Announcement

Collapse
No announcement yet.

In-Progress AMDGPU Updates For Linux 4.20~5.0 Have DC Update, New Polaris ID

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

  • In-Progress AMDGPU Updates For Linux 4.20~5.0 Have DC Update, New Polaris ID

    Phoronix: In-Progress AMDGPU Updates For Linux 4.20~5.0 Have DC Update, New Polaris ID

    Last week AMD sent in their big feature pull request of AMDGPU driver changes to DRM-Next for the Linux 4.20 (or what will likely be Linux 5.0) and since then more changes have been queuing in their work-in-progress branch...

    http://www.phoronix.com/scan.php?pag...WIP-Polaris-DC

  • #2
    I hope they can fix the DisplayPort related mess such as broken virtual terminal (tty) and etc.

    Comment


    • #3
      Originally posted by shmerl View Post
      I hope they can fix the DisplayPort related mess such as broken virtual terminal (tty) and etc.
      Are you talking about the WIP/RC or 4.18? Haven't noticed any DP-issues on my Rx580?

      Comment


      • #4
        Originally posted by Termy View Post
        Are you talking about the WIP/RC or 4.18? Haven't noticed any DP-issues on my Rx580?
        4.19.0-rc4. I have this problem with Vega 56.

        Comment


        • #5
          Originally posted by shmerl View Post

          4.19.0-rc4. I have this problem with Vega 56.
          Not had issues either on my 64 on the the amd-staging (4.19).
          What is it doing/not doing ?

          Comment


          • #6
            Coming back from 2 weeks of vacation I tried once more the current amd-staging-drm-next kernel. The now almost one year old instability (when dynamic power management is active) is still there, going into S3 sleep is completely broken, again, and as a bonus new scary kernel log messages are emitted upon every boot:

            kernel: [drm:gfx_v8_0_ring_test_ib [amdgpu]] *ERROR* amdgpu: IB test timed out.
            kernel: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* amdgpu: failed testing IB on ring 9 (-110).
            kernel: [drmrocess_one_work] *ERROR* ib ring test failed (-110).

            *Sigh*. I'm really looking forward to the new discrete Intel GPUs announced for 2020. Reading the commit logs does not give me any hope amdgpu will eventually become a stable driver, there is so much trial-and-error patching back-and-forth, it almost looks as if the hardware behavior has to be reverse-engineered.

            Comment


            • #7
              Originally posted by dwagner View Post
              Coming back from 2 weeks of vacation I tried once more the current amd-staging-drm-next kernel. The now almost one year old instability (when dynamic power management is active) is still there, going into S3 sleep is completely broken, again, and as a bonus new scary kernel log messages are emitted upon every boot:

              kernel: [drm:gfx_v8_0_ring_test_ib [amdgpu]] *ERROR* amdgpu: IB test timed out.
              kernel: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* amdgpu: failed testing IB on ring 9 (-110).
              kernel: [drmrocess_one_work] *ERROR* ib ring test failed (-110).

              *Sigh*. I'm really looking forward to the new discrete Intel GPUs announced for 2020. Reading the commit logs does not give me any hope amdgpu will eventually become a stable driver, there is so much trial-and-error patching back-and-forth, it almost looks as if the hardware behavior has to be reverse-engineered.
              I wouldn't expect Intels discrete GPU's to be all that wonderful upon arrival. It is a new space for Intel and let face it they don't have a GPU platform right now for writing high performance code thus are not highly skilled in this area.

              Beyond all of that I'm not going to knock AMD as this has been a tough year for systems developers due to Meltdown and Spectra. Tha thas been a huge detour for developer time. All I know is that my Ryzen based laptop gets better with each Fedora update. That is reassuring and means AMD limited developer talent has found time to iron out problems.

              Comment


              • #8
                Originally posted by pete910 View Post

                Not had issues either on my 64 on the the amd-staging (4.19).
                What is it doing/not doing ?
                Switching to tty (DisplayPort connection).

                See https://bugs.freedesktop.org/show_bug.cgi?id=107978

                I'm using Debian's kernel build, not AMD staging. If AMD can't coherently transfer patches to upstream kernel, there is something seriously wrong with the process.

                Comment


                • #9
                  Originally posted by wizard69 View Post
                  I wouldn't expect Intels discrete GPU's to be all that wonderful upon arrival. It is a new space for Intel and let face it they don't have a GPU platform right now for writing high performance code thus are not highly skilled in this area
                  Intel does produce GPUs right now, just integrated into their CPUs, and the Linux drivers for those iGPUs have been very reliable from my experience. Also, I am not looking for high-end performance from GPUs, I just need a stable GPUs to run along with Ryzen CPUs. I would gladly trade a 70% slower amdgpu driver that does not crash the system for the current one.

                  Comment


                  • #10
                    Originally posted by shmerl View Post

                    Switching to tty (DisplayPort connection).

                    See https://bugs.freedesktop.org/show_bug.cgi?id=107978

                    I'm using Debian's kernel build, not AMD staging. If AMD can't coherently transfer patches to upstream kernel, there is something seriously wrong with the process.
                    How do you know it's not a bug in the debian kernel build ?

                    Odd one for sure.

                    Is it the same on rc5 ?

                    Comment

                    Working...
                    X