Re: [RFC net-next v2 4/5] selftests: drv-net: add a kperf runner

[Jakub Kicinski <kuba@kernel.org>]

On Mon,  5 Oct 2026 13:11:13 -0700 Stanislav Fomichev wrote:
> Add a helper to manage kperf servers, run the client, and optionally report
> aggregate throughput and retransmits. Use a random control port and a
> unique PID file, and report server cleanup failures.
> 
> Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
> ---
>  .../testing/selftests/drivers/net/README.rst  |   8 +
>  .../drivers/net/hw/lib/py/__init__.py         |   6 +-
>  .../selftests/drivers/net/lib/py/__init__.py  |   4 +-
>  .../selftests/drivers/net/lib/py/kperf.py     | 161 ++++++++++++++++++
>  4 files changed, 176 insertions(+), 3 deletions(-)
>  create mode 100644 tools/testing/selftests/drivers/net/lib/py/kperf.py
> 
> diff --git a/tools/testing/selftests/drivers/net/README.rst b/tools/testing/selftests/drivers/net/README.rst
> index 1b7264d40b79..4355d5113ec8 100644
> --- a/tools/testing/selftests/drivers/net/README.rst
> +++ b/tools/testing/selftests/drivers/net/README.rst
> @@ -141,6 +141,14 @@ Arguments used to construct the communication channel.
>    for netns - name of the "remote" namespace
>    for ssh - name/address of the remote host
>  
> +kperf
> +~~~~~
> +
> +Performance tests which use kperf expect ``kperf-client`` in ``PATH`` on the
> +local host and ``kperf-server`` in ``PATH`` on both endpoints. Kperf is an
> +optional dependency; tests which require it are skipped when it is not
> +installed.

I don't think this belongs in the README.
README needs to be high signal, otherwise nobody will read.
People running perf tests will quickly figure out kperf by asking LLM