Announcement

Collapse
No announcement yet.

The Linux Kernel Is Close To Enabling "-Wimplicit-fallthrough" By Default

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

  • The Linux Kernel Is Close To Enabling "-Wimplicit-fallthrough" By Default

    Phoronix: The Linux Kernel Is Close To Enabling "-Wimplicit-fallthrough" By Default

    The -Wimplicit-fallthrough compiler flag has been around since GCC 7 for warning over switch fall-through cases where it could lead to potential bugs / unexpected behavior if the programmer inadvertently forgot to add a "break" statement to a case. The Linux kernel is looking to soon enable this warning by default...

    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
    Originally posted by debianxfce View Post
    This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


    No fix from IBM in almost 2 years. Fix is in the bios.
    you know, there are more states than black and white - on and off - yes and no.

    Comment


    • #3
      Originally posted by debianxfce View Post
      This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


      No fix from IBM in almost 2 years. Fix is in the bios.
      This bug is really a bad example of things that went wrong for Ryzen users.

      Comment


      • #4
        Originally posted by debianxfce View Post
        This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


        No fix from IBM in almost 2 years. Fix is in the bios.
        Trolling fail.

        It says right in the article, "In the process a number of bugs have been addressed thanks to this compiler warning, some of which bugs have lasted in the mainline kernel for more than five years."

        Comment


        • #5
          Originally posted by debianxfce View Post
          This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


          No fix from IBM in almost 2 years. Fix is in the bios.
          To be correct the fix is not in the BIOS because the Linux kernel can in fact ignore all settings in bios or be the bios itself. The fix is in fact in the arch support code under hardware quirks.

          This is a bug triggered by IBM RCU code heavy usage of things but there is not a fault in the RCU code causing it. RCU that IBM puts in to the Linux kernel is in fact mathematically proofed this means bugs in RCU is insanely rare. The fault is in fact the AMD Ryzen having a buggy mwait instructions. So there was nothing wrong with the IBM made RCU code so IBM never has to patch that bug because there is nothing for them to patch. The party that has had to make the patch is in fact AMD and first version of patch appears in Linux kernel 5.0 in the arch section of the Linux kernel

          I really do wonder how Redhat developers will cope with the IBM requirements for mathematical proofs this is party why I suspect once the required 12 months is up you could see a lot of Redhat developers jump ship.

          Comment


          • #6
            Originally posted by debianxfce View Post
            This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


            No fix from IBM in almost 2 years. Fix is in the bios.
            Do you contribute anything to free software? If not, maybe don't flame people who do.

            Comment


            • #7
              I'm so used to other programming languages not just having this on by default but being enforced on the programmer.
              C is quite a beast. I'm glad they're moving towards verifiable code, it's the biggest thing L4 fans have been wanting for eons. Myself being one.

              Comment


              • #8
                Originally posted by debianxfce View Post
                This proves that IBM workers have nothing else to do than add /* fall through */ comments to the Linux kernel. This way they are the biggest contributor. No thinking, no innovation, only preventing the Linux desktop success. Here is IBM made RCU bug:


                No fix from IBM in almost 2 years. Fix is in the bios.
                Agreed! Personally, I think IBM synthesized Linus Torvalds as a sleeper agent to make Linux just so it could get popular and crush all our hopes and dreams when they destroy it with Wayland and Systemd.
                I'm not falling for it though, because I recently found this amazing operating system called FreeDOS. It can't run Xfce (big L right there) but it's faster and more lightweight than any IBM-muddled Linux distribution, now I'm using it right now (posting from an Arachne browser, which looks a lot like Xfce, so I'm cool with it)

                Comment


                • #9
                  Originally posted by Ironmask View Post
                  I'm so used to other programming languages not just having this on by default but being enforced on the programmer.
                  C is quite a beast. I'm glad they're moving towards verifiable code, it's the biggest thing L4 fans have been wanting for eons. Myself being one.
                  Yes, it's configured as a fatal error in my Java IDE... fall-through is permitted only for cases that do nothing apart from the fall-through.

                  Comment


                  • #10
                    In Rust switch matching is exhaustive

                    Comment

                    Working...
                    X