Linux 6.9 Sees Invasive & Significant Changes To Workqueues

[coder]

Maybe it's time to rethink the *nix architecture.

There's an original idea!

Linus released the early version of Linux in 1991 and by then Unix was 21 years old.

Today the basic *nix architecture has been around for over 50 years.

The most UNIXy things about Linux are its security model, its filesystem, and the basic kernel/userspace divide. There's a lot else about it that doesn't particularly resemble UNIX proper, or other popular clones from the 80's and 90's.

Back when such things had more meaning, it was said that Linux was something of a mutt between UNIX and BSD, borrowing ideas and APIs from each camp. Since that time, it has developed numerous innovations of its own, placing it in a category of one.

maybe start a new kernel, written in Rust ...

You mean like Redox?

https://www.phoronix.com/news/Redox-OS-Porting-Plans

BTW, language != architecture.

... with no backwards comparability,

Oh, so that's going to be popular!

Just look at the state of hardware support for FreeBSD or Dragonfly. It's not great.

just say this kernel will only work with hardware released in 2025 and newer.

Then it would never hit critical mass.

It just seems that at the current state of about 25 million lines of code in the current kernel, it has become unwieldy

One way to get/keep that number down is to do refactoring and iterative improvements & refinements of the kernel's architecture. If the sort of work being discussed here were never to happen, the source would be much larger and more unwieldy.

Alternatively, I guess you could wait indefinitely for something like HURD that tries to have the best architecture but never actually reaches a usable state?

[rclark]

i do not believe in upgrading or updating Windows until i have no choice.

Choice... That is where Linux/BSD/Redox/etc. come in.... You do have a choice to abandon M$ completely. Win7 was the last for me. Never looked back.

I don't mind 'changing' code IF there is a good reason for it -- even if the code was working. For example, If someone spots a simply change to get 50% more performance from the system, then worth it. On the other hand if it is just because the person doesn't like how it was written, or thinks it needs to be written in Rust (so called modern language), or maybe Basic, Lisp, or whatever... with the same functionality. That is a complete waste of time.

BTW, a new OS would have to have some VERY VERY compelling reason to move to it in this day and age (this isn't the dos, cpm, os/2, unix days of yesteryear). What would compel you or me to move away from say Linux? With Windoze I can think of a lot of reasons to move away from it just from the vendor lock in, money, and virus perspective... But for Linux? Really can't think of anything I need in a 'new' OS that would cause to to jump ship. just an observation.

[Sonadow]

In all fairness, i have stated numerous times that i do not believe in upgrading or updating Windows until i have no choice.

I said i stuck with Win 2k until I bought a Geforce 3 Ti4600 and needed DX8 for new games that I was playing.

I stuck with XP 32-bit until i bought an Athlon64 and needed a 64-bit OS.

I stuck with XP64 until I needed to upgrade to Win 7 and I stuck with Win 7 until I needed to upgrade to Win 10.

I recently upgraded to Win 11 and frankly there are some annoyances that makes me seriously consider wiping the drive clean and going back to Win 10.

For Win 10/11 I don't install any updates except for drivers and the H1/H2 updates after they are fully tested in the wild.

I don't update any of my Linux boxes unless i absolutely have to.

I have never believed in the update for the sake of update because a lot of times there are nasty bugs that aren't discovered until they are in the wild.

Don't waste your time trying to speak sense with that 'V' bot.

[sophisticles]

You mean like Redox?

https://www.phoronix.com/news/Redox-OS-Porting-Plans

This is actually what i had in the back of my mind; in an other thread i said that System76 should start shipping the computers they sell with this.

BTW, language != architecture.

I have heard rumors to this effect.

I only mentioned Rust and architecture together because if you are going to start with a fresh design you might as well use a language that the Biden administration recommends:

https://www.whitehouse.gov/wp-content/uploads/2024/02/Final-ONCD-Technical-Report.pdf

https://media.defense.gov/2022/Nov/10/2003112742/-1/-1/0/CSI_SOFTWARE_MEMORY_SAFETY.PDF

[sophisticles]

If new software has nasty bugs in the wild, same for old software. Good luck fighting computer viruses.

I think you misunderstood what i was saying, I was stating that I don't believe in updating until the early adopters have had the chance to field test the software for me.

[sophisticles]

Choice... That is where Linux/BSD/Redox/etc. come in.... You do have a choice to abandon M$ completely. Win7 was the last for me. Never looked back.

You misunderstood my use of the word choice with regards to not having one with Windows.

What i meant was that as new hardware came out MS made the business decisions to support certain features only with certain Windows versions.

For instance, DX8 was only supported starting with Win XP, so if you had a Geforce 3 or newer and wanted to play DX8 games, then you needed to use Win XP.

BTW, to this day there are lots of open source apps that run way better on Windows than they do on Linux, such as Shotcut, Openshot, Handbrake, list goes on.

BTW, a new OS would have to have some VERY VERY compelling reason to move to it in this day and age (this isn't the dos, cpm, os/2, unix days of yesteryear). What would compel you or me to move away from say Linux? With Windoze I can think of a lot of reasons to move away from it just from the vendor lock in, money, and virus perspective... But for Linux? Really can't think of anything I need in a 'new' OS that would cause to to jump ship. just an observation.

You are making the same mistake that a lot of Linux enthusiast make, namely thinking that people care about the OS they use.

People do not use OSes, they use applications.

It's why I keep saying Linux needs a killer app, that unique, must have, piece of software that is not available any place else.

The reality is that if some Linux project, like Fedora or Ubuntu, decided to create that killer app that everyone wanted but made it exclusive to their distro, they would bury the competition on no time.

Imagine if the KDE project made a closed source game that was crazy good, everyone was talking about it and they said that if you wanted it you had to buy a Slimbook they sell, or you had to at least use Neon, they would kill the competition inside of a year.

I have suggested something like this to the System76 people but they are too ideologist to ever do anything like this.

What keeps me on Windows is MS Office and the fact that it is reliable, anything I write i know will work and the cross platform software runs better on Windows.

What keeps me using Linux is that it is legally free so I am able yo save money in OS licenses for my computers and I like to hedge my bets, no one knows what is going to happen, as crazy as it may seem Microsoft may decide to base their next version of Windows on Linux or BSD and by being using Linux daily i will be able to use such a Windows version easily.

I also like keeping my skill current for a prospective employer, so I make sure to be familiar with both.

[coder]

I was stating that I don't believe in updating until the early adopters have had the chance to field test the software for me.

Yes, and you have similar options with Linux kernel updates.

You are making the same mistake that a lot of Linux enthusiast make, namely thinking that people care about the OS they use.

People do not use OSes, they use applications.

Ah, but applications care about the OS you're using and so do security updates. So, you're forced to upgrade after a while, whether or not you care about the OS. The OS also cares about the hardware, meaning you might have to do a full HW & SW upgrade, when you were otherwise happy with your machine as it was.

The reality is that if some Linux project, like Fedora or Ubuntu, decided to create that killer app that everyone wanted but made it exclusive to their distro, they would bury the competition on no time.

You see some amount of innovation around things that make each distro distinct, such as packaging and now their choice of container technologies.

For more application-level stuff, it's hard to make it distro-specific, since an application that's any good will get ported over to other distros. And if it's not open source, then it probably won't ever gain a lot of traction.

What keeps me on Windows is MS Office and the fact that it is reliable,

It's ubiquitous and it works well enough that most current users stick with it. Using it for my job, I've seen it mess things up enough times that I don't trust any MS Word doc over 50 pages and make sure to maintain it in a version control system since I never know how far back I might need to go if it gets corrupted. MS Excel is a collection of some truly bizarre quirks that they apparently can't fix or change, because it would cause too many long-time users to become disgruntled.

[DumbFsck]

Don't take this as disagreement with all the rest of your post, I basically agree with everything. But;

.

Then it would never hit critical mass.

The only successful modern example was a fully "cold turkey" approach, changing bottom to top, from Isa (sort of, as the Motorola chips and IBM ppc timeline is basically fingerprints in the underground handrail in my memory) to widget toolkit but including everything in between.

I'm talking about the time when apple bought NEXT and threw out nextep and their own OS to "fork" BSD putting it with the Mach micro kernel and making Darwin. Fork gets the quotes because even from it's first release it was more different than similar, not only in code but architecture-wise.

Not that I think Linux is in any sort of place where it needs to or can do the same. Apple and Microsoft if they wanted probably could.

As ssophisticles said, if you have a killer app, they'll come. Even if that killer app is having a trillion dollars, "design" or marketing campaigns.


The thing sophisticles doesn't realise is that every software that is now a "killer app" or a "must play game" got there "by accident" (aside from, you know, what came when digital computing was young. When you went and built a spreadsheet program and yours was the only one, of course it instantly became a must have app, until of course the giants got to copy yours, etc. ). If all it took was a few hundred million dollars and will, unity would make a single game like fortnight to sponsor their engine development, instead of increasing licensing fees. Microsoft would have made paint an alternative to photoshop, or their 3D paint an alternative to CATIA or blender/zbrush. If all it took was a couple tens of billions of dollars we'd be talking to each other inside of Zuckerberg's meta verse (was that the size of the hole? Twenty something billion?).


It basically makes as much sense as telling someone who's sick: "just get better".

[coder]

I basically agree with everything.

Same.
😎

[rclark]

You are making the same mistake that a lot of Linux enthusiast make, namely thinking that people care about the OS they use.

People do not use OSes, they use applications.

True. But my decision (at home) to move away from Windows was motivated by several reasons. One was all the applications we used worked on Linux -- except one (Print Shop at the time). LibreOffice, Firefox, Thunderbird as examples. Second was pocketbook bleeding. Virus protection subscriptions, and OS. No keys, no creating an account with M$, no automatic downloads, no auto boots, no forced DE. All the languages I enjoy programming in are available. There is many many reasons that tipped me to finally going all Linux. No blue screens of death. Not 'anti'. Windows Just not needed in my world.