• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

vendorwatch.org: Who's who in free software friendly hardware companies

n0cmonkey

Elite Member
From vendorwatch.org:
The purpose of this website is to keep tabs on computer hardware vendors and to convince them to provide open documentation for their hardware.

Many hardware vendors are reluctant to provide documentation about their hardware. This is very damaging to consumers because instead of releasing open documentation vendors instead either provide blobs or force those who look at the documentation to sign non-disclosure agreements. For a truly free operating system neither of these options are acceptable.

The goal of this wiki is to allow consumers and developers to discuss their efforts in getting vendors to release documentation.

So here's to blob and NDA free hardware support. :beer:
 
I agree with the concept, but it seems to be a little too much about mudslinging. Instead of concentrating on shame, I think it'd work better if it simply helped people find hardware that worked and eventually it'll gain enough respect that the vendors take it seriously.
 
What is the main obstacle to releasing the documentation? Hardware companies are in the business of HARDWARE so I would think anything that would sell more PCB boards for them is a good thing! Does open documentation actually restrict them? I would think its better because you can hit a bigger market...

I would understand Microsoft not wanting to release its source code because its a SOFTWARE company....but I don't see what a hardware company has to lose
 
Originally posted by: magomago
What is the main obstacle to releasing the documentation? Hardware companies are in the business of HARDWARE so I would think anything that would sell more PCB boards for them is a good thing! Does open documentation actually restrict them? I would think its better because you can hit a bigger market...

I would understand Microsoft not wanting to release its source code because its a SOFTWARE company....but I don't see what a hardware company has to lose

Well what about graphics card manufacturers? If nVidia just open-sourced thier high-end drivers then ATi will copy or benefit from the knowledge while retaining all their knowledge. With companies where there's not much performance gain with going with their competition (like wireless chipset manufactuers) then I wholeheartedly agree.

But I'll still support the companies who try to work with linux in general. For instance, I like Intel a lot now, I believe their wireless cards and graphics cards have open-source drivers, they've really embraced linux. And I'll still support nVidia over ATi because even though nVidia's drivers are not open-sourced, they aren't sh*tty (unlike ATi's).

edit: I forgot to add, I love wikis and think this is a great idea! Thanks for the link!

edit2: Actually I'm not sure if Intel themselves is embracing open-source or its a project that a different group of people are working on with driver support. Someone correct me.
 
Originally posted by: SleepWalkerX
Originally posted by: magomago
What is the main obstacle to releasing the documentation? Hardware companies are in the business of HARDWARE so I would think anything that would sell more PCB boards for them is a good thing! Does open documentation actually restrict them? I would think its better because you can hit a bigger market...

I would understand Microsoft not wanting to release its source code because its a SOFTWARE company....but I don't see what a hardware company has to lose

Well what about graphics card manufacturers? If nVidia just open-sourced thier high-end drivers then ATi will copy or benefit from the knowledge while retaining all their knowledge.
It's not necessarily about the source, it's about the specs. People are willing to write their own drivers if they just know how to talk to the hardware.

That said some of the excuses are intellectual property and, even if it doesn't hurt a company, they could be in a situation where they'd also have to expose someone else's ip, which just makes it a little stickier and slower. Apparently another common reason (that nobody admits to) is that the hardware is really buggy and requires software acrobatics to make it work properly. Rather than display the bugs they'd prefer to release a nicer looking interface in the form of a blob. But I imagine most of the trouble is that companies are just in a mindset where licensing rather than openness is what makes the innovation happen so it's counter-intuitive to give things away for free and they don't see the potential benefits.
 
For instance, I like Intel a lot now, I believe their wireless cards and graphics cards have open-source drivers, they've really embraced linux.

The fact that they require a binary-only firmware that's not redistributable isn't very nice, but it's workable. But the latest round of their cards also requires a binary-only daemon to be running on the machine, pretty crappy if you ask me.

edit2: Actually I'm not sure if Intel themselves is embracing open-source or its a project that a different group of people are working on with driver support. Someone correct me

It's definitely Intel, infact Intel does a lot of Linux work and the wifi stuff is probably the worst stuff they put out.

It's not necessarily about the source, it's about the specs. People are willing to write their own drivers if they just know how to talk to the hardware.

Yes, I would be happy with a decently fast 2D driver that actually works with Xv so I can watch a f'ing video.
 
The fact that they require a binary-only firmware that's not redistributable isn't very nice, but it's workable. But the latest round of their cards also requires a binary-only daemon to be running on the machine, pretty crappy if you ask me.

That's sucks about the binary-only daemon for the wifi card. It doesn't make much sense.. The kernel developers aren't ever going to support a hardware device that is dependant on something odd like that. Were did you get that from? (just curious, I beleive you though.)


Originally posted by: SleepWalkerX
Originally posted by: magomago
What is the main obstacle to releasing the documentation? Hardware companies are in the business of HARDWARE so I would think anything that would sell more PCB boards for them is a good thing! Does open documentation actually restrict them? I would think its better because you can hit a bigger market...

I would understand Microsoft not wanting to release its source code because its a SOFTWARE company....but I don't see what a hardware company has to lose

Well what about graphics card manufacturers? If nVidia just open-sourced thier high-end drivers then ATi will copy or benefit from the knowledge while retaining all their knowledge. With companies where there's not much performance gain with going with their competition (like wireless chipset manufactuers) then I wholeheartedly agree.

But I'll still support the companies who try to work with linux in general. For instance, I like Intel a lot now, I believe their wireless cards and graphics cards have open-source drivers, they've really embraced linux. And I'll still support nVidia over ATi because even though nVidia's drivers are not open-sourced, they aren't sh*tty (unlike ATi's).

edit: I forgot to add, I love wikis and think this is a great idea! Thanks for the link!

edit2: Actually I'm not sure if Intel themselves is embracing open-source or its a project that a different group of people are working on with driver support. Someone correct me.

Intel released specs for their onboard video cards a long time ago, I beleive. I have the feeling that they don't do a whole lot driver-wise to support Linux, but they do release specs. I beleive. Not sure about it. I am not sure about it though.. X.org/Xfree86 folks have had a long history of happily accepting NDA agreements so it's a bit difficult to tell sometimes.

I built my new desktop using Pentium-D proccessor and a ASUStek motherboard with the Intel 945g chipset. I have full 3d and 2d acceleration on it... Get about 700 FPS on glxgears, but it actually performs better then that would indicate in real-world tasks.. It's very good with Blender3d and I get about 15-30fps on enemy territory. I had to compile my own i915_dri.so and i810_drv.so (which is easy now since the modular X stuff) drivers for it to get 3d accel, but it's stable enough. Better then the Nvidia stuff so far stability-wise. Everything on the board is detected well. Sound works, nic works (with Debian unstable, not in "stable" though), video works. Nice and fast.

Still though I am looking to buy a x600 or x800 series ATI video card to try out the r300_dri drivers. I would like to have better performance then what the embedded intel stuff has. Those are suppose to work now, and I am happy enough with the DRI drivers for Intel to try out the DRI stuff for ati.. But I am a bit hesitent due to ATI's being braindead on their part.

Personally I would be very happy if Intel was to release a descrete video card with dedicated memory and such. I think that their actual GPU is probably fairly nice, but that their cards are crippled by having to use shared memory.

Personally I think that Nvidia's and ATI's excuse for being all protective about their card's specs as needing to protect their IP is bull crap.

If free software developers have reverse engineered ATI's cards to the point were they get as least marginally working 3d acceleration then Nvidia and ATI both almost have to have these things reverse engineered to hell. I bet that already Nvidia knows much much more about how ATI's card work then it would ever be nessicary to get good Free software drivers working, and visa versa.

With their drivers they have all these cross-licensed stuff with all this shader technology that is controlled by patents owned by other companies, So I can understand that. Probably a good percentage of their drivers aren't even owned by them anymore. I bet that Nvidia licenses stuff to ATI, ATI licenses stuff to Nvidia. Microsoft licenses stuff to both companies. As does older 3d companies like S3 (I know that both license texture shading stuff). Patents patents patents patents everywere.

If I were to talk to ATI or Nvidia I'd tell them to support free software driver's by releasing specs and any sort of sample code they may think would be very helpfull as well as some technical help when appropriate. They can keep their propriatory drivers for all I care. All the fancy texture shading this and anti-aliasing that they can keep for themselves. I just want good, solid, OpenGL acceleration and everything else I couldn't care less about. Most of all that is specific to this company or that paticular card anyways, so is generally useless to open source-style development were everything needs to be portable. If people need those propriatory features, then let them use the propriatory drivers if they want.

AS far as 2d support goes... I beleive that the ATI r500 and newer series doesn't even have 2d core anymore. That radeon stuff is gone. Probably in ATI's drivers 2d acceleration is translated through the 3d portions of the card. I don't know if all the 2d stuff is gone, but at this moment the absolute best you can expect for those cards in Free software is going to be VESA compatability modes.

Also people figure that's how the Nvidia's drivers work also, that the 2d acceleration is just translated to 3d portions of the cards. In fact when you run Nvidia's drivers your basicly running Nvidia's X server, too. They use a lot of XFree/X.org code in it, but the X developers don't even know exectly how much is their's and how much is nvidia's anymore.

The only reason that Linux does have 3d acceleration with these cards is probably due to the high-end. With the Nvidia Quatro or ATI's FireGL stuff. At least originally. Linux is used in 3d rendering and development in movies, for instance, and those customers are willing to spend a lot of money on even just drivers. 3d wildcats stuff, for instance, has a 3rd party companies like XIG develop Linux drivers for their cards and those drivers can cost several hundred dollars for end users, I beleive. You actually have to purchase the card, then go out and drop more money on the drivers to get it working. Other things like 3d visualization for scientific research and other things like that. I think that if it wasn't for that then there wouldn't be any 3d drivers for nvidia's or propriatory stuff from ATI. Probably now Nvidia and ATI see some consumer demand for these drivers so now they make stuff specificly for games and such, but I don't think it would of happenned originally.
 
Intel doesn't allow redistribution of their firmware. Big :thumbsdown: to them.

These companies don't have to release IP, just the API that drivers use to speak to the hardware. If they did a good job, it should be pretty abstracted.
 
Originally posted by: kamper
I agree with the concept, but it seems to be a little too much about mudslinging. Instead of concentrating on shame, I think it'd work better if it simply helped people find hardware that worked and eventually it'll gain enough respect that the vendors take it seriously.

We already have that: http://www.openbsd.org/i386.html. It hasn't helped. So in addition to that, we can go find out about the companies with parts in our computers. I don't see it as mudslinging, just informing the consumer of which companies care about them.
 
That's sucks about the binary-only daemon for the wifi card. It doesn't make much sense.. The kernel developers aren't ever going to support a hardware device that is dependant on something odd like that. Were did you get that from? (just curious, I beleive you though.)

It was in the announcement made on lkml a few months ago when they released the driver for the new cards.

http://marc.theaimsgroup.com/?l=linux-kernel&m=114082031110775&w=2

From that posting: "...we have placed the regulatory enforcement logic into a user space daemon that we provide as a binary under the same license agreement as the microcode."

 
Originally posted by: Nothinman
That's sucks about the binary-only daemon for the wifi card. It doesn't make much sense.. The kernel developers aren't ever going to support a hardware device that is dependant on something odd like that. Were did you get that from? (just curious, I beleive you though.)

It was in the announcement made on lkml a few months ago when they released the driver for the new cards.

http://marc.theaimsgroup.com/?l=linux-kernel&m=114082031110775&w=2

From that posting: "...we have placed the regulatory enforcement logic into a user space daemon that we provide as a binary under the same license agreement as the microcode."

Sounds awfully not friendly to me...
 
Sounds awfully not friendly to me...

And easier to reverse engineer. If the daemon is just watching/adjusting a sysfs file, it shouldn't be too hard to figure out what the values in sysfs mean.
 
Back
Top