[v3] mptcp: a couple of fixes

[PATCH v3 mptcp-net 0/2] mptcp: a couple of fixes

Paolo Abeni posted 2 patches 1 year, 1 month ago

Diff against v2
Download series mbox

Failed in applying to current master (apply log)

Maintainers: Matthieu Baerts <matthieu.baerts@tessares.net>, "David S. Miller" <davem@davemloft.net>, Eric Dumazet <edumazet@google.com>, Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>

net/mptcp/protocol.c | 54 +++++++++++++++++-----------
net/mptcp/protocol.h |  2 ++
net/mptcp/subflow.c  | 83 ++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 119 insertions(+), 20 deletions(-)

Expand all Fold all

[PATCH v3 mptcp-net 0/2] mptcp: a couple of fixes

Posted by Paolo Abeni 1 year, 1 month ago

dump cover letter to trigger the CI, see the individual
changelog for the details.

Note that patch  2/2 introduces a behavioral change that hits
the packetdrill test close_before_accept.pkt. The msk socket
will be CLOSED after accept if the first subflow is closed
before accept.

v2 -> v3:
 - fix locking in patch 2/2 as spotted by Matttbe

Paolo Abeni (2):
  mptcp: stops worker on unaccepted sockets at listener close
  mptcp: fix accept vs worker race

 net/mptcp/protocol.c | 54 +++++++++++++++++-----------
 net/mptcp/protocol.h |  2 ++
 net/mptcp/subflow.c  | 83 ++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 119 insertions(+), 20 deletions(-)

-- 
2.39.2

Re: [PATCH v3 mptcp-net 0/2] mptcp: a couple of fixes

Posted by Matthieu Baerts 1 year, 1 month ago

Hi Paolo,

On 30/03/2023 21:22, Paolo Abeni wrote:
> dump cover letter to trigger the CI, see the individual
> changelog for the details.
> 
> Note that patch  2/2 introduces a behavioral change that hits
> the packetdrill test close_before_accept.pkt. The msk socket
> will be CLOSED after accept if the first subflow is closed
> before accept.

Thank you for the fixes!

Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>

Now in our tree (fixes for -net) without a few typos spot by
checkpatch.pl --shellcheck.

Please note that I had a few conflicts when applying them to -net. I
hope that's OK, please see below:

New patches for t/upstream-net and t/upstream:
- dddea846b342: mptcp: stops worker on unaccepted sockets at listener close
- 87668b995225: mptcp: fix accept vs worker race
- c6a6d1debd0d: conflict in
t/mptcp-move-first-subflow-allocation-at-mpc-access-time
- e04e8b8fb4b8: conflict in t/mptcp-refactor-mptcp_stream_accept
- Results: 199eed4f57e8..f01e02dfadff (export-net)
- Results: e2a442363ad9..68957ea8fdf9 (export)

Tests are now in progress:

https://cirrus-ci.com/github/multipath-tcp/mptcp_net-next/export-net/20230331T162750
https://cirrus-ci.com/github/multipath-tcp/mptcp_net-next/export/20230331T162750

Cheers,
Matt
-- 
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net