Slashdot Mirror

I'm fairly sure Linux has a generic 802.11 stack

The mac80211 stack? It's not "generic" enough to handle "FullMAC" devices (devices where the Media Access Control (MAC) Sublayer Management Entity (MLME) is handled by the device's hardware or firmware), but, as they say, "FullMAC devices have become scarce". I don't know whether any of the *BSD's 802.11 stacks handle both FullMAC and SoftMAC adapters; if so, those *BSDs have a more generic 802.11 stack, but if, as the mac80211 people note, there aren't many FullMAC devices (which I suspect means "there are few if any newer FullMAC devices", i.e. newer devices will leave the MLME to the OS), that's not a big difference these days.

and I have not idea why it would behave in a significantly different manner.

Because it was developed independently, and its developers had their own ideas about how things should be done (whether those ideas were "this is better", "Not Invented Here", or a combination of the two)?

I'm fairly sure Linux has a generic 802.11 stack

The mac80211 stack? It's not "generic" enough to handle "FullMAC" devices (devices where the Media Access Control (MAC) Sublayer Management Entity (MLME) is handled by the device's hardware or firmware), but, as they say, "FullMAC devices have become scarce". I don't know whether any of the *BSD's 802.11 stacks handle both FullMAC and SoftMAC adapters; if so, those *BSDs have a more generic 802.11 stack, but if, as the mac80211 people note, there aren't many FullMAC devices (which I suspect means "there are few if any newer FullMAC devices", i.e. newer devices will leave the MLME to the OS), that's not a big difference these days.

and I have not idea why it would behave in a significantly different manner.

Because it was developed independently, and its developers had their own ideas about how things should be done (whether those ideas were "this is better", "Not Invented Here", or a combination of the two)?

I'm fairly sure Linux has a generic 802.11 stack

The mac80211 stack? It's not "generic" enough to handle "FullMAC" devices (devices where the Media Access Control (MAC) Sublayer Management Entity (MLME) is handled by the device's hardware or firmware), but, as they say, "FullMAC devices have become scarce". I don't know whether any of the *BSD's 802.11 stacks handle both FullMAC and SoftMAC adapters; if so, those *BSDs have a more generic 802.11 stack, but if, as the mac80211 people note, there aren't many FullMAC devices (which I suspect means "there are few if any newer FullMAC devices", i.e. newer devices will leave the MLME to the OS), that's not a big difference these days.

and I have not idea why it would behave in a significantly different manner.

Because it was developed independently, and its developers had their own ideas about how things should be done (whether those ideas were "this is better", "Not Invented Here", or a combination of the two)?

To the best of my knowledge, the ath5k/madwifi drivers are the only Linux drivers to be ported from the BSDs (OpenBSD/FreeBSD) to Linux. Which other drivers out of the 56 Linux wifi drivers were ported from the BSDs to qualify the "large number of WiFi drivers were written for FreeBSD or OpenBSD and then ported to Linux" statement?

Linux has had its own 802.11 stack called mac802.11 since the 2.6.22 kernel four years ago which was developed by Devicescape. The only driver I know of that carried a (Net)BSD 802.11 stack over to Linux was madwifi which had net802.11, was never mainline and was superseded by ath5k... The madwifi driver never went mainline, nor did its net802.11 stack. Why do you think that the 802.11 stack from a BSD needs copying into a Linux driver when mac802.11 exists?

A little more info, more along the lines of why there is so much generic Broadcom hate and distrust within the Linux community:

http://linuxwireless.org/en/users/Drivers/b43 - If you go down about 3/4 of the page, you'll see that:
Until 9/9/2010, the only drivers for ANY broadcom device were created via reverse engineering. Broadcom provided ZERO support to the b43 developers, and I'm fairly certain they still don't have any proper technical documentation. (Sorry Broadcom, but source code isn't documentation.)
After 9/9/2010 - only THREE chipsets (out of quite many) had any sort of "official" open source driver support for Linux from Broadcom.

Meanwhile, chipsets from other manufacturers (Intel, Atheros, Intersil/Harris, Ralink) have had robust open source support for a VERY long time. For many years, Broadcom WLAN chipsets were completely useless in Linux due to Broadcom's refusal to provide any documentation.

Yes it is crabbing.. The level of functionality is actually really very good.
As to the lack of WiFi, here is a list of adaptors that will work http://linuxwireless.org/en/users/Devices/USB this list is not complete because their are only a few wifi to usb chips on the market so other adaptors will work as well.
Actually all of the USB wifi adaptors I have tired have worked but that is not a large test base.
Just for fun here are two that will work.

http://www.google.com/search?aq=f&gcx=c&ix=c2&sourceid=chrome&ie=UTF-8&q=TEW-664UB#q=TEW-664UB&hl=en&prmd=imvns&source=lnms&tbm=shop&ei=rq6uTtj3A9PbggeSrNWZCg&sa=X&oi=mode_link&ct=mode&cd=6&ved=0CE4Q_AUoBQ&bav=on.2,or.r_gc.r_pw.r_cp.,cf.osb&fp=c7f58bc4f869669a&biw=1433&bih=666

http://www.google.com/products/catalog?q=RNX-N2X&hl=en&prmd=imvns&bav=on.2,or.r_gc.r_pw.r_cp.,cf.osb&biw=1433&bih=666&um=1&ie=UTF-8&tbm=shop&cid=9328323554417061627&sa=X&ei=fK6uTqGqOMyRgQfT4KimDA&ved=0CDMQ8wIwAA

Doing that search also should explain the lack of WiFi on the devices. Even value brands like TrendNet and Rosewill cost almost as much or more than the Pi does.
Finding a working USB adaptor is really just a google search away.
I doubt that this will run open office very well and Java support seems lacking so Eclipse.org and NetBeans are probably not options at this time.
But we are talking about $35 or so. That is less than a tank of gas, a console or PC video game, or most peoples bar tabs.
In a school lab or even in a lot of office settings these can be used as thin clients or even a replacement of a desktop.
For hackers this thing has a good amount of IO 16 bits of GPIO plus I2C and SPI. The The Arduino Uno is about the same price and offers A2Ds but lacks video out, audio, networking, mass storage, a full blown OS..... So for hacking projects this is also really a handy little device.
  If you are looking for a $35 PC well this is your best and probably your only choice.

IMHO, the managers are not aware of the fact that GNU/Linux based operating systems do have a fairly mature collection of device drivers for most wifi devices though this area is still a work in progress. I feel the reaction from the managers were more due to a lack of awareness about the possibilities of FOSS based operating systems to do wifi that caused them to react to you the way they have. On your part you can point them to sites like http://linuxwireless.org/ or conduct a small awareness campaign the the University of Melbourne about the possibilities of using Free/open Source Operating Systems for wireless networking. Since you have identified a need, you can even make that a academic project if you want to.

That is hilarious! I've never had to setup a broadcom card, I guess, but wow, the instructions on http://linuxwireless.org/en/users/Drivers/b43 would make Terry Gilliam proud.

I Love the way the howto treats people like they are the shell; it would be so much better if the computer could talk. You could have that warm feeling of being lambasted by a European Gym Teacher with a hormone imbalance.

not working yet * IEEE 802.11n

That's all 802.11n devices. You know, those things that have been on the market for like 2 or 3 years? From TFA:

Broadcom would like to announce the initial release of a fully-open Linux driver for it's latest generation of 11n chipsets.

My apologies - I didn't know what you had already tried. It looks like I just told you what you already knew.

As for has anything changed, I think you would be better off asking on the Linux wireless mailing list rather than Slashdot as that's where the devs are. Certainly there have been changes made to the Ralink wifi drivers but whether they would help your case is something I would just have to ask someone else or read the source (and time is limited as it stands).

I wasn't trying to shoot down your theory. It's just that I felt that just finding out that you had master mode support would not be enough to know whether it would really work and that something else could be wrong even if master mode worked. I am basically agreeing with your observation that with what you currently know it's not worth pouring more time on this until you learn of a reason why something would have changed.

I would guess the ability get at raw frames is down to firmware support that has a mode that passes packets unchanged to the OS in a known manner rather than additional hardware. This is an extra that most people simply don't need and is another thing that can go wrong so I'd guess some vendors feel no need to expose such a feature.

The Linux wireless drivers page lists which drivers support master/access point mode (see the AP column). The list isn't perfect (the hostap driver definitely supports AP mode :-) but it seems to be a case of omissions. The table also says what form factor the supported chipsets come in (so you can tell which ones you will be able to get in USB form). I'd guess the rt2500usb or p54usb drivers would be your best bet.

Another useful page is the Linux wireless chipset directory which tries to list which cards have which chipsets (there's even a single page table with all the added chipsets but I won't link to it from here). This lets you build a list of boxes with the desired chipsets inside them (finding out whether this is true in reality can in itself can be a fraught process though). The chipset is really the important part in all of this.

I'm not going to point to an Amazon page because I have not bought a USB wifi card with the capabilities you describe from Amazon. I'm in no position to tell you that XYZ USB device on Amazon definitely works as I haven't done it myself. I have used hostapd on Linux and OpenBSD before now on a creaky old Prism 2.5 card and that worked for me but again that's not what you asked.

Finally here's a guide to using hostapd to set a card up in access point mode (just using iwconfig to set master mode is not enough). Googling for hostapd linux will turn up plenty more guides which may be easier to follow.

Good luck!

And dkms doesn't deal with the fact that the Linux kernel APIs (let alone the ABIs) are deliberately not kept stable, so any driver has a restricted window in which it will work.

They don't change the APIs that drastically or often in the kernel and when they do break things, it's not as often as Windows does between versions.

Manu of the user-visible problems in Vista were due to Microsoft changing the driver API for the first time since Windows 2000, and that's been one of the things that's sold people on risking the switch to Linux or OS X.

Microsoft introduced changes to the kernel API in Windows 2000 with service packs that broke graphics support partially until there were new drivers, Windows XP with wireless and graphics, Windows XP SP1 with memory addressing changes in the kernel that broke drivers, Windows SP2 another set of Windows addressing changes with DEP and wireless API changes.

that's been one of the things that's sold people on risking the switch to Linux or OS X.

No it hasn't. I think that's one of the last reasons why people would be using Windows since people don't even know this stuff to begin with. General people don't use a OS for technical merit at all.

I'm running Windows 2000 on my Wintendo. That was released in 1999, and hasn't been supported in years, but apart from Bluetooth devices pretty much every third party driver has Just Worked.

Wouldn't work with my Wacom.

Would a 5 year old Linux system work with a driver released in 2008?

Sure

What would you need to do to make it work?

The only experience I have with this, was with a Slackware install and wireless almost a year ago, I was lazy and grabbed a new kernel off http://www.kernel.org/ and some libraries I knew I didn't have with the wireless driver pack from http://linuxwireless.org/. Compiled, installed, done (really just typing ./configure, make, make install a lot with each thing).

Looking at GoboLinux, it doesn't look even vaguely similar to the NeXTstep style bundles in OS X. It's a linkfarm model, which is also a useful approach, but it's not what I'm talking about.

As I've said, there were some distributions even far closer to doing everything the OS X way, but after a couple of years they simply died out because of lack of interest. Some had a idea of using a kernel and having stable ABI/APIs like Windows/OS X etc.

Just goes to show how that model really mattered to users.

Interesting. I had been wondering about whether to give ath5k a go in place of madwifi. Now that I know it works for someone out there, I'm going to play with it.

Found some decent instructions here. Since the current 2.6.26 kernel on Fedora 9 has ath5k, it seems like this should be a cakewalk (famous last words)...

My wifi hardware is based on the rt2500 chipset series and is quite common on most laptops and until recently were reliable. As far as I remember the drivers were being rewritten for the kernel - which is fine but if it breaks hardware (which until that time had been reliable)
then people should have been made aware of this or even work with the distos for a interm fix.

Try out one of the wireless driver packages from http://linuxwireless.org/ (for hardy http://wireless.kernel.org/download/compat-wireless-2.6/compat-wireless-old.tar.bz2 ).

You will need to install your kernel source headers and the build environment

sudo apt-get install linux-headers-generic build-essential build-common

Then it's simply,

tar jxvf compat-wireless-old.tar.bz2
cd compat-wireless-old
make
sudo make install
sudo make unload
sudo make load

This will install the latest wireless drivers for your system and will not conflict with your distribution's package manager, should you want to remove the install and restore your previous drivers:

Make sure you are in the directory where the wireless driver installer is.

sudo make unload
sudo make uninstall
sudo make load

(It would probably be a good idea to reboot after that).

Normally I would never, ever recommend people compile stuff on Linux, however, in your case, it seems this would be the only way to get good support and this is really a last resort (a resort that you couldn't do under Windows if you ran into this problem).

http://linuxwireless.org/en/users/Drivers/ath5k

Supposedly you need 2.6.25 at least.

You will need to extract the fitting firmware to /lib/firmaware in order to get the new b43 driver working, which is already included into recent kernels. All info is here http://linuxwireless.org/en/users/Drivers/b43. No ndiswrapper needed, that's a stable native driver with good performance. If your distro of choice does not provide recent kernels, move. I recommend sidux.
Greetings,
Chris

A number of Medion laptops have inbuilt wireless based on the ZD1211B/AR5007UG chip on the internal USB bus, supported natively in-kernel with the zd1211rw driver: http://www.linuxwireless.org/en/users/Drivers/zd1211rw

If a larger laptop vendor shipped lots of laptops with this setup, the article being discussed here would be 'satisfied'. We can dream...

My laptop came with a Broadcom 4318 chipset. The support for it is flacky and it only seems to work properly using ndiswrapper. Some days ago I decided I was going to try to buy a USB wifi device that was compatible with Linux. If possible, its drivers had to be already part of the vanilla kernel. To my surprise, those devices exist! They are the ones that have the ZyDas zd1211(b) chipset (the "b" one is better). I thought it was going to be hard to find one of those specific devices, but no. They are present in a wide range of USB wifi devices. I went to the two main malls in my country they had one of those devices each. Piece of cake. Furthermore, a USB dongle can be used in future computers very easily, and don't take power unless they're plugged in.

http://linuxwireless.org/en/users/Drivers/zd1211rw/devices

You only need the device, a vanilla kernel and firmware, which can be downloaded from SourceForge.net, and it's also probably available for your distribution as an official package.

http://sourceforge.net/project/showfiles.php?group_id=129083&package_id=187875

ndiswrapper doesn't require you to do the firmware cutting step, you just supply it the .inf file and make sure the .sys file is in the same directory with it and run ndiswrapper -i file.inf . With bcm43xx you have to, get firmware from http://linuxwireless.org/en/users/Drivers/bcm43xx, cut and install the firmware

# bcm43xx-fwcutter -w /lib/firmware name_of_fw_file

which is derived from the softmac driver or mac80211 driver and I think which is loaded when the bcm43xx module is loaded into the kernel. I'd rather just use the "official" inf and sys files from BCM or the system integrator (compaq in this case).