Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 17 Dec 2004 23:58:58 +0200
From:      Dmitry Ivanov <dimss@solutions.lv>
To:        freebsd-questions@freebsd.org
Subject:   Re: rl0: watchdog timeout
Message-ID:  <200412172358.58286.dimss@solutions.lv>
In-Reply-To: <20041217083524.12a575bc.nico.meijer@zonnet.nl>
References:  <E1CeeVc-00018M-00.gtsy-mail-ru@f12.mail.ru> <200412162023.26041.dimss@solutions.lv> <20041217083524.12a575bc.nico.meijer@zonnet.nl>

next in thread | previous in thread | raw e-mail | index | archive | help
On Friday 17 December 2004 09:35, Nico Meijer wrote:
> I don't (and won't, unless forced) use Intel NICs, so I cannot speak
> of them. I've had the very unpleasant experience of having had to
> deal with a fierce network boost on a RealTek 8139 (don't ask...) in
> a linux box. It meant the nic (and thus the machine) was unavailable
> for 15 minutes. Other machines (which were not mine) which
> experienced the boost, were humming along nicely. So thanks, I'll pay
> $50 extra for the nic.

Is there anything more than high reliability in Intel/3COM NICs? Do they 
produce less interrupts? I've heard rumors that they do some packet 
processing themselves thus offloading CPU. Is that true? If so, where 
can I check that in kernel source?

-- 
...python is just now at 2.4? perl is 3.4 better!



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200412172358.58286.dimss>