Statement by ATI: “For other markets, such as workstation and consumer, performance and feature differentiation are key metrics. Proprietary, patented optimizations are part of the value we provide to our customers and we have no plans to release these drivers to open source. In addition, multimedia elements such as content protection must not, by their very nature, be allowed to go open source.”
No Open Graphics Drivers from AMD
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
a major mistake if you ask me.
ATI doesn’t work that well with linux; NVidia has a clear edge and this way it seems to stay it this way.
ATI’s stability is a big issue as well as long standing bugs.
I will not buy ATI hardware ever again! I can’t believe my own eyes: “…such as content protection must not, by their very nature, be allowed to go open source”.
I will continue to buy ATI because of their great hardware and good drivers.
I have my sarcasm detectors switched on, they caught that one…. hahahaha
“…such as content protection must not, by their very nature, be allowed to go open source”.
I agree, that’s a stupid sentence. There’s no reason DRM software couldn’t be open-sourced, just like open-source versions of PGP doesn’t compromise those using it.
I agree, that’s a stupid sentence. There’s no reason DRM software couldn’t be open-sourced, just like open-source versions of PGP doesn’t compromise those using it.
First you have to understand that content protection relies exclusively on security through obscurity. The whole model is inherently flawed. I’m not talking about the analog hole—that’s a whole different issue. PGP works because you have two types of people, those who you want to have access (usually just one person), and those who you don’t. Alice encrypts a message to Bob and only Bob can read the message because only he has a copy of his private key. With DRM, everyone has to be able to decrypt the content. The goal is to restrict how and when they decyrpt that content. Without locking down the whole system (applications, OS, drivers, and hardware), this is impossible.
EDIT: Don’t take this the wrong way. I oppose DRM. All I’m saying is ATI is right.
Edited 2006-08-12 20:25
EDIT: Don’t take this the wrong way. I oppose DRM. All I’m saying is ATI is right.
Well, they’re right that DRM requires proprietary drivers, but they’re WRONG that DRM is a necessary component of future computing. Of course, I’m sure you agree.
What took you so long?
I already rejected ATI due to their poor driver for Linux (even though I’m using Windows currently: I like to keep my options open).
NVidia driver is better but still closed source, unfortunately.
I will not buy ATI hardware ever again! I can’t believe my own eyes: “…such as content protection must not, by their very nature, be allowed to go open source”.
Neither will I; Wanna know the secret to protecting content? make it affordable so that piracy isn’t an attractive option to the average punter; how about instead if charging NZ$40 for a DVD, charging NZ$25, instead of NZ$35 for a new cd, charge NZ$20.
As for their excuse for no opensourcing, then create an propretary module which the driver loads when available? the MGA driver already does that; if it can’t find the HAL extension, it loads as perusual.
Edited 2006-08-12 23:03
First, both AMD and ATi are bound by SEC restrictions since they are both publically traded corporations that have annouced a pending acquisition. ATi might not even be allowed to make public statements that represent a major change in product licensing, and AMD might not be allowed to comment on ATi’s public statements. IANAL, but I know there’s certain restrictions in place until the acquisition goes through.
nVidia doesn’t seem to have plans to open source their drivers, and they don’t seem to be interested in contributing to the Nouveau project. nVidia’s proprietary drivers might be a bit better than ATi’s, but neither vendor has any plans to open their 3D drivers or contribute to open source 3D driver projects.
AMD might have something to say about ATi plans, but not in light of nVidia’s driver superiority on Linux. Intel is the issue at bar here, since they are making a big push to close the gap between integrated graphics and discrete graphics, and they are committed to supporting all upcoming 3D features in open source drivers. Intel wants to make discrete graphics a niche market for hardcore gamers and graphics professionals.
In the process, the free software community is beginning to fill the resonating chamber with comments about how the availability of open source graphics drivers will effect their CPU purchase options. Intel is positioning itself as the champion of open source 3D graphics drivers, and you can only get their graphics chipsets for Intel platforms with Intel CPUs.
It’s sort of ironic, really, because just a week or two ago, the same resonating chamber was reinforcing the idea that AMD is friendly to open source. Now it appears like AMD is the closed platform, and we are realizing that Intel has been damn friendly to the open source community all along.
Case in point: Look at where the contributions to the Linux kernel are coming from. Which of the big hardware vendors are making the biggest contributions to free software? Without even looking at the commit logs, I’m pretty confident that, among hardware vendors, IBM and Intel are the two biggest contributors to the Linux kernel. AMD is, at best, a distant third.
It’s sort of ironic, really, because just a week or two ago, the same resonating chamber was reinforcing the idea that AMD is friendly to open source. Now it appears like AMD is the closed platform, and we are realizing that Intel has been damn friendly to the open source community all along.
Case in point: Look at where the contributions to the Linux kernel are coming from. Which of the big hardware vendors are making the biggest contributions to free software? Without even looking at the commit logs, I’m pretty confident that, among hardware vendors, IBM and Intel are the two biggest contributors to the Linux kernel. AMD is, at best, a distant third.
How true. Intel is very involved with Linux, and have done great things for Linux. AMD really hasn’t done that much, but they didn’t seem to be in opposition to open source. I guess that has changed now.
I was really hoping that it would turn out the other way with AMD opening up ATI’s drivers. Now I really think I’m going to start edging towards Intel. Their processors are very good these days and I would love to have no hassle drivers for graphics, ethernet, and wireless. All of them are very good now and their graphics cards get better all the time.
assuming the drivers are improved ad work well im still happy with them being binary.
At this point, it is still your choice as to whether to install them or not
If the drivers are to remain binary, there is not real reason for anyone to move from nVidia.
there is not real reason for anyone to move from nVidia
Other than Free Software ethics, there aren’t any. Bye bye NVidia. Bye Bye AMDATI. Welcome Intel.
http://osnews.com/comment.php?news_id=15477
So now it’s: Bye bye NVidia. Bye bye AMDATI. Bye bye Intel. Welcome… ???
Seems like that all the peeps who want to keep their OS open is pretty much about to get f–ked in the ass by everyone in the long run.
Edited 2006-08-13 12:45
http://www.ati.com/products/certified/systems.html
One step at a time…
R&B
so will be a relase for genesi/ppc (for linux, i think)? is a good news.
can i ask you what kind of code cannot be revealed on driver?
and is not possible relase a driver with only 3d function (or also a little avivo use)?
well, i’ d like to see a firegl on ppc one day…
The market share is just too small to let IHVs to take it seriously
It is not that you can’t use an ATI card in linux desktop, just that the drivers are not as feature rich as their windows counterparts.
The future is about content, just like today it is about connected/networked computers.
Without DRM, linux won’t be a strong contenter on the home desktop.
Without DRM, linux won’t be a strong contenter on the home desktop.
yeah, coz everyone knows that your computer will break and not let you do anything without DRM installed…..
Windows users are full of mad dogs shite and they are going to bend over and take it…. as usual
[Without DRM, linux won’t be a strong contenter on the home desktop.
yeah, coz everyone knows that your computer will break and not let you do anything without DRM installed…..
Windows users are full of mad dogs shite and they are going to bend over and take it…. as usual]
Well, today most users pay their internet access and the DSL modem/cable modem boxes are part of the access management scheme. In the future, DRM will be sort of like what DSL/cable modem boxes are today for the internet access, only it’s for the content access in the future.
The whole point of internetworking is to speed the sharing of “content”–information and culture. Your “future” is already here.
“Content” and DRM don’t necessarily go hand-in-hand, and making that assumption is disingenious. There is plenty of culture available that has no technological restrictions on distribution, only social and legal ones.
If DRM is part of the future, then I don’t want any part of it. If that means linux isn’t a contender on the desktop witout it, so much the better — there’s less pressure to screw the user.
Edited 2006-08-12 23:28
Ok, it’d be *nice* if all their drivers were open source, because that would ease the burden of supporting the many operating systems they don’t expend effort on. However, correctly and fully documented specifications for how to program the hardware would make the task of writing open-source (or other licensed) drivers viable.
So, if AMD/ATI (whomever) really don’t have the option or desire to expose their software IP (which is their legal and moral right to choose) that’s definitely a speed bump, but not a road block to using their hardware effectively. If, however, they choose not to release proper documentation required for writing drivers for their hardware, *then* it becomes a road block.
At this time, there isn’t a very significant market in terms of number of sales (real or imagined) that are truly affected by them not releasing their drivers as open source software. However, while that may be true for the number of machines and sales of hardware right now, it is still wise to remember that word of mouth to others that don’t otherwise care about the license or details of drivers may end up becoming significant. Why? Because the reality is that a lot of the people that are interested or use OSS and strongly believe in it are often sufficiently computer-literate that those that are looking for advice will ask them for it (regardless of how knowledgeable the OSS advocate really is, it’s still perceptions that matter most) and take their recommendations with more consideration, and regardless of why that advocate wishes to not suggest a vendor or hardware, will still be influenced by it.
I mean, they force me to change my ATI Radeon 9600Pro to something from nVidia? Or my next computer should contain intel chipset with open source drivers?
Damn… I have to dump my AMD processor too…
I just lost respect to AMD…
Edited 2006-08-12 19:46
I mean, they force me to change my ATI Radeon 9600Pro to something from nVidia? Or my next computer should contain intel chipset with open source drivers?
Your choice is:
-ATI: Good graphic chips, sucky opensource drivers, sucky closed drivers (they’re better than the opensource ones thoug)
-Nvidia: Good graphic chips, no opensource drivers, good closed drivers
-Intel: Not the fastest graphic chip ever, but certainly it’s not slow either. Great open source drivers. Now, they even make CPUs that doesn’t suck and that even surpass AMD.
For me, the choice is clear. I spend 100% of the time using a GUI – I want a GREAT graphics stack, including drivers. Intel also makes opensource drivers for their SATA chipsets and network cards – now that I can buy opensource-friendly hardware, I’m not going back, unless AMD does something with ati…
Edited 2006-08-12 20:06
For me, the choice is clear. I spend 100% of the time using a GUI – I want a GREAT graphics stack, including drivers. Intel also makes opensource drivers for their SATA chipsets and network cards – now that I can buy opensource-friendly hardware, I’m not going back, unless AMD does something with ati…
Same here. It’s time to ditch filthy ATI, AMD, NVidia and SiS (especially SiS
). The next laptop I’m going to buy will contain open-source friendly chips only (Intel).
Same here. It’s time to ditch filthy ATI, AMD, NVidia and SiS (especially SiS
). The next laptop I’m going to buy will contain open-source friendly chips only (Intel).
Given that, when doing every day work, one wouldn’t notice the difference between the ultra high end gear and Intel Media Accelerator 965.
Considering that Ati has its own integrated solution, and same with Nvidia, why go with the other two when Intel has opensourced their specs, and now, there will be stable, secure drivers for all, rather than the mirade of crap drivers that seemed to ge getting pumped out of Ati and Nvidia on a regular basis.
I mean, they force me to change my ATI Radeon 9600Pro to something from nVidia? Or my next computer should contain intel chipset with open source drivers?
Damn… I have to dump my AMD processor too…
I just lost respect to AMD…
I was just thinking the same thing. nVidia’s binary blobs provide excellent support for open-source *NIX OSs, but open-source 3D acceleration in nonexistent. ATI has poor binary support and limited open-source support.
Intel GPUs aren’t the greatest for performance, but their open-source driver support is, as I understand, on-par with nVidia’s binary support. And let’s not forget that the rest of the onboard stuff (sound, ethernet, SATA) is well supported.
After at least five years as an AMD fanboy, I’d consider switching to Intel if AMD doesn’t whip ATI into shape.
Hm, would it really be sensible to consider switching CPUs because of a proprietary graphics driver? There is a sense of logic in this that’s somehow alien to me, but … At the moment Intel might have a slightly better CPU, next year AMD might have a slightly better CPU.
I think ATI’s decision is a mistake – but nVidia is not better at all. And Intel’s graphic’s chip is nice, open source, but can neither be seriously compared to ATI’s nor nVidia’s chip. So?
Well, considering that Intel GPU’s are integrated to the motherboard, and the only motherboards that have integrated Intel GPU’s are those that support only Intel processors, it’s kind of natural that you need to “switch CPU’s” when swithing to Intel GPU hardware. Unless Intel decides to start making graphics cards again, à la i740…
As for myself, ATi and nVidia stopped being choices long ago. I’m currently using Matrox G450 with an SMP Pentium III system, but when the hardware dies, my next system will consist of Intel chips only. Graphics hardware performance is (seriously) irrelevant to me, and I like the quality of Intel Ethernet chips and motherboard chipsets. When I can, at the same time, support a hardware manufacturer who is actively supporting the Free Software movement (at least most of the time), the choice is clear.
…except, Intel’s international business and political practices make it even worse than Microsoft in my eyes.
Yes, they are in fact fairly friendly to _some_ parts of open source, the onboard graphics chipsets for instance is one I’ve given them credit for before. Mind you, even in that latter, I’m not sure why folk are suddenly saying how great Intel’s graphics chips on Linux are, historically, to me, they’ve been tolerable, but hardly great… Being stuck in 16bit colour for instance if you want some modicum of DRI support, and even that not being all that good in terms of performance. They may have gotten better since, but still, they haven’t been that good.
“Or my next computer should contain Intel chipset with open source drivers?”
This drivers are only partially open source. Because “Intel can’t publish in source form,
like Macrovision register stuff and other trade secrets”.
I might be cynical but “Other trade secrets” are strangely similar to DRM.
Of course “It’s optional,
so if you don’t want to use a binary module, you don’t get to use code
written by Intel agents for these features”.
It is nice that it is optional but what features needs this blob and if it gonna change in future(3D support)? So after good coverage in press it is clear that their drivers are not so open. As usual.
http://lkml.org/lkml/2006/8/12/10
ps.
Support for Macrovision ? It should die long ago.
I wish it don’t turn into unmaintained blob like mga HAL.
—
RBEU #1000000000 – Registered Bad English User
It is funny that it is always ATi who gets all the criticism. Partially this is justified, but regarding open source friendlyness NVidia is the real enemy. None of their cards has a decent open source driver.
ATi did until a while ago provide all necessary documentation to develop open source drivers. It is very wrong that they no longer provide this, but NVidia is definately a front that needs more attention.
Edited 2006-08-12 20:25
I’ll be buying an Intel processor with Intel graphics then. I prefer that value.
I don’t care if AMD produces faster chips, it’s less important.
“I don’t care if AMD produces faster chips, it’s less important.”
Not anymore, they don’t. The Intel offerings, from what I’ve heard, totally and absolutely trounce, squash, and obliterate even the best of AMD’s offerings.
If faster is better… Intel’s got the lead again.
The link to the Heise.de article was interesting if you know German. Here are some highlights:
1) ATI Germany consulted with their overseas colleagues to refute rumors circulating in the Linux community that ATI was was either partially, or completely releasing their drivers under an open source license.
2) The actual refutation is the usual doublespeak (given in English, without translation into German – interestingly enough). It stresses ATI’s commitment to open source, while virtually equating it with Red Hat.
3) ATI offered some support in the form of chip documentation for the XFree and Xorg projects, but that stopped with the Radeon 9250. Starting with the 800 series, support became far more sparse, with 3D support affected more than 2D. 3D open source drivers were developed via reverse engineering.
4) ATI support for the latest generation of cards is essentially nil. Indeed the company has blocked Xorg development through the invocation of non-disclosure agreements. Hence, the only driver that works for most Linux distributions is VESA. VESA won’t even allow refresh rate adjustments, let alone 3D support.
5) ATI will keeps its proprietary and partially patented developments to itself. After an English statement (again, not translated into German; do they think Germans are educated, or something?) implying that proprietary drivers equals value to the consumer, the German text states that ATI’s drivers are controversial in the Linux community, possibly illegal, and in violation of the GPL. This is why you don’t find ATI or Nvidia drivers included in most Linux distributions. In addition, ATI didn’t cover itself with glory with its drivers, which were buggy, and six months late for the 1000 series cards.
6) There are certain “content protection” features that ATI may not know how to support in an open source driver. There follows another English quote about “content protection” (a.k.a., DRM) not being open source by definition. This means that, at least in the short term, there will be no resolution of the Linux driver problem for ATI cards. It’s not clear if such an attitude will change significantly upon the completion of AMD’s takeover of ATI. AMD has released no statement answering Heise’s questions in that regard. AMD has been relatively open minded toward open source in the past. The issue of intellectual property and drivers is new territory for AMD, and AMD will have to grapple with the issue.
A few days ago, Intel attracted attention by strengthening its commitment to open source graphics drivers. AMD, ATI, and Intel have all shown active support in the past for open source drivers for SATA, IDE, and audio features on their motherboard chip-sets. The companies often offer driver support to the Linux kernel developers for yet to be released chip-sets.
The article concludes with a, “for more background, click here” link.
Now for my own comment. I found ATI support for earlier cards adequate, and usually stable. If ATI would return to the kind of support offered to Xorg and XFree for their earlier generation cards, I would be satisfied.
ATI’s drivers are controversial in the Linux community, possibly illegal, and in violation of the GPL. This is why you don’t find ATI or Nvidia drivers included in most Linux distributions.
Then they should release drivers for BSD licensed operationg systems instead- we don’t hate good binary drivers
nVidia already released native FreeBSD drivers with good performance.
Edited 2006-08-12 21:09
I will also buy intel CPU for my next computer. As someone pointed out, intel and AMD (CPUs) is always on par speedwise, so that’s not an argument, but freedom IS.
I will not bow to the mighty properiaty kick-ass GPU mainstream. I’ll take that blow, and take a stand for open source. I’ll buy an Intel GPU.
MY future is hopefully free. In every aspect. Not just software. Are you just going to stand there or are you going to help?
Edited 2006-08-12 20:48
you honestly think Intel is now some great champion of freedom? sheesh dude, try to have a memory longer than two months. A world where it’s all “Intel inside” isn’t an ideal I’d care to see.
Proprietary, patented optimizations are part of the value we provide to our customers
What is this supposed to mean??? It’s the funniest sentence I’ve heard in a while. What kind of value do their customers get from the fat that the optimizations are propietary? The fact that they open source their drivers doesn’t mean they will be worse ! Hell, only the contrary can happen. Either they’ll remain as bad as they are now (for Linux, I mean) or they’ll improve due to community support (which is the most likely alternative).
What does ATI win from keeping their drivers closed? I mean, they don’t sell drivers, they sell hardware. Drivers are just a load they have to carry. Why not share that load and at the same time give more value to their customers?
Am I missing something???
is right there: such as content protection must not, by their very nature, be allowed to go open source
No shit sherlock… and the sooner the open source hippies pull their heads out of their asses and realize it, the sooner we can put all this bull behind us and actually have hardware support with ENTHUSIASM from vendors, instead of the ‘oh that’ attitude you quite often encounter… and maybe get rid of this bullshit ‘open source only’ nonsense people seem to have gotten raging chodo’s for of late.
GOD FORBID a hardware manufacturer should want to protect it’s inner workings or programmers want to keep their programming methods to themselves.
AGAIN, PROGRAMMERS – STOP DEVALUATING YOUR WORK with this Open Source nonsense, or WORSE, FLOSS. As always, I can hear the business majors laughing themselves all the way to the bank.
>>By Bending Unit (1.07) on 2006-08-12 19:37:11 UTC in reply to “RE: missed opportunity”
I will continue to buy ATI because of their great hardware and good drivers.<<
You’re kidding about the good drivers, right?
agreed that ati drivers under linux are poor and often difficult to set up.
“Proprietary, patented optimizations are part of the value we provide to our customers and we have no plans to release these drivers to open source.”
The key benefit of someone patenting something (for society and economy) is publishing details in exchange for temporary monopoly. Now we see another corporation going against the spirit of the patent system. The statement above basically says that ATT/AMD is knowingly breaking basic assumptions under which patents are granted – they acknowledged that they are abusing patent system. But what else could you expect from those corporate drones ?
…that since the AMD/ATI merger can hardly have had any effect yet, AMD change their minds. If not, then sorry AMD, bad publicity is exactly that, and the next chips I buy will be Intel.
Sure Duo is longer then X2 in all aspects, thank ATI to proof it.
Some of the code is owned by other companies – so just how nvidia failed open sourcing their drivers a few years ago because of legal troubles this was going to happen to AMD also anyway.
“In addition, multimedia elements such as content protection must not, by their very nature, be allowed to go open source.”
Statements like this should hopefully prove to Linus and people like him that DRM is to be avoided. Of course technology itself cannot hurt anyone – a well-constructed nuclear bomb cannot go off by itself – but the people who masturbate over DRM don’t have anyone’s interests at heart but their own, and have even less interest in open sourcing their technologies.
Ales klar.
I won’t buy neither AMD processor or ATI card any more. The only thing that these guys want is selling us cards with crapy closed-source drivers, written the chepest way they could have done that so they are ashamed to show their code to public. Patented technology? WOW, YOU GUYS (ATI) WANTED SOFTWARE PATENTS, DON”T YOU?! So don’t complain about this. No excuse for me.
DRM in graphics card driver? No way, I’m not buying from them any more. It is too much for me, this world gone mad.
Just plain stoopid M-Fs, really.
Firstly, to those happy with nVidia binary blobs: loading those blobs violates the GPL (yeah, even the v2 Linus is insisting on). nVidia gets away with this because it forces the user to violate the GPL at module load time (hence linking). Now, you may be ok with this from a practical point of view, but it is still Bad, sorry.
Ok, fanboys on all sides: just contemplate, for a minute, just what kind of armageddon awaits ATI if they were to open source their drivers? Go ahead, think about it, I’ll wait. Ok, let’s see if our answers match: none. ATI and nVidia continuously play this assanine game (and the other graphics chip manufacturers follow suit to look tough) of trying to hide everything about their chips, instruction sets, drivers, etc. Wake up people, it’s all pointless excessive corporate protectionism.
If they have real innovations then they already patented them. Look up the patents and everything they are doing is spelled out in black-and-white, so why do they need to hide anything? All the graphics chip manufacturers can see each others patents, so that’s not an issue; what then? Oh, making what? clones? So, you could drop in a card that can dupe nVidia software into working with someone else chipset? If there was money in doing this it would have been done. Certain subsets of the nVidia family have already been completely reverse-engineered. Also, it would be a pretty brave company to think that they could do this and completely skirt nVidia or ATI’s patent portfolio. So this scenario is stillborn.
Oh, the “secret sauce” factor? Like biasing benchmarks and game performance? I guess those who play the games that get special treatment by the GC manufacturers like that, but there really isn’t any reason why open sourcing their drivers should stop this, in fact I would expect that such a practice could be accelerated as more hackers take time to tweak the drivers to get N more FPS in their favorite games. So these patches become public knowledge. Hell, wouldn’t nVidia want to boast that it has a more active hacker community than ATI?
No, all this, in the end, is more of the year-after-year stupidity that these companies put out regarding their driver and chipsets. And it has to stop. Open source the drivers.
Firstly, to those happy with nVidia binary blobs: loading those blobs violates the GPL (yeah, even the v2 Linus is insisting on). nVidia gets away with this because it forces the user to violate the GPL at module load time (hence linking). Now, you may be ok with this from a practical point of view, but it is still Bad, sorry.
That’s not what Linus says. He says the GPL applies to derivative works, and currently he doesn’t believe the nVidia driver to be a derivative work. He also has said that derivative works are currently a legal gray area.
Also, if you weren’t aware, at least in the US, what a copyright holder has to say can have a great effect on the interpretation of a License. This means that Linus’ words can carry legal weight in court. This also means that it is highly unlikely that nVidia drivers are “in violation” or whatever else has been said.
binarycrusader is right, there is no violation.
Just because it says it taints the kernel, is not the same thing.
Although I run only Linux, frankly all I care is whether my graphic drivers work and work well. So far for me, nvidia drivers have been almost flawless on several machines. Whether OSS or not, I don’t really care.
All these people saying how they are jumping ship to Intel before a real driver showing the performance of these other companies is available. Talk about putting the cart before the horse.
Anyone selling their old high end video cards which they now consider crap, post your ebay auction number, I’m interested. 😆
>Firstly, to those happy with nVidia binary blobs: loading those blobs violates the GPL (yeah, even the v2 Linus is insisting on). nVidia gets away with this because it forces the user to violate the GPL at module load time (hence linking). Now, you may be ok with this from a practical point of view, but it is still Bad, sorry.<
This violating the GPL stuff is really getting old, listening to people talk about it that have never read the liscense. Proprietary nvidia and ati drivers do not violate the gpl. They only violate the GPL when the distro includes them in the default install. If I install them myself after the os is installed there is no violation.
Even the use of proprietary closed nvidia and ati drivers in a default install is not a black and white violation of the GPL, it is very much a grey area and has never been tested in court.
Getting old or not, Linus saying so or not is really not the point. Linux is released under GPL, not LGPL. You can’t link to it (statically or otherwise) without having to abide by the terms of the GPL. Linus put the “license check” (to detect “tainting”) as a hack to enable binary-blob drivers to be identifiable. This is to potentially make it possible for Linus to categorize a binary driver as a “gray-area” driver:
http://kerneltrap.org/node/1735
Or
http://www.kernel-traffic.org/kernel-traffic/kt20031226_246.html#1
Also, please note in this thread that there are a lot points made about what is an is not GPL fodder:
a) including the kernel header, do so and you are under the terms of GPL, so assume, WITHOUT PROOF that nVidia and ATI has not done this.
b) Linus himself makes an exception for the _gray_ area modules: they are modules never really “written” for Linux. The specific example he gives is the Andrew File System, or AFS. In the case of AFS it was written BEFORE Linux existed (or at least before the Module interface existed), it is quite old. So the Linux version is a port from another Unix environment. The nVidia and ATI binary drivers are NOT ports from something resembling a Unix-like environment, but a fresh implementation (clearly lifting code from the Windows drivers where they can, of course) specifically written for Linux. Seems pretty _ungray_.
c) Also, Linus goes on to say that binary driver modules ARE derived works!
d) Linus doesn’t pursue this further because of 1) laziness (read for yourself) and 2) remember Linus’ own self description: “I’m just an engineer” – he *claims* to not be interested in the political issues surrounding the GPL (but seems quite vocal, and off the mark, about the GPLv3)
Linus is far more liberal with what he will/will not permit to violate the GPL, but he is probably _wrong_ about this _gray_ area, it would not be the first time Linus has been wrong you know. Here’s a succinct post from Alan Cox:
http://lkml.org/lkml/2005/11/3/55
Binary drivers are ALMOST CERTAINLY in violation of the GPL. If you wish to use them for “practical reasons”, you *can* since nobody will stop you, but you are doing Bad (viz. the GPL). And on top of that there really aren’t any *good* reasons why these chipset manufactures can not or should not open source their graphics drivers.
Proprietary, patented optimizations
If the damn thing is patented it’s supposed to be common knowledge. The whole purpose of the patentsystem is to publish inventions to the world!
It seems like all a company has to do these is release a few statements about their goodwill towards the comunity or throw them a bone or two, and suddenly they are the new hero. Don’t get me wrong, I think Intel’s new move to open their drivers is worth comendation, but does that really mean that we should all turn to denouncing AMD & ATI and making pledges to never buy their hardware in the future? One should always keep history in perspective before making rash statements. It wasn’t so long ago (if i remember correctly) that Intel was in the doghouse because the centrino platform flat didn’t work on linux, just do a simple search and you will find petitions, rantings by various individuals and proposals to boycott intel and the centrino platform. I highly doubt that Intel is actualy making these moves because of goodwill towards opensource or the comunity. They are simply protecting marketshare.
If I personaly was putting together a new machine right now i’d buy an intel processor simply because they are the best speed/price ratio. Amd may something better in a year or so, who knows?
My point is that we should always remember the past when dealing with the present and beware of taking sides too quickly because it’s rare that a company cares about more than marketshare and bottom line and once they have both secure they rarely care about us anymore.
Upside is I won’t be ditching my sli setup for crossfire any time soon and I probably would have.
Big mistake, but Linux’s one, not ATI’s. OS world should reconsider it’s driver strategy. Making driver internal and declaring driver to be derived from Linux does not allow to have normal drivers. Without drivers OSs are just students’ toys, as it used to be for a few decades. Everyone seems to be happy with this position.
API is a point of integration for both drivers and applications. Using documented API should not mean to be derived.
Binary drivers aren’t normal drivers. It works only with stable ABI systems, such as windows (but not beetween all versions).
Open source programs, such as XOrg, take ABI unstable, and change it often, but keeps source API stable. When (as in new XOrg 7.1) ABI changes, every binary drivers that are supposed to work with it don’t. So Manufacturer must compile it’s drivers for new API, and keep old to support older systems.
Stable binary API isn’t really good, becouse it keeps from improving program, and often ends in making ugly hacks to make things working (padding structures with reserver fields, making one variable meaning something different, such things). If only manufacturers take documentation to their products, or open their drivers, it will be no problem, people will recompile drivers when it ABI changes, and distribute it within days.
But when there are no documentation nor open drivers, it is a big problem, and will be bigger, as open source programs will change faster (as XOrg now).
So: documented API from OS side is must. But documented API for hardware is must, too. Without them, manufacturer must support all changing systems that his hardware work with. If there is no monoculture (like in windows world), he can’t.
Big mistake, but Linux’s one, not ATI’s. OS world should reconsider it’s driver strategy. Making driver internal and declaring driver to be derived from Linux does not allow to have normal drivers. Without drivers OSs are just students’ toys, as it used to be for a few decades. Everyone seems to be happy with this position.
It’s virtually impossible for Linux to reconsider its driver strategy (or change its licence to eg. GPLv3): there are over 500 contributors, some of which have died.
And no, many people are unhappy with this position, including me. It’s highly unrealistic to expect everyone to write open-source Linux drivers for every device.
The best way out is probably to use user-space drivers. User-space drivers can be closed source, and the kernel’s user-space interfaces are stable. libusb already lets you do full USB access this way, and there are several ways to get user-space PCI access (bochs’s pcidev kernel module and gelato project at http://lwn.net/Articles/66829/ and http://www.gelato.unsw.edu.au/IA64wiki/UserLevelDrivers). Apparently PCI drivers in user-space are fast and have many other benefits.
The problem is, nVidia’s drivers (and probably ATI’s) dig quite deeply into the kernel, accesses the BIOS and system-management mode, and other horrors.
AFAIK all patents are publically available. Patents are not way to hide something.
Patents can and are often written in obfuscated language. I don’t doubt that in many cases it will be simpler to decompile the NVidia or ATI drivers than to try to read the patents covering them
ATI has stated in the past, and in their actions, that they will never open source their pretty dreadful drivers. On the other hand, it’s been known off the record that AMD is keen to find a way of doing just that.
This just sounds like the normal highly confused message you get when one company takes over another. Don’t read too much into it.
Nvidia have a definite history of crippling various chipset features within their drivers, as opposed to disabling the said features on the actual chips. There used to be a lot of people swapping and tweaking their video BIOS, not just to overclock the default speeds but to introduce new features; presumably the drivers ultimately “enabled” those features.
A new one cropped up just the other day, Nvidia released Forceware 91.45. This driver has introduced Transparency AA into the 6600/6800 geforce chips at very little cost in FPS terms.
BTW, I’m not entirely sure if this feature relies on compatibility tweaks from TweaksRUs.com and the likes, regardless, the point is that such a major feature has always been available in those chipsets – only the security by obscurity used in Nvidia’s driver was holding it back.
I imagine ATI are up to similar tricks. If so, it’s understandable why they keep their drivers closed.
Its nice that some of you can make the choice to move to ‘open’ friendly Intel. For those who work in the visualisation field the choice is about a platform to support the card(s). With the mostly positive reaction of the FreeBSD developers to nVidia’s suggestions, and the near parity of Solaris/Windows driver features – this makes linux a poor 4th choice for PowerWalls, CAVEs, flight simulators, Domes – just as it was beginning to get some traction….
[edit] I forgot Mac OS X – funny, the OS is use most of…so that makes Linux it a poor 5th.
Edited 2006-08-13 14:06
And most of you thought AMD were going to Open Source the drivers! Ha! Y’all should know better than that. They’re no better than nVidia (but at least nVidia still provides decent closed source drivers), so stop adoring AMD as some Open Source God, because they are not!
In addition, multimedia elements such as content protection must not, by their very nature, be allowed to go open source.
They can keep that for themselves for sure. That´s one feature I wouldn´t miss in open drivers.
The answer is simple. No ati in any of my computers ever more. Not that I’ve got any right now
I have one, but it goes with smoke (literally, but it wasn’t fault of graphic card).
And next laptop I’ll buy will have Intel powered graphics, i’m sure. I want to have new XOrg when it is complete, not when manufacturer will recompile his drivers for me.
If AMD does procede in buying out ATI, which it seems like its going to happen at this point, what ATI says doesn’t matter. When AMD owns ATI, if ATI says something and AMD decides differently, AMD is the boss. All this means is that until the merger is over, work will not begin on the open source drivers.
This is one of many reasons why I do not see Linux taking the desktop and even losing ground in the next five years in the server space as well.
is that neither nVidia or ATI obviously even understand the problem domain.
No-one expects them to open-source their own drivers, all we, the consumers, the ones who PAY for their products, want is that they release API documentation so that *others* can write OSS drivers for OSS OS’s.
“even losing ground in the next five years in the server space as well.”
Because 3D graphics and DRM is really important in the server space….
Booo. bad amd. they have a chance to radically improve the quality of their drivers in linux. and they aren’t going to do it. Meanwhile intel is stepping up to the plate. Listen ati/amd, release open source drivers for your videocards. I want the ati driver situation to improve.
Ati’s linux drivers completely suck. they crash. they are worthless. my purchase was worthless. i wont be buying any more ati cards. ill keep testing it. and testing.
im getting close to a new processor purchase in a year or so, and if there aren’t any add in cards with quality open source drivers, then i am definately going with intel. As their integrated video chips have open source drivers, and will improve faster than ati’s drivers will.
I was hoping that AMD would show some guts and open-source their drivers, or at least certain drivers… or provide enough information so that talented programmers could write GPL drivers.
I was also hoping that this would put pressure on nvidia to do a similar thing.
Last but not least, with the AMD takeover, finally it seemed that I would get more *choice* when considering to buy a 3D card.
False hope, or so it seems.
I’m still going to buy nVidia cards for my systems UNLESS AMD/ATI proves they can turn things around (i.e. create *working* drivers for Linux for once).
All of this ‘patent’ stuff leaves a bad taste in my mouth… in the coming years there will probably be a push for this kind of hardware to be DRM’ed… so that you won’t even be able to play games if the DRM chip decides your copy is ‘invalid’ (which might just mean it’s scratched, or is a backup copy… I sure wish I took a copy of my Broken Sword III-disks… the first CD now is… broken).
Here’s hoping the ‘Open Source’ 3D card project takes off.
In the meantime I’ll still need a 3D card to play my favorite games… but I’ll be watching the “Intellectual Property vs. Freedom” battle like a hawk.
After all, in the end, freedom is more important than games. And a card like the ‘Open Source’ 3D card will at least be enough to enable things like XGL to work properly.
Thank you AMD for your stupid decision, from now I will not buy any your components!
“Thank you AMD for your stupid decision, from now I will not buy any your components!”
I am not an AMD fanboy, but I doubt they really care about some nerd running Linux that will not buy their products.
That might depend on how much influence that “nerd running Linux” has on the buying policies of the corporation they work for. 🙂