MIME-Version: 1.0
References: <19508-e5f188183279219b@postgresql.org>
 <CAHGQGwGaC8RNeVP3A4_R1OcHJ-DrXgr2HH2UEYrt-Ueu516Gng@mail.gmail.com>
In-Reply-To: 
 <CAHGQGwGaC8RNeVP3A4_R1OcHJ-DrXgr2HH2UEYrt-Ueu516Gng@mail.gmail.com>
From: Ashutosh Sharma <ashu.coek88@gmail.com>
Date: Fri, 5 Jun 2026 20:59:23 +0530
Message-ID: 
 <CAE9k0Pkr+KKSx4UVmrLp-ZgF1BtwoVCOHOL99MeXr81D8JnX1w@mail.gmail.com>
Subject: Re: BUG #19508: pg_buffercache_pages() crashes the backend with an
 incompatible caller-supplied record definition
To: Fujii Masao <masao.fujii@gmail.com>
Cc: n.kalinin@postgrespro.ru, pgsql-bugs@lists.postgresql.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAE9k0Pkr%2BKKSx4UVmrLp-ZgF1BtwoVCOHOL99MeXr81D8JnX1w%40mail.gmail.com>
Precedence: bulk

Hi Fujii-san,

On Fri, Jun 5, 2026 at 8:49=E2=80=AFAM Fujii Masao <masao.fujii@gmail.com> =
wrote:
>
> On Fri, Jun 5, 2026 at 12:49=E2=80=AFAM PG Bug reporting form
> <noreply@postgresql.org> wrote:
> >
> > The following bug has been logged on the website:
> >
> > Bug reference:      19508
> > Logged by:          Nikita Kalinin
> > Email address:      n.kalinin@postgrespro.ru
> > PostgreSQL version: 19beta1
> > Operating system:   Fedora 44
> > Description:
> >
> > Hello,
> >
> > It appears that pg_buffercache_pages() trusts a caller-supplied record
> > descriptor without verifying that the declared column types match the a=
ctual
> > values returned by the function.
> >
> > The crash is reproducible on the current master branch with a fresh clu=
ster
> > after installing the extension:
>
> Thanks for the report! I could reproduce this as well.
>
>
> > git blame points to the following commit:
> >
> > commit 257c8231bf97a77378f6fedb826b1243f0a41612 (HEAD)
> > Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
> > Date:   Tue Apr 7 16:04:48 2026 +0300
> >
> >     Modernize and optimize pg_buffercache_pages()
>
> Commit 257c8231bf9 changed pg_buffercache_pages() to materialize rows dir=
ectly
> into a tuplestore. As a result, the function started using the caller-sup=
plied
> RECORD descriptor as rsinfo->setDesc, so a mismatched column definition l=
ist
> could cause tuplestore_putvalues() to interpret returned Datums with inco=
rrect
> types.
>
> Before that change, pg_buffercache_pages() exposed its actual tuple descr=
iptor
> to the executor, allowing the executor's existing rowtype checks to rejec=
t
> incompatible definitions with a normal error.
>
> The attached patch restores that behavior while keeping the materialized-=
SRF
> implementation. Thoughts?
>

pg_buffercache_pages uses RETURNS SETOF RECORD whereas other
extensions like pgstattuple define explicit IN/OUT parameters at the
SQL level. Is there a specific reason this pattern was kept, or is it
simply a legacy design that hasn't been modernized? Had we followed
the IN/OUT parameter style, this sort of issue could have been
avoided, no?

--
With Regards,
Ashutosh Sharma.