Horton

Re: Patch to improve performance of replay of AccessExclusiveLock

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: David Rowley <david.rowley@2ndquadrant.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2017-03-07T17:04:12Z
Lists: pgsql-hackers 
Hi,

On 2017-03-08 00:15:05 +1300, David Rowley wrote:
> -static List *RecoveryLockList;
> +/*
> + * RecoveryLockTable is a poor man's hash table that allows us to partition
> + * the stored locks. Which partition a lock is stored in is determined by the
> + * xid which the lock belongs to. The hash function is very simplistic and
> + * merely performs a binary AND against the final 0-based partition number.
> + * Splitting into partitions in this way avoids having to look through all
> + * locks to find one specific to a given xid.
> + */
> +static List **RecoveryLockTable;

Why are we open coding this? That strikes me as a bad plan...

Regards,

Andres

Commits

  1. Move RecoveryLockList into a hash table.