Announcement

Collapse
No announcement yet.

Need help fixing RX460 with potentially bad drivers

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

  • #11
    I think I managed to disable the onboard gpu fully by setting some other thing to auto.
    Then I got the 4.11.9 kernel from here.
    Then I installed this firmware.

    Froze with black screen. Here's the log: https://pastebin.com/sbNzyftb

    I search about the "flip_done timed out" and some suggested booting with video=SVIDEO-1:d flag but I got the same result.

    What else can I update or remove?

    Comment


    • #12
      Originally posted by destingten View Post
      I think I managed to disable the onboard gpu fully by setting some other thing to auto.
      Then I got the 4.11.9 kernel from here.
      It is better to compile your own kernel package for your own hardware than use some third party packages. I just installed Debian testing Xfce to a microsd card to have a test partition and with rx460 firmware installed, it runs fine with stock Debian testing 4.11 kernel.
      Last edited by debianxfce; 07-17-2017, 02:58 AM.

      Comment


      • #13
        Originally posted by debianxfce View Post

        It is better to compile your own kernel package for your own hardware than use some third party packages. I just installed Debian testing Xfce to a microsd card to have a test partition and with rx460 firmware installed, it runs fine with stock Debian testing 4.11 kernel.
        Alright so let me get this right.
        I'll install the Debian testing on another partition (using HD4670, since 460 gives me artefacts and freezes when I tried it)
        Then I'll add deb http://ftp.de.debian.org/debian buster main non-free to /etc/apt/sources.list
        Then I'll run apt-get install firmware-amd-graphics
        Then poweroff, switch to 460 and it should work?

        Comment


        • #14
          Originally posted by destingten View Post

          Alright so let me get this right.
          I'll install the Debian testing on another partition (using HD4670, since 460 gives me artefacts and freezes when I tried it)
          Then I'll add deb http://ftp.de.debian.org/debian buster main non-free to /etc/apt/sources.list
          Then I'll run apt-get install firmware-amd-graphics
          Then poweroff, switch to 460 and it should work?
          Yes, and change the debian main repository from buster to testing. Then it is rolling after 3 years without change.

          Comment


          • #15
            Installed the debian testing, installed the firmware from testing repository.

            Got black screens: https://pastebin.com/raw/3Ba4wHFt

            Tried with the SVIDEO flag, got white screens: https://pastebin.com/raw/MEAUZCc5

            Is this card dead or what?

            Comment


            • #16
              Originally posted by destingten View Post
              Installed the debian testing, installed the firmware from testing repository.

              Got black screens: https://pastebin.com/raw/3Ba4wHFt

              Tried with the SVIDEO flag, got white screens: https://pastebin.com/raw/MEAUZCc5

              Is this card dead or what?
              I installed stock Debian 4.11 kernel to my system, and it is running now fine. I have the firmware-amd-graphics package installed but I have copied latest firmware from here:
              https://people.freedesktop.org/~agd5...ucode/polaris/

              Code:
              screenfetch
                       _,met$$$$$gg.           xfce@carrizo
                    ,g$$$$$$$$$$$$$$$P.        OS: Debian testing buster
                  ,g$$P""       """Y$$.".      Kernel: x86_64 Linux 4.11.0-1-amd64
                 ,$$P'              `$$$.      Uptime: 5m
                ',$$P       ,ggs.     `$$b:    Packages: 2352
                `d$$'     ,$P"'   .    $$$     Shell: bash 4.4.12
                 $$P      d$'     ,    $$P     Resolution: 1920x1200
                 $$:      $$.   -    ,d$$'     DE: XFCE
                 $$\;      Y$b._   _,d$P'      WM: Xfwm4
                 Y$$.    `.`"Y$$$$P"'          WM Theme: Default
                 `$$b      "-.__               GTK Theme: Xfce [GTK2]
                  `Y$$                         Icon Theme: Tango
                   `Y$$.                       Font: Sans 10
                     `$$b.                     CPU: AMD Athlon X4 845 Quad Core @ 4x 3.5GHz [34.0°C]
                       `Y$$b.                  GPU: AMD Radeon (TM) RX 460 Graphics (AMD POLARIS11 / DRM 3.10.0 / 4.11.0-1-amd64, LLVM 4.0.1)
                          `"Y$b._              RAM: 711MiB / 7929MiB
                              `""""

              You had black screen with win10 too, maybe your gpu is broken.
              Last edited by debianxfce; 07-17-2017, 09:25 AM.

              Comment


              • #17
                Originally posted by debianxfce View Post
                You had black screen with win10 too, maybe your gpu is broken.

                Well I figured there was some sort of faulty firmware that was loaded onto the GPU and that's why it didn't boot anywhere. Maybe I need to flash it or something. Or is that volatile and re-added by the OS every time I boot?

                Is there any way to test GPUs for hardware problems? It seems to render things like bios and gnome password/login screen fine, until it doesn't.

                If I Google these errors there seems to be many bug reports about, and very recent at that.
                Code:
                amdgpu: [powerplay] Failed to send Message.
                amdgpu: [powerplay] SMU Firmware start failed!
                amdgpu: [powerplay] Failed to load SMU ucode.
                amdgpu: [powerplay] smc start failed
                [drm:amdgpu_device_init [amdgpu]] *ERROR* hw_init of IP block <amdgpu_powerplay> failed -22
                Code:
                amdgpu: [powerplay] [AVFS] Something is broken. See log!
                amdgpu: [powerplay] Can't find requested voltage id in vdd_dep_on_sclk table!
                Also where is that powerplay or AVFS log supposed to be located?

                Comment


                • #18
                  destingten: "It seems to render things like bios and gnome password/login screen fine, until it doesn't."

                  Try Xfce and lightdm, gnome software is buggy. Make it to boot to desktop automatically by doing this:
                  RealUserName=yourusername
                  sed -i "s/#autologin-user-timeout=\?/autologin-user-timeout=0/" /etc/lightdm/lightdm.conf
                  sed -i "s/#autologin-user=\?/autologin-user=${RealUserName}/" /etc/lightdm/lightdm.conf


                  The first one messages are not normal, but second messages are not harmful. I have them too.

                  [ 3.442022] amdgpu: [powerplay] [AVFS] Something is broken. See log!
                  [ 3.444859] amdgpu: [powerplay] Can't find requested voltage id in vdd_dep_on_sclk table!

                  What brand is your gpu card?

                  Here is amdgpu dmesg messages in system that works:
                  Code:
                   dmesg | grep amdgpu
                  [    3.151371] [drm] amdgpu kernel modesetting enabled.
                  [    3.191627] amdgpu 0000:01:00.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffff
                  [    3.206283] amdgpu 0000:01:00.0: VRAM: 2048M 0x000000F400000000 - 0x000000F47FFFFFFF (2048M used)
                  [    3.206285] amdgpu 0000:01:00.0: GTT: 3072M 0x0000000000000000 - 0x00000000BFFFFFFF
                  [    3.206371] [drm] amdgpu: 2048M of VRAM memory ready
                  [    3.206372] [drm] amdgpu: 3072M of GTT memory ready.
                  [    3.206555] amdgpu 0000:01:00.0: amdgpu: using MSI.
                  [    3.206581] [drm] amdgpu: irq initialized.
                  [    3.376786] amdgpu: [powerplay] amdgpu: powerplay sw initialized
                  [    3.380519] amdgpu 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000000000008, cpu addr 0xffff8802139fa008
                  [    3.380558] amdgpu 0000:01:00.0: fence driver on ring 1 use gpu addr 0x0000000000000018, cpu addr 0xffff8802139fa018
                  [    3.380590] amdgpu 0000:01:00.0: fence driver on ring 2 use gpu addr 0x0000000000000028, cpu addr 0xffff8802139fa028
                  [    3.380617] amdgpu 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000000000038, cpu addr 0xffff8802139fa038
                  [    3.380642] amdgpu 0000:01:00.0: fence driver on ring 4 use gpu addr 0x0000000000000048, cpu addr 0xffff8802139fa048
                  [    3.380670] amdgpu 0000:01:00.0: fence driver on ring 5 use gpu addr 0x0000000000000058, cpu addr 0xffff8802139fa058
                  [    3.380697] amdgpu 0000:01:00.0: fence driver on ring 6 use gpu addr 0x0000000000000068, cpu addr 0xffff8802139fa068
                  [    3.381974] amdgpu 0000:01:00.0: fence driver on ring 7 use gpu addr 0x0000000000000078, cpu addr 0xffff8802139fa078
                  [    3.382038] amdgpu 0000:01:00.0: fence driver on ring 8 use gpu addr 0x0000000000000088, cpu addr 0xffff8802139fa088
                  [    3.382061] amdgpu 0000:01:00.0: fence driver on ring 9 use gpu addr 0x000000000000009c, cpu addr 0xffff8802139fa09c
                  [    3.382904] amdgpu 0000:01:00.0: fence driver on ring 10 use gpu addr 0x00000000000000ac, cpu addr 0xffff8802139fa0ac
                  [    3.382957] amdgpu 0000:01:00.0: fence driver on ring 11 use gpu addr 0x00000000000000bc, cpu addr 0xffff8802139fa0bc
                  [    3.384613] amdgpu 0000:01:00.0: fence driver on ring 12 use gpu addr 0x000000f40082d420, cpu addr 0xffffc90001c5a420
                  [    3.385612] amdgpu 0000:01:00.0: fence driver on ring 13 use gpu addr 0x00000000000000dc, cpu addr 0xffff8802139fa0dc
                  [    3.385643] amdgpu 0000:01:00.0: fence driver on ring 14 use gpu addr 0x00000000000000ec, cpu addr 0xffff8802139fa0ec
                  [    3.442022] amdgpu: [powerplay] [AVFS] Something is broken. See log!
                  [    3.444859] amdgpu: [powerplay] Can't find requested voltage id in vdd_dep_on_sclk table!
                  [    3.448072] amdgpu: [powerplay]
                  [    3.448087] amdgpu: [powerplay]
                  [    4.239992] fbcon: amdgpudrmfb (fb0) is primary device
                  [    4.470747] amdgpu 0000:01:00.0: fb0: amdgpudrmfb frame buffer device
                  [    4.472480] [drm] Initialized amdgpu 3.17.0 20150101 for 0000:01:00.0 on minor 0
                  Last edited by debianxfce; 07-17-2017, 11:18 PM.

                  Comment


                  • #19
                    Originally posted by destingten View Post
                    Well I figured there was some sort of faulty firmware that was loaded onto the GPU and that's why it didn't boot anywhere. Maybe I need to flash it or something. Or is that volatile and re-added by the OS every time I boot?
                    Microcode is volatile and loaded every time you boot by the driver.

                    Code:
                    Jul 17 07:37:53 debian-testing kernel: [ 4.920718] amdgpu 0000:04:00.0: firmware: direct-loading firmware amdgpu/polaris11_smc.bin
                    Jul 17 07:37:53 debian-testing kernel: [ 6.011860] amdgpu: [powerplay] Failed to send Message.
                    Jul 17 07:37:53 debian-testing kernel: [ 6.011865] amdgpu: [powerplay] SMU Firmware start failed!
                    Jul 17 07:37:53 debian-testing kernel: [ 6.014016] amdgpu 0000:04:00.0: firmware: direct-loading firmware amdgpu/polaris11_smc_sk.bin
                    This looks suspicious... I don't *think* Polaris 11 needs two different SMC microcode images loaded (it's supposed to be one or the other AFAIK) but will check. Can you please pastebin the entire dmesg output, and also xorg log if you can get to it easily.

                    Comment


                    • #20
                      Tried with xfce-lightdm
                      added deb http://ftp.de.debian.org/debian testing main non-free
                      installed firmware-amd-graphics

                      Got white screens
                      https://pastebin.com/raw/biu0nPCx

                      Code:
                      Can't find requested voltage id in vdd_dep_on_sclk table!
                      Is that the problem? My PSU is 500W. Maybe the GPU is fried then or something, but it does seem many people have that error.

                      The card is MSI RX 460 OC. When it was working initially I did run into some problems under load, I don't know if it was the cheap adapter I was using, but I underclocked it to normal values and that was stable. Then I switched to Ubuntu and it worked as is, until adapter started turning the screen off repeatedly, so I used the other card for a while and now this one doing all that.

                      Is it possible to flash it or something to get rid of the overclock?

                      Sorry I don't know how to get the dmesg, I can only get into recovery mode with nomodeset so that wouldn't give any info right?

                      Comment

                      Working...
                      X