Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 10 Jan 2017 07:43:15 -0700
From:      Sean Bruno <sbruno@freebsd.org>
To:        Harry Schmalzbauer <freebsd@omnilan.de>
Cc:        freebsd-net@freebsd.org
Subject:   Re: HEADS-UP: IFLIB implementations of sys/dev/e1000 em, lem, igb pending
Message-ID:  <7b048fa1-13bc-2a79-4fbe-6f049b30d65f@freebsd.org>
In-Reply-To: <5874B751.4000808@omnilan.de>
References:  <d6627189-c9ce-fc91-d71a-111e127b0b64@freebsd.org> <092ad9f7-b04c-292f-c626-6ce1956580a8@freebsd.org> <5874B751.4000808@omnilan.de>

next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
--TUhu1uw9kkfiAacp9X6jUMRiq8TtpJCRI
Content-Type: multipart/mixed; boundary="WQXCrKxk9r8kshjXet5Udnoqojabh7Mge";
 protected-headers="v1"
From: Sean Bruno <sbruno@freebsd.org>
To: Harry Schmalzbauer <freebsd@omnilan.de>
Cc: freebsd-net@freebsd.org
Message-ID: <7b048fa1-13bc-2a79-4fbe-6f049b30d65f@freebsd.org>
Subject: Re: HEADS-UP: IFLIB implementations of sys/dev/e1000 em, lem, igb
 pending
References: <d6627189-c9ce-fc91-d71a-111e127b0b64@freebsd.org>
 <092ad9f7-b04c-292f-c626-6ce1956580a8@freebsd.org>
 <5874B751.4000808@omnilan.de>
In-Reply-To: <5874B751.4000808@omnilan.de>

--WQXCrKxk9r8kshjXet5Udnoqojabh7Mge
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

>=20
> Thank you very much for your work!
> I'm unsure if I got it right: stabe/11 won't benefit from IFLIB (not
> that I know waht IFLIB is all about, yet), but driver changes will be
> merged?

Yes, we intend on pulling back IFLIB and the IFLIB'd drivers to
stable/11 for people's use.  The difference will be that the IFLIB
implementation will not be the default here, it will be a KERNCONF
option that you will be able to add during your upgrade process.

>=20
> I'm already using MULTIQUEUE support for em(4) (Hartwell, 82547) since
> 10.2-stable without problems on high saturated single links and
> noticable better performance.
>=20
> Currently I could test a pre-productive igb(4)/Kawela (82576) machine
> utilizing LACP+VLANs to chekck for regression (improvement regarding
> =C2=BBigb stopped distributiong, possible flapping?=C2=AB), but it's st=
able/11.
>=20
> Simply merging r311849 doesn't work, most likely due to the IFLIB
> compatibility reasons you mentioned.
> Have you already prepared a MFC verison?
> What should igb(4) users expect after MFC? (Christmas is over and I
> didn't get SR-IOV for igb(4)/82576, so this its still on my wish list ;=
-))
>=20
> Thnaks,
>=20
> -harry

I suspect end of month for an MFC with all the IFLIB bits.

sean


--WQXCrKxk9r8kshjXet5Udnoqojabh7Mge--

--TUhu1uw9kkfiAacp9X6jUMRiq8TtpJCRI
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----

iQGTBAEBCgB9FiEEuq1GMucSHejSCZfdEgHvyh5yfmQFAlh08wNfFIAAAAAALgAo
aXNzdWVyLWZwckBub3RhdGlvbnMub3BlbnBncC5maWZ0aGhvcnNlbWFuLm5ldEJB
QUQ0NjMyRTcxMjFERThEMjA5OTdERDEyMDFFRkNBMUU3MjdFNjQACgkQEgHvyh5y
fmT6rQf+LLcY+hXoaUbX5pPoGaN1X47q//NRMoj5MHpmWv1uPXZub+qAwdHv5pcx
2XpUk4wXYzW9aTQTaszx8ff94ttiJsGH1u7HH0P0v6qoYYlR4tPWASgs9qnnbp7x
luOnEOu1KQ+xfzpWvsSWygrkSdzLXaQ0EBXAqyDFBQWoDoHljLGCcQdiwSXj8eHB
USH9NgdUNnhDVepDf1mhucY105BN3IVJqm1RCHh78zfE71YXIGUUUlHBxr0Ky+vW
lZ8h72R9cVTrcJ4+XFKZAAmMq5oHGzCeScnplnbrwWI3vrpMQahaJSH1lq29CsXi
Bv2JKnsSjxU4blTLLjJXEx2H0vr0DQ==
=1Og1
-----END PGP SIGNATURE-----

--TUhu1uw9kkfiAacp9X6jUMRiq8TtpJCRI--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7b048fa1-13bc-2a79-4fbe-6f049b30d65f>