Re: Tighten asserts on ParallelWorkerNumber a little bit

From: Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>
To: Tomas Vondra <tomas(at)vondra(dot)me>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Tighten asserts on ParallelWorkerNumber a little bit
Date: 2026-03-13 10:53:29
Message-ID: abPsqcdVKcaK3XNS@ip-10-97-1-34.eu-west-3.compute.internal
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On Thu, Mar 12, 2026 at 07:59:43PM +0100, Tomas Vondra wrote:
> Hi,
>
> While hacking on something I happened to notice a couple asserts on
> ParallelWorkerNumber when collecting shared instrumentation:
>
> Assert(ParallelWorkerNumber <= node->shared_info->num_workers);
>
> This is not quire right, because num_workers is used to size arrays
> indexed by ParallelWorkerNumber. And the comment in parallel.c also
> claims (ParallelWorkerNumber < num_workers).
>
> So AFAICS the assert(s) should be
>
> Assert(ParallelWorkerNumber < node->shared_info->num_workers);

I think that you're right. It looks like the first one has been introduced
by bf11e7ee2e3 and then the others are probably copy/paste.

> I don't think we had issues with this not catching a bug. But it may be
> a bit misleading, so worth fixing and (probably) backpatching.

I do agree.

With that patch applied, the remaining one:

$ git grep "<=.*num_workers"
src/backend/executor/nodeGatherMerge.c: Assert(nreaders <= castNode(GatherMerge, gm_state->ps.plan)->num_workers);

does not need to be fixed, so I think that your patch does not miss any and
LGTM.

Regards,

--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Aleksander Alekseev 2026-03-13 10:55:44 Re: Define DatumGetInt8 function.
Previous Message Peter Eisentraut 2026-03-13 10:43:15 Re: Change copyObject() to use typeof_unqual