Announcement

Collapse
No announcement yet.

AMD Energy Driver Booted From The Linux 5.13 Kernel

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

  • AMD Energy Driver Booted From The Linux 5.13 Kernel

    Phoronix: AMD Energy Driver Booted From The Linux 5.13 Kernel

    While a lot of new features and improvements have been accumulating for the Linux 5.13 kernel with the ongoing merge window, one of the unfortunate aspects of this new kernel is that the AMD Zen CPU energy driver "amd_energy" is indeed being removed...

    https://www.phoronix.com/scan.php?pa...Energy-Removed

  • #2
    I'd propose a compromise of a privileged daemon that can access the info and then provide access to other users. This could be open to all users or limited based on policy. But that would also be outside the kernel itself so maybe it's not a viable solution.

    Comment


    • #3
      Queue birdie's comment condemning the audacity of a puny Linux maintainer who should know and accept his place as the lowly worm that he is, rather than challenge the will of a megacorp like AMD who should promptly be granted absolute freedom to dictate how Linux works because they're actually doing us a favor porting their drivers to our tiny 1% market share OS and we should be thanking them thrice daily for their selfless benevolence.

      In other news, I'm sure this driver will be back in the kernel and properly working (no root-only shenanigans) sooner rather than later.

      Comment


      • #4
        Long story short, if this person maintained any linux distribution, iotop would be removed from this distro's software repo.

        Comment


        • #5
          Originally posted by zxy_thf View Post
          Long story short, if this person maintained any linux distribution, iotop would be removed from this distro's software repo.
          Which would be a shame because iotop is a useful tool.

          Comment


          • #6
            bridgman Could you please provide more insight about this and if a solution to this mess is possible? I know you maybe are more into other AMD hardware than this, but you can ask someone about it or incentivize someone responsible at the drivers of this part of AMD hardware to explain more about this in this community.

            I'm an AMD user, both CPU and GPU. I don't want this to happen, it's a very bad compromise for it. I trust in AMD, I hope a proper solution for this is reached so users can use all AMD hardware features on Linux.

            I'm very sorry to say, but I hope this gets addressed in a proper and satisfactory way for all parties involved (including users of AMD hardware, both individuals and companies). Sincerely, I would sell all my AMD hardware and switch to Intel if I can't monitor my AMD hardware. I'm sure others would think the same. It's a shame, because even EPYC is also used in big iron these days and is becoming quite popular. I want competition of AMD and to succeed even more than now, it's very healthy for the x86 and GPU ecosystem.

            Thanks in advance!

            Comment


            • #7
              Originally posted by zxy_thf View Post
              Long story short, if this person maintained any linux distribution, iotop would be removed from this distro's software repo.
              iotop is not in the kernel

              AMD have fixed their hardware "problems", I hope this frees them for upping their software game.
              Last edited by bug77; 29 April 2021, 06:40 PM.

              Comment


              • #8
                Originally posted by timofonic View Post
                bridgman Could you please provide more insight about this and if a solution to this mess is possible? I know you maybe are more into other AMD hardware than this, but you can ask someone about it or incentivize someone responsible at the drivers of this part of AMD hardware to explain more about this in this community.
                I have pinged the developer to make sure that he is aware of this and to see what discussion if any has happened with our business/marketing folks. No promises that I will be able to report anything back but I have at least given it an initial kick.

                Removing the driver seems premature to me but it seems likely that there have been other discussions that I don't know about.
                Last edited by bridgman; 29 April 2021, 10:37 AM.
                Test signature

                Comment


                • #9
                  I can kinda see why the AMD engineers didn't agree to the suggestion. If you have the means to exploit the vulnerability this driver provides, you also probably have the means to exploit whatever is used to help slightly skew the results, making the "fix" moot.
                  So what I want to know is, do you need this driver in order for the random number generation (or whatever the security feature is) to work? Because if you do, then removing this driver is pointless.

                  Comment


                  • #10
                    Originally posted by zxy_thf View Post
                    Long story short, if this person maintained any linux distribution, iotop would be removed from this distro's software repo.
                    iotop is not part of the kernel.

                    Comment

                    Working...
                    X