Announcement

Collapse
No announcement yet.

Radeon & AMDGPU DRM Fixes Queue Up For Linux 4.2

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

  • Radeon & AMDGPU DRM Fixes Queue Up For Linux 4.2

    Phoronix: Radeon & AMDGPU DRM Fixes Queue Up For Linux 4.2

    While the Linux 4.2 kernel merge window isn't yet over, AMD developers already have amassed a collection of fixes for the Radeon and AMDGPU DRM drivers following the big DRM pull...

    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
    Something in recent releases (a 4.0.x dot release, and maybe Mesa 10.6) broke my suspend and screen energy saving, and I've had to turn both off. And then there are Radeon suspend fixes in 4.2.... looking hopeful.

    Strangely my display energy saving bug only happens on my 144hz monitor over DVI. My other panels suspend / resume fine, my 144hzer has something break in software so that it just outputs black. System is still acting like there is a screen there, and the screen never suspends like its not getting a signal. We'll see if its fixed in one of these newer kernels... it used to work fine!

    Comment


    • #3
      VCE improvements are great. Now I just wish OBS supported VCE.

      Comment


      • #4
        Haven't tried that yet, not sure if this works on open source driver. Hope someone with experience can clarify this XD
        Github repo for the branch. I've implemented some support for AMD VCE. Probably all kinds of broken and memory leaky, but atleast something to build...

        Comment


        • #5
          Originally posted by zanny View Post
          Something in recent releases (a 4.0.x dot release, and maybe Mesa 10.6) broke my suspend and screen energy saving, and I've had to turn both off. And then there are Radeon suspend fixes in 4.2.... looking hopeful.

          Strangely my display energy saving bug only happens on my 144hz monitor over DVI. My other panels suspend / resume fine, my 144hzer has something break in software so that it just outputs black. System is still acting like there is a screen there, and the screen never suspends like its not getting a signal. We'll see if its fixed in one of these newer kernels... it used to work fine!
          Did you try reporting the bugs?

          Comment


          • #6
            Originally posted by extraymond View Post
            Haven't tried that yet, not sure if this works on open source driver. Hope someone with experience can clarify this XD
            As far as I get it's only useful for proprietary driver on Windows.
            On Linux for open source driver you need OpenMAX support.

            Comment


            • #7
              Could you post a link to the radeon bugtracker? I know it's been posted before, but it's 2AM here and I need to get to bed.

              After updating my Arch system last night I've had an issue where upon resume if you changed the volume (ever, no matter how you do it) it starts hanging when you try to interact with windows and things, and the sound dies. At one point there was a kernel BUG and I had to reboot. The only fix is to log-out and log back in. I'm fairly certain it's a radeon/drm bug, since there are all kinds of errors from the driver after resume and audio is being output to my TV via HDMI.

              Won't be able to troubleshoot as I'll be out of town for a while, but here's my system info and log output, copypasta:
              Hardware:
              A10-6800k (iGPU is primary)
              Sapphire r9-270x dual-x 4GB
              Audio through TV via HDMI

              Software:
              KDE Plasma Desktop
              Firefox v38.0.6
              radeon x/mesa drivers from testing/lcarlier [mesa-git] repo
              [. . .]
              Additionally, got a kernel BUG (might as well have been a kernel panic--couldn't do anything but hard reset after) [one time] when I tried to log out after it happened: http://i47.photobucket.com/albums/f1...6/IMG_0267.jpg
              My journal log: https://www.dropbox.com/s/6bm3wigv9n...urnal.txt?dl=0

              Comment


              • #8
                Originally posted by Nobu View Post
                Could you post a link to the radeon bugtracker? I know it's been posted before, but it's 2AM here and I need to get to bed.

                Test signature

                Comment


                • #9
                  Bug 91192 filed, hopefully descriptive enough and in the correct location.

                  Comment


                  • #10
                    I think it sounds like a kernel issue which would be DRI then Radeon - but the dev's will normally fix it up for you. It's usally a good idea to attach the Xorg.0.log and the output of lspci to make it more obvious which card you have, also the output of xrandr (not sure of the options) to they can see the modes available on your monitor - especially if it's a particular mode / frequency that seem to trigger the issue

                    As always try the latest kernel to see if it fixes your issue and be prepared to bisect it

                    Comment

                    Working...
                    X