Content-Type: multipart/mixed; boundary="===============6370295231726861928==" MIME-Version: 1.0 From: Mat Martineau To: mptcp at lists.01.org Subject: [MPTCP] Re: [PATCH v4 0/7] mptcp: just another complete datapath refactor Date: Wed, 18 Nov 2020 16:56:46 -0800 Message-ID: In-Reply-To: cover.1605737932.git.pabeni@redhat.com X-Status: X-Keywords: X-UID: 6806 --===============6370295231726861928== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Wed, 18 Nov 2020, Paolo Abeni wrote: > This iteration drops completely backlog usage, as per Mat suggestions. > Thanks for giving that a try! > The only updated patch WRT the previous iteration is the last one: > 7/7 mptcp: use mptcp backlog. > > Only lighlty tested, but at least it does not explode badly on the first > self-test iteration :) Do you want to test more thoroughly before adding to the export branch, or = apply to export now and follow up with squash-to fixes? I'm ok with the latter. > > Paolo Abeni (7): > mptcp: open code mptcp variant for lock_sock > mptcp: implement wmem reservation. > mptcp: protect the rx path with the msk socket spinlock > mptcp: allocate TX skbs in msk context. > mptcp: avoid a few atomic ops in the rx path > mptcp: simplify mptcp_nospace > mptcp: use mptcp backlog. > > net/core/sock.c | 2 +- > net/mptcp/mptcp_diag.c | 2 +- > net/mptcp/options.c | 47 +-- > net/mptcp/protocol.c | 706 ++++++++++++++++++++++++++++++----------- > net/mptcp/protocol.h | 36 ++- > net/mptcp/subflow.c | 31 +- > 6 files changed, 596 insertions(+), 228 deletions(-) > > -- = > 2.26.2 -- Mat Martineau Intel --===============6370295231726861928==--