[mptcp-next,v3,0/4] mptcp: replace per-addr listener sockets

Message

Florian Westphal Feb. 23, 2022, 11:08 a.m. UTC

Changes since v2:
- drop first patch again, no longer needed
- move announce list check to the new join hook (Paolo)
- add joinhook for timewait sockets
- hide join6 member for !ipv6 build

Changes since v2:
- avoid unneeded sk refcount inc/dec pair
- use more descriptive name for function name
- fix typo in rfc number (Mathieu)
- don't increment listner refcount, caller assumes noref

Not changed:
 - hook location, plain tcp sockets take precedence.
 - per-netns listener

Global listener means other namespaces are affected in case
of backlog overflow. Pernet listeners avoid that, so I don't see
a strong argument for tcp refactoring that would allow global sk.

This series introduces a hidden per-netns pseudo "listener" socket to
handle mptcp join requests with a valid token but that do not match an
existing listening socket.

First patch is a minor preparation patch: MPTCP Join requests packets
that fail to find a suitable socket by means of standard address/port
demultiplexing will be steered to a pseudo-listener, similar to TPROXY
interception.

This pseudo listener isn't bound to an address or port (all zero), so we
need to fetch the port number from the tcp header and not the listener sk.

Patch two adds a stub to the tcp demux code.
This has no functionality, its extra to make tcp datapath change stand out.

Third patch is the bulk work, it adds per netns listener and implements
token-based socket demultiplexing.

Last patch zaps the per-address sockets from mptcp, they are not needed
anymore.

Florian Westphal (4):
  mptcp: prefer ip address in syn skb instead of listen sk bound address
  tcp: add mptcp join demultiplex hooks
  mptcp: handle join requests via pernet listen socket
  mptcp: remove per-address listening sockets

 include/net/mptcp.h    |  15 +++
 net/ipv4/tcp_ipv4.c    |   7 ++
 net/ipv6/tcp_ipv6.c    |  26 +++--
 net/mptcp/ctrl.c       | 229 ++++++++++++++++++++++++++++++++++++++++-
 net/mptcp/pm_netlink.c |  82 +++------------
 net/mptcp/protocol.c   |   2 +-
 net/mptcp/protocol.h   |   4 +-
 net/mptcp/subflow.c    |   9 +-
 8 files changed, 287 insertions(+), 87 deletions(-)

Comments

Mat Martineau Feb. 24, 2022, 12:48 a.m. UTC | #1

On Wed, 23 Feb 2022, Florian Westphal wrote:

> Changes since v2:
> - drop first patch again, no longer needed
> - move announce list check to the new join hook (Paolo)
> - add joinhook for timewait sockets
> - hide join6 member for !ipv6 build
>
> Changes since v2:
> - avoid unneeded sk refcount inc/dec pair
> - use more descriptive name for function name
> - fix typo in rfc number (Mathieu)
> - don't increment listner refcount, caller assumes noref
>

Hi Florian -

Thanks for the v3, I think that covers the review comments on v2 for the 
functionality covered in the previous revs.

My main questions have to do with:

> Not changed:
> - hook location, plain tcp sockets take precedence.
> - per-netns listener
>

Do you have plans for either or both of these? Not sure if you have 
work-in-process or if we (the community) need to figure out a more 
detailed plan for TCP/MPTCP precedence and the degree of TCP changes that 
make sense upstream. Since this week's meeting is happening real soon, we 
can definitely sort it out there.

With the upstream kernel at 5.17-rc5 we should also try to figure out how 
much of the listener and userspace PM changes we want to merge to 
net-next now vs. in the next cycle.

-Mat


> Global listener means other namespaces are affected in case
> of backlog overflow. Pernet listeners avoid that, so I don't see
> a strong argument for tcp refactoring that would allow global sk.
>
> This series introduces a hidden per-netns pseudo "listener" socket to
> handle mptcp join requests with a valid token but that do not match an
> existing listening socket.
>
> First patch is a minor preparation patch: MPTCP Join requests packets
> that fail to find a suitable socket by means of standard address/port
> demultiplexing will be steered to a pseudo-listener, similar to TPROXY
> interception.
>
> This pseudo listener isn't bound to an address or port (all zero), so we
> need to fetch the port number from the tcp header and not the listener sk.
>
> Patch two adds a stub to the tcp demux code.
> This has no functionality, its extra to make tcp datapath change stand out.
>
> Third patch is the bulk work, it adds per netns listener and implements
> token-based socket demultiplexing.
>
> Last patch zaps the per-address sockets from mptcp, they are not needed
> anymore.
>
> Florian Westphal (4):
>  mptcp: prefer ip address in syn skb instead of listen sk bound address
>  tcp: add mptcp join demultiplex hooks
>  mptcp: handle join requests via pernet listen socket
>  mptcp: remove per-address listening sockets
>
> include/net/mptcp.h    |  15 +++
> net/ipv4/tcp_ipv4.c    |   7 ++
> net/ipv6/tcp_ipv6.c    |  26 +++--
> net/mptcp/ctrl.c       | 229 ++++++++++++++++++++++++++++++++++++++++-
> net/mptcp/pm_netlink.c |  82 +++------------
> net/mptcp/protocol.c   |   2 +-
> net/mptcp/protocol.h   |   4 +-
> net/mptcp/subflow.c    |   9 +-
> 8 files changed, 287 insertions(+), 87 deletions(-)
>
> -- 
> 2.34.1
>
>
>

--
Mat Martineau
Intel