netdev - Re: kernel data plane

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAPWQB7ELZLECr7AABOx-2K-yTPGiO1qz6ErFg69eUJt9hxrr5g@mail.gmail.com>
Date:   Tue, 7 Feb 2017 10:03:10 -0800
From:   Joe Stringer <joe@....org>
To:     David Miller <davem@...emloft.net>
Cc:     Pravin B Shelar <pshelar@...ira.com>,
        netdev <netdev@...r.kernel.org>
Subject: Re: kernel data plane

On 7 February 2017 at 09:45, David Miller <davem@...emloft.net> wrote:
>
> Since I see that you guys have marked the kernel data plane of
> openvswitch as experimental and that it "has not been thoroughly
> tested" what is your long term plan for deprecation and eventual
> removal of it from the tree?
>
>         https://github.com/openvswitch/ovs/commit/602e24ee189b76a72fdd26c9ed70c5f6b919a2f3

That's not the kernel dataplane being marked as experimental.
Specifically, the entire sentence states:

"The *userspace-only* mode of Open vSwitch without DPDK is considered
experimental. It has not been thoroughly tested." (my emphasis)

OVS provides a userspace datapath, which is shared between a version
that uses DPDK devices and a version that uses tap devices. Previously
both of these were marked as experimental. Over the past year the
OVS-DPDK developers have been pushing to relax/remove this tag from
the DPDK portion; this patch leaves the experimental tag in place for
the version that does not use DPDK.

>
> If it's experimental and not being tested, I'm not too thrilled about
> new feature patches and whatnot being posted which require developer
> resources to audit and review if the official message is "don't use
> this."

The kernel openvswitch is very much still maintained, tested, and will
continue to be used.