From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
cc: Heikki Linnakangas <heikki.linnakangas@iki.fi>,
        pgsql-committers@lists.postgresql.org
Subject: Re: pgsql: Fix double-release of spinlock
In-reply-to: <20240729164026.yurg37tej34o76uk@awork3.anarazel.de>
References: <E1sYSF2-001lEB-D1@gemulon.postgresql.org>
 <2600757.1722267116@sss.pgh.pa.us>
 <20240729161846.kgiafbe2dtgabxgb@awork3.anarazel.de>
 <2802449.1722270793@sss.pgh.pa.us>
 <20240729164026.yurg37tej34o76uk@awork3.anarazel.de>
Comments: In-reply-to Andres Freund <andres@anarazel.de>
	message dated "Mon, 29 Jul 2024 09:40:26 -0700"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2803948.1722271519.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Mon, 29 Jul 2024 12:45:19 -0400
Message-ID: <2803949.1722271519@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2803949.1722271519%40sss.pgh.pa.us>
Precedence: bulk

Andres Freund <andres@anarazel.de> writes:
> On 2024-07-29 12:33:13 -0400, Tom Lane wrote:
>> I dunno, is that the only extra check that the --disable-spinlocks
>> implementation is providing?

> I think it also provides the (valuable!) check that spinlocks were actua=
lly
> initialized. But that again seems like something we'd be better off addi=
ng
> more general infrastructure for - nobody runs --disable-spinlocks locall=
y, we
> shouldn't need to run this on the buildfarm to find problems like this.

Hmm, but how?  One of the things we gave up by nuking HPPA support
was that that platform's representation of an initialized, free
spinlock was not all-zeroes, so that it'd catch this type of problem.
I think all the remaining platforms do use zeroes, so it's hard to
see how anything short of valgrind would be likely to catch it.

			regards, tom lane