netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Sabrina Dubroca <sd@queasysnail.net>
To: Antonio Quartulli <antonio@openvpn.net>
Cc: netdev@vger.kernel.org, Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Donald Hunter <donald.hunter@gmail.com>,
	Shuah Khan <shuah@kernel.org>,
	ryazanov.s.a@gmail.com, Andrew Lunn <andrew+netdev@lunn.ch>,
	Simon Horman <horms@kernel.org>,
	linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
	Xiao Liang <shaw.leon@gmail.com>,
	willemdebruijn.kernel@gmail.com
Subject: Re: [PATCH net-next v16 07/26] ovpn: introduce the ovpn_socket object
Date: Thu, 9 Jan 2025 12:28:56 +0100	[thread overview]
Message-ID: <Z3-y-AwfTyf924tP@hog> (raw)
In-Reply-To: <9634a1e1-6cc4-45ef-89d8-30d0e50ba319@openvpn.net>

2025-01-06, 00:27:28 +0100, Antonio Quartulli wrote:
> Hi Sabrina,
> 
> On 03/01/2025 18:00, Sabrina Dubroca wrote:
> > Hello Antonio,
> > 
> > 2024-12-19, 02:42:01 +0100, Antonio Quartulli wrote:
> > > +static void ovpn_socket_release_kref(struct kref *kref)
> > > +	__releases(sock->sock->sk)
> > > +{
> > > +	struct ovpn_socket *sock = container_of(kref, struct ovpn_socket,
> > > +						refcount);
> > > +
> > 
> > [extend with bits of patch 9]
> > > 	/* UDP sockets are detached in this kref callback because
> > > 	 * we now know for sure that all concurrent users have
> > > 	 * finally gone (refcounter dropped to 0).
> > > 	 *
> > > 	 * Moreover, detachment is performed under lock to prevent
> > > 	 * a concurrent ovpn_socket_new() call with the same socket
> > > 	 * to find the socket still attached but with refcounter 0.
> > 
> > I'm not convinced this really works, because ovpn_socket_new() doesn't
> > use the same lock. lock_sock and bh_lock_sock both "lock the socket"
> > in some sense, but they're not mutually exclusive (we talked about
> > that around the TCP patch).
> 
> You're right - but what prevents us from always using bh_lock_sock?

TCP detach can sleep, and UDP attach as well (setup_udp_tunnel_sock ->
udp_tunnel_encap_enable -> udp_encap_enable -> static_branch_inc ->
static_key_slow_inc -> cpus_read_lock). UDP detach would also not work
under bh_lock_sock if it really disabled encap on the socket (we end
up in udp_tunnel_encap_enable but that doesn't do anything since encap
is already turned on -- but a "real" detach should disable the encap
and do static_branch_dec).

So attach/detach need to be under lock_sock, not bh_lock_sock.

> > Are you fundamentally opposed to making attach permanent? ie, once
> > a UDP or TCP socket is assigned to an ovpn instance, it can't be
> > detached and reused. I think it would be safer, simpler, and likely
> > sufficient (I don't know openvpn much, but I don't see a use case for
> > moving a socket from one ovpn instance to another, or using it without
> > encap).
> 
> I hardly believe a socket will ever be moved to a different instance.
> There is no use case (and no userspace support) for that at the moment.
> 
> > 
> > Rough idea:
> >   - ovpn_socket_new is pretty much unchanged (locking still needed to
> >     protect against another simultaneous attach attempt, EALREADY case
> >     becomes a bit easier)
> >   - ovpn_peer_remove doesn't do anything socket-related
> >   - use ->encap_destroy/ovpn_tcp_close() to clean up sk_user_data
> >   - no more refcounting on ovpn_socket (since the encap can't be
> >     removed, the lifetime to ovpn_socket is tied to its socket)
> > 
> > What do you think?
> 
> hmm how would that work with UDP?
> On a server all clients may disconnect, but the UDP socket is expected to
> still survive and be re-used for new clients (userspace will keep it alive
> and keep listening for new clients).
> 
> Or you're saying that the socket will remain "attached" (i.e. sk_user_data
> set to the ovpn_priv*) even when no more clients are connected?

Yes. Once attached, it stays attached.

> > 
> > I'm trying to poke holes into this idea now. close() vs attach worries
> > me a bit.
> 
> Can that truly happen?

Actually it can't, so this isn't a concern.

> If a socket is going through close(), there should be some way to mark it as
> "non-attachable".
> 
> Actually, do we even need to clean up sk_user_data? The socket is being
> destroyed - why clean that up at all?

If we allocated some memory to store per-socket info, we need to free
it when we detach or close. There's no generic mechanism to free
sk_user_data since the core can't know where it came from, maybe
kfree() isn't appropriate.

-- 
Sabrina

  parent reply	other threads:[~2025-01-09 11:29 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-19  1:41 [PATCH net-next v16 00/26] Introducing OpenVPN Data Channel Offload Antonio Quartulli
2024-12-19  1:41 ` [PATCH net-next v16 01/26] net: introduce OpenVPN Data Channel Offload (ovpn) Antonio Quartulli
2024-12-19  1:41 ` [PATCH net-next v16 02/26] ovpn: add basic netlink support Antonio Quartulli
2024-12-20 11:00   ` Donald Hunter
2024-12-19  1:41 ` [PATCH net-next v16 03/26] ovpn: add basic interface creation/destruction/management routines Antonio Quartulli
2024-12-20 11:06   ` Donald Hunter
2024-12-19  1:41 ` [PATCH net-next v16 04/26] ovpn: keep carrier always on for MP interfaces Antonio Quartulli
2024-12-19  1:41 ` [PATCH net-next v16 05/26] ovpn: introduce the ovpn_peer object Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 06/26] kref/refcount: implement kref_put_sock() Antonio Quartulli
2024-12-19 17:20   ` Will Deacon
2024-12-31  7:31     ` Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 07/26] ovpn: introduce the ovpn_socket object Antonio Quartulli
2025-01-03 17:00   ` Sabrina Dubroca
2025-01-05 23:27     ` Antonio Quartulli
2025-01-08 10:55       ` Antonio Quartulli
2025-01-09 11:28       ` Sabrina Dubroca [this message]
2024-12-19  1:42 ` [PATCH net-next v16 08/26] ovpn: implement basic TX path (UDP) Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 09/26] ovpn: implement basic RX " Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 10/26] ovpn: implement packet processing Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 11/26] ovpn: store tunnel and transport statistics Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 12/26] ipv6: export inet6_stream_ops via EXPORT_SYMBOL_GPL Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 13/26] ovpn: implement TCP transport Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 14/26] skb: implement skb_send_sock_locked_with_flags() Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 15/26] ovpn: add support for MSG_NOSIGNAL in tcp_sendmsg Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 16/26] ovpn: implement multi-peer support Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 17/26] ovpn: implement peer lookup logic Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 18/26] ovpn: implement keepalive mechanism Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 19/26] ovpn: add support for updating local UDP endpoint Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 20/26] ovpn: add support for peer floating Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 21/26] ovpn: implement peer add/get/dump/delete via netlink Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 22/26] ovpn: implement key add/get/del/swap " Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 23/26] ovpn: kill key and notify userspace in case of IV exhaustion Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 24/26] ovpn: notify userspace when a peer is deleted Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 25/26] ovpn: add basic ethtool support Antonio Quartulli
2024-12-19  1:42 ` [PATCH net-next v16 26/26] testing/selftests: add test tool and scripts for ovpn module Antonio Quartulli
2024-12-20  4:02   ` Jakub Kicinski
2024-12-31  7:42     ` Antonio Quartulli

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Z3-y-AwfTyf924tP@hog \
    --to=sd@queasysnail.net \
    --cc=andrew+netdev@lunn.ch \
    --cc=antonio@openvpn.net \
    --cc=donald.hunter@gmail.com \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=ryazanov.s.a@gmail.com \
    --cc=shaw.leon@gmail.com \
    --cc=shuah@kernel.org \
    --cc=willemdebruijn.kernel@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).