Linux Kernel 6.9 Officially Released (9to5linux.com) 49
"6.9 is now out," Linus Torvalds posted on the Linux kernel mailing list, "and last week has looked quite stable (and the whole release has felt pretty normal)."
Phoronix writes that Linux 6.9 "has a number of exciting features and improvements for those habitually updating to the newest version." And Slashdot reader prisoninmate shared this report from 9to5Linux: Highlights of Linux kernel 6.9 include Rust support on AArch64 (ARM64) architectures, support for the Intel FRED (Flexible Return and Event Delivery) mechanism for improved low-level event delivery, support for AMD SNP (Secure Nested Paging) guests, and a new dm-vdo (virtual data optimizer) target in device mapper for inline deduplication, compression, zero-block elimination, and thin provisioning.
Linux kernel 6.9 also supports the Named Address Spaces feature in GCC (GNU Compiler Collection) that allows the compiler to better optimize per-CPU data access, adds initial support for FUSE passthrough to allow the kernel to serve files from a user-space FUSE server directly, adds support for the Energy Model to be updated dynamically at run time, and introduces a new LPA2 mode for ARM 64-bit processors...
Linux kernel 6.9 will be a short-lived branch supported for only a couple of months. It will be succeeded by Linux kernel 6.10, whose merge window has now been officially opened by Linus Torvalds. Linux kernel 6.10 is expected to be released in mid or late September 2024.
"Rust language has been updated to version 1.76.0 in Linux 6.9," according to the article. And Linus Torvalds shared one more details on the Linux kernel mailing list.
"I now have a more powerful arm64 machine (thanks to Ampere), so the last week I've been doing almost as many arm64 builds as I have x86-64, and that should obviously continue during the upcoming merge window too."
Phoronix writes that Linux 6.9 "has a number of exciting features and improvements for those habitually updating to the newest version." And Slashdot reader prisoninmate shared this report from 9to5Linux: Highlights of Linux kernel 6.9 include Rust support on AArch64 (ARM64) architectures, support for the Intel FRED (Flexible Return and Event Delivery) mechanism for improved low-level event delivery, support for AMD SNP (Secure Nested Paging) guests, and a new dm-vdo (virtual data optimizer) target in device mapper for inline deduplication, compression, zero-block elimination, and thin provisioning.
Linux kernel 6.9 also supports the Named Address Spaces feature in GCC (GNU Compiler Collection) that allows the compiler to better optimize per-CPU data access, adds initial support for FUSE passthrough to allow the kernel to serve files from a user-space FUSE server directly, adds support for the Energy Model to be updated dynamically at run time, and introduces a new LPA2 mode for ARM 64-bit processors...
Linux kernel 6.9 will be a short-lived branch supported for only a couple of months. It will be succeeded by Linux kernel 6.10, whose merge window has now been officially opened by Linus Torvalds. Linux kernel 6.10 is expected to be released in mid or late September 2024.
"Rust language has been updated to version 1.76.0 in Linux 6.9," according to the article. And Linus Torvalds shared one more details on the Linux kernel mailing list.
"I now have a more powerful arm64 machine (thanks to Ampere), so the last week I've been doing almost as many arm64 builds as I have x86-64, and that should obviously continue during the upcoming merge window too."
69? (Score:3, Funny)
Nice!
Re: (Score:1)
6.9!
Re: (Score:1)
Re: 69? (Score:3, Insightful)
It'll be short lived though. Typical...
Re: (Score:3)
Re: (Score:2)
Not quite eye peeling stuff (Score:5, Funny)
I remember when the big update was USB support! I just turned 48 yesterday and it's a little depressing.
Re:Not quite eye peeling stuff (Score:5, Informative)
That's because Linux won. In that old past Linux support was late as compared to Windows, so you'd be waiting for things like USB. Right now new things are supported in Linux first, and long before they reach your hands. USB4 is supported since kernel 5.6 in March 2020, 18 months before Windows.
Re: (Score:1)
Problem is popular device support. For windows, that is what comes early, and for linux late if at all. Linux will probably get some specialist thing early before it's adopted widely. And that's great.
And then your network card, your slightly unusual sound card on the motherboard, etc will just not get drivers and not work. And that will actually suck, unlike "this rare USB 4 device I actually found in the wild will just communicate as USB 3.x gen n instead".
Also did I mention that USB naming convention aft
Re: (Score:2)
Your statements are more indicative of Linux 20 years ago. Popular device support is not as big a problem as in the past. The main dependency is the chipset being used, and manufacturers will generally source the same chipsets. For example, network and sound card commonly use a handful of companies like RealTek, Intel, or Broadcom. Unless the device manufacturer went out of their way to source an obscure chip, Linux device support is not really an issue these days. If anything I found the opposite to be tru
Re:Not quite eye peeling stuff (Score:4, Interesting)
I really wish this was true. But it's been the exact opposite so far. For most of the "kinda not supported for some reason" hardware, you can usually find a driver. If for some weird reason (usually because it's a really old piece of kit) you can't, you can usually either find a driver for older version of windows that will still work... or someone who modded the said older driver to work on your OS.
On linux, you're just SoL. There's not enough people to generate the numbers to care about that sort of thing, and linux backwards compatibility tends to be the one thing in which it absolutely loses to windows as an operating system. That is one area where microsoft has unironically beaten pretty much everyone else, and one of the main selling points of their OS.
And that is why win11 "we don't support chips with no hardware TPM 2.0" was received with so much screeching. It went against basic principles people are used to. And of course, it was pretty much immediately hacked to install without TPM 2.0, and everyone discovered that this was almost certainly a marketing thing. With tech side of microsoft leaving the old support systems in to allow people to bypass the marketing nonsense as usual.
Notably, that is one thing for which windows team really doesn't get enough accolades for in my view. They actually leave the old stuff in, so you can fix a lot of warts caused by marketing and design teams by just re-enabling the old stuff.
Re: (Score:2)
If for some weird reason (usually because it's a really old piece of kit) you can't, you can usually either find a driver for older version of windows that will still work... or someone who modded the said older driver to work on your OS.
I would like to know where you can find older Windows drivers that the manufacturer no longer supports. My Samsung printer scanner software used to work great until HP bought them out. HP does not allow me to load older drivers that used to work.
On linux, you're just SoL. There's not enough people to generate the numbers to care about that sort of thing, and linux backwards compatibility tends to be the one thing in which it absolutely loses to windows as an operating system. That is one area where microsoft has unironically beaten pretty much everyone else, and one of the main selling points of their OS.
Er what? I have found older drivers for Linux all the time. Everything I see is listed on what it works with. Certainly there is more work a Linux user has to do to source the correct driver. The easiest for a Linux user would be if it is in a package. But there ma
Re: (Score:2)
Sure. A good example is my desktop still runs Creative Audigy 2 sound card. It's a PCI card. The original one, not the current PCI-e.
Drivers are unobtainium from Creative. But there's a guy called Daniel_K who hacked together old driver package to work on win 7 and win10 (which even works on win 11!). So it works fine. And I can still run EAX games with proper support.
https://danielkawakami.blogspo... [blogspot.com]
This is what big enough of a community does. It means that there's a relatively much higher chance that some
Re: (Score:2)
What kind of esoteric hardware do you have exactly?
The only hardware I've had any issues with in the past decade have been hardware that's 10+ years old and never had good driver support to begin with - things like early spec n wifi, or niche consumer products with highly specialized capabilities.
It's been many years since I've had any sort of issue with hardware on linux. The last was an AMD motherboard with onboard ethernet. I still have the board, and the ethernet still doesn't work - unsure if it's hard
Re: (Score:2)
See my other post for just one example. There are many others:
https://slashdot.org/comments.... [slashdot.org]
Re: (Score:2)
even today, I have a laptop that's over a year since release, who latest mint LST + newer kernel doesn't support all hardware out of the box. (and less than a year later AMD announces no new improcements for that OLD APU line...)
Still it's a lot better than it has been.... But if I have to go out of my way to get support for current (or behind curve) video....
Re: (Score:2)
You'd be right if it were 2003. It isn't, and you're not.
These days, device manufacturers use linux device support as the basis for their development as often as not: you don't have to write a driver if it's already supported. Cross-platform drivers have been standard for years. They're effectively required as a starting point, due to PC/Mac, so why not do linux too/first in your development environment.
Re: (Score:2)
>so why not do linux too/first in your development environment.
Because doing things costs money, and people don't like to burn money on things that have really awful ROI. As in literally no one sane and even passably intelligent does it. You don't burn your house to keep warm in winter, and neither do I. If your goal is to make drivers, you make them for things that have actually meaningful audiences first like windows and mac, and OS/2, DOS and other exotics are an afterthought. GNU/Linux is somewhere b
Re: (Score:2)
It's a question of what matters. USB 4.0 support in Linux objectively didn't matter in March 2020. It's not an achievement to be proud of if something that isn't on the market is support, but something which is, isn't. Yeah it's great to see native support for USB4.0 in Linux first, but the reality is vendor driven USB4.0 support came to Windows with the very first devices on the market. Native support followed shortly after.
Here we are 4 years later and I (and I will bet you Mars Bar on this) suspect most
Re: (Score:2)
It's a question of what matters. USB 4.0 support in Linux objectively didn't matter in March 2020. It's not an achievement to be proud of if something that isn't on the market is support, but something which is, isn't. Yeah it's great to see native support for USB4.0 in Linux first, but the reality is vendor driven USB4.0 support came to Windows with the very first devices on the market. Native support followed shortly after.
Except Linus is not hoisting some sort of trophy bragging that they did it in Linux first. USB4.0 supports like many other incremental improvement needed to be done sooner or later as it was going to be implemented. So Linux added it. It is a non-issue. I am sure when USB5 get released, it will be added to Linux before it is added to Windows.
In the meantime things which actually matter, like graphics drivers for hardware that is not only available, but is also popular, are a shitshow on Linux. Printer drivers, are a shitshow on Linux (yeah even my Brother printer, but fortunately I am able to use a different model's drivers just fine).
You mean hardware that has lots of proprietary bits are harder to support in an open source system? You don't say! Maybe the problem is not Linux but the manufacturers
Re: (Score:2)
I can use my brother laser MFC without any brother drivers at all if I use it through the network. Scanning and printing are both standards-based in that situation, and the only thing I lose is use of the quick scan buttons on the device. The problem isn't linux, and it isn't manufacturers who don't maintain support; The problem is manufacturers who don't support standard interfaces, and the users who buy their hardware. Samsung is competent at one thing only, making SSDs– and even some of those have
Re: (Score:2)
Re: Not quite eye peeling stuff (Score:2)
Sadly, Brother is now dabbling in toner DRM, so they may also have a dark future. Do extra homework... Sigh.
Re:Not quite eye peeling stuff (Score:4, Interesting)
remember when the big update was USB support! I just turned 48 yesterday and it's a little depressing.
Kids these days...
My first kernel was 0.95a. I only remember that because 0.95b brought in the parallel printer port.
What's funny is I remember that, but can't prove it. Kernel.org goes back to 0.99 patch 10. Which is amusing because it contains entries like:
CHANGES since 0.99 patchlevel 14:
- too many to count, really. Besides, I've lost my notes.
At least Linus didn't blame the dog for eating his homework. :)
And Happy Belated Birthday!
Re: (Score:2)
Thanks, old man. Aaah, that felt good!
Re: (Score:1)
I'm pretty sure you're right since the 0.95c+ release [github.com] (April 9, 1992), more than two weeks after 0.95a [github.com], which came out only 32 years ago, mentions:
"James Wiegand writes initial parallell port printer driver" (sic)
Obligatory (Score:3)
I hope 6.9 doesn't have any backdoors.
Re: (Score:3, Informative)
Nah in 69 the only way in is through the mouth.
Open source Nvidia. (Score:2)
When are we getting an open-source Nvidia driver? I would love to not use a binary blob in the kernel, that is a pain. Having an open-source kernel driver for Nvidia would be amazing.
Re: (Score:1)
Re: (Score:2)
What's up with the open source nouveau driver for NVIDIA cards? I haven't tried it in quite a while.
Re: (Score:2)
What's up with the open source nouveau driver for NVIDIA cards?
It's still primitive and prone to crashes. It's amazing the developers have gotten this far at all, but it's not even remotely ready to be a daily driver.
Re: (Score:3)
That's the same thing they include with the driver, it's just the source release for the part that's required to be open. It uses (as it says on the page) the firmware and user space components from the binary driver. The majority of the driver is still closed.
Linux always make me happy... (Score:1, Interesting)
I just wish the desktop wasn't so fragmented. I feel like X/Wayland/etc always disappoints in some way. Android has been the best interface Linux ever got, because it's consistent and works well. Android desktop strangely isn't satisyfing though. I think Linux needs a new desktop like Android that doesn't require X/Wayland/etc. Maybe one day.
Re:Linux always make me happy... (Score:5, Insightful)
I just wish the desktop wasn't so fragmented
I disagree. The fragmentation is the best aspect of Linux. I don't want a one systemd to rule them all, I want multiple tools that people can select for their application. I don't want a wayland that will bring them all. I want multiple display servers that each person picks because it suits their needs. I don't a Gnome desktop to in the the darkness bind them. I want people who love E16, KDE, XFCE, WindowMaker, i3, and so on to be able to select what they want.
I wholly disagree. I think the people who want sysvinit scripts and the folks who want units from systemd are all awesome. I love the Gnome desktop, the KDE desktop, the titling managers, and the various litestep/gnustep out there. I think the RedHat, the SuSE, the Ubuntu, Slackware, Pop_OS, and so forth are amazing.
I think the wide variety out there makes FOSS be it Linux, GNU Hurd, or any of the BSDs the best place for everyone. Because there isn't a "Linux Desktop (tm)", the Linux desktop is whatever you make it. That's different than anything else out there and is the most compelling feature of open source in my most humble opinion.
I think the hate for systemd is unfounded because we still have sysvinit in a whole lot of distros. I think the hate for wayland is unfounded because we still have lots of Xorg and hells bells, we still have directfb. I think the hate for Gnome is unfounded because we still have more DEs than people actually know about. All the hate that I hear on Slashdot or the bemoaning of "whatever someone thinks is going to replace their favorite tech" it's all unfounded. Shit we still have Delphi and COBOL compilers, I know, that's how I make a bit of my bread and butter. Anyone thinking X11 dies within their lifetime is just wanting to yell at clouds. Because so long as people want to use something, it's there to be used. WindowMaker is still out there, E16 you can compile and run it on directfb today, there's people wanting to remake the IRIX 4dwm, there's folks writing kernels in pure Rust, we've got so many various init systems out there, more flavors of firefox pre and post XUL, and if you want to fire up an IRC server or heck you want to do what the cool kids are doing and fire up a finger server, we've got all of that covered too. That to me is the best thing about all of this.
No I don't think the fragmentation is a bad thing. I think it's one of the biggest strengths of open source.
Re: (Score:3, Funny)
Quick moderators, mod him down because he mentioned something you don't like. We all get to hate this person equally now. #onetruelinux. :-)
Re:Linux always make me happy... (Score:4, Insightful)
I think the hate for systemd is unfounded because we still have sysvinit in a whole lot of distros.
That doesn't make anything like sense.
sysvinit is good. It is simple and it works.
systemd is bad. It is complicated and when it goes wrong you can't figure out what is wrong without using the debugger.
Hate for systemd is logical because sysvinit still works better than systemd, but systemd has infected all the major distributions and we have to deal with it in the real world where we get forced to work with them.
Re: (Score:2)
... systemd has infected all the major distributions and we have to deal with it in the real world where we get forced to work with them.
Agreed. Although systemd's true purpose was successfully shrouded, mostly, it still didn't quite reach the level of acceptance they would have liked.
Actually it was too much. Instead of giving up and signing their datacenter budgets over to the cloudification racket these old stubborn greybeards decided to grin and bear it.
Re: (Score:2)
Hate for systemd is logical because sysvinit still works better than systemd
I will attempt to make this short and sweet. That is not exactly 100% true from what I have seen in the real world. But if you want to hold tight to scripts, far be it for me to convince you otherwise. I work on legacy z/OS and IBM i machines and we have got some HPUX 11i machines in addition to a bunch of images and containers where things like socket activation is really important. (and there is a bunch of Windows servers where I work, but we are not talking about those.) So I have seen both the scri
Re: (Score:2)
systemd solves some real problems and simplifies a lot of things that needed simplification.
I still hate it, and I don't run it on my personal systems. I use Gentoo, in large part because I want to customize a lot of things (including having modifications in /etc/portage/patches that get applied when I upgrade packages), and it has the option of using systemd or not. I do end up writing some fairly complicated init scripts, particularly for networking.
When it comes down to it, my biggest complaint about s
Re: (Score:2)
systemd solves some real problems and simplifies a lot of things that needed simplification.
OK, I'll bite.
Can you name the problems it solves, aside from faster boot time?
What on earth do you see that it simplified?
From what I see, everything it has touched has become markedly more complicated due to its association with systemd. The only benefit it brings is faster booting, and any other "this is a better configuration" type claim is both subjective and often wrong on the basis of negative points.
A system
Re: Linux always make me happy... (Score:2)
Well, it did also bring cgroups... However, those can be manipulated with simple shell commands, so that could have been done by augmenting the init script libraries. So to me the question is what did systemd give us that we couldn't have had some better way.
Re: (Score:2)
No I don't think the fragmentation is a bad thing
Tell me how many ABIs there are, why "binary" is such a naughty word, and why it's sane to have separate package managers for everything. Most of the time I come across an interesting project, the only way I can use it is to download the source from Github and compile, build, and deploy it myself.
Having choices is the result of good technological decisions. Fragmentation is largely the fault of politics and egotistical pissing contests.
ni.ce (Score:2)
ni.ce!