Date: Thu, 10 Feb 2011 17:04:58 -0500 From: Jay Borkenhagen <jayb@braeburn.org> To: Jack Vogel <jfvogel@gmail.com> Cc: freebsd-net@freebsd.org, Sergey Kandaurov <pluknet@gmail.com> Subject: Re: cannot set mtu > 1500 on em0 optical Gig-E Message-ID: <19796.24842.872566.716691@oz.mt.att.com> In-Reply-To: <AANLkTik7TNFr2dqOgcaQrHP8uaDmYijkxYF9HqUgSJi2@mail.gmail.com> References: <19796.13124.805432.342309@oz.mt.att.com> <AANLkTi=VfFbwWhEA-398fSzCQOFqbFg4enNqAt7zPeOP@mail.gmail.com> <AANLkTik7TNFr2dqOgcaQrHP8uaDmYijkxYF9HqUgSJi2@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Jack Vogel writes:
> Thanks Sergey :)
>
[...]
> >
> > Hmm.. Did jumbo frames really work on these cards for you? Intel
> > claims: "Adapters based on the Intel® 82542 and 82573V/E
> > controller do not support Jumbo Frames. These correspond to the
> > following product names:
> >
> > * Intel® PRO/1000 Gigabit Server Adapter
> > * Intel® PRO/1000 PM Network Connection "
> >
> > As seen from code, the behavior was changed to match the
> > description ("keep old 82542 from using jumbo frames") and merged
> > to 6-STABLE before 6.3.
> >
Well, I've got tcpdumps showing DF-bit-set pings succeeding to and
from such a NIC with MTU >> 1500, on a box running 6.2-RELEASE.
(pciconf data below.) In this particular case I have MTU=4470, which
is really good enough for what I need, and DF pings succeed up to
'ping -s 4442'
Maybe Intel thought they couldn't handle MTUs approaching 9k, or maybe
they had problems approaching linerate -- neither of which are
concerns in my application...
em0@pci0:2:0: class=0x020000 card=0x10008086 chip=0x10008086 rev=0x03 hdr=0x00
vendor = 'Intel Corporation'
device = '82542 Gigabit Ethernet Controller'
class = network
subclass = ethernet
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19796.24842.872566.716691>