Announcement

Collapse
No announcement yet.

Gigabyte's ASPM Motherboard Fix: Use Windows

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

  • #76
    Originally posted by agd5f View Post
    It's not really realistic for vendors to go back and validate old hardware for a feature which was "unsupported" at the time. It's a waste of resources since there's no new revenue attached to it.
    New revenue might depend on potential customer's previous experiences with them, so that's not entirely true...

    Comment


    • #77
      Originally posted by Nobu View Post
      Code:
      [    0.142522] pci 0000:02:00.0: disabling ASPM on pre-1.1 PCIe device.  You can enable it with 'pcie_aspm=force'
      [    0.143080] ACPI _OSC control for PCIe not granted, disabling ASPM
      ...must be time to upgrade. I guess I'll send the mobo to the coreboot people and graphics card to the nouveau people when I do. ^_^
      This means your BIOS is OK (or your kernel is not 2.6.38+) but the device itself indicates not to support ASPM. This is usually correct; enabling ASPM on those devices often causes a lot of trouble.
      This message also appears on pre-2.6.38 kernels, going back to at least 2.6.27.

      By the way, you can lookup in lspci which device is "02:00.0".
      Last edited by AlbertP; 10-23-2011, 05:45 AM.

      Comment


      • #78
        Originally posted by AlbertP View Post
        This means your BIOS is OK (or your kernel is not 2.6.38+) but the device itself indicates not to support ASPM. This is usually correct; enabling ASPM on those devices often causes a lot of trouble.
        This message also appears on pre-2.6.38 kernels, going back to at least 2.6.27.

        By the way, you can lookup in lspci which device is "02:00.0".

        Oh so it may be the DEVICE itself that doesn't support ASPM, rather than the mobo. In my opinion it is usually PCIe video cards have issues with ASPM, as putting such devices to sleep could very well leave you with a black screen when it is awoken.

        Comment


        • #79
          ASPM did not exist in the first PCIe specification: power management had to be sorted out by the drivers of those devices. PCIe 1.1 introduced this problematic feature. So there are a lot of cards out which do not yet have any ASPM built-in, and this is properly detected unless you use pcie_aspm=force.

          Comment


          • #80
            I'm using version 2.6.40.6, on Fedora 15. pci device 02:00.0 is my video card--I had verified that a while back, but didn't think it was important enough to edit my post.

            Probably unrelated: my monitor (almost?) always fails to enter suspend mode when I suspend my desktop--sometimes a vt is visible, sometimes the plymouth splash, and sometimes just black. Nothing unusual is ever logged (in kernel.log or xorg.0.log). I tried asking on IRC about it one or two times, but I usually have to leave before I get a reply and my connection is shaky.

            btw, what is the "_OSC control"?

            Comment


            • #81
              ASPM is not supported PCIe 1.0 devices. It was introduced with 1.1.
              This way a lot of older NVidia PCIe cards, at least up to GeForce 6 (not sure about GeForce 7), do not support it.
              When you do not use pcie_aspm=force, Linux correctly detects PCIe 1.0 and disables ASPM. It only affects this device and is not globally: so your BIOS is OK.

              The same thing happens on some (wired & wireless) networking cards, and many other early PCIe devices.

              Comment


              • #82
                Originally posted by deanjo View Post
                Most motherboard manufacturers only list Windows as supported operating systems in their product descriptions and manuals. If it works in another OS, it's a bonus but rarely supported.
                I don't care a bit about what OS they support. What I do care about is that they support the specifications. If they don't that is clearly a bug and should be fixed by them.

                Comment


                • #83
                  Originally posted by AlbertP View Post
                  ASPM is not supported PCIe 1.0 devices. It was introduced with 1.1.
                  This way a lot of older NVidia PCIe cards, at least up to GeForce 6 (not sure about GeForce 7), do not support it.
                  When you do not use pcie_aspm=force, Linux correctly detects PCIe 1.0 and disables ASPM. It only affects this device and is not globally: so your BIOS is OK.

                  The same thing happens on some (wired & wireless) networking cards, and many other early PCIe devices.
                  I have an nVidia GeForce 7900GS I bought in late 2007 and would that fall under 1.0 or 1.1 PCIe specs?

                  Comment


                  • #84
                    If Linux detects a PCIe pre-1.1 device, it is a PCIe pre-1.1 device and thus PCIe 1.0.

                    nVidia does not seem to have PCIe 1.1 cards so the first cards with ASPM were the PCIe 2.0 cards (certain GeForce 8 models).

                    Comment


                    • #85
                      Originally posted by Ragas View Post
                      I don't care a bit about what OS they support. What I do care about is that they support the specifications. If they don't that is clearly a bug and should be fixed by them.
                      That is true. But sadly nobody seems to really care about standards.
                      And some, like MS, create pseudo-standards (office open xml) with 6'000 pages but later not keeping to their own specs.
                      Then, in HW terms, there is ACPI (also made up by MS as one member of ACPI board and their faulty compilers for the tables) or other stuff. In the HW business it is probably often the case "just make W32 work with it and that's it". If Windows doesn't bitch forget the product and get on with the next generation of stuff. Short life/release cycles aren't doing much good here.
                      (free & open) Standards are a good thing, but it would be better for the world if people would be sticking with it.

                      Comment


                      • #86
                        "standards"

                        http://xkcd.com/927/

                        Comment


                        • #87
                          Perhaps Microsoft only helps developing standards like ACPI because they want to be sure sure they can mess up Windows' implementation without breaking it on existing ACPI-compatible computers.

                          Comment


                          • #88
                            since you bumped the thread...

                            Originally posted by AlbertP View Post
                            Perhaps Microsoft only helps developing standards like ACPI because they want to be sure sure they can mess up Windows' implementation without breaking it on existing ACPI-compatible computers.
                            Yes, that's what I understand after reading some old Gates emails. Some words like "We need to make Linux to not work" and things like that. It's all planned.

                            Comment


                            • #89
                              I haven't heard of the motherboard vendors mentioned before, and they didn't come up with I was in the market for a new one. I bought a Gigabyte Z68 one anyway, despite them mailing to me:

                              Thank you for your kindly mail and inquiry. About the issue you mentioned, since our products only support Windows OS, we cannot guarantee Ubuntu to work on our system. We suggest you to install Windows OS to prevent having problems. Sorry for the inconvenience.

                              Regards,
                              GIGABYTE
                              So this is a general company stance and not a one-off thing. It's working out fine so far though, and I didn't think there were any mb vendors who were actually good what they do otherwise.

                              Comment


                              • #90
                                I think Dell is one of the few companies who have released non-server computers (i.e. laptops and desktops) specifically for running linux?

                                Comment

                                Working...
                                X