Received-SPF: pass (google.com: domain of linux-nfs+bounces-1445-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk
MIME-Version: 1.0
From: "NeilBrown" <neilb@suse.de>
To: "Chuck Lever III" <chuck.lever@oracle.com>
Cc: "Jeff Layton" <jlayton@kernel.org>,
 "Lorenzo Bianconi" <lorenzo@kernel.org>,
 "Linux NFS Mailing List" <linux-nfs@vger.kernel.org>,
 "Lorenzo Bianconi" <lorenzo.bianconi@redhat.com>,
 "Jakub Kicinski" <kuba@kernel.org>, "Simon Horman" <horms@kernel.org>,
 "open list:NETWORKING [GENERAL]" <netdev@vger.kernel.org>
Subject: Re: [PATCH v6 3/3] NFSD: add write_ports to netlink command
In-reply-to: <917EC07C-C9D9-4CF2-9ACB-DCA2676DFF67@oracle.com>
References: <cover.1705771400.git.lorenzo@kernel.org>,
 <f7c42dae2b232b3b06e54ceb3f00725893973e02.1705771400.git.lorenzo@kernel.org>,
 <9e3ae337dcf168c60c4cfd51aa0b2fc7b24bcbfb.camel@kernel.org>,
 <170595930799.23031.17998490973211605470@noble.neil.brown.name>,
 <Za7zHvPJdei/vWm4@tissot.1015granger.net>, <Za-N6BxOMXTGyxmW@lore-desk>,
 <85b02061798a1b750a87b0302681b86651d0c7a3.camel@kernel.org>,
 <Za-9P0NjlIsc1PcE@lore-desk>,
 <3f035d3bc494ec03b83ae237e407c42f2ddc4c53.camel@kernel.org>,
 <ZbDdzwvP6-O2zosC@lore-desk>,
 <8fabd83caa0d44369853a4040a89c069f9b0f935.camel@kernel.org>,
 <917EC07C-C9D9-4CF2-9ACB-DCA2676DFF67@oracle.com>
Date: Fri, 26 Jan 2024 09:44:01 +1100
Message-id: <170622264103.21664.16941742935452333478@noble.neil.brown.name>

On Thu, 25 Jan 2024, Chuck Lever III wrote:
>=20
>=20
> > On Jan 24, 2024, at 6:24=E2=80=AFAM, Jeff Layton <jlayton@kernel.org> wro=
te:
> >=20
> > On Wed, 2024-01-24 at 10:52 +0100, Lorenzo Bianconi wrote:
> >> [...]
> >>>=20
> >>> That's a great question. We do need to properly support the -H option to
> >>> rpc.nfsd. What we do today is look up the hostname or address using
> >>> getaddrinfo, and then open a listening socket for that address and then
> >>> pass that fd down to the kernel, which I think then takes the socket and
> >>> sticks it on sv_permsocks.
> >>>=20
> >>> All of that seems a bit klunky. Ideally, I'd say the best thing would be
> >>> to allow userland to pass the sockaddr we look up directly via netlink,
> >>> and then let the kernel open the socket. That will probably mean
> >>> refactoring some of the svc_xprt_create machinery to take a sockaddr,
> >>> but I don't think it looks too hard to do.
> >>=20
> >> Do we already have a specific use case for it? I think we can even add it
> >> later when we have a defined use case for it on top of the current serie=
s.
> >>=20
> >=20
> > Yes:
> >=20
> > rpc.nfsd -H makes nfsd listen on a particular address and port. By
> > passing down the sockaddr instead of an already-opened socket
> > descriptor, we can achieve the goal without having to open sockets in
> > userland.
>=20
> Tearing down a listener that was created that way would be a
> use case for:

Only if it was actually useful.
Have you *ever* wanted to do that?  Or heard from anyone else who did?

NeilBrown


>=20
> > Do we ever want/need to remove listening sockets?
> > Normal practice when making any changes is to stop and restart where
> > "stop" removes all sockets, unexports all filesystems, disables all
> > versions.
>=20
>=20
>=20
> --
> Chuck Lever
>=20
>=20
>=20