[MPTCP] Re: [PATCH v2 mptcp-next] mptcp: add receive buffer auto-tuning

[Christoph Paasch <cpaasch at apple.com>]

[-- Attachment #1: Type: text/plain, Size: 1516 bytes --]

Hello Davide,

On 06/03/20 - 17:08, Davide Caratti wrote:
> On Wed, 2020-06-03 at 16:39 +0200, Florian Westphal wrote:
> > Davide Caratti <dcaratti(a)redhat.com> wrote:
> > > hello Florian,
> > > 
> > > net-next will not be accepting patches for a while - but I would like to
> > > fix this on some tree, so that periodic tests can be run over there. So, I
> > > will rebase the fallback rework on top of the rx buffer auto-tuning patch.
> > 
> > Do you plan to submit your patches to net or net-next when it reopens?
> 
> net-next, because the patch doesn't change functionality (moreover, the rx
> support for infinte maps is splitted into another patch). It just
> "improves" fallback to tcp.
> 
> > My plan wrt. autotuning is to send the selftest patch for net, and the
> > autotune one for net-next.  So, to me it would make more sense for your
> > work to go into mptcp-next first.  I would then apply the delta from
> > Christoph and send a new version for mptcp-next.
> 
> ok. Then, I just need to re-send [1] targeting mptcp-next (or Matthieu can
> change it with pwclient?), and let it settle for some days while reviews /
> further troubles occur. Correct?

I don't know if you saw this one, but your patch seems to reveal another
issue when MPTCP tries to connect to itself
(https://github.com/multipath-tcp/mptcp_net-next/issues/35).

AFAICS, it does not trigger because of your change, but rather just triggers
the WARN you added in subflow_data_ready.



Christoph