“Over two years ago a group was founded to reverse engineer the Broadcom Wireless LAN chipsets to provide Linux drivers. This chipset is used by many OEMs, for example in Apple’s AirPort Extreme in Power- and iBooks, Linksys’ WAP and WRT series of consumer grade wireless routers, various laptops from Acer, Dell, Gateway, HP and others and many more external and internal devices, including CardBus cards. That work has now come (.pdf) to a first milestone as there now is a free (GPL2 or later) Linux driver for a variety of these chipsets.”
This will benefit so many open source projects who are struggling to support Broadcom chipsets. I am waiting for FreeBSD and OpenBSD to adopt this.
Well, due to license issues, the BSD folks will have to reverse engineer the reversed engineered driver but at least they got some source code to take a peek at (which may or may not count as a derivate work depending on how strict you see things). Nut agreed, I can’t wait to see them get these chips to work. A lot of Powerbook users that insist on replacing Mac OS X with Linux (or BSD) will be very glad to hear this.
Well, due to license issues, the BSD folks will have to reverse engineer the reversed engineered driver
Uh? They won’t need to reverse engineer the Linux driver at all, seeing as the source is available. All they will need to do is re-implement the functionality using the Linux driver source as an example.
Well, call it poor wording on my part, but that is actually waht I meant.
Uh? They won’t need to reverse engineer the Linux driver at all, seeing as the source is available. All they will need to do is re-implement the functionality using the Linux driver source as an example.
Its highly dependent on what licence the code is under. If they are not satisfied, then they’ll re-write their version based finding how the Linux works. (and re-release under BSD license and clean up the code).
This is what they’ve done with the Ralink RT2500 series chipset.
Since this Broadcom reversed engineered driver is under GPL, expect another year (or two) until BSDs get a stable driver under BSD license.
Its a bit late for me, as I use a pair of WRT54G routers with a third-party firmware to fulfill my wireless requirements.
Well, due to license issues, the BSD folks will have to reverse engineer the reversed engineered driver
No, they will just use the reverse engineer specification that the GPL driver developers used as well.
Those that do the reverse engineering do not write the driver but a paper others can use to write one.
This makes sure the driver developers do not have knowledge about the closed source driver’s internals.
You realise that you know nothing right? BSDs don’t “adopt” GPL code into their kernel.
I know that BSDs dont adopt GPLed code. But like DevL said now BSDs have the source code to look at which can give them a good insight to reverse engineer it.
Holy crap! This is huge.
I (and many others) have been waiting for such a driver for at least 2-3 years now! The Linksys router 3rd-party-firmware groups will have a hell of a good time with this now that they’re not limited to using the Broadcom-provided binary module. 🙂
Edited 2005-12-05 21:26
I’ve got a shiny Powerbook, and the only thing to keep me tied to OSX (as nice as it is, it’s still not free and doesn’t have the gazillions of applications that Debian/ Ubuntu/ whateverGNULinuxDistroYouLike) is Airport Extreme support…
..and now…
wow!!! Do you know if there are instructions anywhere?
GR
Indeed a plus for PPC Linux. I always thought this was a disencentive to buying a 12″ Apple notebook. [Which is not to say it’s not great news for those with x86 hardware but generally there are alternatives with other wireless chipsets] Perhaps I’ll pick up a late model iBook after the intel switch – after prices have gone through the floor.
Most probably Apple’s next gen will be based on the centrino chipset for which intel already has a linux driver.
You might want to try Darwinports and/or Fink. It kind of diminishes the “gazzillions of applications” argument quite a bit…
While that does help a lot, it’s still not quite there.
I have 2 desktops running gentoo and an iBook. The ONLY reason I dont use linux on the iBook (I would probably go with ubuntu on it) is because of the wireless driver problem. If this gets fixed in the next ubuntu, WOHOO!
No, I’m not a stickler for stupid patent laws and stuff that hamper the progress of human society. I AM worried about Apple getting all sue-happy and a bunch of good FOSS programmers getting up to their necks in fines…
Don’t worry – the necessary information was not stolen but reverse engineered. There is a HUGE legal difference. Hopefully Broadcom will learn a lesson from this and finally provide open source drivers for their products.
Besides, Apple aren’t the one to do any suing, Broadcom would be the one (IF they even had a case, which they don’t).
oh dont worry apple will find a way to get involved, they love sueing and bullying people
Been using Broadcom chipset wireless driver under Zeta ever since Zeta released.
In Zeta is it a NATIVE driver or is it using a wrapper around a windows driver? Does it support WEP? WPA? Would it work on PPC?
Didn’t think so.
So this is not a Yawn, this is big news, not just for Linux but for all other Alt-OSes.
Well done to them. I don’t have any of those devices but have suffered setting them up for others in the past. Another victory to OSS. Anyone with legal issues can get bent.
so when are they integrating this w/ ubuntu… 🙂
One of the reasons that there was no opensource driver released was that open source software would be able to change the operation of the wireless cards to do things like tread on unauthorized frequencies.
The broadcom was limited in operation by the software..
“The broadcom was limited in operation by the software..”
I wouldn’t be surprised seeing that change in the future.
Maybe in the future it’ll mean more OSS drivers, instead of binary. But anyone hoping this kind of work will mean a “free for all” on the spectrum, may soon be disappointed.
“One of the reasons that there was no opensource driver released was that open source software would be able to change the operation of the wireless cards to do things like tread on unauthorized frequencies. ”
That’s a cop-out exscuse by Broadcom.
Funny how companies like Atheros and Ralink have given out their API spec’s for exactly the same kind of wireless cards.
“One of the reasons that there was no opensource driver released was that open source software would be able to change the operation of the wireless cards..
That’s a cop-out exscuse by Broadcom.
Funny how companies like Atheros and Ralink have given out their API spec’s for exactly the same kind of wireless cards.
Doh Atheros is a BINARY DRIVER!!!
Doh Atheros is a BINARY DRIVER!!!
There is open source atheros driver -> http://ath-driver.org/ It’s in early stage but it does not use binary HAL.
Or you can just use OpenBSD’s one, using their open source HAL.
Or you can just use OpenBSD’s one, using their open source HAL.
Yes, but what I’ve read on a blog of ath-driver’s developer OpenBSD version is missing a lot of functionality.
http://www.blogger.com/comment.g?blogID=14134766&postID=11230211684…
you base this on a cheesy no name blog entry.
This blog entry was written by Mateusz Berezecki — person who reverse engineered whole binary HAL from madwifi driver. I think that he is the proper person to judge the quality of OpenBSD driver.
“Doh Atheros is a BINARY DRIVER!!!”
Not on the BSD’s.
“That’s a cop-out exscuse by Broadcom.
Funny how companies like Atheros and Ralink have given out their API spec’s for exactly the same kind of wireless cards.”
I wouldn’t put money on that. Atheros and Ralink could have secured their limits either in the hardware, or firmware (or both). Broadcom could have gone with soft-limits set in software. Both have advantages and disadvantages. Anyway you argue it, any product sold in the US has to meet government rules.
OS X is pretty, and some of the apps are quite good, but overall I’ve often found Apple’s design choices to be a hindrance to my productivity, particularly in the areas of window manager behavior and keyboard control. I’ve been really tempted to switch to linux on my powerbook, but I couldn’t give up sleep mode and wireless support. Sleep mode works now, and it looks like wireless support will be coming before too long! Maybe it’s time to dust off that Linux partition again…
They have released an amazing open source (bit pricey) G550 PCIe graphics card. But the HAL for tv out (I think) is closed source. Maybe this is the next target.
I have a Compaq V2305CA but I don’t know what wireless chipset it is. Could someone tell me how to check on WinXP.
i have a compaq v2405 and have been runing ubuntu for weeks with the ndistwrapper with my broadcom card. Runs perfectly! Whats the benifit of switching?
NDIS wrapper only works on Intel based boxes not on my ibook so now I can have a wireless linux lptop…
good news…
i have a compaq v2405 and have been runing ubuntu for weeks with the ndistwrapper with my broadcom card. Runs perfectly! Whats the benifit of switching?
1) ndiscrapper sucks → http://acx100.sourceforge.net/ndis_cludge.html
2) it presumably will not work with kernel 2.6.16+ → http://lwn.net/Articles/160138/
I have waited so long for this…. I had completely given up.
huge props to the guys who stuck to the task.
Good job, guys.
OpenBSD on the wireless routers from Linksys would be great!
Exactly what I was thinking! OpennBSD wireless support is great right now and is the perfect platform for a wireless router/firewall. Hopefully someone will use this as a reference for an OpenBSD driver soon so I don’t have to buy a much more expensive WRAP box fore this.
Thanks for making this possible. I am running Ubuntu on my iBook. I have nothing against OS X, but Ubuntu is a lot better if you are a developer IMHO.
Now they will see even less of a point opening up since without doing any effort, they get linux support !
Looks like this chipset is ubiquitous so a lot of people can’t avoid it but I’d rather see the linux community throwing their weight behind the likes of Ralinktech who have provided a GPL driver for their wares, which has since been taken over by serialmonkey’s people.
Expressing what a wonderful thing it is to have a driver and how regrettable it was that they trailed so far behind their competitors in having a driver for Linux.
Remind them that in the future, they could simply release the specs or their own driver and not be relegated to the kiddie table.
Another hurdle is cleared. I really hope Ubuntu puts this into their version of Linux quickly.
WOW! this is huge. Nice work all!
I wanted to ask a newbie question here. Please, don’t flame me. I have an Acer laptop with a wireless card. lspci lists the following information for it:
00:0b.0 Network controller: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller (rev 02)
Is this driver supposed to work with this card? The original webpage mentions that their intention was to reverse-engineer the Broadcom 4301 chipset only. The specs page mentions several PCI and CardBus cards, but mine doesn’t have a known brand that I can check against that list. And finally, the driver page mentions BCM 43xx (that would include BCM 4318, I suppose). Is this a targetted card?
Apart from that, while ndiswrapper seems to work fine with the card, I would be very glad if I had a native Linux driver. Specially because ndiswrapper is a hack (a good one, though, kudos to the ndiswrapper people) and I’m not sure how long the hack will remain functional. Kudos too for the people who reverse-engineered the chipset. Good job.
Can someone explain to me, because I must be really stupid and naive, how a company benefits by NOT releasing specifications about their hardware such that a group of folks can broaden (no pun intended) the application and use of its products in various platforms. It makes no sense to me to deliberately restrict the development and adoption of what could potentially be another [future] market.
I have to conclude this is just short-sightedness. Where am I going wrong in this assessment?
“I have to conclude this is just short-sightedness. Where am I going wrong in this assessment?”
They say that life is the best teacher, so rather than ask us (whom are biased anyway). Why don’t you produce an expensive piece of hardware (expensive to design. expensive to produce. expensive to maintain.), and release the drivers and let us know how it turns out? If all else fails, at least you will no longer be “naive”.
I did not mean to imply that companies should give their IP away for free. I understand the resources invested in developing a product – although not specifically in wireless cards. Any my response presumed that Broadcom did not see the ROI in developing Linux drivers – hence my suggestion to release sufficient specs to enable third part development. The resultant drivers could be a cheap, easy way for Broadcom to get a feel for demand and size up the market opportunity, while precluding support investment. After all, they could always claim “Not our drivers – you’re on your own.”
I still see this as shortsightedness – especially in a crowded, cut-throat margin business. There is everything to gain and nothing to lose by supporting Linux long term. That’s my unbiased view, whether you accept it or not.
Broadcom sells chips. The open source geek isn’t their customer. The company that buys chips by the pallet are their customers (minimum for one company was lots of a thousand). Open drivers aren’t going to do them any good. The companies that buy the chips might, but we’re yelling at Broadcom instead of the companies that make something with those chips, now aren’t we?
“There is everything to gain and nothing to lose by supporting Linux long term.”
Debatable, but anyway there’s more to the world than just “numbers”.
Good point, except that maybe this will draw enough attention to Linux where they will be motivated to put out their own driver whereas in the past they simply ignored it.
Ain’t gonna happen, kiddies, for one very simple reason. The docs are licensed under GPL. If you want to see this put into the oBSD tree, it needs a Free license such as BSD or ISC.
http://www.openbsd.org/policy.html
The easiest way to get code into OpenBSD is to get it committed to NetBSD first.
And vice versa. And by committing to FreeBSD, etc etc.
Or was your reply a vague version of the old “OpenBSD” steals from NetBSD” bullcrap?
Hmmm…I wouldn’t call it stealing but after a while the “sync with NetBSD” commit messages would seem to bring into question the originality of OpenBSD’s efforts.
Uhu. I hope netbsd isnt using OpenSSH or the Ralink driver or the Atmel driver or systrace etc etc. If they did it could kinda bring into question the originality of NetBSD’s efforts….
“I AM worried about Apple getting all sue-happy and a bunch of good FOSS programmers getting up to their necks in fines…” You shouldn’t be worried in that sense. (1) Apple doesn’t care. (2) The only one who would care is Broadcom.
Or do we have to wait more?
About fscking time! You have no idea the amount of pain caused from not being able to use wifi in *BSD/Linux on the PowerBooks.