pg.ddx.io pgsql-committers@postgresql.org mailing list archivehelp / color / mirror / Atom feed
pgsql: Get rid of WALBufMappingLock 2+ messages / 1 participants [nested] [flat]
* pgsql: Get rid of WALBufMappingLock @ 2025-02-17 02:25 Alexander Korotkov <akorotkov@postgresql.org> 0 siblings, 0 replies; 2+ messages in thread From: Alexander Korotkov @ 2025-02-17 02:25 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Get rid of WALBufMappingLock Allow multiple backends to initialize WAL buffers concurrently. This way `MemSet((char *) NewPage, 0, XLOG_BLCKSZ);` can run in parallel without taking a single LWLock in exclusive mode. The new algorithm works as follows: * reserve a page for initialization using XLogCtl->InitializeReserved, * ensure the page is written out, * once the page is initialized, try to advance XLogCtl->InitializedUpTo and signal to waiters using XLogCtl->InitializedUpToCondVar condition variable, * repeat previous steps until we reserve initialization up to the target WAL position, * wait until concurrent initialization finishes using a XLogCtl->InitializedUpToCondVar. Now, multiple backends can, in parallel, concurrently reserve pages, initialize them, and advance XLogCtl->InitializedUpTo to point to the latest initialized page. Author: Yura Sokolov <y.sokolov@postgrespro.ru> Co-authored-by: Alexander Korotkov <aekorotkov@gmail.com> Reviewed-by: Pavel Borisov <pashkin.elfe@gmail.com> Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/6a2275b8953a4462d44daf001bdd60b3d48f0946 Modified Files -------------- src/backend/access/transam/xlog.c | 176 +++++++++++++++++------- src/backend/utils/activity/wait_event_names.txt | 2 +- src/include/storage/lwlocklist.h | 2 +- 3 files changed, 132 insertions(+), 48 deletions(-) ^ permalink raw reply [nested|flat] 2+ messages in thread
* pgsql: Get rid of WALBufMappingLock @ 2025-04-02 09:45 Alexander Korotkov <akorotkov@postgresql.org> 0 siblings, 0 replies; 2+ messages in thread From: Alexander Korotkov @ 2025-04-02 09:45 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Get rid of WALBufMappingLock Allow multiple backends to initialize WAL buffers concurrently. This way `MemSet((char *) NewPage, 0, XLOG_BLCKSZ);` can run in parallel without taking a single LWLock in exclusive mode. The new algorithm works as follows: * reserve a page for initialization using XLogCtl->InitializeReserved, * ensure the page is written out, * once the page is initialized, try to advance XLogCtl->InitializedUpTo and signal to waiters using XLogCtl->InitializedUpToCondVar condition variable, * repeat previous steps until we reserve initialization up to the target WAL position, * wait until concurrent initialization finishes using a XLogCtl->InitializedUpToCondVar. Now, multiple backends can, in parallel, concurrently reserve pages, initialize them, and advance XLogCtl->InitializedUpTo to point to the latest initialized page. Author: Yura Sokolov <y.sokolov@postgrespro.ru> Co-authored-by: Alexander Korotkov <aekorotkov@gmail.com> Reviewed-by: Pavel Borisov <pashkin.elfe@gmail.com> Reviewed-by: Tomas Vondra <tomas@vondra.me> Tested-by: Michael Paquier <michael@paquier.xyz> Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/bc22dc0e0ddc2dcb6043a732415019cc6b6bf683 Modified Files -------------- src/backend/access/transam/xlog.c | 234 +++++++++++++++++++----- src/backend/utils/activity/wait_event_names.txt | 2 +- src/include/storage/lwlocklist.h | 2 +- 3 files changed, 189 insertions(+), 49 deletions(-) ^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2025-04-02 09:45 UTC | newest] Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2025-02-17 02:25 pgsql: Get rid of WALBufMappingLock Alexander Korotkov <akorotkov@postgresql.org> 2025-04-02 09:45 pgsql: Get rid of WALBufMappingLock Alexander Korotkov <akorotkov@postgresql.org>
This inbox is served by DDX for PostgreSQL; see mirroring instructions for how to clone and mirror all data and code used for this inbox