No announcement yet.

Linux x86 Ready To Remove Its Old 32-bit a.out Support

  • Filter
  • Time
  • Show
Clear All
new posts

  • Linux x86 Ready To Remove Its Old 32-bit a.out Support

    Phoronix: Linux x86 Ready To Remove Its Old 32-bit a.out Support

    Going along with the recent patches to stop building a.out support for Linux's Alpha and m68k architecture ports as the last of the CPU architectures that were still building the kernel with the support enabled, developers are ready to remove the x86 a.out support outright...

    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
    What even is the point of this? Is there more to this than a relatively small 350-ish line diff?


    • #3
      removing useless maintenance burden, reducing attack surfaces for unknown exploits, diminished risk that other legacy internal kernel interfaces can't be scraped because there is a bunch of old stuff seemingly using it (but actually useless themselves), etc

      also try to imagine the big picture...

      if you apply the "useless but harmless >> stays in the codebase" thought process widely, not just to this specific case, then you're not talking about 350-ish lines of code, you're talking about an endless growth of gibberish/legacy code which at some point threatens making the useful stuff hard to locate, modify, distinguish from useless parts, etc

      notably this kind of story about removing old crust isn't infrequent here in phoronix, so a quick search might give an idea of the real volume steadily removed... and there are kernel development statistics for removed lines of code too (though that includes LOC removal of all sorts, not just due to removed features)

      there are also stories in other phoronix articles about real issues where eg: the kernel maintainer of the driver for an old hardware stops being able to keep maintaining it, and nobody else is available to step up for years... and here's a tricky part: if you let things come to that point where nobody alive is involved/understands that bit of the kernel well and has time to ponder on it, who then will be able to accurately decide if it's ok to remove?

      all that being said, the linux kernel is a paradise for old stuff being supported for a long time, compared to windows, android, iOS, as well as device-specific OSs made by their manufacturers, which is awesome especially after realizing just how much stuff is actually supported by linux nowadays


      • #4
        Michael, can you be anymore desperate for clicks.


        • #5
          I really hope aliens don't check Phoronix forums to measure human intelligence.

          I feel dumber everytime I foolishly follow the link at the bottom of the article to get a better insight of the situation.


          • #6


            • #7
              I didn't realise it was still supported.


              • #8
                I hereby complain. I don't even know why, but it's too strange no one complained until now, so let's fix it.


                • #9
                  Originally posted by marlock View Post
                  ...this kind of story about removing old crust isn't infrequent...
                  cruft, not crust


                  • #10
                    Originally posted by Old Grouch View Post

                    cruft, not crust
                    Old crusts are just as terrible.