Bickering Continues About NVIDIA Using DMA-BUF

[blackiwid]

A mix of both open and closed seems to truly be "pragmatic".

Its maybe in short term pragmatic for the user but not for linux itself... even that its allowed to write such drivers and release it... closed source drivers I mean, does harm really hard the reputation of linux. Why is that so... what happens if a noob tries out linux... he says aih I get asked if I want to install the "better" or faster nvidia drivers at the first start... oh yes of course... tahts the normal way windows does it that way so that must be the normal thing to do everywhere...

What happens next it causes massive problems instabilities... even black screens after boot in some cases etc. What happens next... the user deinstalls linux und whines everywhere how bad linux is. When this Closedsource drivers would be kind of banned not allowed to make any interface changes to the kernel and would not be allowed to be easily installed through such restricted-driver-management tools... the user would boot intel pc perfekt experince... linux is great... they would boot amd hardware... no problems... ok not so fast so amd is to blame. then they would boot nvidia stuff (and that would be especialy true in the time before nouvou drivers got released) they would see a black screen or some vesa drivers or something... they would say... ohh nvidia sucks because they as only company dont release any drivers to linux.

Then nobody who would want to do anything with linux would buy nvidia card... Nvidia is the weekest market player anyway today... as grafic-chip/card seller... they would hopefully get bankrupt... or they would release a driver finaly...

long story short point... the mega-buggy closed source drivers harms the reputation of linux... anybody will blame linux for each bug you get using this because peoples head are streamlined with windows-world-thinking...

[johnc]

Its maybe in short term pragmatic for the user but not for linux itself... even that its allowed to write such drivers and release it... closed source drivers I mean, does harm really hard the reputation of linux. Why is that so... what happens if a noob tries out linux... he says aih I get asked if I want to install the "better" or faster nvidia drivers at the first start... oh yes of course... tahts the normal way windows does it that way so that must be the normal thing to do everywhere...

What happens next it causes massive problems instabilities... even black screens after boot in some cases etc. What happens next... the user deinstalls linux und whines everywhere how bad linux is. When this Closedsource drivers would be kind of banned not allowed to make any interface changes to the kernel and would not be allowed to be easily installed through such restricted-driver-management tools... the user would boot intel pc perfekt experince... linux is great... they would boot amd hardware... no problems... ok not so fast so amd is to blame. then they would boot nvidia stuff (and that would be especialy true in the time before nouvou drivers got released) they would see a black screen or some vesa drivers or something... they would say... ohh nvidia sucks because they as only company dont release any drivers to linux.

Then nobody who would want to do anything with linux would buy nvidia card... Nvidia is the weekest market player anyway today... as grafic-chip/card seller... they would hopefully get bankrupt... or they would release a driver finaly...

long story short point... the mega-buggy closed source drivers harms the reputation of linux... anybody will blame linux for each bug you get using this because peoples head are streamlined with windows-world-thinking...

While I recognize many of the problems you're referring to, and have had some of them happen to me, I don't think these problems are inherent to closed source (see OS X, Windows, Android, etc. as counter-examples). Most of these problems appear to be due to a lack of testing, bugs, or ridiculous design decisions made outside of the driver... typically in OSS or distro mechanisms.

For example, with a stock install of Ubuntu 12.10, the user can not install the NVIDIA driver without crippling his desktop. Why? Because Canonical has been mailing it in for quite a few releases these days. Why should the concept of a "closed source" driver be blamed for what is clearly an error from the OS vendor? Lightdm was another situation. If it attempted to bring up the display manager before Xorg was loaded... kaboom... the user was screwed with a blank screen. How is that the fault of a closed driver? I had a fun time troubleshooting that one. In some of these cases it's impossible to believe that there was any testing at all.

I also take exception with the idea that NVIDIA is the weakest market player. First, they have the best Linux driver (fact). Second, they are actually in strong footing in the GPU market. They have far superior GPUs to those of Intel and AMD as a company isn't exactly enjoying financial bliss.

And, with all of the frustrations I experience with Linux (which fortunately aren't too often), it's hardly ever NVIDIA that I'm shaking my fist at.

Dare I say, the big problem with "Linux" (from a noob perspective) is that these distributions are all beta (or even alpha) releases. There is NO quality assurance being applied, and that's just blunt and to the point. Look at how many times there are versions released with show-stopping bugs and regressions that would have been picked up had a formal regression testing phase been followed. Even an LTS like 12.04 shipped with major cluster-f's. They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works.

[Calinou]

While I recognize many of the problems you're referring to, and have had some of them happen to me, I don't think these problems are inherent to closed source (see OS X, Windows, Android, etc. as counter-examples). Most of these problems appear to be due to a lack of testing, bugs, or ridiculous design decisions made outside of the driver... typically in OSS or distro mechanisms.

For example, with a stock install of Ubuntu 12.10, the user can not install the NVIDIA driver without crippling his desktop. Why? Because Canonical has been mailing it in for quite a few releases these days. Why should the concept of a "closed source" driver be blamed for what is clearly an error from the OS vendor? Lightdm was another situation. If it attempted to bring up the display manager before Xorg was loaded... kaboom... the user was screwed with a blank screen. How is that the fault of a closed driver? I had a fun time troubleshooting that one. In some of these cases it's impossible to believe that there was any testing at all.

I also take exception with the idea that NVIDIA is the weakest market player. First, they have the best Linux driver (fact). Second, they are actually in strong footing in the GPU market. They have far superior GPUs to those of Intel and AMD as a company isn't exactly enjoying financial bliss.

And, with all of the frustrations I experience with Linux (which fortunately aren't too often), it's hardly ever NVIDIA that I'm shaking my fist at.

Dare I say, the big problem with "Linux" (from a noob perspective) is that these distributions are all beta (or even alpha) releases. There is NO quality assurance being applied, and that's just blunt and to the point. Look at how many times there are versions released with show-stopping bugs and regressions that would have been picked up had a formal regression testing phase been followed. Even an LTS like 12.04 shipped with major cluster-f's. They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works.

Then use Debian Stable and whine about how old it is.

Also, people aren't paid for making Linux distros and the show-stopping bugs almost always have workarounds. Get over it, choose what you want.

[Ramiliez]

Dare I say, the big problem with "Linux" (from a noob perspective) is that these distributions are all beta (or even alpha) releases. There is NO quality assurance being applied, and that's just blunt and to the point. Look at how many times there are versions released with show-stopping bugs and regressions that would have been picked up had a formal regression testing phase been followed. Even an LTS like 12.04 shipped with major cluster-f's. They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works.

Umm don't you know the ancient saying that MS Windows becomes usable after releasing first service pack? OS X has same birthing difficulties...

[Bernard Swiss]

I don't think NVIDIA made such a request though. I think what they asked for was the label "EXPORT_SYMBOL" rather than "EXPORT_SYMBOL_GPL" on the DMA-BUF functions, so that they can make invocations on those functions from their driver as I presume they do with other cases. But whether or not there's a legitimate legal difference between these two markers is highly questionable it seems.

You think it's running in user space?
Nope... then it must be running in kernel space. Fancy that.
The re-labelling of the symbols is in fact -- as others have pointed out -- merely clarifying the matter. That's why the lawyers liked the idea so much; it saves on having to wade through the GPL legalese to figure out what developers are and aren't allowed to do with those items.

However, and I don't know if it's considered pragmatic or something else, but Linus has acknowledged that the Linux ecosystem should make a way for proprietary software, including proprietary drivers, to participate.

And your point is?
Although Linus is definitely and explicitly an "Open Source" guy, and not a "Free Software" guy (he considers "users" too stupid to merit or benefit from FOSS -- it's for developers like himself), he still somehow finds nVidia extremely difficult to work with (the worst, in fact). Why do you suppose that is? Could it be that nVidia is trying to reap the benefits, but not fulfil their consequent responsibilities?

The problem I have with your criticism of proprietary software is that you're saying that a software developer such as myself does not have a right to his intellectual property, and that providing a product to an end-user should necessarily entail the developer giving up his IP rights. Whether or not that's truly best for the end-user is questionable, and I think it depends on the software. I do think there are examples where it's clearly superior for the end-user, and other examples where it's inferior. A mix of both open and closed seems to truly be "pragmatic".

(emphasis mine)

Hell no! Don't spout such nonsense in my name -- I don't appreciate it!

I would say, that a software developer such as yourself does not have a right to other people's "intellectual property". And that providing a product to an end-user should entail the developer giving up his IP rights is simply something that the GPL most definitely does not require.

If the code is actually yours, you get to release it under whatever terms you want. Of course, if the code isn't actually yours, you do have to conform to the license terms of the actual owner -- whether that code is under the GPL or whether it's under some proprietary license.

- - - -

PS:
If you don't like the terms of the GPL, then don't use GPL code, write your own damn code, or licence someone else's. That someone else will likely want cash and/or other material considerations, but BSD code is under terms you would probably find congenial -- you won't have to pay a dime for it, and you won't have to share it with anyone to whom you give the resulting software.

PPS:
There are times when the smart thing to do is to use a proprietary license. There are times when the smart thing to do is to use a BSD license, or MIT license, or something similar -- but there are many advantages to the GPL approach -- and much of Linux's success is due precisely to it's being licensed under the GPL, which encourages a degree of collaboration that other licenses either simply don't, or even discourage.

[johnc]

Umm don't you know the ancient saying that MS Windows becomes usable after releasing first service pack? OS X has same birthing difficulties...

So in Windows if I install a video driver the system becomes unbootable?

When did that start?

[blackiwid]

And, with all of the frustrations I experience with Linux (which fortunately aren't too often), it's hardly ever NVIDIA that I'm shaking my fist at.

Dare I say, the big problem with "Linux" (from a noob perspective) is that these distributions are all beta (or even alpha) releases. There is NO quality assurance being applied, and that's just blunt and to the point. Look at how many times there are versions released with show-stopping bugs and regressions that would have been picked up had a formal regression testing phase been followed. Even an LTS like 12.04 shipped with major cluster-f's. They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works.

thats the problem, you dont shake your fist at Nvidia while you should do... ok maybe a bit canonical is here guilty too... because they make the users beliving that they could give support for closed source drivers... They should not give any support at all... or at least put some big warning labels in this installer tools that they give no guarantee because they cant and if something goes wrong send email to blabla@nvidia and whine there because they suck...


And to the update thing, yes its basicly in beta-state... but windows is in pre-alpha when it releases, and I never heard of that somebody sucessfuly made a real update no sp, from windows x to windows x+1 was it not even that they dont allow it at all for some versions vista ->7? or something and if they "support" it it just installs both oses basicly over its other... like copy 1000 new files to the old 700...

So your lightdm problem with update is basicly a luxus-problem because in windows you would just not be able... and even if that in some rare cases works perfectly... they do manage such updates all what 3-5 years? its a bit easier than do it all 6 months... and they just dont fix security holes because they dont want to risk something does go wrong... I think thats a big price for having a somewhat stable experince... then there are the installs... install windows over usb-stick is less easy than a linux...

Thats the problems you "noobs" all compare linux to a pervect virtual os... and flame then because its not perfect... instead of compare it to windows...


"They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works." yes that sounds like you want to use debian... so its your fault for picking a somewhat bleeding edge distri... ok there are more bleeding edge distries out there... but ubuntu is still very bleeding edge... but if you lets say wait one year after the release of a ubuntu lts it should be pretty stable... like a windows is also first somewhat bugfree after sp1 (aka rc1) state...

I mean windows xp was after some years pretty good.... but then they released vista... total crap... totaly buggy, against this most bugs in ubuntu are very harmless... at least a lts version... they took even money for this alpha version... then they released windows 7 and now its somewhat usable... now whats next windows 8 gets very very bad feedback... basicly most people even hate it more than vista... they had in vista btw big regressions the boottimes did get after what 7 years of development time (at least they had that long time if they wanted to) it boot did take way longer than the old windows xp.

And then there is another problem, you compare a basic vanilla os like windows with basicly the hole opensource world... you would not blame microsoft if photoshop would have a bug, but basicly you did just blame Linux or Ubuntu for bugs in additional software... just because they make it easy installable for you...

[blackiwid]

tried to edit but to slow... I just googled about update paths... for windows users I did not find it for windows 8 but for windows 7:

http://blog.this.at/post/2009/08/07/...Windows-7.aspx

it says basicly that in most cases you have to reinstall windows you are not able to upgrade it...

so you really think that ubuntu has bad upgrade problems? so you would be more happy if ubuntu just would say... upgrades are not allowed so you run into 0 problems... because fresh installs at least from lts should normaly work...

and if you install a normal windows on a pc like on mine I have also no working network... thats also a big problem... the only reason less people maybe whine about install problems of windows is because they buy it preinstalled... or they get a cd with drivers from it... and basicly the company that sells the hardware look after it that the user run not into big problems... its a bit hard to blame ubuntu because the hardware-selling companies dont give the same support for linux/ubuntu...


so yes the binary drivers are not the only problem on the "blaming the wrong people" problem... but they are one of it... but on the other side... the amount of "noobs" using linux is not that important... except maybe for companies that try to earn money with proprietary software for linux...

[johnc]

Thats the problems you "noobs" all compare linux to a pervect virtual os... and flame then because its not perfect... instead of compare it to windows...

Believe me, if I thought Windows was better, I would use Windows [for things other than gaming]. The truth is, many things about Windows just drive me up the wall... with the primary difference being that there's usually no good workaround. Granted, I have enough technical skill with Linux that I can get around any problems I find. I don't know if I'm a noob since I've been using it in some capacity (school, work, home) since before the turn of the century.

But I'm not even complaining much about bugs per se. I write software, so I know the realities and I've never been immune to bugs myself. My primary criticism was about major regressions. (I.e., it worked here, they made a change, didn't test it, pushed it out in a release, and boom...) But even here, the point I was making was to refute your argument that NVIDIA is to blame for the woes of the noobs. In my experience it's not the driver that's the problem (though it is certainly not bug-free), but rather all the other stuff in the distro that causes the big headaches. So I don't understand why you were saying that NVIDIA is turning off new users to Linux.

"They have to slow down and not release a single change until it has been thoroughly tested. That's just how professional software works." yes that sounds like you want to use debian... so its your fault for picking a somewhat bleeding edge distri... ok there are more bleeding edge distries out there... but ubuntu is still very bleeding edge... but if you lets say wait one year after the release of a ubuntu lts it should be pretty stable...

I don't really want to use Debian because I don't want a 5-year-old kernel and Firefox 3.6. I do like the bleeding edge kernel and software packages, but am a bit of a late adopter for Ubuntu (11.04 on one machine, but 12.04 on the other). But my case isn't what's important. I can get around any problem I face. I'm getting to your point. If somebody says, "Hey, let me try out this Linux stuff... Anybody have any recommendations?" Typically someone will recommend Ubuntu to the person (though I'm finding this shifting towards Mint as a recommendation), since it's been historically the most new user-friendly. So the person is obviously going to get the latest version of Ubuntu, not something 12-18 mos old.

Aaaaaaaaaaaaaaaand then the person goes into Jockey and selects "NVIDIA driver" and then reboots to a black screen. Who wouldn't consider this a major hurdle to adoption? It's not about me or NVIDIA. I run into a lot of these types of postings on forums.

And then there is another problem, you compare a basic vanilla os like windows with basicly the hole opensource world... you would not blame microsoft if photoshop would have a bug, but basicly you did just blame Linux or Ubuntu for bugs in additional software... just because they make it easy installable for you...

Well I put "Linux" in quotes and called out Ubuntu in specifics merely to bring attention to the desktop aspect of a Linux distro, since this was the context of your criticism. The Linux kernel, on the other hand, is the perfect example of how it should be. A change is signed off, there's a pull request, then massive testing... and NO regressions. And no way is a show-stopping bug making it through to the release version.

Nor am I particularly blaming any piece of OSS software... I'm just saying that it's no less susceptible to bugs as, say, a proprietary driver. And even here, I only criticized distro decisions rather than actual software projects.

And all of this is from the context of what would cause frustration in a noob, not necessarily my personal complaints, though I think my experiences serve as legitimate points.

I hate to call out Canonical so much in this thread because IMO they put out the best distro. It's just some of the things I see make me face-palm and then SMH.

And, to be fair, I publicly shook my fist at NVIDIA when a driver update broke my suspend-to-ram. But that's the only problem the driver has ever given me.

[johnc]

so you really think that ubuntu has bad upgrade problems? so you would be more happy if ubuntu just would say... upgrades are not allowed so you run into 0 problems... because fresh installs at least from lts should normaly work...

I haven't criticized the upgrade path.

In the one example I gave I was referring to a fresh install of 12.10.