Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 26 Feb 2007 14:55:23 +0100
From:      Andre Oppermann <andre@freebsd.org>
To:        Kip Macy <kip.macy@gmail.com>
Cc:        freebsd-net <freebsd-net@freebsd.org>, Jack Vogel <jfvogel@gmail.com>
Subject:   Re: improved TSO interface needed
Message-ID:  <45E2E6CB.8030707@freebsd.org>
In-Reply-To: <b1fa29170702252328o470ce804n893093b6362c4464@mail.gmail.com>
References:  <b1fa29170702242255i323077e8t3e5cfe696431c50b@mail.gmail.com>	 <45E19B54.9060007@freebsd.org>	 <b1fa29170702250641w3b365a97u62f066087d1bffe8@mail.gmail.com>	 <45E1A3B4.7090002@freebsd.org>	 <2a41acea0702252053v2357b5f5tefbcf58375be1a2f@mail.gmail.com>	 <b1fa29170702252110h3217bf82pdc9a3b46561b1671@mail.gmail.com>	 <2a41acea0702252325k4ec64d59m95a63bebaf6ea30c@mail.gmail.com> <b1fa29170702252328o470ce804n893093b6362c4464@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Kip Macy wrote:
>> 64K is the max, and I believe that is taking into account the headers.
>> What size are you seeing that is larger?  Even if an mbuf chain comes
>> down that large, it does NOT form a single packet, it is packetized by
>> the hardware into MTU size, as I said, on the wire, ie as the receiver
>> sees it, its just a lot of packets. The only question is whether the
>> transmit hardware can handle it. Do you have eivdence of some
>> hardware that supports TSO in the driver that cant??
> 
> I see this when torture testing a couple of different 10GigE cards
> running at line rate.

Your reported overshooting of 10 bytes is due to a buglet that didn't
appear until now.  Please try this fix:

  http://people.freebsd.org/~andre/tso_fix-20070226.diff

-- 
Andre




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?45E2E6CB.8030707>