Date: Mon, 2 Apr 2001 02:57:40 +0200 (CEST) From: Bert Driehuis <driehuis@playbeing.org> To: freebsd-stable@freebsd.org Subject: Re: Network performance question Message-ID: <Pine.BSI.4.21.0104020242170.4679-100000@c1111.nl.compuware.com> In-Reply-To: <20010402023800.B75063@mail.webmonster.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 2 Apr 2001, Karsten W. Rohrbach wrote:
> mike, driver stuff can be implemented in not just one way or style of
> code as we all know. the freebsd approach is to have readable code, no
> awkward hacks (well, errrhm, almost ;) and a structure that meets the
> specifications of the hardware and, hopefully, is extendible for newer
> hw versions. the linux approach is to have a flying penguin touching the
> ground with the tip of his left foot, trying not to crash into the next
> obstacle - and so is the source. the linux community tends to fix driver
> problems on certain hardware with evil hacks and, hell, at least it
> _seems_ to work ;-)
I should probably shut up at some stage, but I couldn't let this one
pass :-)
Evil hacks are required from time to time. I still remember the rift
between the BSDI community and FreeBSD, because BSDI refused to
implement bounce buffers for the ISA Adaptec devices because they were
Evil. FreeBSD chose to be evil, and won a fair amount of BSDI users over
on these grounds.
Hardware is inherently buggy -- once it ships, you have no easy recourse
like recommending firmware updates; if a soldering iron is all that can
help the end user, the end user is screwed if the hardware has a bug.
I'm glad you mentioned the i82559 chip as an approved one, because I'm
currently dealing with a vendor that happened to miswire them, and broke
100/full autonegociation by doing so.
I'm not a big Linux fan, but if it does the job on a machine that
FreeBSD throws its hands up on, what am I to recommend? Buy new
hardware?
I've been through this loop often enough to recommend top notch hardware
from vendors that actually can tell a 100pF capacitor from a 100uF one
(which is apparently what broke those Intel based boards), but if a
workaround exists that doesn't jeopardize the users of Decent Hardware,
let a thousand workarounds blossom!
If life were easy, we'd all be out of a job. I'd far rather concentrate
on getting stuff to work than on bashing the competitor for bashings
sake.
{Free,Net}BSD have the edge in driver development for having a clear,
bus-agnostic driver development model. I never stop being amazed at how
cleanly most hardware bugs can be worked around if the basics are taken
care of.
Cheers,
-- Bert
--
Bert Driehuis -- driehuis@playbeing.org -- +31-20-3116119
If the only tool you've got is an axe, every problem looks like fun!
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSI.4.21.0104020242170.4679-100000>