Announcement

Collapse
No announcement yet.

Debian's MIPS64EL CPU Port Is At Risk Due To Declining Hardware Access

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

  • #11
    Originally posted by OneTimeShot View Post
    Can’t they build this in qemu? It doesn’t really make sense to support architectures that do not have a good emulator.
    What would be the point, if less and less people use those machines anyway?

    Comment


    • #12
      Originally posted by rene View Post
      cool, guess one day everyone will just use T2 Linux ;-) https://t.ly/giNHP
      Hopefully T2 something else than Linux.

      I think, by now, you should be well-informed Linux is, by far, not the end game or even sustainable long term.

      The codebase is a mess, the lack of driver APIs causes a lot of overhead, and maintainers routinely touch code all over the place in drivers and not test their changes, routinely breaking them. I am pretty sure you're familiar with it.

      Comment


      • #13
        Originally posted by ayumu View Post

        Hopefully T2 something else than Linux.

        I think, by now, you should be well-informed Linux is, by far, not the end game or even sustainable long term.

        The codebase is a mess, the lack of driver APIs causes a lot of overhead, and maintainers routinely touch code all over the place in drivers and not test their changes, routinely breaking them. I am pretty sure you're familiar with it.
        you sound like me, do you happen to be one of my YT followers and know about my secret micro kernel dreams? ;-)

        Comment


        • #14
          Originally posted by iustinp View Post

          Sigh. Feel free to step up and maintain it yourself then. It's soo easy to comment from the sidelines, no?

          (A Debian maintainer, who has had to deal with obscure arch support once or twice)

          I have zero interest in MIS64EL, but if I did, I wouldn't be using the excuse "I can't find enough hardware to run build processes". I would still need to test on real hardware, but for everything else, I'd cross-compile or run a build process in a QEMU instance...

          Heck, am I the only person who worked on the 8 different architectures that WinCE needed?

          Comment


          • #15
            Originally posted by CommunityMember View Post
            I believe the Debian system administration team requires their official buildd machines to be physical systems (along with other common datacenter requirements, so, essentially, "server class" hardware).
            So you're saying server hardware with MIPS CPUs is/was a thing that you could actually buy as a commercial product? Could you share some links or product names? I only know MIPS from SoCs in embedded hardware (mostly networking stuff running OpenWRT) and from the Nintendo 64. I've always wondered what the use cases for Debian's MIPS ports actually were.

            Comment


            • #16
              Originally posted by ayumu View Post

              Hopefully T2 something else than Linux.

              I think, by now, you should be well-informed Linux is, by far, not the end game or even sustainable long term.

              The codebase is a mess, the lack of driver APIs causes a lot of overhead, and maintainers routinely touch code all over the place in drivers and not test their changes, routinely breaking them. I am pretty sure you're familiar with it.
              AI-guided kernel optimizations are 1-3 years away. If implemented correctly a lot of the maintenance could be handled by an LLM for human final review.

              Comment


              • #17
                Originally posted by rene View Post

                you sound like me, do you happen to be one of my YT followers and know about my secret micro kernel dreams? ;-)
                I am following you on YT, sure. Although I found out all these things independently ages ago.

                I happen to own a Blade 2000, maintained some related patches popular in Gentoo over a decade ago.

                Linux is unsustainable, plainly a waste of developer time by design. Imagine where we would be if we stopped trying to make Linux work and put that effort into furthering other operating systems instead, be it DragonflyBSD, Genode, Haiku or the newish seL4 runtime efforts.

                Comment


                • #18
                  accidental double post from mobile browser

                  Comment


                  • #19
                    Originally posted by OneTimeShot View Post


                    I have zero interest in MIS64EL, but if I did, I wouldn't be using the excuse "I can't find enough hardware to run build processes". I would still need to test on real hardware, but for everything else, I'd cross-compile or run a build process in a QEMU instance...

                    Heck, am I the only person who worked on the 8 different architectures that WinCE needed?
                    Oh my gosh. NEC MobilePro 790?! 😍😫

                    Did that thing ever get a working Debian MIPS Version, or was it too RAM-limited?

                    (NOW you have me curious about WinCE's 8x architectures)

                    Comment


                    • #20
                      If these CPU are still manufactured thus the same manufacturers should the ones interested in having solid linux support to begin with, and providing a couple of machines to the Debian project, supporting Debian means that also Ubuntu will be benefit.

                      Comment

                      Working...
                      X