LCA: The state of the Nouveau project
The Nouveau project is an effort to develop a set of free 3D drivers for NVidia chipsets. NVidia has long annoyed the free software community with its refusal to release free drivers or programming information for its video chipsets. The Nouveau folks have had enough of that, and they are doing something about it. Dave Airlie used his slot at linux.conf.au to talk about the project and its current status.
Nouveau got its start in February 2005, though serious work did not begin until June of that year. The project was announced at FOSDEM 2006, at which point others started to help. There are currently about six developers doing serious work on Nouveau.
The project is relying on reverse engineering for the information needed to write free drivers. To that end, the developers have put together a set of tools. At the top of the list is renouveau, which is designed to reveal the commands sent to the card in response to specific operations. Using the existing binary drivers, renouveau sets up a context, then scans the process's mappings until it finds the command FIFO. It then requests an operation and sees how the FIFO changes. With enough operations, a pretty good idea of how the adapter is programmed to specific ends can be had. This was not a trivial tool, and the better part of a year was put into its development.
Renouveau is useful for examining the FIFO, but it doesn't help with reads and writes to I/O registers. For that, there's another set of tools, starting with valgrind-mmt - a version of valgrind designed to trap I/O memory operations. Libsegfault is a modified version of mmap() which doesn't actually do the mappings as the caller would like; it traps the subsequent segmentation faults and dumps out the operations. There is another tool, called kmmio, which performs a similar task for register operations done in kernel space. Finally, the project uses a BIOS tracer which runs BIOS code in x86emu and traps I/O register accesses.
All of the information obtained from these tools is supplemented with hints from the old, free nv driver. There is also, says Dave, information "which shouldn't be there" to be found on some Russian web sites.
Where has all of this information led the project? Basic tasks, like the allocation of instance RAM and FIFO initialization, are working. Hardware context switching works - on little-endian machines. There is 2D support derived from the nv driver; it offers basic EXA and RandR 1.2 support. On the 3D front, the Mesa TCL (transform, clipping, and lighting) driver mostly works. Textures and objects do not, however. It is possible to run glxgears on nv4x chips. It has taken some time to get to this point, but Dave thinks that things will start to move a lot faster from here.
The next milestone would be to run Quake 3. That is, says Dave, an obligatory step on the roadmap. Getting there will involve texture support, a better memory manager, and better locking in the kernel DRM code. The developers (Dave in particular) are aiming for RandR 1.2 multi-head support. Once all of this is in place, the nouveau driver will have reached a reasonably capable state.
There are a lot of people asking when this will be; Dave says that the project's IRC channel is often overwhelmed by spectators looking for news. There is no wish to push the code out ahead of its time; among other things, that would nail down the API between the kernel and the X server, making things harder to change. The current hope is to have some sort alpha release toward the end of 2007.
For people wanting to help, Dave had a simple message: they need developers. There's not much for people who can't work on driver code to do at this point. Graphics drivers, he says, are not as hard as people think. Finally, he addressed the issue of the $10K pledge for the project. It rather took the developers by surprise; they had not endorsed this drive, and had held some doubts as to whether it would be successful. How the pledge money will be handled is still being worked out; it looks like it will mostly be used for hardware purchases.
Lack of support for 3D video adapters has stalled the community for years;
there has been a long wait in the hope that the vendors would come to their
senses. That wait is just about over. The Nouveau project (along with
various others) shows that we have the resources to figure out how our
hardware works, even in the face of complex devices and uncooperative
vendors. It would be better if we did not have to take things into our own
hands this way, but it is nice to see how well we can do it when the need
arises.
Index entries for this article | |
---|---|
Kernel | Device drivers/Nouveau |
Kernel | Nouveau |
Kernel | Reverse engineering |
(Log in to post comments)
ATI drivers welcome developers too
Posted Jan 17, 2007 8:40 UTC (Wed) by tajyrink (subscriber, #2750) [Link]
Go Nouveau. Btw, open source ATI drivers would welcome developers, too. Even though the basic and even quite advanced functionality is there, there are some rather serious problems with some specific cards - I'd guess too many of the current developers have working cards :)
There also used to be autodetection problems, and to some extend still is (MonitorLayout needs to be specified for some). But if you happen to be an X300 - X800 owner who found out that Driver "ati" , the wrapper for other ati drivers, does not work at all for you ("radeon" does, but you have to specify it manually), or that any 3D, eg. glxgears, gives you an error message please try out one of the newest daily Ubuntu releases at http://cdimage.ubuntu.com/daily-live/ - no, Herd2 is not new enough except for the latter problem. Those two bugs affecting some percentage of 9500 - X800 owners were fixed recently in mesa/drm (and kernel 2.6.20) and the ati DDX driver.
Paying nVidia
Posted Jan 17, 2007 10:10 UTC (Wed) by dion (guest, #2764) [Link]
Has anyone ever asked nVidia how much money they want to release their own parts of the driver code?
Their excuse to not free the code has always been that there is some third party code that they cannot release, if that's the only reason then they ought to be able to take that code out and release their own code.
I can fully understand nVidia not wanting to document the hardware, because I'm pretty sure that documentation doesn't exist other than as lines of code in their driver, so any documentation written would be wrong or at least obsolete.
The Pledge(tm) quite quickly found 1200 people willing to spend $10 eac without any sort of idea of what the money would be used for, so I don't think it would be too unlikely to be able to find a much larger sum (like a million) if there was an agreement in place with nVidia that they would release their driver code.
Paying nVidia
Posted Jan 17, 2007 11:24 UTC (Wed) by alvherre (subscriber, #18730) [Link]
An nVidia executive said some time ago that the reason they didn't open the code was that open source developers are not smart enough for something as complex as a 3D graphics driver, which I found a bit insulting. So if a million dollars was to be gathered, I'd rather give it to Dave Airlie and the Nouveau folks than nVidia itself — I already gave nVidia some money for its hardware and it refused to give me the code (not that I actually asked, but ...); so let's support the open source supposedly-not-smart-enough guys this time and see how dumb they turn out to be (or not).
Paying nVidia
Posted Jan 17, 2007 12:33 UTC (Wed) by mattmelton (guest, #34842) [Link]
...in nVidia's defence, I can see why a marketing person would say that. My department has had a few applicants to nVidia before. The telephone interview process is phenomenal - you are asked to cite and cross reference everything from advance vector theory to the current research on pixel and vertex shaders you see coming out of eastern physics labs. As an MSc student (computer games systems) studying what I can only call the tip of the iceberg, linux kernel and device driver writing is quite far off the beaten track.
In my opinion, the hierarchy of programmers goes something like this...
nvidia programmers -> pixel/vertex shader programmers -> kernel / assembly programmers -> image processing / neural network / database programmers -> C programmers -> ruby/perl programmers -> scripters -> web programmers
(over generalized, please don't flame)
This line of thought leaves quite a gap however... what about people who have been in the 3d hardware industry, or PhD students who have spent fantastic amounts of time and new theories? I think nVidia, are, in a way right, but again so myopic - it's painful.
The last thing nVidia wants is people to gain that extra competitive step into the 3d hardware industry that it had over 3DFX in the late 1990s... because we all know how that went.
Paying nVidia
Posted Jan 17, 2007 12:47 UTC (Wed) by gravious (guest, #7662) [Link]
what are you smoking dude?
hierarchy of programmers?
everybody knows it goes like this
systems architect -> senior team lead -> senior proggy -> junior proggy (or some such crap hierarchical management-speak employment-agency rubbish)
_or_
(torvalds, kernighan, bellard... other guru status proggies) -> everyone else
_or_
if you put your brain on you'll realise that there are smart programmers in every field - i guess if you think that the 'web programmers' that wrote gmail are the bottom of the pile why don't you do one better, eh? any piece of software once it becomes sufficiently complex requires smarts. that said, i believe that great c/asm programmers are probably a step ahead of the rest of us by necessity.
_or_
have i completely missed you and you were being funny? i hope so...
Paying nVidia
Posted Jan 17, 2007 13:37 UTC (Wed) by ekj (guest, #1524) [Link]
There's great, good, average, poor and sucky programmers in every field.The separation ain't along lines of what they work on, nor along which language they use to develop. (ok, so not that many great programmers prefer qBasic)
There are people doing abolutely brilliant work in Javascript, and there are people pounding out stupid, buggy, unreadable, unmaintainable code in Assembler.
Paying nVidia
Posted Jan 17, 2007 15:21 UTC (Wed) by rknop (guest, #66) [Link]
If you're going to talk about a hierarchy of programmers, or any other sort of virtual schlong-measuring contest, it's obligatory to reference this:
http://www.brunching.com/images/geekchart.pdf
I'm sure that real time device drivers consider themselves superior to Nvidia card driver programmers consider themselves superiour to kernel hackers consider themselves superior to hard-core scientific algorithm developers consider themselves superior to complier developers consider themselves superior to lisp purists consider themselves superior to Richard Stallman consider themselves superior to LAMP programmers consider themselves superior to reverse engineers consider themselves superior to BSA lawyers consider themselves superior to everybody else.
I mean, making a hierarchy like that, about who's smarter than whom, is just asking for ridicule.
-Rob
Paying nVidia
Posted Jan 17, 2007 16:10 UTC (Wed) by marduk (subscriber, #3831) [Link]
Even if your statements are true, and it has been dutifully debated so I won't even bother to do that here, all nVidia has to do is release the technical details of how to *program* the card, not how the card was written. I'm sure there are more than enough developers in the OSS community that know how to read technical specifications. You don't, for example, have to know how Intel designs and builds Xeons in order to program one now do you?
Paying nVidia
Posted Jan 18, 2007 10:07 UTC (Thu) by gyles (guest, #1600) [Link]
In the case of the Xeon all the instructions fed to it are from the known x86 instruction set (plus SSE etc.). The Xeon then translates these to micro-ops and executes them. Thus knowing the instruction set reveals nothing about the internals of the chip.
The nVidia chip is designed to optimise the whole chain, and does not have to adhere to a published interface at the hardware level. To maintain performance there cannot be the same sort of translations as the Xeon performs. Thus knowing the language fed to the hardware is more revealing about the hardware itself.
Paying nVidia
Posted Jan 17, 2007 18:27 UTC (Wed) by iabervon (subscriber, #722) [Link]
It sounds to me like nVidia programmers are good at things relevant to graphics, but that doesn't mean they can write code that doesn't perform badly or incorrectly on a 1024-processor, 256-graphics-card IA-64 machine. Most likely, they can do a better job of translating OpenGL calls into card primitives. But there's no reason this needs to be done in the kernel. I'd bet that kernel developers are better than nVidia programmers at arranging bus transactions.
There's a certain level of general coding skill that you need to have to do things of a certain complexity. And then there are an assortment of specific skills you need to do particular tasks. I'd guess that the open source community has the best general coding skills, and is okay at a lot of specific skills. That should be sufficient to let nVidia focus on writing code which is in userspace and part of a MIT/X-licensed program. And, also, allow open source programmers who happen to be good at graphics stuff to take a crack at driving some common hardware.
Paying nVidia
Posted Jan 18, 2007 3:03 UTC (Thu) by mattmelton (guest, #34842) [Link]
ugh.
I asked for no flames. The LWN community has let me down. A previous LWN edition with a letter to the editor points to a downturn in LWN comment quality recently. I concur. (this IS a flame)
I was not trolling or flaming intentionally. I gave my honest opinion as someone in the gaming industry. If anyone has studied physics or maths at a masters level, and then engaged in 3d software development, I welcome a rebuttal.
Interestingly, there's an issue I've come to recognise between business owners and doctors/surgeons/consultants since medicine is in my family.. It's quite a simple issue: business men/doctors/surgeons/consultants don't like to be criticised. I feel I've criticised people who *think* they can successfully perform at the same level as an nVidia employee (or, frightened those who aspire). I apologise if I belittled anyone - specially the abstract XML gurus out there (etc) - by my labelling. I directed nothing at no one. I merely stated what I saw in the employment market place.
There are people who can perform at a distinctive level no matter their programming 'label'. While I did not acknowledge this, I regret not mentioning it.
Paying nVidia
Posted Jan 18, 2007 4:25 UTC (Thu) by jwb (guest, #15467) [Link]
You posted flame bait and were flamed. Why are you surprised? You can't undo a flame (nor flame bait) with a parenthetical clause. For the same reasons, you can't head off an argument by ending your argumentative post with "I don't want to argue about this."
Paying nVidia
Posted Jan 18, 2007 4:54 UTC (Thu) by modernjazz (guest, #4185) [Link]
I have a PhD in physics and I write 3d software (or even n-d software,take that!). Expertise is partly a matter of having some ability, but
perhaps more a matter of devoting the time to master a subject---and once
you've done so, it usually seems quite easy and straightforward. (As a
mathematician would say, "trivial!") But I don't know a damn thing about
writing even bare-bones HTML, so I guess I'm both a genius and a dunce in
your hierarchy.
My personal view is that hierarchies like the one you proposed tend to be
more confusing than helpful: there is a range of talent in every field,
and the standard deviation within a community often exceeds the
difference in means between communities. Yes, some fields are
unapproachable without a certain baseline ability, and that tends to
guarantee that most of the people working in them are at least decent.
But one also finds that absolutely terrific people pop up in the
strangest of places. Unfortunately, it's harder to remember this if you
also maintain a mental hierarchy.
I guess it's part of the power of open source: an acknowledgement that,
sometimes, the person with the best idea may be "out there" and without
any obvious credentials, other than the code s/he writes.
Paying nVidia
Posted Jan 18, 2007 13:57 UTC (Thu) by lysse (guest, #3190) [Link]
I don't concur about the comment quality - but if I did, I'd be citing this comment as an example. First of all, you post something inflammatory and then expect it not to be challenged; secondly, when it is challenged you complain that because you asked to get away with it, you should have been let off; thirdly, you state that you'll only listen to different opinions from a class of people whose qualifications you (in your magnanimity) recognise - which is just plain intellectual snobbery. You say that professionals don't like to be criticised - but you've failed to demonstrate any ability to respond to criticism yourself.
At best, that's an extremely immature way to conduct yourself - and I know, having been guilty of it when I was a kid - and at worst it's trollish behaviour. Please don't do it any more.
Paying nVidia
Posted Jan 21, 2007 22:39 UTC (Sun) by mattmelton (guest, #34842) [Link]
I disagree with what you've said. And this is why.
Your type of comment and attitude is something I am not familiar with at LWN. With the exception of a few inter-developer debates spreading over from LKML, if there was a non-hostile post which did not troll or flame, people posted positive or non-negative criticism at the very least.
You have not. Instead you take initiative to flame me and my response and my applolgy. You are NOT constructive. You are very negative.
Unlike yours (yes, it's personal as you took the time to flame me), my initial post was clear and informative.
* I study where someone has applied for a post at nVidia.
* I have been told in no uncertain words what to expect and what level I am supposed to be at before apply to a company like nVidia
* I have my foot at the bottom rung of a very tall ladder.
* And I gave my opinion on where I see programmers standing.
If you have an issue with what I said, then by all means please reply with something along the lines of, "I disagree with your hierarchy - i don't believe its right to..." etc
Getting passed the flare of what people write is hard nowadays. The problem is, Lysse, that I believe you're really not annoyed about the nVidia issue on which I commented on; you're not annoyed at the perspective I chose to talk from; and you're not annoyed by the way I apologised to the people I mislabelled in my rather sweeping comment. I honestly think you're annoyed about the sweeping hierarchy statement and that I sided with nVidia. I believe you disliked my comment so much - to the degree that it personally insulted you - and that any attempt by myself to correct it has only caused you to insult me directly.
Could this be a deep rooted dislike or jealousy of "plain intellectual snobbery" in you? Maybe it's hard to refrain from constructively building a comment when "plain intellectual snobbery" becomes a front, and this is what rushed you to insult me, rather than the issue I was commenting on.
I could go on about how "plain intellectual snobbery" actually adds to ones experience and ability; and that qualifications (plain intellectual snobbery) are nothing to scoff at. But again, I'm sure that "plain intellectual snobbery" might get in the way of what I'm trying to say, and that you'd attack me again.
I've just re-read what I've said, and I still don't see it as anything more than a neutral opinion - one I only ever so lightly support I might add.
If the rolls were reversed, and you had posted my comment (I'm not proclaiming you ever would here), I would have either
* not felt the need to reply
* replied constructively, adding new information into the debate
* replied without trolling or flaming, but offering an opinion that perhaps I was a little too generic.
You did none of these. You are flaming, and you are trolling. You are inciting my to reply (I'm a fool for doing so) and adding to what I've grown to dislike in LWN comments.
Paying nVidia
Posted Jan 22, 2007 5:33 UTC (Mon) by bronson (subscriber, #4806) [Link]
mattmelon, please stop. You seem to be posting the very type of comment you claim to not like to see. You complain bitterly about the level of discourse on LWN and call someone a troll in that very same message.
If you're joking, sir, I salute you. The irony is most impressive.
Paying nVidia
Posted Jan 25, 2007 11:35 UTC (Thu) by lysse (guest, #3190) [Link]
Likewise, when I step in a pile of, er, second-user dog food in my local park, I curse the dog because I believe the creature personally insulted me.
*sigh* I hate to say it, but I believe the time has come for an "ignore user" feature.
Paying nVidia
Posted Jan 18, 2007 20:01 UTC (Thu) by dw (guest, #12017) [Link]
Both your comments (in addition to this one) offer no meaningful or on-topic content at all. Furthermore, your original comment is rude and insulting, "random type X is better than Y DONT CONTRADICT ME!". Life just doesn't work that way.
Paying nVidia
Posted Jan 17, 2007 18:23 UTC (Wed) by dion (guest, #2764) [Link]
I wasn't talking about getting the nVidia sources in stead of supporting the Nuveau project, I imagine that getting most of the sources from nVidias driver would be e great help for the free driver project:)
I do hope that Nuveau can get there in the end just by reverse engineering, but it's clearly a huge job that can use every bit of help.
Paying nVidia
Posted Jan 18, 2007 5:46 UTC (Thu) by BrucePerens (guest, #2510) [Link]
All of the computer graphics effects houses are using Linux. Pixar, Lucasfilm/ILM, Dreamworks, etc. They live in fear that nVidia will walk away from the market, leaving them with nothing to run but that new Intel chipset. They have people as smart as the ones at nVidia, and ones who know how to code for that sort of hardware.I wrote microcode for the Pixar image computer. Mostly 2D point processes, not 3D rendering. But I'm hardly the only one who has worked in this.
Bruce
Paying nVidia
Posted Jan 25, 2007 18:37 UTC (Thu) by anton (subscriber, #25547) [Link]
They live in fear that nVidia will walk away from the market, leaving them with nothing to run but that new Intel chipset.Has ATI/AMD vanished from this universe? I recently bought an X850XT which does 3D nicely with free drivers (also tried an X800GTO and an X550 successfully). ATI are pretty bad nowadays, but at least they gave us the info for the R200, and AFAIK this then helped with the R300 and R400 (don't buy R500 (X1xxx), though).
Also, I would expect the film people to do all their rendering in software, no need for 3D hardware.
Paying nVidia
Posted Aug 16, 2007 18:20 UTC (Thu) by MenTaLguY (guest, #21879) [Link]
The authoring tools generally take advantage of hardware acceleration to work with the assets in realtime; the final software-only rendering is non-realtime. However, with the increased programmability of GPUs, there's increasing potential for hardware acceleration there too.
Paying nVidia
Posted Jan 17, 2007 18:35 UTC (Wed) by jospoortvliet (guest, #33164) [Link]
nvidia most likely is not willing to release the sourcecode because thereare many patented algorithms in there, and they are afraid to get sued. of
course they could spend a lot of time and money on checking and rewriting
code, but i guess the FOSS world is just not persuasive enough...
Paying nVidia
Posted Jan 17, 2007 18:39 UTC (Wed) by dion (guest, #2764) [Link]
Well, that's where the Mega Dollar comes in:)
Paying nVidia
Posted Jan 18, 2007 4:43 UTC (Thu) by chromatic (guest, #26207) [Link]
I've paid NVidia enough already for a working driver, and the last time I bought one of their video cards was in 1998. They go out of their way to provide a solution that is completely unacceptable to me; I can't see any reason to give them any more of my money.
Paying nVidia
Posted Jan 27, 2007 7:14 UTC (Sat) by beoba (guest, #16942) [Link]
What message would this send to other hardware companies? "Hold out long enough and we'll give you free money?"
LCA: The state of the Nouveau project
Posted Jan 17, 2007 11:15 UTC (Wed) by johoho (subscriber, #2773) [Link]
this sounds like very interesting stuff.
However, I wonder if there's a risk that at some point in time NVidia - or a another copyright holder for that matter - might sue the project for using non-reverseengineered information. After all, it just *might* be possible that the information from the russian websites was obtained illegally.
let's hope the project will be able to proof that they wrote it by using only "real" reverse-engineering techniques.
Wiktor
LCA: The state of the Nouveau project
Posted Jan 17, 2007 12:25 UTC (Wed) by ibukanov (subscriber, #3942) [Link]
In ancient Rome there was a law about a buyer who bought a stolen thing. The law stated that the buyer is innocent and can keep the thing if during the purchase he made reasonable efforts to check that the thing was obtained by legal means. Effectively the original owner can only sue the seller, not its customers.
This law survived pretty much through out the history. So if Nouveau team was made enough efforts that the information they use was obtained legally and does not include in the sources a code of unknown origin to avoid copyright violation allegations, then NVidea can not sue them and should go for the Russian web site operators instead.
This is, of cause, a theory, and courts especially in US may disagree, but even in SCO vs IBM case SCO tries to bring copyright charges against IBM, not the contract breach, so they can sue later the people who used IBM's code.
LCA: The state of the Nouveau project
Posted Jan 18, 2007 2:22 UTC (Thu) by DonDiego (guest, #24141) [Link]
Nowadays the law is different: You cannot aquire property of a stolen good. It remains the property of the original owner. You are free to sue the person that sold it to you for damages, though. This is valid at least for German law.
LCA: The state of the Nouveau project
Posted Jan 18, 2007 14:48 UTC (Thu) by ibukanov (subscriber, #3942) [Link]
Well, that rule AFAIK is applicable to contracts between business parties in most western countries. I.e. if one party while providing a service to its customers breaches the contract, then the other party can only sue the first party, not its customers.
But even with real stolen goods it would be interesting to know what police in Germany would do in the following case. Suppose a shop owner picked 20 identical TV sets, not 10 as stated in the contract, from distributor's warehouse. Then he sold all 20 TVs with normal marked prices before the theft was discovered. Will the law enforcement confiscate 10 TVs from people who bought? If so how they decide which ones?
Ownership of stolen property
Posted Jan 19, 2007 17:12 UTC (Fri) by giraffedata (guest, #1954) [Link]
The law of who owns something after it gets stolen and eventually passed to an innocent buyer is just a matter of policy. Neither rule is more moral than the other; it's just a question of which rule reduces theft most efficiently. So it's likely to vary a lot between jurisdictions and over time.
But throughout modern times, English/American law has been that the theft victim remains the owner. There are various exceptions. One is where there is a recorded title, such as with a car or land.
The TV picking case is complicated and the answer depends on the exact facts of the case. It could be that there was no theft -- the TV shop took title to all 20 TVs and thus was able to pass it on to the consumers and the distributor has only the contract dispute. This would be similar to the case where the TV shop takes only the 10 TVs agreed upon, but then doesn't pay for them. Not a theft.
If the facts are such that the TVs were stolen, then the consumers do have to return them to the distributor. The police wouldn't confiscate them, because 1) it's not their job and 2) civil cases like this are normally resolved with money.
LCA: The state of the Nouveau project
Posted Jan 17, 2007 13:17 UTC (Wed) by sanjoy (guest, #5026) [Link]
Copyright covers the expression (e.g. the code), not the ideas (e.g. not the register information or algorithms). So even if trade secrets were obtained in dubious ways (who might themselves face civil or criminal penalties), others can use this information.
If the information in question were source code, then others have to be very careful to show that they did not look at the code, otherwise they risk copyright infringment suits, which can be ruinous even if you are totally in the clear.
Patents are also another story, but by definition their information is already made public in the patent filing, so it's not a question of how the information was obtained (but whether you can use it).
Usual disclaimers that I'm not a lawyer, so this for sure ain't legal advice.
LCA: The state of the Nouveau project
Posted Jan 17, 2007 14:35 UTC (Wed) by mattdm (subscriber, #18) [Link]
I'm not a lawyer either, but this actually doesn't work with trade secrets, which are a separate area of intellectual property law from copyright. If you know something was only available wrongly, you can't use it.
LCA: The state of the Nouveau project
Posted Jan 17, 2007 15:47 UTC (Wed) by emk (subscriber, #1128) [Link]
If you know something was only available wrongly, you can't use it.
Well, if you're responsible for the wrong-doing, then you're liable. But as far as I know, once the information becomes public, then the trade secret is pretty much moot, at least as far as innocent third-parties are concerned. (I'd check with a lawyer before acting on that assumption, though. And rules are presumably different in different jurisdictions.)
Trade non-secrets
Posted Jan 23, 2007 0:16 UTC (Tue) by Max.Hyre (subscriber, #1054) [Link]
Back in college (a few decades ago), I actually took a course in copyright, patent, trade secret, &c. (And, boy, do they have roughly nothing in common.)The law said if you've been granted access to a trade secret under an NDA, you can be held accountable (like, sued), if you disclose said secret. The law, however, also recognized the impossibility of putting the toothpaste back in the tube—once the information is out, it isn't a secret, and can no longer be treated as such.
(So why would anyone use secrecy instead of patent? Because there's no time limit. So long as you can keep it secret, even beyond the twenty years allotted to patents, it's yours alone. That's why Coca-Cola hasn't patented its recipe.)
Nouveau copyright defense
Posted Jan 19, 2007 17:23 UTC (Fri) by giraffedata (guest, #1954) [Link]
Has there ever been a case where original code accidentally looked enough like other code that a court found it was more likely than not that it was copied?
I don't think so. I don't think Nouveau has anything to worry about.
LCA: The state of the Nouveau project
Posted Jan 17, 2007 12:36 UTC (Wed) by pointwood (guest, #2814) [Link]
Best of luck - I hope they get more developers helping out. Good open source drivers for widely used hardware can only be a good thing!
LCA: The state of the Nouveau project
Posted Jan 17, 2007 18:42 UTC (Wed) by ahornby (subscriber, #3366) [Link]
Is there any information to be gleaned from the old Utah-GLX driver sources? They have 3D working on TNT/GeForce era cards I believe.
LCA: The state of the Nouveau project
Posted Jan 17, 2007 21:31 UTC (Wed) by bboissin (subscriber, #29506) [Link]
They are listed in the slides.
So Long Open Graphics
Posted Jan 18, 2007 3:53 UTC (Thu) by bryanr (guest, #25324) [Link]
One caveat to Nouveau is that, if successful, it could thwarta market for completely open, vendor-documented video cards,
such as the one designed by the Open Graphics Project.
So Long Open Graphics
Posted Jan 18, 2007 10:03 UTC (Thu) by gnb (subscriber, #5132) [Link]
On the other hand it spreads the risk. The OGP may deliver a viable openvideo card (I would certainly like to see that happen) but they may fail
regardless of the Nouveau effort. Free, good quality, nvdia drivers might
work against them long term but I would guess they are well over a year
from the point where anyone will be making a choice between buying OG
product and nvidia, and there's no guarantee that they'll get there,
despite a lot of good work.
So Long Open Graphics
Posted Jan 19, 2007 2:21 UTC (Fri) by drag (guest, #31333) [Link]
Well the thing that is nice about the OGP is that they are not depending on desktop users.. They've pretty much assumed that was going to be a non-starter from the get-go.
The first design, the FPGA, should appeal towards hobbiests, very custom hardware chips, and other such things. It's open and programmable with some logic that is freely aviable. People who would like to maybe try CPU designs, or make a very high speed encoder for HDTV compression. Something like that.
From that perspective it is a open programmable PCI card with a huge chip and a massive amount of onboard ram for relatively cheap. This should appeal to some people since it's the only one of it's type and it's logical that there would be some demand for it. It should be very usefull.
Then the second idea is that after the 3d logic has been worked out then you produce a ASIC chip that they would be able to sell it to embedded developers who need or at least would rather have a open design for themselves to use. A open 3d graphics chip that they can freely integrate with a custom embedded devices should be attractive.
Also it's worth keeping in mind that although there are things like Intel graphics and such they are NOT open hardware. The drivers are open, but they are developed behind closed doors. This is certainly better then what AMD or Nvidia does, but it's not ideal.
BTW, I don't know if it's been covered here but the first prototype of OGD1 is now a reality!
http://wiki.duskglow.com/tiki-index.php?page=OGD1&PHP...
They have pictures of it and everything...
256MB of Memory.
2 dual link DVI connectors (for dual high resolution displays)
Xilinx Sparten XC3s4000 FPGA for GPU.
The Lattice XP10 FPGA for PCI controller.
100 pin IDC bus expander
and all sorts of other fun stuff.