SteamOS Has Its Own Graphics Compositor

[benhutchings]

Is it really -rt?

Debian kernel maintainer here.

SteamOS appears to be using linux 3.10 - PREEMPT_RT_FULL (unsurprisingly) with a heavy amount of patching (282 patches for -rt in the 'all' architectures/folder, alone). The kernel is also using aufs and they seem to be sitting on some bug fixes for upstream on top of that.

Some of the -rt related hacks they are using i have seen (in one case, i am using the same patch for ntrig)... It looks like they have gone to a lot of effort getting the kernel just right for their needs.

I haven't looked at the distribution yet, but the package listing included linux-image-3.10-3-amd64 which matches the stock Debian kernel (from a few months back). We have a single source package for regular and -rt kernels, and although the -rt patches are in debian/patches/features/all/rt/ they are not used for all kernels.

[Aleve Sicofante]

Reputation is everything for a new project especially in the FOSS arena. Do you think ReiserFS was dropped by many because of technical hurdles? Ubuntu was chosen early on by Valve before they started doing stuff that really started ticking off the community. I wouldn't doubt that they would change their mind had they released steam today.

Well, all that is pure speculation. They still recommend Ubuntu if you want to install their client on a Linux desktop. Why would they if they don't like it? They DO like it, but not for the reasons you think.

Want some more speculation? They went with a barebones distro that's not completely disconnected to Ubuntu (were they started). They will wait and see for a few years. If Ubuntu succeeds in mobile, it will succeed on the desktop too and Mir will stabilise. Then Valve might take another look. If Ubuntu fails in mobile, Canonical will shut down and Ubuntu will have a very dark future. It's as simple as that. Nobody cares about Mir vs Wayland in the real world.

They couldn't care less about reputation among Linux geeks. Their demographics target is the general public, not Linux geeks. They don't give a damn about Linux itself either, it's just a tool to move away from Microsoft and build their own appliance, like Tivo did. They are going their own way with the kernel and the compositor and they know this will bring tons of criticism from the usual suspects -the same that bash Ubuntu every other day for that very same reason- and they couldn't care less. Guess which distro is more popular among non-geeks? Exactly. Valve love Ubuntu for that and that's why they recommend it to its desktop users, not because is technically this or technically that.

So I bet Valve bashing will start in a few weeks/months, when Linux geeks understand these facts. Steam and Ubuntu users won't care, as we don't care today with every geek's moan and whine on every Linux site.

[duby229]

Well, all that is pure speculation. They still recommend Ubuntu if you want to install their client on a Linux desktop. Why would they if they don't like it? They DO like it, but not for the reasons you think.

Want some more speculation? They went with a barebones distro that's not completely disconnected to Ubuntu (were they started). They will wait and see for a few years. If Ubuntu succeeds in mobile, it will succeed on the desktop too and Mir will stabilise. Then Valve might take another look. If Ubuntu fails in mobile, Canonical will shut down and Ubuntu will have a very dark future. It's as simple as that. Nobody cares about Mir vs Wayland in the real world.

They couldn't care less about reputation among Linux geeks. Their demographics target is the general public, not Linux geeks. They don't give a damn about Linux itself either, it's just a tool to move away from Microsoft and build their own appliance, like Tivo did. They are going their own way with the kernel and the compositor and they know this will bring tons of criticism from the usual suspects -the same that bash Ubuntu every other day for that very same reason- and they couldn't care less. Guess which distro is more popular among non-geeks? Exactly. Valve love Ubuntu for that and that's why they recommend it to its desktop users, not because is technically this or technically that.

So I bet Valve bashing will start in a few weeks/months, when Linux geeks understand these facts. Steam and Ubuntu users won't care, as we don't care today with every geek's moan and whine on every Linux site.

So, to make that whole rant shorter..... "I love Canonical (for non-technical reasons) and you should too, damn it! Bash SteamOS instead assholes!"

[mrugiero]

If Ubuntu succeeds in mobile, it will succeed on the desktop too and Mir will stabilise. Then Valve might take another look. If Ubuntu fails in mobile, Canonical will shut down and Ubuntu will have a very dark future.

I don't know where you get all of this from.
Ubuntu succeeding in mobile doesn't imply Ubuntu succeeding in desktop, as they are different beasts. Do you see many desktops using Android?
The same for the failure. KDE is not very successful in mobile, while it's pretty much successful (well, as a Linux desktop is successful, I mean) in desktop.
Also, even if Canonical shuts down, Ubuntu might live on. It wouldn't be the first community distro able to survive.

They couldn't care less about reputation among Linux geeks. Their demographics target is the general public, not Linux geeks.

This part, I agree A LOT.

They are going their own way with the kernel and the compositor and they know this will bring tons of criticism from the usual suspects -the same that bash Ubuntu every other day for that very same reason- and they couldn't care less.

Not at all. Ubuntu gets bashed mostly for breaking compatibility. A small compositor and a few RT patches on the kernel? Meh, great for them. If they try to mainline them, even better, but even if they don't they don't harm anyone.

Guess which distro is more popular among non-geeks? Exactly. Valve love Ubuntu for that and that's why they recommend it to its desktop users, not because is technically this or technically that.

I agree.

So I bet Valve bashing will start in a few weeks/months, when Linux geeks understand these facts. Steam and Ubuntu users won't care, as we don't care today with every geek's moan and whine on every Linux site.

There will probably be some bashing, but it will be negligible compared to Ubuntu, as there is no legit reason to be concerned here.

[Aleve Sicofante]

I don't know where you get all of this from.
Ubuntu succeeding in mobile doesn't imply Ubuntu succeeding in desktop, as they are different beasts. Do you see many desktops using Android?

Ubuntu is already "succeeding" in the desktop (KDE is not, by any measure, even close), if by "success" we count "being the most used Linux desktop", which is clearly far from a success in the whole scheme of things, where it's as big as a statistical error. My point is very easy to understand if you are following Canonical's plans and why it's a whole different case from Android.

Canonical is betting the house on a unified ("converged") OS that would drive all diferent devices, from mobiles to tablets to desktops to TVs. The biggest bet there is obviously mobile (the cloud is very important for Canonical, but that part of the company is doing as good as it gets and it doesn't target consumers). If Ubuntu succeeds in mobile, Canonical will see two consequences: revenue from mobile carriers and hardware manufacturers (which might finnaly turn the company profitable) and a bigger adoption of their desktop BECAUSE it's just the same OS sharing a big number of applications (unlike Android). Apple got a lot of laptop sales from its success in mobile and, while OS X is being iOSified slowly but steadily, it's a long shot from Ubuntu's convergence model. (Chrome OS is getting a lot of attention lately, with Chromebooks, but I wouldn't say it comes from Android's success, but Google's.) Ubuntu will, from version 14.10 on, be as a matter of fact the same OS on mobile and desktop, with apps adapting themselves ("responsively") to the form factor they're running on, if properly coded. If this is well publicized along with a success in mobile, the desktop will benefit from a convergent environment which no other mobile OS offers as of today. (Although both Apple and MS are trying to unify their OSs for every platform, they have to fight a backwards compatibility Ubuntu has not need to address, not at least at the same level, mostly thanks to being pretty much irrelevant for 99% of desktop users today.)

On a side note, I don't believe Ubuntu would last a single quarter as a community edition. Maybe a few years ago, but not now. But that's just my personal perception.

Regarding what's expecting Valve from Linux geeks, wanna bet they're so proprietary and so "their own way" that all this love will turn into hate in a few months? Can't wait...

[shmerl]

Mir Linux future? Ubuntu's may be, not Linux one. Valve would be pretty stupid by investing any time in Mir. Their choice of Debian vs Ubuntu clearly shows they aren't stupid.

[duby229]

Hahaha typical Canonical irrational hater. Valve will make a grave mistake not supporting Mir. Mir is essential to Linux future. Do you think Ubuntu is so popular because it's bad? Hahaha! Stop with the hallucinations. There is nothing even close to Ubuntu in the Linux amateur bedroom brogrammers. The only way Linux will work out is if some big corporations gets behind it and do not listen to the linux community. The community is Linux's worst feature. Ban the community and linux becomes viable.

Wow, You just made the dumbest ever list....

[mrugiero]

Ubuntu is already "succeeding" in the desktop (KDE is not, by any measure, even close), if by "success" we count "being the most used Linux desktop", which is clearly far from a success in the whole scheme of things, where it's as big as a statistical error. My point is very easy to understand if you are following Canonical's plans and why it's a whole different case from Android.

It's still arguable, as it's changing a lot. Ubuntu's success on desktop might be for things different than what they are doing in their try for succeeding on mobile.

Canonical is betting the house on a unified ("converged") OS that would drive all diferent devices, from mobiles to tablets to desktops to TVs. The biggest bet there is obviously mobile (the cloud is very important for Canonical, but that part of the company is doing as good as it gets and it doesn't target consumers). If Ubuntu succeeds in mobile, Canonical will see two consequences: revenue from mobile carriers and hardware manufacturers (which might finnaly turn the company profitable) and a bigger adoption of their desktop BECAUSE it's just the same OS sharing a big number of applications (unlike Android). Apple got a lot of laptop sales from its success in mobile and, while OS X is being iOSified slowly but steadily, it's a long shot from Ubuntu's convergence model. (Chrome OS is getting a lot of attention lately, with Chromebooks, but I wouldn't say it comes from Android's success, but Google's.) Ubuntu will, from version 14.10 on, be as a matter of fact the same OS on mobile and desktop, with apps adapting themselves ("responsively") to the form factor they're running on, if properly coded. If this is well publicized along with a success in mobile, the desktop will benefit from a convergent environment which no other mobile OS offers as of today. (Although both Apple and MS are trying to unify their OSs for every platform, they have to fight a backwards compatibility Ubuntu has not need to address, not at least at the same level, mostly thanks to being pretty much irrelevant for 99% of desktop users today.)

That's kind of a lot of assumptions of how consumers work. Again, I'm not saying it won't happen, I'm saying it's hardly a fact that it will. A possibility? Hell, yeah, it is. A fact? Nay.

On a side note, I don't believe Ubuntu would last a single quarter as a community edition. Maybe a few years ago, but not now. But that's just my personal perception.

A huge percentage of the software Ubuntu uses is maintained by the community, so my bet is just the opposite.

Regarding what's expecting Valve from Linux geeks, wanna bet they're so proprietary and so "their own way" that all this love will turn into hate in a few months? Can't wait...

Bet all you want, but I'm basing only on facts. Fact is, they are currently not diverging in any concerning way from normal distros. If they do in the future, well, they do in the future, right now it isn't the case. The bashing will come IF they do.

[ninez]

Debian kernel maintainer here.

I haven't looked at the distribution yet, but the package listing included linux-image-3.10-3-amd64 which matches the stock Debian kernel (from a few months back). We have a single source package for regular and -rt kernels, and although the -rt patches are in debian/patches/features/all/rt/ they are not used for all kernels.

Hi Ben, I recognize your name ... I'm no expert on Steam/SteamOS, but they are shipping both non-rt and PREEMPT_RT_FULL for *amd64*. (ie: not for some ARM SoC or something like that - one could think _maybe it was for some ARM Soc - but AMD64+PREEMPT config kills any notion of that being exclusively the case) ...

But rather than re-posting the same thing from anouther thread (on this subject); http://phoronix.com/forums/showthrea...d=1#post380925 ... I point out a few of the relevant files. (but there are others that i didn't point out)...

[Aleve Sicofante]

Bet all you want, but I'm basing only on facts. Fact is, they are currently not diverging in any concerning way from normal distros. If they do in the future, well, they do in the future, right now it isn't the case. The bashing will come IF they do.

No, you're not "basing only on facts". "Any concerning way" is pure opinion and that -not the true fact that they are indeed diverging and going solo, exactly the same as Ubuntu- is as much assuming as you say I do. The importance of the concern is yet to be analyzed in depth. We are confronting just a beta. Let's talk when SteamOS reaches release status. And let's not forget that every single piece of code that makes SteamOS what it is -i.e. the Valve Steam client- is closed source. I'm eager to see how the "community" keeps bashing Canonical for releasing open source code and live peacefully with Valve's vendor lock-in through their Steam client... LOL

Of course we're discussing what will happen in the future and making assumptions (we still haven't got crystal balls do we?). All I'm saying is your "facts" and my "assumptions" are probably of the same nature: educated guesses. You have your reasons and I have mine. You think their diverging is non-critical and won't provoke Linux fanatics; I happen to think the opposite. You think my vision of Canonical's future is "a lot of assumptions" but I don't make them in a vacuum. There are lots of facts sustaining my assumptions. You just need to follow Canonical in the press everyday and you'll come to pretty much the same conclusions.

The fact that many parts of Ubuntu are maintained by the community doesn't mean a thing. Ubuntu's "community" has been basically dismantled long ago, by Canonical no less. The parts that the community maintains for Ubuntu aren't just for Ubuntu only. Once the leader shuts down, Ubuntu will probably break up in a myriad of hobby derivatives and factually disappear from the landscape in the real world (of course those myriad derivatives will show up in Distrowatch, and there'll be fans of each claiming that's the real successor, and blah, blah, blah...). In the meantime, non-basement ex-Ubuntu users will look for another company backed distro.