosh Bapat, Amit Kapila
Discussion: https://postgr.es/m/Yz2hivgyjS1RfMKs%40depesz.com
Discussion: https://postgr.es/m/85fff40e-148b-4e86-b921-b4b846289132%40vondra.me
Backpatch-through: 13
In commit 08c0d6ad6 which introduced "rainbow" arcs in regex NFAs,
I didn't think terribly hard about what to do when creating the color
complement of a rainbow arc. Clearly, the complement cannot match any
characters, and I took the easy way out by just not building any arcs
at all in the complement arc set. That mostly works, but Nikolay
Shaplov found a case where it doesn't: if we decide to delete that
sub-NFA later because it's inside a "{0}" quantifier, delsub()
suffered an assertion failure. That's because delsub() relies on
the target sub-NFA being fully connected. That was always true
before, and the best fix seems to be to restore that property.
Hence, invent a new arc type CANTMATCH that can be generated in
place of an empty color complement, and drop it again later when we
start NFA optimization. (At that point we don't need to do delsub()
any more, and besides there are other cases where NFA optimization can
lead to disconnected subgraphs.)
It appears that this bug has no consequences in a non-assert-enabled
build: there will be some transiently leaked NFA states/arcs, but
they'll get cleaned up eventually. Still, we don't like assertion
failures, so back-patch to v14 where rainbow arcs were introduced.
Per bug #18708 from Nikolay Shaplov.
Discussion: https://postgr.es/m/18708-f94f2599c9d2c005@postgresql.org
This fixes a set of race conditions with cumulative statistics where a
shared stats entry could be dropped while it should still be valid in
the event when it is reused: an entry may refer to a different object
but requires the same hash key. This can happen with various stats
kinds, like:
- Replication slots that compute internally an index number, for
different slot names.
- Stats kinds that use an OID in the object key, where a wraparound
causes the same key to be used if an OID is used for the same object.
- As of PostgreSQL 18, custom pgstats kinds could also be an issue,
depending on their implementation.
This issue is fixed by introducing a counter called "generation" in the
shared entries via PgStatShared_HashEntry, initialized at 0 when an
entry is created and incremented when the same entry is reused, to avoid
concurrent issues on drop because of other backends still holding a
reference to it. This "generation" is copied to the local copy that a
backend holds when looking at an object, then cross-checked with the
shared entry to make sure that the entry is not dropped even if its
"refcount" justifies that if it has been reused.
This problem could show up when a backend shuts down and needs to
discard any entries it still holds, causing statistics to be removed
when they should not, or even an assertion failure. Another report
involved a failure in a standby after an OID wraparound, where the
startup process would FATAL on a "can only drop stats once", stopping
recovery abruptly. The buildfarm has been sporadically complaining
about the problem, as well, but the window is hard to reach with the
in-core tests.
Note that the issue can be reproduced easily by adding a sleep before
dshash_find() in pgstat_release_entry_ref() to enlarge the problematic
window while repeating test_decoding's isolation test oldest_xmin a
couple of times, for example, as pointed out by Alexander Lakhin.
Reported-by: Alexander Lakhin, Peter Smith
Author: Kyotaro Horiguchi, Michael Paquier
Reviewed-by: Bertrand Drouvot
Discussion: https://postgr.es/m/CAA4eK1KxuMVyAryz_Vk5yq3ejgKYcL6F45Hj9ZnMNBS-g+PuZg@mail.gmail.com
Discussion: https://postgr.es/m/17947-b9554521ad963c9c@postgresql.org
Backpatch-through: 15
The current code calls array_eq() and does not provide FmgrInfo. This commit
provides initialization of FmgrInfo and uses C collation as the safe option
for text comparison because we don't know anything about the semantics of
opclass options.
Backpatch to 13, where opclass options were introduced.
Reported-by: Nicolas Maus
Discussion: https://postgr.es/m/18692-72ea398df3ec6712%40postgresql.org
Backpatch-through: 13
Commit 5a2fed911 had an unexpected side-effect: the parallel worker
launched for the new test case would fail if it couldn't use a
superuser-reserved connection slot. The reason that test failed
while all our pre-existing ones worked is that the connection
privilege tests in InitPostgres had been based on the superuserness
of the leader's AuthenticatedUserId, but after the rearrangements
of 5a2fed911 we were testing the superuserness of CurrentUserId,
which the new test case deliberately made to be a non-superuser.
This all seems very accidental and probably not the behavior we really
want, but a security patch is no time to be redesigning things.
Pending some discussion about desirable semantics, hack it so that
InitPostgres continues to pay attention to the superuserness of
AuthenticatedUserId when starting a parallel worker.
Nathan Bossart and Tom Lane, per buildfarm member sawshark.
Security: CVE-2024-10978
The SQL spec mandates that SET SESSION AUTHORIZATION implies
SET ROLE NONE. We tried to implement that within the lowest-level
functions that manipulate these settings, but that was a bad idea.
In particular, guc.c assumes that it doesn't matter in what order
it applies GUC variable updates, but that was not the case for these
two variables. This problem, compounded by some hackish attempts to
work around it, led to some security-grade issues:
* Rolling back a transaction that had done SET SESSION AUTHORIZATION
would revert to SET ROLE NONE, even if that had not been the previous
state, so that the effective user ID might now be different from what
it had been.
* The same for SET SESSION AUTHORIZATION in a function SET clause.
* If a parallel worker inspected current_setting('role'), it saw
"none" even when it should see something else.
Also, although the parallel worker startup code intended to cope
with the current role's pg_authid row having disappeared, its
implementation of that was incomplete so it would still fail.
Fix by fully separating the miscinit.c functions that assign
session_authorization from those that assign role. To implement the
spec's requirement, teach set_config_option itself to perform "SET
ROLE NONE" when it sets session_authorization. (This is undoubtedly
ugly, but the alternatives seem worse. In particular, there's no way
to do it within assign_session_authorization without incompatible
changes in the API for GUC assign hooks.) Also, improve
ParallelWorkerMain to directly set all the relevant user-ID variables
instead of relying on some of them to get set indirectly. That
allows us to survive not finding the pg_authid row during worker
startup.
In v16 and earlier, this includes back-patching 9987a7bf3 which
fixed a violation of GUC coding rules: SetSessionAuthorization
is not an appropriate place to be throwing errors from.
Security: CVE-2024-10978
If a CTE, subquery, sublink, security invoker view, or coercion
projection references a table with row-level security policies, we
neglected to mark the plan as potentially dependent on which role
is executing it. This could lead to later executions in the same
session returning or hiding rows that should have been hidden or
returned instead.
Reported-by: Wolfgang Walther
Reviewed-by: Noah Misch
Security: CVE-2024-10976
Backpatch-through: 12
Commit ac04aa84a put the shutoff for this into the planner, which is
not ideal because it doesn't prevent us from re-using a previously
made parallel plan. Revert the planner change and instead put the
shutoff into InitializeParallelDSM, modeling it on the existing code
there for recovering from failure to allocate a DSM segment.
However, that code path is mostly untested, and testing a bit harder
showed there's at least one bug: ExecHashJoinReInitializeDSM is not
prepared for us to have skipped doing parallel DSM setup. I also
thought the Assert in ReinitializeParallelWorkers is pretty
ill-advised, and replaced it with a silent Min() operation.
The existing test case added by ac04aa84a serves fine to test this
version of the fix, so no change needed there.
Patch by me, but thanks to Noah Misch for the core idea that we
could shut off worker creation when !INTERRUPTS_CAN_BE_PROCESSED.
Back-patch to v12, as ac04aa84a was.
Discussion: https://postgr.es/m/CAC-SaSzHUKT=vZJ8MPxYdC_URPfax+yoA1hKTcF4ROz_Q6z0_Q@mail.gmail.com
