From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============8223984650582609513==" MIME-Version: 1.0 From: Paolo Abeni To: mptcp at lists.01.org Subject: [MPTCP] [PATCH mptcp-next 0/3] mptcp: re-enable snd buf autotune Date: Fri, 08 Jan 2021 12:50:00 +0100 Message-ID: X-Status: X-Keywords: X-UID: 7284 --===============8223984650582609513== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable It was lost in a recent refactor. With great power come great responsability, so we need a more accurate packet scheduler decision, or b/w usage for assimetric links will be sub-optimal. Patches 2 && 3 are the minimal changes I was able to cook somehow addressing the above = There are still some sporadic self-test failures for assimetric links. Looks like they are caused by retransmissions/receiver dropping packets due to MPTCP level OoO check. I'll try to address that asap, but I still consider this series for inclusion, so that it can stage a bit in the export branch. Note: patch 3 re-introduces the workqueue usage for the datapath (yes, it makes me feel sick, sad and tired :(. I was unable to avoid it. There is weird and convoluted schema to possibly avoiding (again) the workqueue usage. If you don't feel too scaring using a fake NAPI instance to delegate actions to other mptcp subflows, I can try to cook that... = Paolo Abeni (3): mptcp: re-enable sndbuf autotune mptcp: do not queue excessive data on subflows. mptcp: schedule work for better snd subflow selection net/mptcp/protocol.c | 85 +++++++++++++++++++++++++------------------- net/mptcp/protocol.h | 20 +++++++++++ net/mptcp/subflow.c | 10 +++++- 3 files changed, 77 insertions(+), 38 deletions(-) -- = 2.26.2 --===============8223984650582609513==--