Announcement

Collapse
No announcement yet.

A Proper Solution To The Linux ASPM Problem

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

  • A Proper Solution To The Linux ASPM Problem

    Phoronix: A Proper Solution To The Linux ASPM Problem

    At long last, it looks like there is an adequate solution to the Active State Power Management (ASPM) problem in the Linux kernel , a.k.a. the well-known and wide-spread power regression in the Linux 2.6.38 kernel, which has been causing many laptops to go through significantly more power than they should. This is not another workaround, but rather a behavioral change in the kernel to better decide when the PCI Express ASPM support should be toggled.

    http://www.phoronix.com/vr.php?view=16682

  • #2
    What a relief. I am happy today

    Comment


    • #3
      Finally.

      Actually I was planning to e-mail LKML with my solution to the problem, a white-list of ASPM supported devices on computers with a proper BIOS, which could be collected through Phoronix Test Suite.

      Comment


      • #4
        Other regressions

        [quote]
        That's one regression now addressed, with only several more power management issues left in the kernel...
        [\quote]

        Did you find those other regressions with git-bisect?

        Comment


        • #5
          'Bout fuckin' time! Seriously, this should have been fixed way earlier. If you introduce a regression, you revert to previous version, too bad the kernel devs don't get that...

          Comment


          • #6
            Originally posted by LinuxID10T View Post
            'Bout fuckin' time! Seriously, this should have been fixed way earlier. If you introduce a regression, you revert to previous version, too bad the kernel devs don't get that...
            That's what they did do. They introduced a power regression to get rid of a stability regression. It was the lesser of two evils until they came up with a proper solution ie try any copy what windows does

            Comment


            • #7
              Michael what are the other power regressions you mention in the article?

              Comment


              • #8
                Originally posted by 89c51 View Post
                Michael what are the other power regressions you mention in the article?
                I'm not permitted to publicly talk about them at the moment.
                Michael Larabel
                http://www.michaellarabel.com/

                Comment


                • #9
                  Originally posted by Michael View Post
                  I'm not permitted to publicly talk about them at the moment.
                  Huh? Is some kind of vendor who forbids this???

                  Comment


                  • #10
                    Originally posted by Michael View Post
                    I'm not permitted to publicly talk about them at the moment.
                    weird.

                    at least can you answer if they are taken care of???

                    Comment


                    • #11
                      Originally posted by 89c51 View Post
                      at least can you answer if they are taken care of???
                      In the mainline kernel, no, I haven't seen any patches.
                      Michael Larabel
                      http://www.michaellarabel.com/

                      Comment


                      • #12
                        Originally posted by FireBurn View Post
                        That's what they did do. They introduced a power regression to get rid of a stability regression. It was the lesser of two evils until they came up with a proper solution ie try any copy what windows does
                        I hardly see it being the "lesser of two evils." Considering the issue only effected a very, very small percentage of people, they killed battery life for anyone with a laptop. So no, I don't see it as the lesser of two evils. My laptop went from 5 hours of battery life to 2.5 hours, how is that not more serious considering how many people it effects?

                        Comment


                        • #13
                          This is so widespread and well known that the only person in the universe who bitches about it is.... Michael from Phoronix. For everybody else, its "meh."

                          Comment


                          • #14
                            The change in the Linux 2.6.38 kernel disabled ASPM unless the BIOS advertised support for it, but it turns out a vast number of systems supporting ASPM do not actually advertise it from the BIOS
                            What about PCIe hotplug support? I've tested hotplugging on a few Intel desktop and server motherboards with an ExpressCard adapter board. Strangely I could get hotplug going on a Core2Duo board (with PCIEHP force option) but not on a Corei7 board.

                            The only way I could get ExpressCard hotplug working reliably was using ACPI hotplugging on laptop ExpressCard slots...

                            Comment


                            • #15
                              Originally posted by droidhacker View Post
                              This is so widespread and well known that the only person in the universe who bitches about it is.... Michael from Phoronix. For everybody else, its "meh."
                              I am pretty sure other people bitch about it too. As to the kernel devs reaction to Michael's article wasn't great. I will admit, Michael does sometimes blow an issue out of proportion, however, as a laptop user, in this case I completely agree with him. It was a pretty damn major regression and the kernel devs sat there scratching their asses for months rather than fixing the damn issue. I know there was questions as to how to go about fixing the issue, but in all reality, it was placed on the backburner. The fact that a mere 60 lines of code fixed this goes to show how easy it should have been.

                              Comment

                              Working...
                              X