What's New in Linux 6.5?

An anonymous reader shared this report from 9to5Linux: Just a couple of days after celebrating its 32nd anniversary , Linus Torvalds announced today the final release of the Linux 6.5 kernel series as a major update that introduces several new features, updated and new drivers for better hardware support, and other changes.

After seven weeks of RCs, Linux kernel 6.5 is here with new features like MIDI 2.0 support in ALSA, ACPI support for the RISC-V architecture, Landlock support for UML (User-Mode Linux), better support for AMD "Zen" systems, as well as user-space support for the ARMv8.8 memcpy/memset instructions. Also new in Linux 6.5 is Intel TPMI (Topology Aware Register and PM Capsule Interface) support for the power capping subsystem and a TPMI interface driver for Intel RAPL, and the "runnable boosting" feature in the EAS balancer to improve CPU utilization for specific workloads.

This release also improves SMP scheduling's load balancer to recognize SMT cores with more than one busy sibling and allows lower-priority CPUs to pull tasks to avoid superfluous migrations, and improves EXT4 file system's journalling, block allocator subsystems, and performance for parallel DIO overwrites.

Re:

[ArchieBunker]

It seems like the kernel 2.x days lasted forever. Wonder why the major version numbers are advancing much faster?

Why the faster version number changes?

[davidwr]

It's the new scheduler :)

Re:6.5? Is this a joke?

[Tim the Gecko]

2.x was around a long time, but in 2011 Linus posted "So there it is. Gone are the 2.6.<bignum> days, and 3.0 is out" [indiana.edu] and kicked off the new scheme.

Re:

[sg_oneill]

Theres no real reason. Linus just changed the way he was advancing the numbers, and there really isn't another reason, at least according to torvalds.

Re:

[kriston]

Completely by the whim of the Creator.

Honestly, he should change to Month/Year dates instead of changing major/minor version numbers at his sole whim.

Re:

[kriston]

Or Year/Month, whatever.

because

[Anonymous Coward]

Software projects tend to ossify into a version scheme where changing the big number implies cataclysmic changes. 2.0 brought SMP and kernel modules ... so if that's what constitutes a change big enough for a major number, nothing else will ever justify a major number change. So years and years went by with 2.x.y version numbers that were utterly opaque where it was impossible to tell whether a kernel "seemed" old without memorizing release dates. Linus recognized the problem and started bumping the major

Re: 6.5? Is this a joke?

[OrangeTide]

Windows 8.1 was technically Windows NT 6.3. By this arbitrary standard, Linux is only a few years behind Windows 11 (NT 10.0)

Re:

[kriston]

I like how the Windows NT version number mysteriously jumped to 10 when Windows 10 came out.

Kinda like Sun Solaris 10 actually being SunOS 5.x under the covers.

Re:

[OrangeTide]

I work at a place where marketing and engineering have different numbering schemes for chips. When friends ask me about some consumer product I have to look up what is in it because I can never remember off the top of my head. I feel like version number inflation is a common pathology among marketing, in their defense they are reacting to human psychology in order to get people to buy stuff.

Re:

[tlhIngan]

It seems like the kernel 2.x days lasted forever. Wonder why the major version numbers are advancing much faster?

What does it matter? The older kernels are still supported.

The maintenance of kernels changed strategies around the 3.x era. Now there are more maintained kernels - you can run the bleeding edge 6.5, but there are older kernels for the various ports out there. If you

It used to be everyone had to be on the latest kernels and you bring everything forward - if you were on 2.x, you had to follow 2.x.

Re:

[backslashdot]

Yes it's a joke. you got it. lolololololol. Ver 6.5 for the lolz

Re:

[vbdasc]

Novell Netware never went beyond version 6.5 though, so Linux is now in good company.

Re:

[NoWayNoShapeNoForm]

Linux is better cuz it's not encumbered by a lawsuit-happy Corporate owner as in Novell.

Re:

[sconeu]

On the gripping hand, Novell defended Linux from SCO, and pretty much eviscerated the SCOundrels, proving that they had NO copyrights.

Phoronix

[bill_mcgonigle]

https://www.phoronix.com/revie... [phoronix.com]

Yes, but does it run Linux ...

[davidwr]

... on a Commodore 64 [slashdot.org]??? :)

Glad to see ALSA still developed

[paralumina01]

PulseAudio has always been terrible for me. Random beeps. Terrible background static. Randomly drops output and have to restart the server.

Re:

[caseih]

Alsa is the standard for sound drivers on Linux, so yes it is seeing active development and maintenance. PulseAudio requires Alsa.

Re:

[gweihir]

Well, what do you expect from Poetterix? Guy is so incompetent, he had to go to Microsoft.

Re:Glad to see ALSA still developed

[drinkypoo]

Switch to pipewire with wireplumber, it replaces both pulse and JACK and also supports ALSA clients (gives them per-app volume control.)

Submit a pull request

[davidwr]

what progress has been made on excising naughty words from the code and the comments.

Dunno. If you see anything that can be improved, code it, test it, and submit it.

Was BTRFS updated to support RAID-5/6?

[Lady Galadriel]

It has been forever that BTRFS was going to support write hole-less RAID-5/6. Like more than 7 years. But, checking the following site still shows RAID-5/6 as unstable.
BTRFS - Status [readthedocs.io]

So, back to the competition, the one that must not be named unless flame wars are desired.

Re: Was BTRFS updated to support RAID-5/6?

[RightwingNutjob]

Emacs has raid support now?!

Re:

[Schoenlepel]

If only it had an editor...

Re:

[NoWayNoShapeNoForm]

If only it had an editor...

That's what 'vi' is for :)))))

Re:

[WereCatf]

No, the write-hole appears to still be present. Personally, I am keeping an eye on bcachefs as a possible replacement, since it seems to have the benefit of not having that issue. It's also got other benefits, like e.g. it doesn't require an out-of-tree module, like ZFS does. I've also found a rather big issue with Btrfs where you can hard-crash the Btrfs-driver in the kernel by using mergerfs with individual Btrfs-formatted drives/partitions (tested with kernels 5.15, 5.16 and 6.2 under Ubuntu), so I've be

Re:

[drinkypoo]

it doesn't require an out-of-tree module, like ZFS does.

What do you see as a drawback of an out of tree module?

Re:

[WereCatf]

it doesn't require an out-of-tree module, like ZFS does.

What do you see as a drawback of an out of tree module?

Not me, personally, but rather kernel-devs seem to have quite a lot of animosity towards such, including ZFS. Also I've seen some random negative ramblings from distro/package maintainers about it. I've never bothered to really read what all the hubbub was about as it's not really relevant for my interests, but my impression is that it has a tendency of causing quite a lot of friction in getting stuff fixed/improved and such.

Re:

[drinkypoo]

Somehow ZFS seems to work great with Ubuntu (you can select root on encrypted ZFS as an install time option) and fine with Debian (you have to do the install on ZFS manually, but it works once you do so.) I've been using it for some time now through several kernel upgrades and one ZFS upgrade without any problems whatsoever.

If other distributions have trouble integrating it, perhaps they should read some of the documentation on the OpenZFS website.

Re:

[WereCatf]

I don't necessarily disagree with you. I haven't used ZFS myself, but I've just been looking into it as I was planning to possibly switch my servers to it from Btrfs until bcachefs matures some more. Also, I was thinking of throwing together a TrueNAS Scale setup from leftover parts and that uses ZFS as well, so I've got a reasonable excuse to learn all about it.

Re:

[gweihir]

Use md raid? You know, like a sane person?

Re:

[dunkelfalke]

RAID 5/6 is, for large modern hard drives, unsafe on any file system. I personally prefer using RAID 1c3 for both data and metadata on BTRFS.

Re:

[JackieBrown]

Same. Honestly, BTRFS is wonderful to work with. It replaced me using mergerfs with snapraid a while ago.

Linux CNC needs real-time functions as a switch

[BrendaEM]

A real-time switch would eliminate the need for a custom-built kernel for low-latency operations such as music recording and CNC. Distros that use these are often running older kernels, creating security issues, as well, as missing features found in fresh kernels. If it low-latency operation were just a switch, no custom compiling would need to be done.

Re:

[wed128]

It's really not that simple from an implementation perspective. Linux is rather inappropriate as an RTOS; it was very much designed for throughput over latency. If RT performance is really important, i'd choose another OS such as VxWorks or QNX.