Re: Latches vs lwlock contention

Thomas Munro <thomas.munro@gmail.com>

From: Thomas Munro <thomas.munro@gmail.com>

To: vignesh C <vignesh21@gmail.com>

Cc: pgsql-hackers <pgsql-hackers@postgresql.org>, Yura Sokolov <y.sokolov@postgrespro.ru>, Andres Freund <andres@anarazel.de>

Date: 2023-03-04T19:50:30Z

Lists: pgsql-hackers

Attachments

v3-0001-Allow-palloc_extended-NO_OOM-in-critical-sections.patch (text/x-patch) patch v3-0001
v3-0002-Provide-SetLatches-for-batched-deferred-latches.patch (text/x-patch) patch v3-0002
v3-0003-Use-SetLatches-for-synchronous-replication-wakeup.patch (text/x-patch) patch v3-0003
v3-0004-Use-SetLatches-for-SERIALIZABLE-DEFERRABLE-wakeup.patch (text/x-patch) patch v3-0004
v3-0005-Use-SetLatches-for-heavyweight-locks.patch (text/x-patch) patch v3-0005
v3-0006-Don-t-re-acquire-LockManager-partition-lock-after.patch (text/x-patch) patch v3-0006

On Sat, Jan 28, 2023 at 3:39 AM vignesh C <vignesh21@gmail.com> wrote:
> On Tue, 1 Nov 2022 at 16:40, Thomas Munro <thomas.munro@gmail.com> wrote:
> > Here's an attempt at that.  There aren't actually any cases of uses of
> > this stuff in critical sections here, so perhaps I shouldn't bother
> > with that part.  The part I'd most like some feedback on is the
> > heavyweight lock bits.  I'll add this to the commitfest.
>
> The patch does not apply on top of HEAD as in [1], please post a rebased patch:

Rebased.  I dropped the CV patch for now.

Commits

Fix crash if LockErrorCleanup() is called twice
- 51a0382e8d87 18.0 landed
Split ProcSleep function into JoinWaitQueue and ProcSleep
- 3c0fd64fec8e 18.0 landed
Move TRACE calls into WaitOnLock()
- 6ae0897e4241 18.0 landed
Set MyProc->heldLocks in ProcSleep
- 0464f25b6aa1 18.0 landed
Fix comment in LockReleaseAll() on when locallock->nLock can be zero
- 1fe0466cf2cb 18.0 landed
Allow a no-wait lock acquisition to succeed in more cases.
- 2346df6fc373 17.0 cited
Teach planner about more monotonic window functions
- 456fa635a909 16.0 cited