Horton

Re: Improve LWLock tranche name visibility across backends

Sami Imseih <samimseih@gmail.com>

View thread View on postgr.es
From: Sami Imseih <samimseih@gmail.com>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Andres Freund <andres@anarazel.de>, Bertrand Drouvot <bertranddrouvot.pg@gmail.com>, Rahila Syed <rahilasyed90@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2025-08-25T21:37:21Z
Lists: pgsql-hackers 
> On Fri, Aug 22, 2025 at 03:01:53PM -0500, Sami Imseih wrote:
> > I kept the local array to serve consecutive reads and to avoid having to
> > take a shared lock on shared memory every time GetLWTrancheName is
> > called. A new LWLock to protect this array is required.
>
> I'm not seeing why we need this cache anymore.  This is an append-only
> list, so we could instead keep a backend-local copy of LWLockCounter that
> gets updated as needed.  As long as the ID is less than our backend-local
> counter, we can go straight to the shared array.  If it is greater, we'll
> have to first update our counter, which should be rare and inexpensive.

When we lookup from shared array only, we need to take a shared lock
every lookup. Acquiring that lock is what I am trying to avoid. You
are saying it's not worth optimizing that part, correct?

--
Sami

Commits

  1. test_dsa: Avoid leaking LWLock tranches.

  2. Teach DSM registry to ERROR if attaching to an uninitialized entry.

  3. Add a test harness for the LWLock tranche code.

  4. Revert recent change to RequestNamedLWLockTranche().

  5. Move dynamically-allocated LWLock tranche names to shared memory.

  6. Prepare DSM registry for upcoming changes to LWLock tranche names.

  7. Add GetNamedDSA() and GetNamedDSHash().