The OSS Solution to the Linux Wi-Fi Problem 204
tobs writes "Matt Hartley of MadPenguin.org fame has published an open source way of solving the Linux Wi-Fi problem. He writes, "For intermediate to advanced users, who are willing to track down WiFi cards based on chipsets, live without WPA in some instances or have opted to stick with Ethernet, buying a new notebook for the sake of improved wireless connectivity may seem a little overkill.
When a new user faces problems jumping through the NDISWrapper hoops, tracking down WiFi cards from HCLs and other related activities, the end result is almost always the same — they give up. What so many of us, as Linux users, fail to grasp is that projects like OpenHAL are critical to long-term development. The education on what to expect and what not to expect remains a complete load of hot air when articles claim how easy it is to setup wireless Internet on Linux machines. It's downright misleading."
Weird... (Score:5, Insightful)
Re:Scientist's viewpoint (Score:3, Insightful)
I agree (Score:3, Insightful)
Take the time to buy the right hardware... (Score:5, Insightful)
I just installed Fedora 7, and I am managing multiple wireless networks with NetworkManager, no configuration at all. Zilch.
Of course, I have a 5 year old Dell. People think they can buy whatever hardware they want and just have it work. No. You have to be selective. That's why my 3D desktop runs on Intel video.
Buy companies that support open source from the beginning, dammit, or other companies will never see the use of providing drivers or specs PERIOD.
It's one of the three big weaknesses (Score:3, Insightful)
Are the three things I get embarassed talking about when trying to promote Linux to non-technical friends and family. All they want it "to do stuff". As the article mentions, they won't spend time fiddling with drivers, checking if the hardware will/might/won't work.
They have a real expectation that they can plug in whatever they choose to a PC and it will just work. This is their experience of (modern) MS and they won't accept any less from an alternative.
Until peripherals become seemlessly operable ordinary people will steer clear of Linux.
Until the applications (and I mean video playing in particular) just work, with no drama and no crashes (Kaffeine, why do you insist on popping up messages saying "The specified file or URL was not found", when you're playing it?) we're backing a loser.
Re:What's the REAL Solution though? (Score:5, Insightful)
Re:Take the time to buy the right hardware... (Score:4, Insightful)
then support that operating system by buying hardware that it is allowed to interact with by the vendor. activism is also sometimes pragmatic you know.
Re:There's Still a Problem With Wireless? (Score:2, Insightful)
What I meant is that many people are in different situations. Many people who are not considering Linux this week might in couple of months.
This means that many people will wake up some day, want to install Linux and realize that Linux can or cannot manage their hardware. They will react in different ways : fix the problem by using the ndiswrapper, installing other hardware, go back to Windows or OS X etc.
So, buying the right hardware from the start is only an option for people like us who already know they want Linux. Of course it helps as it rewards the good hardware makes who are Linux friendly, but it does not solve the problem instantly.
As TFA said, OEM might bring a big solution : DELL/HP wants to offer Linux laptop, so they choose compatible hardware. Then they want to use the same components across the line of products, so they ditch a few incompatible components.
This brings benefits on two sides :
- The hardware/chipset makers then realize they need to be selected for the Linux to avoid being excluded from markets bigger than the Linux market that have suddenly become tied to the Linux market
- The people who bought the non-Linux computers, got Linux compatible hardware anyway which makes their potential switch easier.
So, there are many big stories playing out here and I can understand why people would want to discuss them on Slashdot.
Major Pain (Score:2, Insightful)
I had read that WiFi has "solved" in the latest release of Ubuntu and have long been wireless in my home network, even for the desktop machines.
After trying all the non-NDISWRAPPER options, I finally used that tool and was able to get WiFi up and running, but even with that, it fails to initialize properly about half the time and I have to manually restart networking.
Combined with not having support for the latest NVIDIA drivers available through the package manager and having to recompile the drivers after a kernel security patch, this would have been an utter failure if I was new to this. NVIDIA is partially to blame as well, since they could well make their drivers have a safe mode that will work with cards released after the drivers, but the 8X series of cards has been out for how long and the driver still isn't in the package manager?
The lack of fail-safe mode in X after all these years is just insane. Fortunately, we shouldn't have to wait too much longer for that to be a mainstream patch.
Re:I agree (Score:3, Insightful)
Editing files has nothing to do with it. Generally, Under Linux, Wifi falls in three catagories. Those that do work. Those that work with NDISwrapper, and those that don't work. Those that work with NDIS wrapper NDIS wrapper installs the drive for you. Those that work out of the box will simply work out of the box. Those that don't work will sit there and stare at you and do nothing. There is a minor special exception for the BCM 43xx, you have to install their firmware first using something called the bcm43xx-fwcutter. But most distroes automate this.
So stop Trolling.
Re:We need new laws (Score:3, Insightful)
Today the hardware is almost immaterial to a "hardware" product - it is the inner workings of the firmware, the driver and such that are where both the bulk of the engineering time and bulk of the "value" are. There are few, if any, secrets in silicon today.
So a manufacturer puts a lot of effort into developing a new product in the US or EU. If the functionality of the hardware/firmware is then required to be disclosed it is trivial to make the same hardware product elsewhere and compete head-to-head without any real R&D cost. Sure the Linux community and a few hackers might be better off, but at what price?
Today, the only effective way to compete against Chinese manufacture is to have the hardware, firmware and software talking behind the scenes. The firmware interface to the hardware isn't disclosed and the driver that talks to the firmware isn't disclosed. Absolutely, the hardware can be duplicated but without the firmware and driver the device requires an equivalent amount of development effort.
Yes, that put the Chinese manufacturer on an equal footing with the US or EU manufacturer. Instead of how they would much prefer it where all the "hard" problems are solved in the US and they Chinese get to just make cheap knock-offs.
Where was development for USB hubs done? Where are they made today? Compare this to video cards - how many 3D cards are distributed by Chinese manufacturers? Sure, they are all imported with "Made in China" stickers but they are made for US, EU and Canadian manufacturers that own the firmware and drivers.
Re:What's the REAL Solution though? (Score:3, Insightful)
Frankly, its a shame that you can't get the equivalent of a PCI (or PCI-X) "wireless bridge". I would love a DD-WRT box that went into my system, and managed all aspects of my networking for me, addressable via some kind of internal IP address scheme.
This would give you all sorts of cool abilities; control it via your browser or any sort of "internal" application (something like Apple's airport stuff).
Hell, even given basic engineering skills this wouldn't take more than 3-4 chips, one for the "ethernet" card, one for the "bridge", one for RAM, and maybe one for ROM, if you didn't network "boot" the bridge.
package it up (Score:3, Insightful)
Ndiswrapper works very well, once it's set up. Kudos to the team for their efforts.
Re:What's the REAL Solution though? (Score:4, Insightful)
Ah, but the firmware on the wireless card is running, effectively, at a higher privilege level than your kernel --- it can do things totally outside the kernel's control. Even if you are legally allowed to redistribute the image, how do you know what it's doing? Given that all your network traffic is passing through that thing, and that it's got complete unsupervised control over all the radio bandwidth it can eat, and that on some interfaces (such as PCI) they can even access host memory... there's a lot of scope for malicious behaviour. Without source, they can't be audited. That's what I mean by the binary blob problem.
(The firmware source code probably includes lots of deeply patented and proprietry frequency-hopping and radio control software, which the FCC would be deeply unamused to have people play with; most likely there's also going to be a third-party embedded operating system, too, to make it all go. It would probably be a legal nightmare to release source.)
(You're right in that there's not much difference between uploaded firmware and firmware in ROM --- it's just a variation of the same problem.)
Re:Take the time to buy the right hardware... (Score:3, Insightful)
On what planet exactly? End user distros have been auto-detecting hardware for at least several years now. On average I'd even say that given a bare computer plus a Windows CD and a Ubuntu CD you're more likely to have everyting work on Ubuntu.
And just recently, I bought an ASUS motherboard, and tried to install Slackware 12.0 on it. And you know what happened? The kernel hangs after reporting the serial driver message. Sure, I could trace through the code to find the likely culprit.....
Or you could plug your Asus motherboard name/model into Google along with the words Linux and Slackware and probably have your answer in a few seconds. Better still, you could do this before buying the darn thing. If you buy from major online sites like Newegg you can search their user comments/reviews for the word Linux as well.
But a large part of driver installation on Linux is a manual process; on my 2003 Toshiba laptop, I end up recompiling the kernel to get the sound to work.
If you don't want to be recompiling kernels etc why on earth are you using Slackware? I can't get a standard size sheet of plywood into my wife's little Subaru hatchback either but that doesn't imply some sort of design deficiency in the Subuaru, it just wasn't designed for that.
If you want an idea of how good the Linux hardware detection/support is just burn yourself a copy of Knoppix and try it out on a few machines.
Re:What's the REAL Solution though? (Score:3, Insightful)