[SOLVED] No signal on monitor after kernel update, amdgpu

[ralienpp]

I have the following set-up:

• HP T630 thin client, with integrated AMD graphics and DisplayPort input
• I use a DP-HDMI adapter to connect a TV to this computer
• Running Linux Mint 21.3 with automatic updates enabled
• Timeshift backups are disabled

It seems that a number of kernel updates were applied throughout the uptime of the system. At some point it was rebooted and a new kernel was loaded - the display does not work. When the system boots - I see the regular boot output on the console, but when the desktop session is about to start- the entire screen goes green, then dark, then green again, then "no signal". This happened in the past and I solved it by reverting to an older version of the kernel.

However, this time something is different. Either there were multiple kernel updates while the computer was online (the uptime was close to a year), hence the "keep previous version of kernel around" principle doesn't help, because after the reboot the "previous kernel version" was not a version that was actually loaded and used in the past.



I examined my apt history to check what kernel versions were installed in the past and went for something old, e.g., 5.15.0-78-generic. If I install it and boot into that - the graphical desktop is loaded, but it operates at a ridiculously low resolution and is practically unusable. I cannot attempt to change the resolution because the GUI menus in Mate are so big that there's no way to navigate to the settings.


After reading these forums I decided to go ahead and figure out how to do it properly with new kernel versions. The solutions described by others imply booting the kernel with `amdgpu.dc=1`, but this has no effect in my case. I therefore ask for assistance in understanding and addressing the root cause of the problem.



As a starting point, the output of `inxi -Fxxxrz`, when executed over SSH, is:

Code: Select all

System: 
  Kernel: 5.15.0-101-generic x86_64 bits: 64 compiler: gcc v: 11.4.0 Console: pty pts/1
    DM: LightDM 1.30.0 Distro: Linux Mint 21.3 Virginia base: Ubuntu 22.04 jammy
Machine:
  Type: Desktop System: HP product: HP t630 Thin Client v: N/A serial: <superuser required>
    Chassis: type: 3 serial: <superuser required>
  Mobo: HP model: 8158 v: A01 serial: <superuser required> UEFI: AMI v: M40 v01.14
    date: 01/04/2023
CPU:
  Info: quad core model: AMD Embedded G-Series GX-420GI Radeon R7E bits: 64 type: MT MCP
    smt: enabled arch: Excavator rev: 1 cache: L1: 320 KiB L2: 2 MiB
  Speed (MHz): avg: 1554 high: 1792 min/max: 900/2000 boost: enabled cores: 1: 1792 2: 1396
    3: 1593 4: 1436 bogomips: 15969
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
Graphics:
  Device-1: AMD Wani [Radeon R5/R6/R7 Graphics] vendor: Hewlett-Packard driver: amdgpu v: kernel
    ports: active: DP-1 empty: DP-2 bus-ID: 00:01.0 chip-ID: 1002:9874 class-ID: 0300
  Device-2: Logitech C920 HD Pro Webcam type: USB driver: snd-usb-audio,uvcvideo
    bus-ID: 1-1.4:3 chip-ID: 046d:0892 class-ID: 0102 serial: <filter>
  Display: server: X.org v: 1.21.1.4 with: Xwayland v: 22.1.1 compositor: marco v: 1.26.0
    driver: X: loaded: amdgpu,ati unloaded: fbdev,modesetting,vesa gpu: amdgpu tty: 158x51
  Monitor-1: DP-1 model: LG (GoldStar) TV serial: <filter> res: 1920x1080 dpi: 42
    size: 1150x650mm (45.3x25.6") diag: 1321mm (52") modes: max: 1920x1080 min: 720x400
  Message: GL data unavailable in console. Try -G --display

Some potentially-relevant lines from the kernel log, `dmesg | grep amdgpu`:

Code: Select all

[    0.000000] Command line: BOOT_IMAGE=/boot/vmlinuz-5.15.0-101-generic root=/dev/mapper/vgmint-root ro amdgpu.dc=1 quiet splash
[    0.086441] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-5.15.0-101-generic root=/dev/mapper/vgmint-root ro amdgpu.dc=1 quiet splash
[    5.199024] [drm] amdgpu kernel modesetting enabled.
[    5.216331] amdgpu: Topology: Add APU node [0x0:0x0]
[    5.216722] amdgpu 0000:00:01.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
[    5.231788] amdgpu 0000:00:01.0: amdgpu: Fetched VBIOS from ROM BAR
[    5.231796] amdgpu: ATOM BIOS: 113-C75100-X28
[    5.232128] amdgpu 0000:00:01.0: vgaarb: deactivate vga console
[    5.232226] amdgpu 0000:00:01.0: amdgpu: VRAM: 1024M 0x000000F400000000 - 0x000000F43FFFFFFF (1024M used)
[    5.232233] amdgpu 0000:00:01.0: amdgpu: GART: 1024M 0x000000FF00000000 - 0x000000FF3FFFFFFF
[    5.232400] [drm] amdgpu: 1024M of VRAM memory ready
[    5.232404] [drm] amdgpu: 3072M of GTT memory ready.
[    5.236184] amdgpu: hwmgr_sw_init smu backed is smu8_smu
[    5.239588] amdgpu: smu version 18.01.00
[    5.544483] kfd kfd: amdgpu: Allocated 3969056 bytes on gart
[    5.544897] amdgpu: Topology: Add APU node [0x9874:0x1002]
[    5.544903] kfd kfd: amdgpu: added device 1002:9874
[    5.544923] amdgpu 0000:00:01.0: amdgpu: SE 1, SH per SE 1, CU per SH 8, active_cu_number 6
[    5.548133] fbcon: amdgpudrmfb (fb0) is primary device
[    5.548141] amdgpu 0000:00:01.0: [drm] fb0: amdgpudrmfb frame buffer device
[    5.575395] [drm] Initialized amdgpu 3.42.0 20150101 for 0000:00:01.0 on minor 0
[   11.641623] snd_hda_intel 0000:00:01.1: bound 0000:00:01.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])

And `lspci -nn | grep VGA` says:

Code: Select all

00:01.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Wani [Radeon R5/R6/R7 Graphics] [1002:9874] (rev 88)

The display is a 42" TV that normally works at 1920x1080, the computer is used as a "family box" for watching films and doing video calls. Although it cannot be used as a desktop now, I can still access it over SSH or use the self-hosted services running on it (e.g., SMB, web server, etc.).

I am looking forward to your tips.

[SMG]

Graphics:
Device-1: AMD Wani [Radeon R5/R6/R7 Graphics] vendor: Hewlett-Packard driver: amdgpu v: kernel
ports: active: DP-1 empty: DP-2 bus-ID: 00:01.0 chip-ID: 1002:9874 class-ID: 0300

I am looking forward to your tips.

Please see AMD Stoney and Wani graphics and problems with 5.15.0-85 and newer 5.15 kernels which is a sticky topic at the top of the Graphics Cards & Monitors forum.

[ralienpp]

Thank you for the quick reply and for the pointers. Also, it is a shame that I didn't find those existing topics myself, even though I did a search before writing my message.


For the benefit of archaeologists looking into this matter, these are the key points and quirks I dealt with:

• I installed kernel 6.5.0-26.
• This didn't work out initially because I installed the package linux-image-6.5.0-26-generic and assumed that must be all.
• However booting into this kernel did not do the trick: video still didn't work, WiFi didn't work either, and probably a bunch of other things too.
• Other packages need to be installed as well: linux-modules-extra-6.5.0-26-generic linux-headers-6.5.0-26-generic linux-headers-generic-6.5.0-26 linux-image-generic-hwe-22.04 . This list may be redundant, I am merely documenting what worked for me.
• The reason why downgrading to 5.15.0-84 didn't work is because of the way I installed it - just the linux-image package, without all the aforementioned ones.
• If one were to install a kernel via mintupdate it would have taken care of this - but since I was operating in "recovery mode", I was doing it by hand and at that time I had no awareness of the need for having the other packages installed too.

[SMG]

[*]If one were to install a kernel via mintupdate it would have taken care of this - but since I was operating in "recovery mode", I was doing it by hand and at that time I had no awareness of the need for having the other packages installed too.

There are what I refer to as two different recovery modes. One is a blue screen recovery mode where the system is not working properly and only terminal access is available.

The other can be accessed from the Advanced Options in the grub menu and then selecting a kernel with (recovery mode) in the line item. In that recovery mode, one can access Update Manager to install the kernel which does install the additional required kernel packages.

Glad to hear you now have it working.