netdev
[Top] [All Lists]

Re: [Bonding-devel] Re: [RFC][bonding] Improve VLAN support on top of bo

To: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Subject: Re: [Bonding-devel] Re: [RFC][bonding] Improve VLAN support on top of bonding
From: Dan Hollis <goemon@xxxxxxxxx>
Date: Tue, 15 Jul 2003 11:16:38 -0700 (PDT)
Cc: Shmulik Hen <shmulik.hen@xxxxxxxxx>, bond-devel <bonding-devel@xxxxxxxxxxxxxxxxxxxxx>, linux-net <linux-net@xxxxxxxxxxxxxxx>, linux-netdev <netdev@xxxxxxxxxxx>, "David S. Miller" <davem@xxxxxxxxxx>, Jeff Garzik <jgarzik@xxxxxxxxx>, Jay Vosburgh <fubar@xxxxxxxxxx>, Amir Noam <amir.noam@xxxxxxxxx>, Noam Marom <noam.marom@xxxxxxxxx>, Tsippy Mendelson <tsippy.mendelson@xxxxxxxxx>
In-reply-to: <3F144466.8010003@candelatech.com>
Sender: netdev-bounce@xxxxxxxxxxx
On Tue, 15 Jul 2003, Ben Greear wrote:
> Dan Hollis wrote:
> > On Tue, 15 Jul 2003, Ben Greear wrote:
> >>I'd consider ignoring the HW accel unless you can prove it actually helps
> >>performance to a noticeable degree.  I have never seen results of any 
> >>benchmarking
> >>related to this...
> > For gigabit ethernet, it makes a *H*U*G*E* difference.
> I'm curious to see numbers.  The VLAN shim is only inserting
> a small shim header, at at most shifting the first part of the packet
> when sent a pre-built packet.
> Maybe the hw-accel turns on tcp checksumming or something too??

That is exactly what it does. hw tcp checksumming helps a LOT at gbe rates

-Dan
-- 
[-] Omae no subete no kichi wa ore no mono da. [-]


<Prev in Thread] Current Thread [Next in Thread>