Linux To Get "Extended LTS" Releases, Kernel Support For Six Years

[birdie]

More like morally crappy business practices...
By cutting product lifecycle short manufacturers can artificially "kill off" mobile devices that use their hardware components.
How could IPs even contribute to this?
The chipmaker already sold his product, and an open driver will not enable reverse-engineering of it.

How many businesses do you run, an excellent economist who's hid himself at Phoronix forums? Maybe you own some of Fortune 500 companies? No?

It's amazing that Linux fanboys (it equally applies to all fanboys) always forget to take a reality check.

[darkbasic]

How many businesses do you run, an excellent economist who's hid himself at Phoronix forums? Maybe you own some of Fortune 500 companies? No?

It's amazing that Linux fanboys (it equally applies to all fanboys) always forget to take a reality check.

How many Fortune 500 companies do you own instead? None? Then please stop saying bullshit.

[RussianNeuroMancer]

Don't feed birdie

[RelaxTrolls]

Actually I'm not against the frantic pace of the Linux kernel development but only if it's made as microkernel-y as possible so that the core has rock solid APIs for working with various modules..

Well, as you know; linux is monolithic, so there is no 'made as microkernel-y as possible". As for Android, you can think of this news (and Project treble), as making android more modular, decoupled and isolated from the HALs, kernel and SOC.

Or maybe I just love QNX ;-) Who knows. It was never clear to me why Google hadn't attempted to buy QNX from whoever its current owners were, then open source it and use it for Android. A high performance microkernel with excellent features. Robust and stable as hell.

It's fairly common knowledge that BlackBerry owns QNX and has owned it for close to a decade. It's seems, given the history of android (even before google aquired it), as to why google didn't try to acquisition QNX; given Blackberry's popularity and dominance in the market at the time... QNX still has a life in the auto industry and medical industry, but it's dead on mobile... regardless of any technical merits it may have.

[wizard69]

More like morally crappy business practices...

There us nothing crappy about protecting you business. It is a practice that existed well before the tech world evolved. Trade secrets help you survive the competition. It doesnt matter if you are the local baker, a contractor, a manufacture of soap or even an airplane. Information shared is information that can be used against you.

Now i wish this wasn't so because my biggest frustration with Linux revolves mostly around drivers for the chipset installed in my machine. However i undersstand the need to eat as survival in industry requires having an advantage.

By cutting product lifecycle short manufacturers can artificially "kill off" mobile devices that use their hardware components.

In the tech world most hardware is kilked off due to obsolescence .

How could IPs even contribute to this?

Are you being purposefully dense here???

The chipmaker already sold his product, and an open driver will not enable reverse-engineering of it.

Why wouldnt it? Seriouslt if you know what the registers fo you can easily go from there to design the required hardware.

You do realize that most of the stuff that's on that list is the fault of companies and has nothing to do with the linux community, right?
The author even admits it in some instances...

Fuchsia is not, will not, and cannot be a solution for the problems found in linux.
The whole point of having linux as a desktop OS is that no single company or party can railroad the development of the entire OS to fit their specific needs only. Sure, this causes problems like slower, more controversial development of standard components (Wayland, systemd), but also ensures that there are a lot of interested parties invested in the development itself.
Fuchsia on the other hand will be Google's pet project, for Google's needs. Even if it takes of at all as a mobile and desktop OS, it is just going to be another corporate OS that will offer no real protection to users or to contributors other than Google.

But that is only my opinion on the matter, of course.

Opinion should have at least some basis in fact, yours doesnt.

[wizard69]

How many Fortune 500 companies do you own instead? None? Then please stop saying bullshit.

In this case it doesnt matter if a person runs a one man shop or has 100,000 working for him. In the end secrets protect a company from the realities of competition. Keeping a company viable means being able to oay the bills.

[Hi-Angel]

Why wouldnt it? Seriouslt if you know what the registers fo you can easily go from there to design the required hardware.

It's mind-boggingly hard for it's too low-level. Look for example: http://kohei.us/wp-content/uploads/2...ss-diagram.png it's a diagram of how LibreOffice loads a file. It's extensively commented, and still not easy to understand. Now try to understand it without comments. And without function names. And without c-code — in a pure assembler but with values being named. That's the level you're talking of: a bunch of registers, and values bouncing between them. You know what values are for, but it is very hard to make up the diagram from out there.

I might be exaggerating, but not too much. This is the reason why GPUs have documentation. And, there actually is a plenty of good documentation already for many GPUs, why wouldn't you just take some of it, and design your chip based off them? Releasing docs for one more or one less GPU wouldn't change anything from the perspective of someone trying to start making own GPUs. But it would ease very much the life of developers and enthusiasts — potential customers FTW.

[starshipeleven]

It would be better if Google would enforce open hardware drivers... I really cannot understand why these cheap devices need closed driver blobs at all. Unfortunately all the ARM devices are a big mess regarding drivers and firmware blobs.

Embedded hardware usually has a very fast development-release-EOL scheme.
They don't have time to make quality drivers at all, anything they make won't be upstreamable at all anyway.

But it's not impossible to move to open drivers, imho it's more of a mindset issue and inertia (switching over to making open drivers means usually rewriting from scratch because their blob isn't using any linux infrastructure like Mesa/Gallium and usually has third-party licenses in it).

[starshipeleven]

As much as this list is hated in the Linux community, the community is little by little embracing the problems and solutions outlined in it. Kinda amazing.

That list isn't hated, it's just mostly flamebait bullshit because it blames Linux for things it should blame hardware manufacturers, and blows stuff out of proportion for the sake of throwing shit around.

[starshipeleven]

Intellectual Property.

In some imaginary world it probably doesn't exist and everyone is broke. In the real world inventors want to put bread in their mouths instead of starving to death.

And how is that related to the fact that hardware manufacturers make most of their money selling you hardware while drivers are usually free?