Ryan McBride requests the help of the OpenBSD community in convincing Texas Instruments to change the license of their firmware for the ZCX100 802.11b chipset. Theo de Raadt makes a similar request directed at Intel. A success story from Theo de Raadt in using this tactic on Adaptec. This is also important for Linux and the rest of the BSDs. I had to “pirate” my Prism’s firmware files in order to make the pcmcia card work with my Linuces. I don’t see the point of keeping these firmware files bound.
OpenBSD looking for better licenses on binary firmware for WiFi
2004-10-27 OpenBSD 37 Comments
Are you serious?! This is blackmail. All this gives OpenBSD part of it reputation and with “leadership” like that what should we expect?
It shows the companies that there _are_ customers out there waiting to drop a few bucks on a piece of hardware. All the company has to do to get that money is help out a little bit.
Why doesn’t <insert OS here> get better hardware support? Because companies don’t provide anything. Why don’t companies provide anything? Because they don’t believe there is a market. We’re trying to show these companies that there _is_ a market.
If I understand it all correctly, all they have to do is allow distribution of the firmware.
if this won’t help in creating a driver than what is this for? Is this image for flashing the bios of a chip to upgrade it? does not having this somehow worsen the quality the device in OpenBSD?
Jens: I do not consider this blackmail since there is no real harm to the company besides people just not buying their product and maybe a bit of spam. Though just add “BSD” to the list of words like “b1gg3r p3n1s” and it’ll all get blocked.
The firmware is like software that runs on the card. Without it, the card will be useless. The firmware used to be included on a prom or something on the card, but companies want to produce these cheaper so they move the firmware off board to be loaded by the driver.
BTW, I am not asking about open sourcing that code. I just want it to be redistributed freely on the web or by OSes, so alternative OS users can find them easily and legally. Or, even better, to allow driver writers to include these firmware s to their code.
I hope, someone will bugging on NVIDIA next for nForce chipest.
Lots of postings with bad grammar (jens, bsdrocks). Can’t even understand some of them..
In the end they moved the firmware off of PROM chips on the device, because it makes it easier to upgrade the firmware. In any case it would be nice if they at least allowed a binary license for redistribution. Which would not only help other OS’s, but help them in that it would be less tempting to hack it and create new firmware.
I think the biggest problem with the nVidia chipsets are the ethernet controllers. It isn’t a firmware issue, but a documentation one. Get some names and emails together, and maybe something will happen.
I find it very annoying when i install BSD or Linux on a machine and it can’t see a particular device. Previously i would just think well the driver hackers aren’t there yet but now i know who to blame. I always look on the box of any new hardware to see if it supports at least linux.
As someone who works in a Big Evi^H^H^H Fluffy Corporation answering phones, Theo’s story rings true to me – the single most important thing to a call centre is volume. One call is small beans even it’s an important problem. One hundred calls about the same thing get priority up the wazoo.
@anonymous (jccbi.gov) – a perfectly working free driver for nvidia onboard networking already exists and has done for several months, it’s called forcedeth and most recent distributions include it.
@smartpatrol – even looking at the box doesn’t always help, especially in the case of wireless network cards. I’ve bought two completely different cards with the same problem here, a Netgear wg311 and an SMC card. Both cards have ‘v1’ and ‘v2’ versions. In both cases, the two versions are based on *different chipsets*, one with much better Linux support than the other. In both cases, the two versions are sold in *identical boxes*: the only way to tell the difference is to examine the physical card itself. This is ridiculous behaviour…
I’m aware of the forcedeth driver. I’m not sure what the status is for *BSD though, since I refuse to purchase nVidia motherboards because of the nic issue.
ooh, good point. The BSD guys are good, though, I’d be amazed if they hadn’t adapted forcedeth by now.
TI uses GNU/Linux in some of their VLCT chip testers. Hopefully, they will listen and change the license.
Bill Paul wrote an NDIS compatibility layer for FreeBSD that allows users to use Windows drivers for non-ported wifi cards and it uses the binary firmware files included in vendor distributed driver packages. I assume obsd dev has heard of this.. but maybe not? I can understand wanting the firmware to be free, but barring that, an approach like ndisulator would seem to be the next best thing. At least it gives users the opportunity to use their hardware while dev communicates with vendors. It also doesn’t require including proprietary code in the base…
That’s ironic, because nvidias next chipset, the nForce4, will have an hardware firewall:
Something like that will probably never make it into OpenBSD, Theo believes in free software.
The 350GB(?) chipset had a built in firewall too. I’d rather trust PF than something I can’t control from OpenBSD.
Good call CaptainPinko:
“Jens: I do not consider this blackmail since there is no real harm to the company besides people just not buying their product and maybe a bit of spam. Though just add “BSD” to the list of words like “b1gg3r p3n1s” and it’ll all get blocked.”
How many companies pay for market reseach? How much do they pay for these reports? Well, we now can cut out the middle man and produce results for said company X.
1) People want to buy their products; however they cant use them due to the implementation of hardware and software.
2) Just provide the info that would be on the Proms.
What would be the result.
1) It wont cost them anything.
2) No IP violations
3) More people would buy their hardware (this would be a selling point)
Everyone wins: Consumers get the products and the business get the revenue over their competitors who dont provide the data.
If there is a negative aspect to doing this, please point out the potential repercussions. Just curious.
There’s always the legal thing hanging over the head of wireless. You’re not permitted to just let anyone use whatever frequency they want (Which is why the Atheros driver comes with a closed binary module).
Sometimes it’s just not possible for people to show the source even if they wanted to.
I think 6 years ago we asked our user community to contact Adaptec in the same way.
Six years ago, 1998, it was not illegal to send spam, unsolicited emails or just email someone to death of his/her mailbox. Unethical- may be, illegal- no.
Today, by recommending to spam someone, or by doing it, you might get in serious legal trouble. Can you (financially) afford it?
We asked our users to email the 30-some email
addresses which we had attempted dialogue with over the year. Five hours later I had a phone call from Adaptec asking me to stop our users from mailing them. I said it was `beyond my control’.
So, did you ask users to stop emailing them, or not? Why not? May be they, your users, will listen to you and stop?
They said that a few of the people were HR staff. I said it was `beyond my control’.
You mean, for a year you attempted a dialogue with wrong people, and when you find out these people rightfully ignore your requests- decided to keep mum and not to ask your users stop spamming them?
Sorry, but asking HR person to provide you with code sources is, well, not right. Not realizing it for a year is, well, not bright.
Some people are freedom fighters, some people are just terrorists. Sometimes freedom fighters become terrorists and don’t even realize it. So sad…
Releasing BINARY firmware blobs under a license that allows the OpenBSD developers to distribute it on the install cds will not affect that.
Reverse engineering the firmware would probably cause more problems.
This isn’t spam. We’re not selling anything, we’re letting a company (whose job it is to make money by providing products to consumers) what we (as consumers) want.
Bill Paul wrote an NDIS compatibility layer for FreeBSD that allows users to use Windows drivers for non-ported wifi cards and it uses the binary firmware files included in vendor distributed driver packages. I assume obsd dev has heard of this.. but maybe not? I can understand wanting the firmware to be free, but barring that, an approach like ndisulator would seem to be the next best thing.
Someone pointed out that Theo likes software to be free – true, but actually that’s not the main issue. OpenBSD is famously paranoid about security, and every line of code is security-audited. Therefore, an exploit in the Windows driver (however remote the possibility) would result in an exploit in OpenBSD. This is something that Theo won’t compromise about. Every driver in OpenBSD must have the source code available so that it can be checked for security. Anything less just won’t be acceptable to the OpenBSD team.
I’d be willing to bet it’s more of a freedom issue than a security issue.
In regards to being “free” – yeah, there’s no doubt that’s a concern, but ndisulator code is under BSD license. Would refusal of this code suggest obsd dev getting into the business of not allowing proprietary code to *run* on their platform? Ndisulator only enables the translation of windows drivers… It’s similar to saying samba shouldn’t be allowed as a port because it deals with proprietary specs….
Honestly, I can’t see them embracing either, only because of the philosophical weight of their position – they specifically want open-source drivers, and using something like NDISulator would not get them that, and would only prolong the larger issue. I can respect that. I just brought it up for conversation sake. OTOH, it would be a nice-to-have if someone developed an unofficial port.
It might not be allowed in ports either. Putting the firmware in ports was suggested, and denied. An unofficial port might be popular though. I’d prefer to live without certain brands of hardware if it means I can’t use it the way I want to.
Just an add on:
I really meant an unofficial “patch”. Ndisulator is a module, and the original freebsd cut was roughly 13 files + a userland utility. See freebsd’s code: /usr/src/sys/compat/ndis.
Ok, ok. I’ll stop now.
Putting the firmware in ports was suggested, and denied.
That’s understandable. The procedure for using ndisulator successfully requires the user to import the driver files from the vendors windows installation cd. If users want to use unsupported hardware, they should have to do the dirty work. But it’s quite easy to use and pretty seamless aside from the manual import. *shrug*
Just for all you people who keep saying that binary only distribution is “ok” for firmware, it’s not. If there’s a bug in that firmware you can’t fix it. It’s exactly the same as if there were a bug in a binary only driver: You can’t fix it. You can’t audit it. You can’t trust it. If the code were perfect they wouldn’t make it flash upgradable, they’d burn it into a ROM.
Russian (Canadian) Guy, please be careful! You will summon the OpenBSD army, err…platoon, to action. “I am just the sweet South African Theo raised on my own whits and need only my CD OS sales to survive.” Dan Dodge (Canadian and QNX) did it right. Theo is a spoiled child who thinks the world owes him something for writing software.
Wake up OSNews!
To all you shmucks who keep posting nonsense here, this is not, I repeat NOT, a call to make vendors release their firmare uner an OSS license , it’s about the redistribution license. It has NOTHING to do with the ndis wrappers and i’d be very surprised if the ndis wrappers ships with license breaking firmware.
All that is asked is that the binary firmware to be licensed in such a way that it is freely redistributable by OSS operating systems. It has NOTHING to do with the source code of the firmware. This does not only benefit OpenBSD but, as the original item says, ALL OSS operating systems. It would be nice of people could lie their irrational hatred for Theo to rest and try to actually understand this issue and what it’s about
To the guy who said it is spam, you’re an idiot.
Something like that will probably never make it into OpenBSD, Theo believes in free software.
+ stability, native solutions. Though he’s not asking the firmware to become BSD licensed.
@ Russian Guy
Unfortunately you haven’t argued this is ‘spam’. ‘Spam’ is unsollicitated bulk mail. For one, every mail send by each person is unique. At least that is the intention. They mail every person because they’re searching for the person who they have to contact.
I don’t like it when hardware doesn’t work no matter what OS runs -or is intended to run- on that hardware.
spam’s not that easy to define. there are those of us who still think of it as badly cross-posted usenet messages…
@dpi: Unfortunately you haven’t argued this is ‘spam’. ‘Spam’ is unsollicitated bulk mail.
“How do you define “spam”?
The definition of “spam” is a tricky issue, with as many strongly held opinions as many other age old questions such as “the number of angels who can dance on the head of a pin” and “chicken versus egg.” For example, many define spam as unsolicited electronic mail sent in bulk. Others believe “bulkness” is irrelevent, it’s merely a matter of whether the message sent was solicited. Still others debate the importance of whether the message was commercial in nature.
So, take simple quiz, are you spammer or not, if:
1. You initiated flood of unsolicited emails.
2. You know that your target audience is minority of all people you sent emails to- rest is collateral damage.
3. When people who want to op-out from this emailing contacted you and asked to stop emails- you refused to cooperate.
@By Anonymous (IP: —.jccbi.gov)
This isn’t spam…we’re letting a company what we (as consumers) want.
Wrong! You are not consumers. Consumers buy products. You are potential consumers, may be consumers, might be consumers.
Also, you know what is the difference between a government office providing services to citizens and company providing products and services to consumers? A company can refuse to sell products or provide services to potential consumers if it does not want to.
If you were a consumer you would have bought that product, contacted company support department (not ambush some unsuspecting HR person who no doubt felt terrorized by friendly emails from Open Source geeks), and if support does not give you want you want- you just return their product to the store where you bought it and take your business elsewhere.
Sounds fair for me. How about you?
Well, i guess most people who would define this as spam would agree with me the 419 scammers who are commercial, lying and mass-mailing than a form like this one.
Also i’m not so sure if this is unsollicited. Because they’re basically searching for the right person to contact.
By my definition of ‘unsollicited’, it would mean i could prosecute someone who sends me a message i prefer not to see no matter who sets any importance to that message. Say its a takedown request from the MPAA or RIAA after which ones argues “i don’t want to receive mail from them” after which it is defined as spam.
What i’ve seen though is that in such cases as the above, a reply is made with ‘please don’t contact me futher, since that is regarded as spam’. I’ve done that too [but not with *AA] and it works fine.
PS: I think this has actually been asked to Theo (IIRC he replied that the actions of others are ‘beyond his control’). He could at least forward that e-mail to the list and ask to seize sending to that person. I also think there are possibly better solutions such as phone or RL contact. Which reminds me, if this would happen and you’d phone that ‘helpdesk’ or so with multiple persons that would be okay.
How do you think they got the HR people’s email addresses? Someone else in the firm gave them out. How are they supposed to know they’re not the right address? Do you think the TI people gave them a nice list of the email address of everyone in the firm along with their job description? Of course frikking not. They tried to get in touch through the proper channels and got nowhere. Through persistence they were given other forms of contact, which they also tried. It’s not *their* fault if the company gave out the addresses of the wrong people, is it now?
We are consumers. We may choose to buy another product because of the level of support offered by that company, but we are still consumers. I’m letting that company know that I would buy their products if they offered better support for my needs.