ci: openbsd: Increase RAM disk's size
Its size was ~3.8GB before, which sometimes was not enough. OpenBSD CI task
often were failing due to no space left on device. Increase the RAM disk size
to ~4.6 GB.
Author: Nazir Bilal Yavuz
Discussion:
https://postgr.es/m/CAN55FZ2XVVPJRJmGB2DsL3gOrOinWh=H
Mark shared buffer lookup table HASH_FIXED_SIZE
StrategyInitialize() calls InitBufTable() with maximum number of entries that
the buffer lookup table can ever have. Thus there should not be any need to
allocate more element after initialization. Hence mark the hash table as fixed
sized.
Author: A
ci: openbsd: Increase RAM disk's size
Its size was ~3.8GB before, which sometimes was not enough. OpenBSD CI task
often were failing due to no space left on device. Increase the RAM disk size
to ~4.6 GB.
Author: Nazir Bilal Yavuz
Discussion:
https://postgr.es/m/CAN55FZ2XVVPJRJmGB2DsL3gOrOinWh=H
bufmgr: Use consistent naming of the clock-sweep algorithm
Minor edits to comments only.
Author: Greg Burd
Reviewed-by: Tomas Vondra
Reviewed-by: Andres Freund
Discussion:
https://postgr.es/m/70c6a5b5-2a20-4d0b-bc73-eb09dd62d...@getmailspring.com
Branch
--
master
Details
---
https
scalability, but that having the freelist would require
dedicated complexity in several upcoming patches. As we have not been able to
find a case benefiting from the freelist...
Author: Greg Burd
Reviewed-by: Tomas Vondra
Reviewed-by: Andres Freund
Discussion:
https://postgr.es/m/70c6a5b5-2a20-4d0b
aio: Stop using enum bitfields due to bad code generation
During an investigation into rather odd aio related errors on macos, observed
by Alexander and Konstantin, we started to wonder if bitfield access is
related to the error. At the moment it looks like it is related, we cannot
reproduce the f
aio: Stop using enum bitfields due to bad code generation
During an investigation into rather odd aio related errors on macos, observed
by Alexander and Konstantin, we started to wonder if bitfield access is
related to the error. At the moment it looks like it is related, we cannot
reproduce the f
Hi,
On 2025-08-19 15:08:08 +, Bertrand Drouvot wrote:
> Hi,
>
> On Tue, Aug 19, 2025 at 10:28:29AM -0400, Andres Freund wrote:
> > Hi,
> >
> > On 2025-08-19 14:01:52 +, Bertrand Drouvot wrote:
> > > On Tue, Aug 19, 2025 at 12:31:14PM +0200, Álvaro He
> attached
> fix it.
That can't be the reason - it's reached from src/test/Makefile
> SUBDIRS = perl postmaster regress isolation modules authentication recovery
> subscription
Greetings,
Andres Freund
ci: Simplify ci-os-only handling
Handle 'ci-os-only' occurrences in the .cirrus.star file instead of
.cirrus.tasks.yml file. Now, 'ci-os-only' occurrences are controlled
from one central place instead of dealing with them in each task.
Author: Andres Freund
Reviewed-b
ci: Simplify ci-os-only handling
Handle 'ci-os-only' occurrences in the .cirrus.star file instead of
.cirrus.tasks.yml file. Now, 'ci-os-only' occurrences are controlled
from one central place instead of dealing with them in each task.
Author: Andres Freund
Reviewed-b
ci: Simplify ci-os-only handling
Handle 'ci-os-only' occurrences in the .cirrus.star file instead of
.cirrus.tasks.yml file. Now, 'ci-os-only' occurrences are controlled
from one central place instead of dealing with them in each task.
Author: Andres Freund
Reviewed-b
ci: Simplify ci-os-only handling
Handle 'ci-os-only' occurrences in the .cirrus.star file instead of
.cirrus.tasks.yml file. Now, 'ci-os-only' occurrences are controlled
from one central place instead of dealing with them in each task.
Author: Andres Freund
Reviewed-b
ci: Simplify ci-os-only handling
Handle 'ci-os-only' occurrences in the .cirrus.star file instead of
.cirrus.tasks.yml file. Now, 'ci-os-only' occurrences are controlled
from one central place instead of dealing with them in each task.
Author: Andres Freund
Reviewed-b
unning tests for the
Commitfest app.
Backpatch this back to PG 15, even though PG 15 does not have any manually
triggered task. Keeping the CI infrastructure the same seems advantageous.
Author: Andres Freund
Co-authored-by: Thomas Munro
Co-authored-by: Nazir Bilal Yavuz
Reviewed-by: Nazir Bil
unning tests for the
Commitfest app.
Backpatch this back to PG 15, even though PG 15 does not have any manually
triggered task. Keeping the CI infrastructure the same seems advantageous.
Author: Andres Freund
Co-authored-by: Thomas Munro
Co-authored-by: Nazir Bilal Yavuz
Reviewed-by: Nazir Bil
unning tests for the
Commitfest app.
Backpatch this back to PG 15, even though PG 15 does not have any manually
triggered task. Keeping the CI infrastructure the same seems advantageous.
Author: Andres Freund
Co-authored-by: Thomas Munro
Co-authored-by: Nazir Bilal Yavuz
Reviewed-by: Nazir Bil
unning tests for the
Commitfest app.
Backpatch this back to PG 15, even though PG 15 does not have any manually
triggered task. Keeping the CI infrastructure the same seems advantageous.
Author: Andres Freund
Co-authored-by: Thomas Munro
Co-authored-by: Nazir Bilal Yavuz
Reviewed-by: Nazir Bil
unning tests for the
Commitfest app.
Backpatch this back to PG 15, even though PG 15 does not have any manually
triggered task. Keeping the CI infrastructure the same seems advantageous.
Author: Andres Freund
Co-authored-by: Thomas Munro
Co-authored-by: Nazir Bilal Yavuz
Reviewed-by: Nazir Bil
'd like to get CI to become green again, and a proper fix (in
meson) will presumably take longer.
Suggested-by: Andres Freund
Author: Nazir Bilal Yavuz
Reviewed-by: Jacob Champion
Reviewed-by: Peter Eisentraut
Reviewed-by: Tom Lane
Discussion:
https://postgr.es/m/CAN55FZ1RuBhJmPWs3
'd like to get CI to become green again, and a proper fix (in
meson) will presumably take longer.
Suggested-by: Andres Freund
Author: Nazir Bilal Yavuz
Reviewed-by: Jacob Champion
Reviewed-by: Peter Eisentraut
Reviewed-by: Tom Lane
Discussion:
https://postgr.es/m/CAN55FZ1RuBhJmPWs3
'd like to get CI to become green again, and a proper fix (in
meson) will presumably take longer.
Suggested-by: Andres Freund
Author: Nazir Bilal Yavuz
Reviewed-by: Jacob Champion
Reviewed-by: Peter Eisentraut
Reviewed-by: Tom Lane
Discussion:
https://postgr.es/m/CAN55FZ1RuBhJmPWs3
'd like to get CI to become green again, and a proper fix (in
meson) will presumably take longer.
Suggested-by: Andres Freund
Author: Nazir Bilal Yavuz
Reviewed-by: Jacob Champion
Reviewed-by: Peter Eisentraut
Reviewed-by: Tom Lane
Discussion:
https://postgr.es/m/CAN55FZ1RuBhJmPWs3
Add very basic test for kill_prior_tuples
Previously our tests did not exercise kill_prior_tuples for hash and gist. For
gist some related paths were reached, but gist's implementation seems to not
work if all the dead tuples are on one page (or something like that). The
coverage for other index t
Reduce ExecSeqScan* code size using pg_assume()
fb9f955025f optimized code generation by using specialized variants of
ExecSeqScan* for [not] having a qual, projection etc. This allowed the
compiler to optimize the code out the code for qual / projection. However, as
observed by David Rowley at th
meson: add and use stamp files for generated headers
Without using stamp files, meson lists the generated headers as the dependency
for every .c file, bloating build.ninja by more than 2x. Processing all the
dependencies also increases the time to generate build.ninja.
The immediate benefit is th
aio: Fix assertion, clarify README
The assertion wouldn't have triggered for a long while yet, but this won't
accidentally fail to detect the issue if/when it occurs.
Author: Matthias van de Meent
Discussion:
https://postgr.es/m/CAEze2Wj-43JV4YufW23gm=uwr7lkj+p0ykctkhxnm1rwfc+...@mail.gmail.com
aio: Fix assertion, clarify README
The assertion wouldn't have triggered for a long while yet, but this won't
accidentally fail to detect the issue if/when it occurs.
Author: Matthias van de Meent
Discussion:
https://postgr.es/m/CAEze2Wj-43JV4YufW23gm=uwr7lkj+p0ykctkhxnm1rwfc+...@mail.gmail.com
Remove long-unused TransactionIdIsActive()
TransactionIdIsActive() has not been used since bb38fb0d43c, in 2014. There
are no known uses in extensions either and it's hard to see valid uses for
it. Therefore remove TransactionIdIsActive().
Discussion:
https://postgr.es/m/odgftbtwp5oq7cxjgf4kjkmy
Add pg_assume(expr) macro
This macro can be used to avoid compiler warnings, particularly when using -O3
and not using assertions, and to get the compiler to generate better code.
A subsequent commit introduces a first user.
Reviewed-by: Tom Lane
Discussion:
https://postgr.es/m/3prdb6hkep3dugl
Use pg_assume() to avoid compiler warning below exec_set_found()
The warning, visible when building with -O3 and a recent-ish gcc, is due to
gcc not realizing that found is a byvalue type and therefore will never be
interpreted as a varlena type.
Discussion:
https://postgr.es/m/3prdb6hkep3duglhs
aio: Combine io_uring memory mappings, if supported
By default io_uring creates a shared memory mapping for each io_uring
instance, leading to a large number of memory mappings. Unfortunately a large
number of memory mappings slows things down, backend exit is particularly
affected. To address th
aio: Combine io_uring memory mappings, if supported
By default io_uring creates a shared memory mapping for each io_uring
instance, leading to a large number of memory mappings. Unfortunately a large
number of memory mappings slows things down, backend exit is particularly
affected. To address th
aio: Fix reference to outdated name
Reported-by: Antonin Houska
Author: Antonin Houska
Discussion: https://postgr.es/m/5250.1751266701@localhost
Backpatch-through: 18, where da7226993fd4 introduced this
Branch
--
REL_18_STABLE
Details
---
https://git.postgresql.org/pg/commitdiff/95163c
aio: Fix reference to outdated name
Reported-by: Antonin Houska
Author: Antonin Houska
Discussion: https://postgr.es/m/5250.1751266701@localhost
Backpatch-through: 18, where da7226993fd4 introduced this
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/f20a347e1a613
a separate patch.
Thanks a lot to Alexander and Konstantin for all the debugging help.
Reported-by: Tom Lane
Reported-by: Alexander Lakhin
Investigated-by: Andres Freund
Investigated-by: Alexander Lakhin
Investigated-by: Konstantin Knizhnik
Discussion:
htt
Remove accidentally added meson.build
Commit b4363fc66e6 added an empty meson.build to 13, presumably due to a
conflict resolution while backpatching. Remove it.
Discussion:
https://postgr.es/m/b7angm4wz2ln2wsqvsb3bwl4pqa4xxgcawn36o5kweu5pgpl6s@ofhhrgl5ouwb
Branch
--
REL_13_STABLE
Details
aio: Fix possible state confusions due to interrupt processing
elog()/ereport() process interrupts, iff the log message is < ERROR and the
log message will be emitted. aio's debug messages are emitted via ereport(),
but in some places the code is not ready for interrupts to be processed.
Fix the
Hi,
On 2025-04-29 13:45:44 +, Bruce Momjian wrote:
> doc: clarify new behavior of max_files_per_process
I don't think this really is an improvement - if postmaster is called with a
bunch of files already opened, they aren't opened by postmaster, but still are
ignored...
Greetings,
Andres
aio: Fix crash potential for pg_aios views due to late state update
pgaio_io_reclaim() reset the fields in PgAioHandle before updating the state
to IDLE or incrementing the generation. For most things that's OK, but for
pg_get_aios() it is not - if it copied the PgAioHandle while fields were being
aio: Improve debug logging around waiting for IOs
Trying to investigate a bug report by Alexander Lakhin made it apparent that
the debug logging around waiting for IO completion is insufficient. Fix that.
Discussion:
https://postgr.es/m/h4in2db37vepagmi2oz5vvqymjasc5gyb4lpqkunj4eusu274i@37jpd3c2
Fix bug allowing io_combine_limit > io_max_combine_combine limit
10f66468475 intended to limit the value of io_combine_limit to the minimum of
io_combine_limit and io_max_combine_limit. To avoid issues with interdependent
GUCs, it introduced io_combine_limit_guc and set io_combine_limit in assign
Hi,
Walther, adding you because of the failure on buildfarm animal dogfish. The
main issue is something unrelated to your animal (except it was the only one
to catch it), but there's one oddity:
On 2025-04-08 10:36:32 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On April 8,
Hi,
On 2025-04-08 17:17:21 +0200, Wolfgang Walther wrote:
> Andres Freund:
> > But I'm rather bewildered - I can't actually build with the professed
> > version
> > of ninja, it errors out:
> >
> > ninja-1.9.0 src/bin/psql/sql_help.c
> > ninja:
Hi,
On April 8, 2025 10:05:48 AM EDT, Tom Lane wrote:
>Daniel Gustafsson writes:
>>> On 7 Apr 2025, at 23:18, Tomas Vondra wrote:
>>> Add support for basic NUMA awareness
>
>> dogfish is a little bit upset it seems.
>
>Hm, but why isn't anything else?
I assume it's a question of degree of para
Increase BAS_BULKREAD based on effective_io_concurrency
Before, BAS_BULKREAD was always of size 256kB. With the default
io_combine_limit of 16, that only allowed 1-2 IOs to be in flight -
insufficient even on very low latency storage.
We don't just want to increase the size to a much larger hardc
extension version is needed, as it was already increased this release
by ba2a3c2302f.
Author: Nazir Bilal Yavuz
Reviewed-by: Andres Freund
Reviewed-by: Aidar Imamov
Reviewed-by: Joseph Koshakow
Discussion:
https://postgr.es/m/CAN55FZ0h_YoSqqutxV6DES1RW8ig6wcA8CR9rJk358YRMxZFmw
aio: Make AIO more compatible with valgrind
In some edge cases valgrind flags issues with the memory referenced by
IOs. All of the cases addressed in this change are false positives.
Most of the false positives are caused by UnpinBuffer[NoOwner] marking buffer
data as inaccessible. This happens e
localbuf: Add Valgrind buffer access instrumentation
This mirrors 1e0dfd166b3 (+ 46ef520b9566), for temporary table buffers. This
is mainly interesting right now because the AIO work currently triggers
spurious valgrind errors, and the fix for that is cleaner if temp buffers
behave the same as sha
6%3A49
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=hippopotamus&dt=2025-04-07%2018%3A28%3A12
Greetings,
Andres Freund
read_stream: Fix overflow hazard with large shared buffers
If the limit returned by GetAdditionalPinLimit() is large, the buffer_limit
variable in read_stream_start_pending_read() can overflow. While the code is
careful to limit buffer_limit PG_INT16_MAX, we subsequently add the number of
forwarde
aio: Avoid spurious coverity warning
PgAioResult.result is never accessed in the relevant path, but coverity
complains about an uninitialized access anyway. So just zero-initialize the
whole thing. While at it, reduce the scope of the variable.
Reported-by: Ranier Vilela
Reviewed-by: Noah Misch
docs: Reframe track_io_timing related docs as wait time
With AIO it does not make sense anymore to track the time for each individual
IO, as multiple IOs can be in-flight at the same time. Instead we now track
the time spent *waiting* for IOs.
This should be reflected in the docs. While, so far,
meson: Flush stdout in testwrap
Otherwise the progress won't reliably be displayed during a test.
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/kx6xu7suexal5vwsxpy7ybgkcznx6hgywbuhkr6qabcwxjqax2@i4pcpk75jvaa
Backpatch-through: 16
Branch
--
REL_17_STABLE
Details
---
https://g
Hi,
On 2025-03-18 11:03:36 -0400, Tom Lane wrote:
> Andres Freund writes:
> > aio: Add core asynchronous I/O infrastructure
>
> Some of the buildfarm is mildly unhappy with this.
> So far I see
>
> ayu | 2025-03-18 13:08:04 | aio_callback.c:83:12: warning:
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
Add errhint_internal()
We have errmsg_internal(), errdetail_internal(), but not errhint_internal().
Sometimes it is useful to output a hint with already translated format
string (e.g. because there different messages depending on the condition). For
message/detail we do that with the _internal()
bufmgr: Improve stats when a buffer is read in concurrently
Previously we would have the following inaccuracies when a backend tried to
read in a buffer, but that buffer was read in concurrently by another backend:
- the read IO was double-counted in the global buffer access stats
(pgBufferUsage)
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
Add test for HeapBitmapScan's broken skip_fetch optimization
In the previous commit HeapBitmapScan's skip_fetch optimization was removed,
due to being broken in not easily fixable ways. Add a test that verifies we
don't re-introduce this bug if somebody tries to re-add the feature.
Only add the t
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
Remove HeapBitmapScan's skip_fetch optimization
The optimization does not take the removal of TIDs by a concurrent vacuum into
account. The concurrent vacuum can remove dead TIDs and make pages ALL_VISIBLE
while those dead TIDs are referenced in the bitmap. This can lead to a
skip_fetch scan retur
tests: Fix incompatibility of test_aio with *_FORCE_RELEASE
The test added in 93bc3d75d8e failed in a build with RELCACHE_FORCE_RELEASE
and CATCACHE_FORCE_RELEASE defined. The test intentionally forgets to exit
batchmode - normally that would trigger an error at the end of the
transaction, which t
tests: Cope with WARNINGs during failed CREATE DB on windows
The test added in 93bc3d75d8e sometimes fails on windows, due to warnings like
WARNING: some useless files may be left behind in old database directory
"base/16514"
The reason for that is createdb_failure_callback() does not ensure th
tests: Cope with io_method in TEMP_CONFIG in test_aio
If io_method is set in TEMP_CONFIG the test added in 93bc3d75d8e fails,
because it assumes the io_method specified at initdb is actually used.
Fix that by appending the io_method again, after initdb (and thus after
TEMP_CONFIG has been added b
aio: Minor comment improvements
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/usbwzckj7q3jhfx3ann3nrfnukmupbs35axvq5zfyeo6nvrzrm@onjhxs2du4st
Branch
--
master
Details
---
https://git.postgresql.org/pg/commitdiff/e19dc74491e68e1df05c99c4937929cef3092a3b
Modified Files
---
aio: Add README.md explaining higher level design
Reviewed-by: Noah Misch
Reviewed-by: Heikki Linnakangas
Discussion:
https://postgr.es/m/uvrtrknj4kdytuboidbhwclo4gxhswwcpgadptsjvjqcluzmah%40brqs62irg4dt
Discussion:
https://postgr.es/m/20210223100344.llw5an2akleng...@alap3.anarazel.de
Discussi
aio: Add test_aio module
To make the tests possible, a few functions from bufmgr.c/localbuf.c had to be
exported, via buf_internals.h.
Reviewed-by: Noah Misch
Co-authored-by: Andres Freund
Co-authored-by: Nazir Bilal Yavuz
Discussion:
https://postgr.es/m
md: Add comment & assert to buffer-zeroing path in md[start]readv()
mdreadv() has a codepath to zero out buffers when a read returns zero bytes,
guarded by a check for zero_damaged_pages || InRecovery.
The InRecovery codepath to zero out buffers in mdreadv() appears to be
unreachable. The only kn
docs: Add acronym and glossary entries for I/O and AIO
These are fairly basic, but better than nothing. While there are several
opportunities to link to these entries, this patch does not add any. They will
however be referenced by future patches.
Reviewed-by: Noah Misch
Discussion: https://pos
aio: Add pg_aios view
The new view lists all IO handles that are currently in use and is mainly
useful for PG developers, but may also be useful when tuning PG.
Bumps catversion.
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/uvrtrknj4kdytuboidbhwclo4gxhswwcpgadptsjvjqcluzmah%40brqs62
Hi,
On 2025-03-31 19:23:57 +0300, Aleksander Alekseev wrote:
> Hi Andres,
>
> > aio: Add core asynchronous I/O infrastructure
> >
> > [...]
>
> I noticed that aio.c references README.md:
>
> ```
> * aio.c
> *AIO - Core Logic
> *
> * For documentation about how AIO works on a higher leve
Noah Misch
Reviewed-by: Nazir Bilal Yavuz
Co-authored-by: Andres Freund
Co-authored-by: Thomas Munro
Discussion:
https://postgr.es/m/uvrtrknj4kdytuboidbhwclo4gxhswwcpgadptsjvjqcluzmah%40brqs62irg4dt
Discussion:
https://postgr.es/m/20210223100344.llw5an2akleng...@alap3.anarazel.de
Discussion:
tely.
- It can be beneficial to defer starting new IOs until we can issue multiple
IOs at once. That however requires non-trivial heuristics to decide when to
do so.
Reviewed-by: Noah Misch
Co-authored-by: Andres Freund
Co-authored-by: Thomas Munro
Branch
--
master
Details
---
Enable IO concurrency on all systems
Previously effective_io_concurrency and maintenance_io_concurrency could not
be set above 0 on machines without fadvise support. AIO enables IO concurrency
without such support, via io_method=worker.
Currently only subsystems using the read stream API will tak
read_stream: Introduce and use optional batchmode support
Submitting IO in larger batches can be more efficient than doing so
one-by-one, particularly for many small reads. It does, however, require
the ReadStreamBlockNumberCB callback to abide by the restrictions of AIO
batching (c.f. pgaio_enter
bufmgr: Implement AIO read support
This commit implements the infrastructure to perform asynchronous reads into
the buffer pool.
To do so, it:
- Adds readv AIO callbacks for shared and local buffers
It may be worth calling out that shared buffer completions may be run in a
different backend
Let caller of PageIsVerified() control ignore_checksum_failure
For AIO the completion of a read into shared buffers (i.e. verifying the page
including the checksum, updating the BufferDesc to reflect the IO) can happen
in a different backend than the backend that started the IO. As
ignore_checksum
aio: Add WARNING result status
If an IO succeeds, but issues a warning, e.g. due to a page verification
failure with zero_damaged_pages, we want to issue that warning in the context
of the issuer of the IO, not the process that executes the completion (always
the case for worker).
It's already po
pgstat: Allow checksum errors to be reported in critical sections
For AIO we execute completion callbacks in critical sections (to ensure that
AIO can in the future be used for WAL, which in turn requires that we can call
completion callbacks in critical sections, to get the resources for WAL
io).
Fix mis-attribution of checksum failure stats to the wrong database
Checksum failure stats could be attributed to the wrong database in two cases:
- when a read of a shared relation encountered a checksum error , it would be
attributed to the current database, instead of the "database" represen
localbuf: Track pincount in BufferDesc as well
For AIO on temporary table buffers the AIO subsystem needs to be able to
ensure a pin on a buffer while AIO is going on, even if the IO issuing query
errors out. Tracking the buffer in LocalRefCount does not work, as it would
cause CheckForLocalBuffer
aio, bufmgr: Comment fixes/improvements
Some of these comments have been wrong for a while (12f3867f5534), some I
recently introduced (da7226993fd, 55b454d0e14). This includes an update to a
comment in FlushBuffer(), which will be copied in a future commit.
These changes seem big enough to be wor
aio: Implement support for reads in smgr/md/fd
This implements the following:
1) An smgr AIO target, for AIO on smgr files. This should be usable not just
for md.c but also other SMGR implementation if we ever get them.
2) readv support in fd.c, which requires a small bit of infrastructure wor
Fix "‘static’ is not at beginning of declaration" warning
b98be8a2a2a used "const static" instead of "static const". We normally use the
latter form.
Discussion:
https://postgr.es/m/z4mc2hzecahyq3paupfsouhuupmzmgum45md3k5my6bmo7gvn7@z5j26doqamqy
Branch
--
master
Details
---
https://git
Hi,
On 2025-03-27 00:15:53 +, Andres Freund wrote:
> aio: Add io_method=io_uring
FWIW, the BF failure on culicidae looks to be independent to me
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=culicidae&dt=2025-03-27%2000%3A16%3A30
The relevant line from the log is:
2025-03-2
aio: Add io_method=io_uring
Performing AIO using io_uring can be considerably faster than
io_method=worker, particularly when lots of small IOs are issued, as
a) the context-switch overhead for worker based AIO becomes more significant
b) the number of IO workers can become limiting
io_uring, how
aio: Add liburing dependency
Will be used in a subsequent commit, to implement io_method=io_uring. Kept
separate for easier review.
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/uvrtrknj4kdytuboidbhwclo4gxhswwcpgadptsjvjqcluzmah%40brqs62irg4dt
Branch
--
master
Details
---
ht
aio: Rename pgaio_io_prep_* to pgaio_io_start_*
The old naming pattern (mirroring liburing's naming) was inconsistent with
the (not yet introduced) callers. It seems better to get rid of the
inconsistency now than to grow more users of the odd naming.
Reported-by: Noah Misch
Discussion: https://
aio: Be more paranoid about interrupts
As reported by Noah, it's possible, although practically very unlikely, that
interrupts could be processed in between pgaio_io_reopen() and
pgaio_io_perform_synchronously(). Prevent that by explicitly holding
interrupts.
It also seems good to add an assertio
aio: Pass result of local callbacks to ->report_return
Otherwise the results of e.g. temp table buffer verification errors will not
reach bufmgr.c. Obviously that's not right. Found while expanding the tests
for invalid buffer contents.
Reviewed-by: Noah Misch
Discussion: https://postgr.es/m/202
Redefine max_files_per_process to control additionally opened files
Until now max_files_per_process=N limited each backend to open N files in
total (minus a safety factor), even if there were already more files opened in
postmaster and inherited by backends. Change max_files_per_process to contro
aio: Change prefix of PgAioResultStatus values to PGAIO_RS_
The previous prefix wasn't consistent with the naming of other AIO related
enum values. It seems best to rename it before the users are introduced.
Reported-by: Melanie Plageman
Discussion:
https://postgr.es/m/caakru_yb+jzqpnsguxcb0gbi
smgr: Hold interrupts in most smgr functions
We need to hold interrupts across most of the smgr.c/md.c functions, as
otherwise interrupt processing, e.g. due to a < ERROR elog/ereport, can
trigger procsignal processing, which in turn can trigger smgrreleaseall(). As
the relevant code is not reentr
Hi,
On 2025-03-18 11:26:15 -0400, Tom Lane wrote:
> Andres Freund writes:
> > I wonder if we should instead either ask those buildfarm animals to be
> > disabled or have the warning manually disabled. I don't think it's a good
> > investment on our part to work
meson: Flush stdout in testwrap
Otherwise the progress won't reliably be displayed during a test.
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/kx6xu7suexal5vwsxpy7ybgkcznx6hgywbuhkr6qabcwxjqax2@i4pcpk75jvaa
Backpatch-through: 16
Branch
--
master
Details
---
https://git.post
meson: Flush stdout in testwrap
Otherwise the progress won't reliably be displayed during a test.
Reviewed-by: Noah Misch
Discussion:
https://postgr.es/m/kx6xu7suexal5vwsxpy7ybgkcznx6hgywbuhkr6qabcwxjqax2@i4pcpk75jvaa
Backpatch-through: 16
Branch
--
REL_16_STABLE
Details
---
https://g
Greetings,
Andres Freund
1 - 100 of 1772 matches
Mail list logo
