Horton

Re: A potential memory leak on Merge Join when Sort node is not below Materialize node

Peter Geoghegan <pg@bowt.ie>

View thread View on postgr.es
From: Peter Geoghegan <pg@bowt.ie>
To: David Rowley <dgrowleyml@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Ronan Dunklau <ronan.dunklau@aiven.io>, Önder Kalacı <onderkalaci@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-09-29T05:13:06Z
Lists: pgsql-hackers 
On Wed, Sep 28, 2022 at 9:59 PM David Rowley <dgrowleyml@gmail.com> wrote:
> select b from t1 order by b offset 1000000;
>
> Master:
> latency average = 344.763 ms
>
> Patched:
> latency average = 268.374 ms
>
> about 28% faster.

That's more like it!

-- 
Peter Geoghegan

Commits

  1. Allow nodeSort to perform Datum sorts for byref types

  2. Restrict Datum sort optimization to byval types only

  3. Make nodeSort.c use Datum sorts for single column sorts

  4. Fix actual and potential double-frees around tuplesort usage.

  5. Allow avoiding tuple copy within tuplesort_gettupleslot().