[PATCH net 00/10] mptcp: more misc. fixes for v6.8

Matthieu Baerts (NGI0) posted 10 patches 2 months ago
Failed in applying to current master (apply log)
net/mptcp/diag.c                                |  3 ++
net/mptcp/options.c                             |  2 +-
net/mptcp/pm_userspace.c                        | 10 +++++
net/mptcp/protocol.c                            | 52 ++++++++++++++++++++++++-
net/mptcp/protocol.h                            | 21 +++++-----
tools/testing/selftests/net/mptcp/diag.sh       | 30 +++++++++++++-
tools/testing/selftests/net/mptcp/mptcp_join.sh | 33 ++++++++++------
tools/testing/selftests/net/mptcp/mptcp_lib.sh  |  4 +-
8 files changed, 128 insertions(+), 27 deletions(-)
[PATCH net 00/10] mptcp: more misc. fixes for v6.8
Posted by Matthieu Baerts (NGI0) 2 months ago
This series includes 6 types of fixes:

- Patch 1 fixes v4 mapped in v6 addresses support for the userspace PM,
  when asking to delete a subflow. It was done everywhere else, but not
  there. Patch 2 validates the modification, thanks to a subtest in
  mptcp_join.sh. These patches can be backported up to v5.19.

- Patch 3 is a small fix for a recent bug-fix patch, just to avoid
  printing an irrelevant warning (pr_warn()) once. It can be backported
  up to v5.6, alongside the bug-fix that has been introduced in the
  v6.8-rc5.

- Patches 4 to 6 are fixes for bugs found by Paolo while working on
  TCP_NOTSENT_LOWAT support for MPTCP. These fixes can improve the
  performances in some cases. Patches can be backported up to v5.6,
  v5.11 and v6.7 respectively.

- Patch 7 makes sure 'ss -M' is available when starting MPTCP Join
  selftest as it is required for some subtests since v5.18.

- Patch 8 fixes a possible double-free on socket dismantle. The issue
  always existed, but was unnoticed because it was not causing any
  problem so far. This fix can be backported up to v5.6.

- Patch 9 is a fix for a very recent patch causing lockdep warnings in
  subflow diag. The patch causing the regression -- which fixes another
  issue present since v5.7 -- should be part of the future v6.8-rc6.
  Patch 10 validates the modification, thanks to a new subtest in
  diag.sh.

Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
Davide Caratti (1):
      mptcp: fix double-free on socket dismantle

Geliang Tang (3):
      mptcp: map v4 address to v6 when destroying subflow
      selftests: mptcp: rm subflow with v4/v4mapped addr
      selftests: mptcp: join: add ss mptcp support check

Matthieu Baerts (NGI0) (1):
      mptcp: avoid printing warning once on client side

Paolo Abeni (5):
      mptcp: push at DSS boundaries
      mptcp: fix snd_wnd initialization for passive socket
      mptcp: fix potential wake-up event loss
      mptcp: fix possible deadlock in subflow diag
      selftests: mptcp: explicitly trigger the listener diag code-path

 net/mptcp/diag.c                                |  3 ++
 net/mptcp/options.c                             |  2 +-
 net/mptcp/pm_userspace.c                        | 10 +++++
 net/mptcp/protocol.c                            | 52 ++++++++++++++++++++++++-
 net/mptcp/protocol.h                            | 21 +++++-----
 tools/testing/selftests/net/mptcp/diag.sh       | 30 +++++++++++++-
 tools/testing/selftests/net/mptcp/mptcp_join.sh | 33 ++++++++++------
 tools/testing/selftests/net/mptcp/mptcp_lib.sh  |  4 +-
 8 files changed, 128 insertions(+), 27 deletions(-)
---
base-commit: b0b1210bc150fbd741b4b9fce8a24541306b40fc
change-id: 20240223-upstream-net-20240223-misc-fixes-1630cd6b3b0a

Best regards,
-- 
Matthieu Baerts (NGI0) <matttbe@kernel.org>