Horton

Re: LogwrtResult contended spinlock

Alvaro Herrera <alvherre@2ndquadrant.com>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Andres Freund <andres@anarazel.de>
Cc: pgsql-hackers@lists.postgresql.org, Jaime Casanova <jaime.casanova@2ndQuadrant.com>
Date: 2021-01-30T01:47:56Z
Lists: pgsql-hackers

Attachments

So I tried this, but -- perhaps not suprisingly -- I can't get it to
work properly; the synchronization fails.  I suspect I need some
barriers, but I tried adding a few (probably some that are not really
necessary) and that didn't have the expected effect.  Strangely, all
tests work for me, but the pg_upgrade one in particular fails.

(The attached is of course POC quality at best.)

I'll have another look next week.

-- 
Álvaro Herrera                            39°49'30"S 73°17'W

Commits

  1. Remove bogus assertion in pg_atomic_monotonic_advance_u64

  2. Add XLogCtl->logInsertResult

  3. Operate XLogCtl->log{Write,Flush}Result with atomics

  4. Split XLogCtl->LogwrtResult into separate struct members

  5. Introduce atomic read/write functions with full barrier semantics.

  6. Reduce the number of GetFlushRecPtr() calls done by walsenders.

  7. Remove most volatile qualifiers from xlog.c