This series collects a few follow-up for the backlog refactor, with some
of them posing as fixes just to confuse the enemy.

Targeting net-next as the issues addressed are very old, the change
is quite invasive and, as mentioned, based on the BL refactor.

Overall this:
- introduces a tracepoint similar to tcp useful to track down the issues
  addressed here
- avoid rx drop in corner case scenarios, improving tput stability and
  peak tput over fast link
- refactor the rcv space and rtt estimator, overall making DRS more
  correct and avoiding rcv buffer drifting to tcp_rmem[2], which in
  turn makes the tput more stable and less bursty

Note that after this upstream series will be merged:

https://patchwork.kernel.org/user/todo/netdevbpf/?series=1025209

we want to introduce similar changes in mptcp, too.

I'm sorry for the high frequency spamming; trying to complete this
stuff before ENORESOURCES. Please ignore the previous 2 revisions.
---
v6 -> v7:
   - re added the patches dropped in the last 2 iterations

v5 -> v6:
  - dropped the buggy patch 3/5

v4 -> v5:
  - dropped patch 4/6; it proved to be useless together with new version
    of patch 4/5
  - refactor rtt estimator: use sliding window instead of dropping "too
    high" values

v3 -> v4:
  - dropped already merged patch 1
  - update comment in patch 6/6

v2 -> v3:
  - use __entry->family in patch 1/7
  - more verbose commit message for patch 6/7

v1 -> v2:
  - new patch 1/7
  - old patch 2/4 has been split in 3: 3/7, 4/7 and 5/7
  - refactor both the rcv space init and rtt estimator

Paolo Abeni (6):
  trace: mptcp: add mptcp_rcvbuf_grow tracepoint
  mptcp: do not account for OoO in mptcp_rcvbuf_grow()
  mptcp: fix receive space timestamp initialization.
  mptcp: consolidate rcv space init
  mptcp: better mptcp-level RTT estimator
  mptcp: add receive queue awareness in tcp_rcv_space_adjust()

 include/trace/events/mptcp.h |  80 ++++++++++++++++++++++++
 net/mptcp/protocol.c         | 116 ++++++++++++++++++-----------------
 net/mptcp/protocol.h         |  44 ++++++++++++-
 net/mptcp/subflow.c          |   2 -
 4 files changed, 183 insertions(+), 59 deletions(-)

-- 
2.51.1

Hi Paolo,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal (except selftest_mptcp_join): Success! ✅
- KVM Validation: normal (only selftest_mptcp_join): Success! ✅
- KVM Validation: debug (except selftest_mptcp_join): Success! ✅
- KVM Validation: debug (only selftest_mptcp_join): Success! ✅
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/19531247842

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/9db02a7e6d33
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=1025709


If there are some issues, you can reproduce them using the same environment as
the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-normal

For more details:

    https://github.com/multipath-tcp/mptcp-upstream-virtme-docker


Please note that despite all the efforts that have been already done to have a
stable tests suite when executed on a public CI like here, it is possible some
reported issues are not due to your modifications. Still, do not hesitate to
help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (NGI0 Core)

Hi Paolo, Mat,

On 20/11/2025 09:39, Paolo Abeni wrote:
> This series collects a few follow-up for the backlog refactor, with some
> of them posing as fixes just to confuse the enemy.
> 
> Targeting net-next as the issues addressed are very old, the change
> is quite invasive and, as mentioned, based on the BL refactor.
> 
> Overall this:
> - introduces a tracepoint similar to tcp useful to track down the issues
>   addressed here
> - avoid rx drop in corner case scenarios, improving tput stability and
>   peak tput over fast link
> - refactor the rcv space and rtt estimator, overall making DRS more
>   correct and avoiding rcv buffer drifting to tcp_rmem[2], which in
>   turn makes the tput more stable and less bursty

Thank you for the new improvements and reviews!

I guess the answer to my question in patch 5/6 will be that nothing
needs to be changed, so I decided to apply this series (hopefully) to
ease things. Worst case, I can easily amend/revert patches!

> Note that after this upstream series will be merged:
> 
> https://patchwork.kernel.org/user/todo/netdevbpf/?series=1025209
> 
> we want to introduce similar changes in mptcp, too.

(ah, that's where I saw that message, and just noticed that was no urgency!)


Now in our tree (feat. for net-next):

New patches for t/upstream:
- 1ac40c258962: trace: mptcp: add mptcp_rcvbuf_grow tracepoint
- 744ffd37f43a: mptcp: do not account for OoO in mptcp_rcvbuf_grow()
- 2c093fc5ab1b: mptcp: fix receive space timestamp initialization
- f2a2ee211112: mptcp: consolidate rcv space init
- 436eef292ce0: mptcp: better mptcp-level RTT estimator
- 4d83ab6b697c: mptcp: add receive queue awareness in tcp_rcv_space_adjust()
- Results: 1fea9a6bd10f..36ddaf26cab4 (export)

Tests are now in progress:

- export:
https://github.com/multipath-tcp/mptcp_net-next/commit/51795655931ed517a9362868f07690f569417880/checks

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.