[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20200706.124746.191843852351699222.davem@davemloft.net>
Date: Mon, 06 Jul 2020 12:47:46 -0700 (PDT)
From: David Miller <davem@...emloft.net>
To: matthieu.baerts@...sares.net
Cc: netdev@...r.kernel.org, cpaasch@...le.com,
mathew.j.martineau@...ux.intel.com, kuba@...nel.org,
shuah@...nel.org, mptcp@...ts.01.org,
linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next] selftests: mptcp: capture pcap on both sides
From: Matthieu Baerts <matthieu.baerts@...sares.net>
Date: Mon, 6 Jul 2020 14:44:08 +0200
> When investigating performance issues that involve latency / loss /
> reordering it is useful to have the pcap from the sender-side as it
> allows to easier infer the state of the sender's congestion-control,
> loss-recovery, etc.
>
> Allow the selftests to capture a pcap on both sender and receiver so
> that this information is not lost when reproducing.
>
> This patch also improves the file names. Instead of:
>
> ns4-5ee79a56-X4O6gS-ns3-5ee79a56-X4O6gS-MPTCP-MPTCP-10.0.3.1.pcap
>
> We now have something like for the same test:
>
> 5ee79a56-X4O6gS-ns3-ns4-MPTCP-MPTCP-10.0.3.1-10030-connector.pcap
> 5ee79a56-X4O6gS-ns3-ns4-MPTCP-MPTCP-10.0.3.1-10030-listener.pcap
>
> It was a connection from ns3 to ns4, better to start with ns3 then. The
> port is also added, easier to find the trace we want.
>
> Co-developed-by: Christoph Paasch <cpaasch@...le.com>
> Signed-off-by: Christoph Paasch <cpaasch@...le.com>
> Signed-off-by: Matthieu Baerts <matthieu.baerts@...sares.net>
Applied, thank you.
Powered by blists - more mailing lists