Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Jarod Wilson <jarod@redhat.com>
To:     linux-kernel@vger.kernel.org
Cc:     Jarod Wilson <jarod@redhat.com>
Subject: [PATCH net-next v3 0/2] bonding/balance-alb: support VMs behind bridges better
Date:   Thu, 27 May 2021 16:00:49 -0400
Message-Id: <20210527200051.1871092-1-jarod@redhat.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

I've been further educated on a use case, where a bridge sits on top of
a bond, with multiple vnetX interfaces attached to virtual machines,
also acting as ports of the bridge. Each leg of the bond goes to a
different switch, but there is NO mlag/vpc in play, the bonding driver
has to handle traffic that loops back appropriately to avoid breaking
transmission. Rather than adding some sort of mac filtering to
balance-xor mode, we switched to using balance-alb, which already does
some of this, and with the tweaks provided in this series, empirically
seems to behave as desired in actual operation.

v2 attempts to support srcmac-only hashing via a modparam instead of by
adding yet another hashing mode, as well as cleaning up and clarifying
commit messages.

v3 omits two patches from the series that prove unnecessary when all
slaves get promiscuity propagated to them.

Jarod Wilson (2):
  bonding: add pure source-mac-based tx hashing option
  bonding/balance-alb: put all slaves into promisc

 Documentation/networking/bonding.rst | 13 +++++++++++++
 drivers/net/bonding/bond_main.c      | 23 +++++++++++++++--------
 2 files changed, 28 insertions(+), 8 deletions(-)

-- 
2.30.2