Date: Thu, 13 Jun 2019 14:00:14 -0400
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Oleksii Kliukin <alexk@hintbits.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: upgrades in row-level locks can deadlock
Message-ID: <20190613180014.GA9775@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <40AE3BAD-CE9B-4A34-9DB8-3017A66653C3@hintbits.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

On 2019-Jun-13, Oleksii Kliukin wrote:

> Makes sense. For the symmetry I have included those that perform lock
> upgrades in one session and those that doesn’t, while the other sessions
> acquire locks, do updates or deletes. For those that don’t upgrade locks the
> test checks that the locks are acquired in the correct order.

Thanks for the updated patch!  I'm about to push to branches 9.6-master.
It applies semi-cleanly (only pgindent-maturity whitespace conflicts).

The [pg11 version of the] patch does applies to 9.5 cleanly ... but the
isolation test doesn't work, because isolationtester was not smart
enough back then.  Since there have been no previous reports of this
problem, and to avoid pushing untested code, I'm going to refrain from
back-patching there.  My guess is that it should work ...

In 9.4 there are quite some severe conflicts, because 27846f02c176 was
not back-patched there.  (The bug number "#8470" still floats in my
memory from time to time.  Shudder)

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services