MIME-Version: 1.0
References: <11505016-8cf3-4691-b996-7faed99b7877.xiyuan.zr@alibaba-inc.com>
In-Reply-To: <11505016-8cf3-4691-b996-7faed99b7877.xiyuan.zr@alibaba-inc.com>
From: Robert Haas <robertmhaas@gmail.com>
Date: Mon, 29 Jul 2024 14:10:27 -0400
Message-ID: <CA+TgmobgW-hyqJwVO+J-Lf+o51Uw8ays1swoHz1Jtb4QX2S7=g@mail.gmail.com>
Subject: Re: Detach shared memory in Postmaster child if not needed
To: Rui Zhao <xiyuan.zr@alibaba-inc.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>, hlinnaka <hlinnaka@iki.fi>, 
	tgl <tgl@sss.pgh.pa.us>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: <https://www.postgresql.org/message-id/CA%2BTgmobgW-hyqJwVO%2BJ-Lf%2Bo51Uw8ays1swoHz1Jtb4QX2S7%3Dg%40mail.gmail.com>
Precedence: bulk

On Mon, Jul 29, 2024 at 5:57=E2=80=AFAM Rui Zhao <xiyuan.zr@alibaba-inc.com=
> wrote:
> Prior to PG16, postmaster children would manually detach from shared memo=
ry
> if it was not needed. However, this behavior was removed in fork mode in
> commit aafc05d.

Oh. The commit message makes no mention of that. I wonder whether it
was inadvertent.

> Detaching shared memory when it is no longer needed is beneficial, as
> postmaster children (like syslogger) don't wish to take any risk of
> accidentally corrupting shared memory. Additionally, any panic in these
> processes will not reset shared memory.

+1.

> The attached patch addresses this issue by detaching shared memory after
> fork_process().

I don't know whether this is the correct approach or not, but
hopefully Heikki can comment.

--=20
Robert Haas
EDB: http://www.enterprisedb.com