Slashdot Mirror
Linux Centrino Driver Update
Edy52285 writes "An article on News.com talks about how Intel has been, and still is, dragging on releasing their Linux drivers for Centrino. Intel is reluctant to release its drivers as open source since doing so would reveal secrets about their wireless hardware. Linux in currently unable to take advantage of Centrino's wireless networking devices, without, that is, prying $20 from your thin wallet to buy Linuxant's DriverLoader (discussed in an earlier story). Will Swope (Intel's General Manager of Software and Solutions Group) said in an interview said "What I believe will happen is we will end up having a Linux compatibility driver that is not open source at first, then designing future drivers in such a way that they are open source but will not expose intellectual property," Intel seem to be taking its time on releasing the drivers, and even in the article, there is a lack of any commitment on a date or under what conditions the drivers will be released." Also, someone pointed out that it's worth checking out ndiswrapper for the driver.
Hardware details -- it's like a chef not wanting to talk about his latest recipie, because that's the big secret. Sure, you and I probably don't have the cookware (hardware fab plant), but other restaurants (AMD and Qualcomm) would probably be very interested.
Not true. I'm using the open-source ndiswrapper project together with the win32 drivers, and it works, although a bit buggy. See here
...I won't buy a "Centrino" laptop. That's fine, since Apple's laptops are looking more appealing anyway, and still run Linux. Some of those new AMD offerings in mobile computing, as well as Tranmeta's installation in some of the Sony lines make them nice options as well.
Do not look into laser with remaining eye.
Why should anyone be surprised that a company that makes its money off of proprietary designs should be at odds with a movement to wrest control away from proprietary vendors?
Isn't this why Stallman insists on running only Free software?
I have been pwned because my
Intel PRO/Wireless Lan (Centrino)
For more info:
http://ndiswrapper.sourceforge.net/
Controlling a microwave radio with bits on a data bus from a digital processor issuing an interrupt signal is an invention and highly protected intellectual property. Please respect that. Jobs are at stake.
http://ndiswrapper.sourceforge.net/
This is an open source implementation that allows linux users to load their windows drivers and use their WiFi cards.
Its still very new, but there has been some success with the centrino chipset, as well as Admtek, Atheros and Broadcom cards.
Why don't they do like nVidia, release a pre-compiled binary driver core and an open-source, compilable interface, which hopefully will manage to unify all diferences between different kernel versions and distros ?
Alvie
We had exactly this problem.
Our solution was to write a proprietry driver, and then write a wrapper for this to interface it to the kernel. Release the wrapper under the GPL, then release our proprietry software as closed source.
Only buy opensource supported products. The demand
will drive the market. This is also what we would need to do as soon as the PC gets locked up with
the new Award Bios. Demand has to be so low that it
will just about drive the home PC vendors out of business. then and only then DRM will be dropped.
I'm stocking up on some hardware now, that way if my
desktop or firewall does die, I can build a new one.
Much of the actual work of the Centrino wireless hardware is performed in software, much like the "Winmodems" that were so widespread a few years ago (and, I guess, still are - does anyone make a modem with an actual UART on it anymore?) Intel is hesitant to provide the information that will allow people to write a driver for Linux, because that information would necessarily provide 100% of the software engineering necessary for someone else to create a Centrino-like hardware solution.
Blogging Weight Loss, Distance Education, and more at verlin.com
Think like a capitalist and vote with your wallet.
Until they have a proper Linux driver, buy an AMD based system instead.
AFAIK, the radio emissions from the wireless card are regulated (by the FCC in the US?) so as to avoid interference with other spectrum users. Much of this regulation is acheived through the close-source drivers.
Using modified drivers, it would be possible to make the card emit different frequencies or more power, thereby violating the usage licence.
Other chipmakers, I presume. So that nobody could produce an alternative wireless card to go with a Pentium M processor or some such.
But wouldn't anyone who's capable of designing and producing his own chipset be able to dissect the Centrino architecture and reengineer it, either by careful blackbox testing or by actually taking a microscope and looking at the chips? Am I way off mark here?
But if it's not other chipmakers they are protecting this from, if it actually is a software issue, then they are simply dancing to the tune of Microsoft due to whatever behind-the-scenes agreement they have with them.
(typing this on a Centrino-based WinXP laptop)
Indeed.
However, when I decided to purchase a decent wireless card , I would've liked to have been able to use it under Linux without paying extra. When you spend nearly UKP50 on the card, a discount on the Linuxant driver (at the very least!) would have been a nice gesture.
"Intel is hesitant to provide the information that will allow people to write a driver for Linux, because that information would necessarily provide 100% of the software engineering necessary for someone else to create a Centrino-like hardware solution."
Well, that sucks for them. Perhaps they should have built a real wireless device rather than taking away CPU time for something that is best handled by a seperate device.
This revealed, do most linux users even want a Centrino-based laptop?
Do not look into laser with remaining eye.
In all honesty, I can't see what is so special about Centrino that Intel wants to keep it so secret.
It's a freaken' wireless chipset and a power efficient CPU. It's not like no one else makes them.
What I believe will happen is we will end up having a Linux compatibility driver that is not open source at first, then designing future drivers in such a way that they are open source but will not expose intellectual property
So in other words, Intel is considering open source projects in the future. Isn't this news to get a little excited about?
How often in the past have companies brushed aside Linux? Many, many times. It gives me a bit of a fuzzy feeling inside to see guys like this being honest and forthright towards the Linux community.
I know in the short term it would be great if they would give us a bit more respect, but look a little further down the road...big companies are feeling pressure to do things the open-source way.
M-Systems' DiscOnChips are very nice flash chips which we use in some ThinClients. While there is support for those in the 2.4.x kernel tree it never worked for us. So we took M-Systems drivers.
.o file which is the real driver which does the real work. This way you can build kernel modules for you favourite kernel with M-Systems not releasing any "critical" source code.
:-)
.o file which links together a kernel module. Et voila: Neutrino support for every kernel without releasing the real source code !
Now they seem to be in a similar boat: they don't like to give out their intellectual property. Their solution is what looks like a driver stub and a binary
This practice means that you can't compile the driver into the kernel, you have to build a module (since the GPL does not allow building that propietary driver into the GPL'ed kernel, but allows non-GPL'ed kernel modules since they are not part of the resulting program or so... at least this what I recall Linus saying about that subject).
But having a module does the job as well, using an initrd we can boot from M-Systems DoC perfectly (in Real Mode they are accessible like a harddisk). The extra-effort is worth it since in our experience they are a lot more reliable than Flash IDE Chips, and reliablity is an important factor in embedded systems like ThinClients
Intel could do it the same way: release a driver stub and a binary
I know that the word is taboo around here, but isn't this precicely what (hardware) patents are for? From what I understand, they are pretty easy to get.
Intel has been using linux to bring up there new products for years. If they want to protect there ip then the least they can do is release a driver in the manner of the nvidia driver is release. Sure I would perfer a total open source driver but baby steps would be ok for now.
Please respect that. Jobs are at stake.
Well, they were. The last of the engineers were fired last Thursday.
Business isn't willing to pay for products, innovation and careers, so we get brands, mortgage commercials and layoffs.
Read this post about using the Windows NDIS centrino driver (and other drivers) on FreeBSD, using the "NDISulator" (a.k.a. "Project Evil"). See this post for details on Project Evil. And unlike the linuxant thing, this is free.
They take time. What would be the point of finally shelling out centrino 5yrs down the road? And backroyalties would be pointless because the patented technology would be so outdated.
The whole Centrino bit is a textbook monopolist tactic called a tying agreement. Intel can skirt around it because its still offering the pentium-m, but with no marketing support. The general customer is really confused and assumes that if the laptop does not have the centrino sticker, its not the best one.
This revealed, do most linux users even want a Centrino-based laptop?
Hells yeah! I'm within a year of replacing my old PIII 550Mhz Compaq laptop, which has been a trusty and faithful machine until recently but is now starting to give me hardware problems.
My next machine would be a Dell 300M running SUSE because it's ultra-portable, but thanks to Intel dragging their feet my next machine will probably be a G5 powerbook running Fink. Actually, Dell gets part-credit. Their recent quality control problems have made me suspect the reliability of their hardware.
That's the way the market works. Hey Intel, thanks for playing, but this ball just went over the fence!
If this were Usenet, I'd killfile the lot of you.
Trying to obscure hardware by only handing out binary-only drivers and hiding the API from the average programmer does not help at all against professional counterfeiting / industrial espionage. But it's quite amusing to see a company like Intel play the security-by-obscurity song.
They should know better.
next generation of laptops considering Broadcom & Philips have already cooked up
their own even lower power chipset.
I won't make any claims on the validity of these numbers{---Google Cache
Since i couldn't find the Yahoo Article they mention
- $12 a chipset
- 97% less power consumption than Intel Centrino in standby mode
- 70% less transmit power consumption
- 90% less receive power consumption
- 802.11g "not that far away"
~And this was October 2003
[Fuck Beta]
o0t!
These card are relatively inexpensive. There's no particular reason to pick a Centrino laptop because of the built-in WLAN support.
-------
Warning: Slashdot may contain traces of nuts.
But wouldn't it be obvious that your stealing their IP? If Intel found out, they can just sue them. I'm sure the tech is patented so AMD would have to license it to use it anyway.
I've been a Linux zealot since 1995, but just two annoying things have forced me to spend 90% of my time booted into my WinXP partition on my Panasonic toughbook:
:-/
- swsusp is not reliable. Sorry, but I can't be patient when my fucking laptop hangs on the 2nd or 3rd resume. Cold booting and shutting down is just too damned slow, so I rarely bother anymore.
- lack of Centrino support. Bastards at Intel! I would not have purchased this laptop if I knew I would have gotten shafted on Linux support -- especially when I was under the impression Intel was Linux-friendly!)
Oh, and I guess a 3rd problem has begun to rear its ugly head now that I'm getting into video capture and editing via firewire. Namely driver support and applications.
Ah, but I'll never give up Linux on the server OR my main desktop.
I track known Slashdot scumbags on my foes list!
Are a pain in the ass. I noticed this the day a colleague tried to install Windows 1900 on a "Made for XP" notebook. OEM drivers were nowhere to be found, and all the new drivers refused to install because they were "optimized for XP." The CDs that came with the machine were only "disk image restore" CDs. so re-installing the OS was impossible.
Linux (Red Hat 9), of course, installed without so much as an extra line feed, and supported each and every device perfectly. This was a fairly new notebook as well. It was amazing.
Can't figure out why manufacturers go out of their way to make it difficult for people to work with their own computers the way they want. Centrino should be supported, especially with notebooks being as expensive as they are.
Business isn't willing to pay for products, innovation and careers, so we get brands, mortgage commercials and layoffs.
Since then I haven't wasted a single second searching for drivers or wrestling with hardware to get it to work. Sleep and restore works 100% of the time. Bluetooth and wireless LAN are bulletproof. I'd almost forgotten what it was like until I read this article.
Intel's entire strategy over the last 10 years has been precisely to move as many functions as possible into the CPU. This enables them to justify selling processors with far more horsepower than anybody needs for word processing or browsing, and it lets them assert control and influence over a much larger fraction of the hardware market.
That's why they keep adding more multimedia-oriented units to their architecture; it's also why they designed the P4's memory architecture to be mainly good at streaming blocks of video data.
Their strategy has been relatively successful up to now. There's just no way that they would design a totally stand-alone wireless solution to be tightly marketed with their CPUs.
In fact, just from the Centrino marketing material, you'd get the impression that the CPU itself is handling the wireless functions. Perhaps they plan to move that logic into a future mobile CPU chip.
Distrust begets distrust. Secrets beget snooping. If someone (Intel) is going to be so damned hypocritical and lavish in Linux's support of it's product lines (especially the nice early Itanium support while Microsoft was getting is OS finished) they had better not complain when someone "hacks" a solution out of the chip.
It's like the who DVD-CSS mess. Linux people just wanted to be able to watch DVD's without runnning Windows. What resulted was a hack that made convertion of DVD's into cheap Divx copies easy and painless.
It feels like dating someone who never trusts you, never earns your trust (or respect) and goes hysterical when you don't behave exactly how they want. Reminds me of an ex-girlfriend, frankly.
Murray Todd Williams
The problem lies in Intels inherent desire to eat spare cpu cycles. Why? Because the more cpu cycles wasted on things better handled in hardware the more incentive to upgrade your cpu.
Those spare cycles could do something better than doing the hardwares work. Microsoft wants to have it all in windows if they can. That way they can tie the whole platform to windows cementing the monopoly on desktops. MS and Intel have had their jousts and Intel have always folded under the pressure. Intels project to make hardware more platform agnostic was stopped by MS who saw a threat to their Wintel Symbios.
There is nothing stopping eg. device drivers from being implemented much lower down like in the actual hardware, talking only in pre standardized APIs. Whats stopping that great innovation that would put a stop to driver problems and make it much more easy to develop new products?
Guess once!
HTTP/1.1 400
But competing restaurants (AMD, Qualcomm) have the resources to send some samples of the meal to a chemical analysis lab (disassemble the object code), and learn the secrets that way.
I've simply removed from my laptop the Intel Pro Wireless 2100 WiFi Mini-PCI card and replaced it with a Atheros 802.11a/b/g chipset. The Multiband Atheros Driver for Wifi (aka MadWifi) is well supported under Linux. It even has a great FAQ.
The card I bought is an IBM 11a/b/g Wireless LAN MiniPCI Adapter (IBM Part Number: 31P9701), and works flawlessly under REHL3.
If that's the real reason, how come they release them as binaries? Reverse engineering will reveal exactly the same information (eh, sorry, intellectual property) as inspecting the source code, although it takes a little bit more work. So, does anyone think Intel gave the real reason?
I got a Dell Inspiron 8600 with Centrino technology and am typing on it now. It has a 1.4GHz Pentium-M processor. I got it for the battery life, I can do average (not idle) tasks for over 4 hours straight on the single 72 watt-hour battery (a second is available). I tried and successfully put a Knoppix CD in the drive and it booted up fine. It works but I probably can't get the same longer battery life. I'm pretty happy with this laptop, and squeezing another hour out of a battery on an operating system I don't primarily use isn't going to make me regret my decision.
--It only takes 20 years for a liberal to become a conservative without changing a single idea.
(AMD and Qualcomm) would probably be very interested.
Those guys are probably the least interested. Their engineers know exactly how to make similar devices. The only ones that might be interested are some third world country's bootleg industry. And they can/will reverse engineer the devices anyhow.
All these "Oh, we can't release the specs, that would reveal our secrets!", are pretty full of it. There are very, very few hardware/software solutions that aren't widely known. It gets really silly when companies such as NVidia refuse to release info to the XFree community, due to their hardware secrets. For heaven's sake! Even the insides of such (more or less) proprietary devices as the PS2, the GameCube, and so on are well known...
I imagine that one reason that Intel doesn't want to release these details is because the driver has too much control over the device. If as much work as it seems is done by the processor, then that probably means you could force the chip to do some strange things. The most obvious ones that come to mind are a) increase power (although I can't really see why that would be a problem), b) sniff to your heart's content, and c) try a DOS attack on any nearby networks by saturating the airwaves with crap.
Intel doesn't want to risk being associated with these kinds of things (and you know if they released an open source driver, someone would).
This still doesn't however totally explain their not releasing a closed-source driver...
Combination - fun iPhone puzzling
Odds are Intel does know better than you armchair engineers.
Having worked there up until a few years ago, I'd say that most of them agree with the "armchair engineers" and think it's silly to try to hide stuff this way. But there are (lets see, where's that Post Anonymously button?) flaming morons in various management positions (mostly marketing) that are totally clueless. AMD, et al. have the resources to disassemble the binary only drivers anyway, so the only thing you are doing is slowing the adoption by technically oriented users, but they can not / will not see this.
We even had people like Linus, ESR, BP, etc. come out and do dog & pony shows about why it's a Good Idea to open things like this up, but the only thing that seems to be working is a gradual process of selective retirement of the morons. (Intel's culling process to rid itself of the clueless can best be described as "brutal".) Saddly (since I still have stock & and friends in Intel) there is a fair voluntary exodus of the cluefull as well.
-- Anonomous Coward
P.S. The funniest part of the dog & pony show was when one of the PHPs listed among the downside of open sourcing the "fact" that it would piss off MS.
The legal department people who were there were not clueless and came down on the poor idiot like a ton of lead. From the hurt look on his face I think he expected them to side with him.
Have Intel invented the WinWiFi?
Didn't anybody learn from the WinPrinter and WinModem farces?
Maybe they don't want people to have access to it because the hardware could be used for other things than just wireless LAN.
Depending on the hardware, who knows maybe someone could even implement GSM/PCS on it. Whatever may be the case, having access to hardware like this would allow people to play around with it.
What is SDR?
GNU SDR implementation
When I emailed people at Intel over this matter, they eluded to the fact that releasing the source to a centrino wireless driver would violate FCC rules. Basically, the radiomodem in centrinos are totally programmable. That means, if you have the code, you can broadcast whatever you want on whatever frequency you want. You can violate FCC rules, and the FCC doesn't like that. Therefore, to be FCC certified, the user can not be able to change certain parts of the modem.
It sounds plausible, but they also could have been blowing smoke.
It is possible that as with some Atheros-based WLAN cards (the D-Link DWI-G650 Bx for example), the radio in Intel's Centrinos is software-controlled. This means that its frequency and power can be changed to just about anything using software alone. Open-source drivers for something like this are out of the question -- the FCC would not be impressed.
Atheros' ended up releasing a binary-only driver... kernel-tainting and all. If the Centrino radio controllers are also software-based, you can expect a binary-only driver as well.
Perhaps they plan to move that logic into a future mobile CPU chip.
yes they do. and into everything
Err, the whole point of "Centrino" is marketing, pure and simple.
The term "Centrino" is a 100% pure marketing term. There is absolutely ZERO technology connected to it, it just means that you are using an Intel Pentium M processor with a an Intel motherboard chipset and an Intel wifi chip.
The trick behind all this though is that if you combine those three elements then Intel will give you MUCHO-$$$ for marketing purposes. Last year Intel gave out $300 million to the likes of Toshiba and Dell to market their Centrino laptops I would not be at all surprised if it turned out that it was CHEAPER to add in an Intel WiFi chip than to have no wifi chip at all once you factor in the advertising bonuses. So that $12 Broadcom chip could well be $14 or $15 more expensive than an Intel one.
[Fuck Beta]
o0t!
Based on their (lack of) responsiveness so far, I would not recommend them. I have switched to using the madwifi driver (with a different wireless card).
Owners of IBM Thinkpad X31's (and perhaps other IBMs, too; I don't know) cannot use this solution without a BIOS upgrade.
2 .2/0147.html for a more complete explanation)
The reason is, IBM's BIOS actually checks that the Mini-PCI wireless card is one of several "acceptable" cards, and will refuse to boot if it is not. The acceptable ones are the Intel Centrino card, and a a Cisco card. With a BIOS upgrade, apparently an IBM sourced Atheros-based card (model number 31P9701) will work. (see http://www.ussg.iu.edu/hypermail/linux/kernel/031
-- -pjk Perry Kundert perry@kundert.ca http://kundert.2y.net
I don't really buy this. I've worked with 900 MHz and 2.4 GHz software controlled radio chipsets from both Chipcon and Rockwell. Nobody ever does any background check on me or anything. Nobody has ever made me sign something saying I won't program it to interfere with other equipment. Nobody has ever said they would need to audit any firmware/software I wrote for it. Nobody has ever even charged me for a dev kit. They throw a dev kit at me, and say "Have fun!"
Granted, these are the chipsets all by themselves, and not something a typical user could do anything with except fry with static.
And of course, my final products have to meet the specs in order to get FCC approval. But did you know that the firmware on the unit which the FCC tests doesn't have to be the same firmware shipped in production units! In fact, it's common practice to give the FCC testing labs special testing firmware. It doesn't even need to be fully functional. All it needs to do is the unit's "primary function." For instance, there are many handheld radio transmitters with keypads out there. Keypad scanning circuits tend to be really noisy, espcially on some microcontrollers. So you have special test firmware that starts transmitting when you press a key, and stops scanning the keyboard once you press that key. The FCC tester doesn't start measuring the emissions until the device it transmitting (it's "primary function"). Since the keypad has stopped scanning, it's no longer generating that noise and passes the test. Yes, the FCC tester is fully aware that they using "test-only" firmware!
Maybe it is because I'm not a market droid, but what good is a product to a company when they are too afraid to sell it?
Not flamebaitish at all, it's actually a very good question.
First off, a 1.5GHz Pentium M will run circles around a 2.0GHz Celeron. Actually it will beat the pants off a 2.8GHz Celeron, but the Celeron is perhaps a bad example because that chip REALLY stinks! The current Celerons (1.7GHz through to 2.8GHz, basically a castrated bastard-child of the regular Pentium4) are absolutely abysmal performers, so it doesn't take much to beat them; AMD's $35 Duron processors running at 1.6GHz will usually match or beat the 2.8GHz Celeron.
Simply put, there are two main methods of designing a fast processor; the "brainiac" model where the chip does a lot of work per clock cycle but doesn't clock as high, and the "speed demon", which doesn't do much per clock cycle but runs at very high clock speeds.
The Pentium4 is very much a "speed demon" design, which is why it clocks nearly twice as high as most other chip produced on a similar manufacturing technology. The Pentium-M takes more of the "brainiac" style of design, so it's harder for Intel to clock it to high speeds, but it does more work per clock cycle.
In reality, the Pentium M doesn't really run at slower clock speeds than many other CPUs, it currently tops out at 1.7GHz. For comparison the Athlon64 is running at 2.2GHz, the PPC 970 (aka G5) at 2.0GHz, the Power4+ at 1.7GHz, the Itanium2 at 1.5GHz, Alpha EV7 at 1.25GHz, UltraSparc III at 1.2GHz, etc. Really the only odd-ball is the Pentium4, which currently clocks up to 3.2GHz. Despite the wide range in clock speed though, in the end all of these chips are in the same general ball-park in terms of performance.
Now, there are a LOT of factors that influence the overall speed of a processor, and even a quick summary of them could easily take dozens of pages, but it's already well documented in books and on the web if you're interested. Suffice it to say that a Pentium-M is usually about as fast as a 2.2 to 2.6GHz P4, though individual applications can vary wildly.
This doesn't exactly mean that clock speeds are irrelevant, a 1.7GHz Pentium-M is still going to be faster than a 1.3GHz Pentium-M, it's just that clock speed is only one small part of the whole picture. I like to equate it to the displacement of a engine. All else being equal, a 4.0L engine will give you a faster car than a 3.0L engine. However, it's certainly possible to build a 3.0L engine that will produce more horsepower than a totally different design of 4.0L engine (F1 cars manage to pump ~900bhp out of a 3.0L engine, while most 4.0L engines you're likely to see in production cars produce only ~300bhp). What's more, the peak horsepower number doesn't tell the full picture of engine performance and it certainly doesn't tell you how fast the car as a whole would be. Similarly, for any given processor core, higher clock speeds will give you more performance. On the other hand, two different cores can haver very different performance at different clock speeds, and certainly other components like the video cards, memory and hard drive can all have a big impact on the overall performance of the system.
When you get down to it, it's simply a matter of design decisions and trade-offs. The Pentium-M was designed to offer good performance and low power, and it succeeds VERY well (I'm a big fan of the Pentium-M processor, even if I rather dislike the "Centrino" marketing program). The P4 was designed for the highest overall performance at a reasonable price-point. As a result, the top-end P4s are faster than the top-end Pentium-M chips, and probably always will be. However, the Pentium-M at 1.7GHz consumes only about 25W. A similar performance P4, even in it's low-power laptop version (the "Mobile Pentium4-M", not to be confused with either the "Mobile Pentium-M" or the "Mobile Pentium4", and some people say AMD's names are confusing!) consumes 35W at 2.5GHz. Meanwhile the regular desktop Pentium4 (and also the "Mobile Pentium4") conume 61W. That 2.0GHz mobile Celeron processor you mentioned comes in at 32W.