[CUWiN-Dev] [mattw@seattlewireless.net: [Fwd: [Roofnet-hackers]
Fwd: [Madwifi-users] future directions]]
David Young
dyoung at pobox.com
Sun Feb 27 05:06:33 CST 2005
This e-mail (below) is making the rounds. My own personal interpretation
is that Linux is 2-3 years behind [*] the BSDs in its generic 802.11
support, and it will remain so until Linux kernel developers get their
collective heads out of their collective asses, and pick up the 802.11
subsystem that is called 'net80211' in BSD.
Part of the "major work going on outside the madwifi CVS repository"
that Sam alludes to may be virtual AP support, which he is working on.
VAPs are pretty neat by themselves, but what I think is especially cool
is that the VAP code sets a foundation for things like OPN, MultiNet,
et cetera.
Since the net80211 code can be re-used for all kinds 802.11 cards, a
small advance in net80211 can yield improvements in N drivers, without
N-times the effort. In the BSDs, there are several drivers under the
net80211 system:
ADMtek
AMD PCnetMobile
Atheros
Cisco/Aironet 802.11b
Intel 802.11a/b/g
Intel 802.11b
Intersil Prism II / Lucent WaveLAN / Apple Airport
Intersil Prism54
Ralink
Realtek
TI ACX-100
In Linux, there are just two under net80211:
Atheros
Realtek
One finds that in the remaining Linux drivers, there is an astounding
amount of code duplication. Code duplication means more defects, bigger
object size, inconsistent operation, greater maintenance difficulty,
and a tremendous duplication of programming effort.
Dave
[*] NetBSD is about a year behind FreeBSD---I haven't re-imported in
that long. Today I spent 8+ hours merging the latest code from FreeBSD.
Work still remains. It's going more smoothly than I expected!
----- Forwarded message from Matt Westervelt <mattw at seattlewireless.net> -----
Date: Sat, 26 Feb 2005 19:07:13 -0800
From: Matt Westervelt <mattw at seattlewireless.net>
To: David Young <dyoung at pobox.com>
Subject: [Fwd: [Roofnet-hackers] Fwd: [Madwifi-users] future directions]
Date: Sun, 20 Feb 2005 15:20:59 -0500
From: John Bicket <jbicket at amsterdam.lcs.mit.edu>
To: roofnet-hackers at amsterdam.lcs.mit.edu
Subject: [Roofnet-hackers] Fwd: [Madwifi-users] future directions
----- Forwarded message from Sam Leffler <sam at errno.com> -----
Subject: [Madwifi-users] future directions
From: Sam Leffler <sam at errno.com>
To: madwifi-users at lists.sourceforge.net
Date: Fri, 18 Feb 2005 20:48:52 -0800
Some folks may have noticed that I've been scarce. Michael Renzmann
asked me to post something explaining what's going on so people are
equally informed.
I started working on support for Atheros hardware ~3 years ago. My
motivation was that I wanted to build mesh networks for community
wireless using freely available components. At the time the only
accessible devices were made with Prism/Intersil parts. I considered
these too limiting (read antiquated) and so hooked up with Atheros to
bring support for their hardware to the open source community. My work
initially was only for BSD systems. The plan was to enlist someone in
the Linux community to do a driver for Linux. We approached numerous
people but were unable to locate someone so I agreed to do a driver and
this is where madwifi came from.
Madwifi has grown in fits and starts based on my available time and
funding (I'm self-employed). Various people and companies have
contributed along the way and each is gratefully acknowledged. While
Atheros has been great in providing access to materials and people,
contrary to what many folks think a large part of the time I've spent on
this project has been unpaid.
So where's this going? My original motivation for working on madwifi was
to make the hardware accessible to as wide a community as possible. Past
that I hoped to help the Linux community by bringing over the net80211
work from the BSD world. There's been an obvious need for a
device-independent 802.11 protocol implementation in Linux and I saw
net80211 as an excellent starting point. Recently however it's become
clear that the Linux kernel developers have no interest in leveraging
this work in any way shape or form. As a result, given that I don't use
Linux on a day-to-day basis, I have zero incentive to continue being
involved with the madwifi project and so have backed away.
Where does this leave madwifi? Michael Renzmann has stepped up to help
keep things moving forward. He's brought in a number of folks who look
to be doing great work and I really appreciate this. I cannot speak for
Atheros but I believe they have a strong commitment to insure Linux
users have good support for their products. The bottom line is that I
think things will improve with some new blood in the project; especially
folks that use Linux as their main system and track the kernel
development more closely. I will try to help the developers as my time
permits and will assist in bringing in new hal releases.
In terms of technical direction I think it's important that for Atheros
hardware support to be most accessible to the Linux community it work
well with the effort going on to create a "generic 802.11 stack" (as
they describe it). Unfortunately the existing code that's been adopted
is far from usable with Atheros hardware and has some design decisions
that make it painful (if not impractical) to support advanced features
such as SuperG. Trying to switch over to this code is a significant job
that when done will provide no tangible benefit (IMO). Further, there is
major work going on outside the madwifi CVS repository to extend the
functionality of the current code that will make switching to this
generic stack a step backwards. Nonetheless I still believe that the
right thing for the madwifi community is to have a code base that fits
into Linux. I tried to help this happen but it didn't work out.
Hopefully others will be able to see this through.
While I'm curtailing my involvement with madwifi I plan to continue work
on wireless networking support in FreeBSD. I expect madwifi users will
benefit from this work.
Sam
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Madwifi-users mailing list
Madwifi-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/madwifi-users
----- End forwarded message -----
----- End forwarded message -----
--
David Young OJC Technologies
dyoung at ojctech.com Urbana, IL * (217) 278-3933
More information about the CU-Wireless-Dev
mailing list