Announcement

Collapse
No announcement yet.

The Linux 6.7 Merge Window Is Massive With Many New Features

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • The Linux 6.7 Merge Window Is Massive With Many New Features

    Phoronix: The Linux 6.7 Merge Window Is Massive With Many New Features

    The Linux 6.7 merge window has been downright exciting with additions like Nouveau GSP support and the Bcachefs file-system being added. It's also been downright massive as one of the largest merge windows in recent history in terms of code changes. Here's some statistics of the Linux 6.7 merge window ahead of today's Linux 6.7-rc1 release...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Did I miss something, or little.big cluster scheduling improvements by Intel devs once again didn't make it? Same for the missing realtime pieces?

    Comment


    • #3
      meh. Mostly gpu stuff that's inherently large. No RT? No console revamp?
      Scheduler changes in 6.6 were far more interesting.

      Comment


      • #4
        Originally posted by milkylainen View Post
        No console revamp?
        i'm still waiting for a modern kmscon alternative T.T

        Comment


        • #5
          Originally posted by Quackdoc View Post
          i'm still waiting for a modern kmscon alternative T.T
          Serious question, because I just don't know. Is there a general sense that it makes sense to get the virtual terminal stuff out of kernel space (i.e. as currently is for the most part) and into userspace (i.e. what kmscon is supposed to do)? And if you get the VT stuff into kernel space, what do the userspace components communicate with? Do the interact directly with the kernel API and system calls?

          I'm asking these question with the best knowledge I have, may have gotten some terms wrong, but genuinely curious. I have heard about kmscon over the years, know it is an attempt to create a KMS-based virtual terminal implementation, but that really not going anywhere these days. So curious if otherwise still value in the concept. Thanks as always.

          Comment


          • #6
            Originally posted by ehansin View Post

            Serious question, because I just don't know. Is there a general sense that it makes sense to get the virtual terminal stuff out of kernel space (i.e. as currently is for the most part) and into userspace (i.e. what kmscon is supposed to do)? And if you get the VT stuff into kernel space, what do the userspace components communicate with? Do the interact directly with the kernel API and system calls?

            I'm asking these question with the best knowledge I have, may have gotten some terms wrong, but genuinely curious. I have heard about kmscon over the years, know it is an attempt to create a KMS-based virtual terminal implementation, but that really not going anywhere these days. So curious if otherwise still value in the concept. Thanks as always.
            as far as I know, the major limiting factor right now is that there is no real virtual terminal that will work. iirc all of the stuff needed for a userland virtual terminal to replace vt. and I think you can use cage + foot to get a decent experience for an alternative to kmscon. I myself want to really try it, I find that a lot of wayland servers need the entire ecosystem (gbm and xkbdcommon etc) and if possible, I really want to escape those. so I will be waiting quite a while lol.

            more info on it can be found here. a good starting place, iirc it's a wee bit dated now but it should be fine to get more or less somewhat caught up https://www.reddit.com/r/linux/comme...g_vtn_in_2023/

            Comment


            • #7
              Originally posted by ehansin View Post

              Serious question, because I just don't know. Is there a general sense that it makes sense to get the virtual terminal stuff out of kernel space (i.e. as currently is for the most part) and into userspace (i.e. what kmscon is supposed to do)? And if you get the VT stuff into kernel space, what do the userspace components communicate with? Do the interact directly with the kernel API and system calls?

              I'm asking these question with the best knowledge I have, may have gotten some terms wrong, but genuinely curious. I have heard about kmscon over the years, know it is an attempt to create a KMS-based virtual terminal implementation, but that really not going anywhere these days. So curious if otherwise still value in the concept. Thanks as always.
              The kernel VT is obsolete and needs a user-space VT replacement. But there is no agreement on the best user-space VT or how to integrate it with the system.

              The user-space VT uses KMS and DRM device nodes and ioctls to talk to the kernel. It also uses protocols like TTY, consolekit, and utmp to talk to other user-space components.
              ‚Äč

              Comment


              • #8
                I find it pretty sad that the in-kernel console is being removed. There should be a way to interact with the kernel even if userspace blows up.

                Comment


                • #9
                  maybe v6.7 will be an LTS release instead of v6.6.

                  Comment


                  • #10
                    With all these new changes I cant wait for other people to test it first

                    Comment

                    Working...
                    X