Forgot your password?
typodupeerror
GNU is Not Unix Open Source Hardware

FSF Certifies Atheros-Based ThinkPenguin 802.11 N USB Adapter 85

Posted by Soulskill
from the going-free dept.
gnujoshua writes "You may recall that last Fall, the LulzBot AO-100 3D printer was awarded the use of the Free Software Foundation's Respects Your Freedom certification mark. Today, the FSF announced certification of the ThinkPenguin TPE-N150USB, Wireless N USB Adapter, which uses the Atheros ARAR9271 chip. The FSF's RYF certification requirements are focused on the software (not the hardware designs) of a product, which in this case was primarily the device firmware and ath9k-htc module in the Linux-libre kernel. (Disclosure: I work for the FSF.) There's also a cool story that is within this story... which is that the firmware for the Atheros AR9271 chipset was released as a result of a small device seller (ThinkPenguin) striking a deal with a large electronic device manufacturer (Qualcomm Atheros) to build a WLAN USB adapter that shipped with 100% free software firmware. This deal was possible largely because two motivated Qualcomm Atheros employees, Adrian Chadd and Luis Rodriguez, made the internal-push to get the firmware released as free software."
This discussion has been archived. No new comments can be posted.

FSF Certifies Atheros-Based ThinkPenguin 802.11 N USB Adapter

Comments Filter:
  • by acariquara (753971) on Tuesday April 30, 2013 @08:39PM (#43596697) Journal

    Looks like the price of freedom is pretty steep.

    It's not like we are starved for wifi dongles that "just work" on Linux without NDISWrapper. We're not in 2003.

    eg: http://dx.com/p/802-11n-150mbps-wifi-wlan-wireless-network-usb-adapter-53538 [dx.com] $10 bucks including shipping, and there are TONS cheaper than this. I just looked for one that specifically said "Linux compatible".

    • by alantus (882150) on Tuesday April 30, 2013 @08:51PM (#43596767)

      Looks like the price of freedom is pretty steep.

      It's not like we are starved for wifi dongles that "just work" on Linux without NDISWrapper. We're not in 2003.

      eg: http://dx.com/p/802-11n-150mbps-wifi-wlan-wireless-network-usb-adapter-53538 [dx.com] $10 bucks including shipping, and there are TONS cheaper than this. I just looked for one that specifically said "Linux compatible".

      As a matter of fact, I wish it "just worked".
      I have one of these dongles, and last time I tried to use it I was hitting this bug:
      https://lists.ath9k.org/pipermail/ath9k-devel/2011-November/007467.html [ath9k.org]

      • by adri (173121)

        Now that the firmware source is open and the UART wiring instructions are public, there's enough basic stuff there to figure it out.

        We're digging up instructions for JTAG debugging.

      • by Anonymous Coward on Tuesday April 30, 2013 @10:04PM (#43597127)

        As a matter of fact, I wish it "just worked".
        I have one of these dongles, and last time I tried to use it I was hitting this bug:
        https://lists.ath9k.org/pipermail/ath9k-devel/2011-November/007467.html [ath9k.org]

        This is a bug from November 2011 is against the same chipset. But, it was a module that was loading proprietary firmware.

        The firmware was released as free software within the past couple of months. So, now instead of the ath9k kernel module folks having to treat the firmware as a black box, they can file bugs [github.com] and submit patches [github.com] to the firmware itself.

        The reason we certified this device is because it carries freedom to the user. Not arbitrary freedoms, but the specific freedoms to run the program, share it with others, make modifications to the source code, and share modified versions of the source. With this freedom, a user can not only work with others to find and eliminate bugs, but they can find ways to adapt and improve the software so as to squeeze the most they can out of the device. The same can't be said with the 2011 adapters that shipped with this chipset.

        And, this isn't just idle speculation. Already we have seen a fair bit of cooperation between this firmwares lead developers and the ath9k module maintainers. I would be very surprised if the almost two-year old bug you pointed to still exists. But if it does, then at least I know you and I can easily reach out to a trustworthy community of free software hackers to explain to us the problem.

        • by Zalgon 26 McGee (101431) on Tuesday April 30, 2013 @10:53PM (#43597401)

          If I can spend 10% and it works, I'm happy.

          http://dx.com/p/ultra-mini-nano-usb-2-0-802-11n-b-g-150mbps-wi-fi-wlan-wireless-network-adapter-black-71905 [dx.com]

          Therefore, I am happy.

          Enjoy your purity. I'll enjoy my $48.60 in leftover money.

          • Re: (Score:2, Insightful)

            by Arker (91948)

            Enjoy making the world a crappier place, retarding the progress of science, and generally screwing the world up for your kids.

            I dont know how much you lose when you are offline for an hour, but it would take me at least that long to drive in to replace one of these things, and that's more than the dongle costs however you look at it. So as I see it you are penny-wise but dollar (and otherwise) foolish.

            • by Anonymous Coward

              Ok, great argument. But if money is on the line, why are you using random unsupported driver that was thrown on to some FTP site? Shouldn't you have considered this before your NIC blew-up due to some obscure technical issue?

              And I'm sure your employer is paying you to work, not shop for obscure USB dongles because Linux is too shitty to support your laptop hardware properly. The economically rational thing to do would be to replace you with a Mac user who has a unix system that came from the factory with wo

              • Re: (Score:2, Insightful)

                by Anonymous Coward

                Your "economically rational" thing ends up working like this: You buy the Apple product that definitely 100% works no problem. It doesn't work. Apple says "Oh, really? Too bad". You buy something else, you write off the enormous cost of the non-solution, you sigh at the people insisting that won't happen. Really. This is how big IT projects mire themselves, they rely on some God-like single supplier (Apple in your case) and then get bitten because the God-like supplier doesn't give a shit. Freedom means you

          • Enjoy your purity. I'll enjoy my $48.60 in leftover money.

            Good point, we should be prepared to do anything for money.

            • Reminds me of this joke:

              A man walks into a bar and sits down. A woman walks up to him and starts talking. He says "Would you sleep with me for $1 million?". "Sure", she replies. "How about $10,000?", he counters. "What do you think I am, a hooker?", she retorts. He responds, "We've already established that. Now we're just negotiating."

              P.S. I will admit that I might have gotten the second number wrong.
          • Enjoy your purity. I'll enjoy my $48.60 in leftover money.

            Every man has his price. Now we know yours.

    • by Arker (91948) on Tuesday April 30, 2013 @09:11PM (#43596857) Homepage

      Freedom isnt about cheap and it never was.

      A dongle that 'just works' today with a particular binary wont necessarily work tomorrow on a different machine or after a simple recompile with different options, let alone after a major software upgrade.

      At the moment this appears to be the only properly supported wireless dongle on the market. It should be no surprise it's a little more expensive than the junk.

      • Sure, there's a point to buying quality hardware, but at the same time, why is buying a $54 dongle and keeping it for a long time better than buying a $20 one today and buying an improved one for $20 sometime in the future.

        This isn't 2004, you really don't have to search for laptops/wireless dongles that support Linux, its a rarity if they don't support Linux.
        • by Arker (91948) on Tuesday April 30, 2013 @11:08PM (#43597479) Homepage

          Sure, there's a point to buying quality hardware, but at the same time, why is buying a $54 dongle and keeping it for a long time better than buying a $20 one today and buying an improved one for $20 sometime in the future.

          Primarily because doing so sends a clear signal to suppliers that we ARE willing to pay extra to get something done right.

          Secondarily because buying the "improved one" should be done on my timescale and for my reasons, not forced because I have a piece of junk that wont work properly.

          This isn't 2004, you really don't have to search for laptops/wireless dongles that support Linux, its a rarity if they don't support Linux.

          To the contrary, although it is not 2004 and some things have improved, I still count one single dongle that actually supports GNU/Linux properly. One.

          Supporting one or many binary distributions of GNU/Linux does not constitute proper support. Meeting the criteria [fsf.org] for this particular certification does.

          • by drinkypoo (153816)

            Primarily because doing so sends a clear signal to suppliers that we ARE willing to pay extra to get something done right.

            But that's not true. Actually, you're willing to pay extra to get something done over right. You only have to pay again because they did it wrong the first time. I'll support the manufacturer who does it right the first time, and doesn't expect me to pay them to do it again the way they should have done it once.

        • by adri (173121) on Tuesday April 30, 2013 @11:12PM (#43597495) Homepage Journal

          The people that want to do dirty hacks, like mesh or TDMA offload on the USB NIC.

          Or even improved hostap support.

          Or an experimental platform for ${THING_YOU_HAVENT_THOUGHT_OF_YET}.

          Yes, you can buy cheaper NICs. Same as buying cheaper anything. But here's a USB NIC with a well-understood wifi part (AR9285 on-die) and now open firmware with open tools to fiddle with the thing. If the FSF and manufacturers manage to ship a million units, great. I'm happy just knowing that people are doing interesting stuff with it. Doubly so if I haven't thought of it yet. Triply so if it's cool and turns out to be transferrable to the other Atheros wifi hardware out there.

          • Srsly. Since when does a piddly $54 scare away a bona fide nerd. Having tools like this around when you have an epiphany on the crapper is priceless. I mean, come on, any of you could wipe your ass with $54 and still have plenty where that came from...

            Ohhh...wait....I just realized something

            ::cue Willy Wonka meme face::

            some of them think playing D&D and working at Burger King makes them nerds...how adorable
            • Srsly. Since when does a piddly $54 scare away a bona fide nerd. Having tools like this around when you have an epiphany on the crapper is priceless. I mean, come on, any of you could wipe your ass with $54 and still have plenty where that came from...

              I was going to buy FOUR of them to hack and play around with small devices like the Raspberry Pi. Balked at the $216 price tag and never came back. Don't judge.
              We have $25 single-board computers [raspberrypi.org] and $25 OpenWRT-enabled routers WITH USB SUPPORT (TP-Link WR703N) [dx.com].

              A $54 Wifi dongle is, no matter how you put it, downright stupid.

        • I will buy a $54 dongle if it means that I can include it within my embedded devices and not have to worry about the next linux kernel update supporting it.
        • by idontgno (624372) on Wednesday May 01, 2013 @10:58AM (#43600573) Journal

          Here's irony: the difference between the $20 dongle you bought and threw away and the $20 dongle you replaced it with (and the next $20 dongle you buy to obsolete the second one) may just be in firmware. Firmware that, if you'd paid the money up-front, you could have flashed from open-source repositories and had the exact same features... for $0 extra.

          BTW, the entire premise that you have to constantly, obsessively, upgrade hardware is foolish. Just thought you should know.

          • by IANAAC (692242)

            BTW, the entire premise that you have to constantly, obsessively, upgrade hardware is foolish. Just thought you should know.

            Thanks for saying that.

            It bugs me to no end that we mindlessly dump perfectly good electronics into our landfills all for the sake of having the latest and greatest, rather than improved functionality (which, most of these hardware upgrades don't provide).

      • by Coren22 (1625475)

        I am curious, does this mean better Kismet support? Can we now do wardriving without $300 adapters or do spectrum analyzing to pick the best frequency to run our wireless on to not interfere with the neighbors?

        http://www.kismetwireless.net/ [kismetwireless.net]

        • by adri (173121)

          The AR9280 and later NICs support spectrum analyser mode, as a diagnostic thing.

          Yes, this includes the AR7010+AR9280, AR7010+AR9287, and the AR9271 (which is an AR9285 wifi core.)

    • Re: (Score:3, Insightful)

      by Anonymous Coward

      Cheap Wifi dongles are very problematic and have a history of issues as alantus suggests. Having a completely open Wifi dongle is a Very Good Thing (tm) as many (or all) of these issues will be moot. Plus, if something stops working correctly the device firmware is out there to troubleshoot.

    • by Anonymous Coward

      I've used this for years with Linux and it's the best wifi card I've ever used: http://www.amazon.com/Alfa-AWUS036H-Upgraded-Wireless-Long-Rang/dp/B000QYGNKQ

  • Master Mode (Score:5, Informative)

    by Anonymous Coward on Tuesday April 30, 2013 @08:58PM (#43596803)

    one thing i noticed last week on ThinkPenguin is that their adapters generally support Master Mode making it easy to build access points. i don't think this can be said for most adapters requiring ndiswrapper.

    • by adri (173121)

      The underlying wifi parts are the same as the PCI(e) chips of that era; the same basic features are supported.

    • by laing (303349)
      Whether or not it supports master mode is the first thing I looked for on the product page. It doesn't say whether it does or not anywhere. If it did, I would buy one today.
      • Re:Master Mode (Score:5, Informative)

        by adri (173121) on Tuesday April 30, 2013 @10:15PM (#43597187) Homepage Journal

        The wifi part of the hardware does. One of the reasons we opened up the firmware was to let people at it to make it better at supporting master mode.

        The NIC has a small embedded CPU to act as a PCIeUSB gateway and a small amount of RAM to run code and buffer frames. The problem with master mode is the amount of RAM that you need for each associated station. So there's been discussion about moving some of the stuff done in the NIC CPU (transmit aggregation, rate control) into the host, so the NIC itself doesn't need to store (that much|any) per-station state.

        • by Arker (91948)
          So what exactly happens, with the provided firmware, when the onboard RAM is exhausted?
          • Re:Master Mode (Score:4, Informative)

            by adri (173121) on Wednesday May 01, 2013 @01:21AM (#43598029) Homepage Journal

            Things then crash. :-)

            Embedded software looks different to your Linux/FreeBSD kernel development. There's fixed buffers allocated for things. Once those buffers are full, everything stops until they're not full.

            If you want more information please subscribe to the ath9k firmware list and ask questions there. I'd rather everyone benefit from the answers!

            • by Arker (91948)

              I probably wouldnt recognise your typical modern *nix development, I havent had the slightest thing to do with development since the 90s. It was just a question of curiosity. That said, the stuff I vaguely remember probably looks more like what is done now in embedded systems than elsewhere I would guess.

              Hopefully it doesnt really crash, it just quits accepting new clients? If it really crashes that does not sound like good design. If it quits accepting new clients, then in effect the available RAM produces

      • As far as I know, every ath9k device supports AP mode. I have one and run hostapd on it.

        More info: http://wireless.kernel.org/en/users/Drivers/ath9k [kernel.org]
        • AFAIK, there have been no USB dongles out there with support for 802.11n Master mode in Linux. Sounds like this open firmware is progress!
  • by Anonymous Coward

    USB 1x1, 2.4GHz only? That's five bucks on eBay. Couldn't it at least be a dual-band radio?

    Apparently the cost of freedom is not being able to connect to my network, I'm in a highrise and there are 40 APs fighting over three channels that I can see from where I sit, 5GHz is the ONLY way to get things done.

    And, hey, why not fight for a 802.11ac radio? 802.11n is last year's tech, and these guys are going to have to go another ten rounds with Atheros if they want to be able to make another model. What was the

    • by adri (173121) on Tuesday April 30, 2013 @11:14PM (#43597507) Homepage Journal

      The FSF decided to investigate this AR9271 part. I'm not sure why.

      The AR7010+AR9280 NICs are dual-band. There's AR7010+AR9283 NICs that are 2x2 2.4ghz only. The AR7010+AR9287 NICs are also 2x2 2.4GHz only but support a few newer things (like short-GI in 20MHz mode, and generally better behaviour all around.)

      Hopefully the FSF certifies the AR7010 based firmware devices too. But, they've chosen this one and I'm glad they saw it through.

      I don't know if there's a hardware list that shows the dual-band ath9k_htc hardware. But it's out there, somewhere.

      • by gnujoshua (540710)

        The FSF decided to investigate this AR9271 part. I'm not sure why.

        The reason is because RYF certification is not simply done on software for a given chipset. RYF certification works by us entering into a formal agreement with a company that sells hardware. The agreement states that the company can display the RYF certification mark on all products that pass our testing and certification process so long as that company agrees to meet various requirements [fsf.org]; we agree to do a limited amount of promotion on the product (press release, listing the product on our site, etc). In t

    • by Arker (91948) on Tuesday April 30, 2013 @11:52PM (#43597667) Homepage

      Unfortunately, at the moment, the manufacturers perceive the proprietariness of their products as a value. You see how much this costs as is?

      But it's real. Every bit is there, driver, firmware, documentation. This thing will be supported as long as there is one old hacker that has one and doesnt like to replace a working part.

      And honestly, I know, I like having the latest and greatest when I can too, but can you please quit shitting on those less fortunate? USB 1.1 is 12mbps and there are a lot of people trying to work on less than that. I have the best service available in my area and it would not be a bottleneck in my system. (Not that I run critical systems on wireless anyway, it's ethernet, but if I needed to run something wireless the USB 1.1 throughput limit wouldnt slow me down.)

      Last years tech fully and truly available is infinitely better than this years tech locked away where I can never see it, even if I did supposedly buy the hardware. And hopefully this will lead to the manufacturers starting to figure this stuff out and doing more of it.

      • by tftp (111690) on Wednesday May 01, 2013 @03:25AM (#43598433) Homepage

        This thing will be supported as long as there is one old hacker that has one and doesnt like to replace a working part.

        First, a USB device is designed to be easily replaceable. Second, imagine that we have open sourced design of Pentium II. How much interest would that generate? In just a year or two it might be hard to find a hacker who'd want to deal with obsolete stuff - and WiFi stuff gets obsolete faster than you can put the credit card back into the wallet.

        It is certainly possible that someone, somewhere, buys a COTS consumer-level part, sticks it into the server, and then 10 years later is unwilling to replace the whole module when it fails. But that's what people deal with every single day in the industry and elsewhere - things fail and they need replacement. I would be far more concerned that the hardware of this dongle fails 10 years later - where would you get a replacement then?

        This whole approach appeals to too few people. Most are pragmatists. A pragmatic approach means that when the thing fails, there will be money and resources to replace whatever needs replacement. If no money is available, then I guess the project is not that important, after all.

        I do not know what business would be attracted by this specific dongle just on the basis that it is documented. This whole concept is way above the pay grade of pretty much everyone who works in IT. It is not even feasible, in most cases, for an IT guy to start a science project to debug a problematic device. This is handled by simple replacement of what doesn't work. This method offers fixed and predictable duration of repair. Hacking a driver, on a live system ... well, there are crazier things to do, but not too many.

        There is only one useful function that is directly fulfilled by this product - and that is creation of completely free computing systems. Days are coming (perhaps not tomorrow, but who knows?) when RMS's dark prophecy [gnu.org] materializes in laws and COTS hardware like WinRT, that denies you, the owner, the right to use the equipment as you see fit. There are F/OSS designs of the CPU and other key blocks already. This is another addition to the collection. Perhaps the hardware will be obsoleted and not available anymore (quite soon, actually, considering that every new IC has about 6 to 9 months on the market before it is obsoleted and replaced with something else.) But the principles of operation may be useful if one wants to build a free computing system.

        This function - a free computer - is very important. However, just as nearly all things that are good for the society (and the soul,) there is very little financial reward for doing good deeds. I understand pretty well how much labor went into development of the hardware, MCU software, and the PC software to make the thing work. I do some of that, now and then, for living. This is a good thing to do; but expect no monetary reward. The cost of the device is high, and only a handful of devotees will invest. (There are many devotees, but not too many will support F/OSS by buying the device.) I, for one, simply have no need for such a product - all my computers have built-in WiFi, not that I use it much anyway. Cable is more reliable, and has no interference from neighbors, and nosy Google cannot intercept it easily.

      • by chill (34294)

        Uh, you misread part of the complaint.

        USB 1x1 isn't the same as USB 1.1. The latter is a spec. The former refers to the antenna configuration. The poster was taking issue with the lack of MIMO and single-band radio, not speed of the USB bus.

        http://www.networkworld.com/newsletters/2009/110909wireless2.html [networkworld.com]

        I'm not disputing your points, just correcting your error.

  • by Arker (91948) on Tuesday April 30, 2013 @11:33PM (#43597597) Homepage
    Good to see a wifi device with free firmware. Freaking finally. The long-term implications of this are probably greater than even I imagine.
  • I have used free software starting with inews on the extra tape archive on the SunOS (not Solaris) 1.X and 2.X 1/4" QIC cartridge boot tapes.

    The FSF with its devotion to making everything it touches be free of copyright restrictions and to require you to publish any code that touches the FSF code has relegated itself to irrelevance. A $54 USB WIFI adaptor. Whoopdedo.

    I have moved on. Most of the open source software i use has a far less restrictive Berkeley or Apache style license.

    • by adri (173121)

      .. you mean, how the firmware is clearbsd licenced?

      Except for the Tensilica runtime, that's MIT licenced?

      Except for two GPLv2 files from ECoS? Which state that only those two files fall under GPL, not the rest of the stuff it's compiled with?

      I'm pretty sure it's free like you wish. You can create a closed source derivative of the firmware if you so choose.

  • Was it the Chadd? No, the Chadd was great! Kudos dude.
  • Too bad Atheros makes the lowest end garbage pile of chips, especially wireless ones. They're in all the cheapest laptops. They fail all the time, generate 1990's era hardware errors, and have the worst driver support imaginable.
    • by adri (173121)

      Atheros make high end and low end chips. It's up to the manufacturers as to what they choose. They choose price. Sad, but true.

      You can buy the higher-end 2x2 and 3x3 devices. The unit prices are more than the low end chips.

      Driver support? It's up to the company you bought the laptop from, not Atheros. Atheros only makes the chips. We don't make the NICs or the rest of the device. Especially in the windows world, vendors have a habit of doing 'strange ass shit' here and there. Please don't blame QCA for the

  • I picked up a Think Penguin wifi adapter for my machine and am happy to support the cause because it is a worthy goal. It is not about "purity" as some put it, rather for technology as a whole to advance in a stable, more beneficial direction for all. Patent laws were originally used to encourage inventors to share ideas without fear of loss of credit. Those patent laws were subsequently perverted by corporations so they can litigate people out of their inventions and claim benefits for their own interests.
  • I applaud the idea and implementation of the RYF certification; its nice to know that the software/firmware/drivers etc... needed to run a given piece of hardware are "verified" FOSS meaning that it is going to work just about everywhere (sometimes by default, like merged into Linux kernels, sometimes with a little work such as if someone wants to modify the damn thing to work in a special capacity). I'm even willing to a pay a reasonable amount more for RYF certified software/hardware. In an age where ev

Only God can make random selections.

Working...