Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Open Source Operating Systems Linux

Linux 5.5 Released (kernel.org) 32

jrepin writes: Linus Torvalds has announced Linux 5.5 release, codenamed as Kleptomaniac Octopus.The latest version of the open source operating system kernel brings RAID1 with 3- and 4- copies to btrfs filesystem, ext4 gets direct I/O via iomap together with fscrypt supporting smaller block sizes, and you can now use SMB as root filesystem. AMD OverDrive overclocking is now supported on Navi GPUS, wake-on-voice on newer Google Chromebooks is now supported. Added was a Logitech keyboard driver. KUnit is a new unit testing framework for the kernel. There are many more new features which you can read about on Kernel Newbies changelog page. For downloads visit The Linux Kernel Archives.
This discussion has been archived. No new comments can be posted.

Linux 5.5 Released

Comments Filter:
  • by Major Blud ( 789630 ) on Monday January 27, 2020 @09:44AM (#59660462) Homepage

    and you can now use SMB as root filesystem

    Does this mean that the root file system can reside on a samba share? I take it the /boot volume would still have to reside on the local machine?

    • Re:SAMBA (Score:5, Informative)

      by fred6666 ( 4718031 ) on Monday January 27, 2020 @09:51AM (#59660486)

      You can possibly load the kernel over the network too (PXE). But yes, it means the root file system can now be a samba share. It was already possible with NFS.

      • Yeah that's what I thought, thanks.

      • in the memory of a GPU on another PC, that is connected via PCI-over-Ethernet?

        (Yes, all these things are real. :)

        • Why not? BOOTP lets you just specify files. If you've mapped your PCIoE GPU memory space to a file system, you're just a configuration file away from a good Friday night.
        • Sounds like a cool thing to try.

          Almost 20 years ago, I got to try out a Linux machine (pc) to run the local X server for our chip design software on the HP ux servers (instead of using Exceed or so), and set up my office colleague with a boot floppy, running an ftp server that could pass my systems drive as root file system, substituting config files for target specific ones.

          One morning we found out about the true stability of Linux, as my colleague had come in slightly before me, and found his system d

  • Is this a nod at NROL-39?
  • Are Linux releases usually Noun Noun pairs or should this be Kleptomaniacal Octopus?

  • hmm, sounds like something can be hacked to listen in on people's devices 24/7/365 i bet the CIA/FBI/NSA are all drooling over that new feature
    • Amazon has it down to a science already, and the spooks knew how long before Alexa became part of the Amazon universe. It's been accepted as inevitable by the Linux community, because voice commands are considered a widely used input now.
    • SELinux code has been in Linux for a long time.

      Do you know anyone trustworthy who audited it? Cause I'm not trusting those fuckers blindy even if they'd say the sun is round.

      INB4 blackeyers trying to contain their comforting house of "all is well" cards. :D

  • ZFS (Score:1, Offtopic)

    by slickwillie ( 34689 )
    I'm guessing no kernel support for ZFS?
    • Re:ZFS (Score:5, Informative)

      by Kjella ( 173770 ) on Monday January 27, 2020 @01:22PM (#59661530) Homepage

      The situation hasn't changed much, if you're not in the main tree you're not supported and ZFS has never been there. Recently there was a kernel API cleanup where a function no longer had any internal users and so they removed it, they've broken AMD/nVidia's blob drivers many times this way and the policy is they don't care. They'll refactor the internal workings of the kernel when and how they want and won't offer any stable APIs/ABIs for the kernel, downstream modules will simply have to keep up. This is to avoid cruft, discourage blobs and encourage driver developers to work with them to get code mainlined.

      The OpenZFS project wanted a more recognized position but Linus basically shot them down, it's out of the question to merge it into the kernel tree because of the licensing issue and it's out of the question to start paying attention to code not in the kernel tree. And because they came across as choosy beggars who wanted the perks of being in the kernel without giving Linus any code he could use he got a bit snarky which was a bit unnecessary but his position has been the same for close to 30 years now, you can use my kernel if I can use your code to make it better. The OpenZDF project can't make this quid pro quo because of the CDDL, so he gave them the cold shoulder.

      So the OpenZFS team did what they've always had to do, fix their code to match whatever the kernel gives them now. Now I'm sure they feel that it sucks because they're probably not the ones who have a problem with the GPL license, but I don't think the kernel team is intentionally trying to trip them up. Maybe they were expecting something else but I'm guessing that Linus has learned that if you give someone an inch it'll become a whine feast about any future changes too and when legacy interfaces can be deprecated/removed. The "if you're not in the kernel tree you don't exist" policy is at least consistent and simple.

      • Barking to the wrong tree.
        Go complain to Larry, not to Linus.

  • You know: How many
    * complete rewrites (Think "Perl 6" aka Perl 2 1.0 ;),
    * compatibility-breaking fundamental architectural changes,
    * compatibility-keeping feature additions,
    * bug fixes and patches
    did it have?

    Because I know everything since 3.0 (incl.) is so arbitrary, it is just plain insulting.

    And yes, version numbers are supposed to converge towards the smallest possible increases and then be finished. That's how you know you haven't been infected with feature creep.

    • Wait... are you sure youâ(TM)re not describing every OS?

    • by Kjella ( 173770 )

      Linux hasn't had a feature-based release since version 2.6 back in 2003, you simply don't replace a significant amount of a mature kernel in one go. I agree that it's gone completely arbitrary and he might as well just use the Ubuntu numbering now and call it version 20.01 because it's released in January 2020. At least you'd have an idea how much wall time has passed, because they don't really say anything else.

  • I upgraded to 5.xxx on Linux Mint recently and my PC blew chunks, exhibiting major display issues that could not be resolved.

      I had to roll back to the 4.xxx series and remove all the 5.xxx kernels.

    • Maybe don't use Linux Mint, then.
      • Maybe don't use Linux Mint, then.

        Why?

        Other than this one issue, I've literally never had a problem with Mint.

        It was fairly easily resolved, so why should I switch, and to what?

        • You should switch because someone on the internet told you to.
          You should switch to MacOS because if you run into problems there, trying to do thing A they'll gladly tell you either that thing A isn't a problem, you're just imagining your problem, or, your problem occurs due to you holding it wrong (or variations thereof) or you shouldn't want to do thing A at all.

          In any case, you'll not be back complaining about Linux trouble, so it'll be a win-win.

    • This issue bite one of my PC too. It used LTS HWE kernel, so it was updated from 5.0 to 5.3, and suddenly NVidia driver was broken.

      I had to remove HWE kernels, and install vanilla LTS kernel (4.15) instead (and the related Xorg packages).

      • The older nVidia drivers (like 390) that are needed to support their older chipsets do not compile with 5.3 kernels. I went the nouveau route and am still happy.
    • I'm using it on Debian without any problems.

The 11 is for people with the pride of a 10 and the pocketbook of an 8. -- R.B. Greenberg [referring to PDPs?]

Working...