This morning I awoke to my PC unable to login.
At first, I wondered if I'd run something in DOSBox that had corrupted the disk... After finding out that was likely not the case, by using recovery console to SMART check and fsck my disks; I set about setting up proprietary EFI compatible memtest86. Nothing wrong with RAM either. I'm running a 2015 HP z640 server.
After a lot of digging, I found that disabling modesetting worked, but now my graphics were in the toilet. My main card is a xfx 5500XT 8GB GDDR5, so I figured that would hopefully not be a problem. I Powered down, removed the Radeon HD5450 1G DDR3, switched my second desktop screen to connect via DisplayPort->HDMI. Now my machine works.
This must be recent because I've had the PC setup with both GPU's since getting the new GPU over a year ago. The Computer is mostly off while I work as it sucks power (I think close to 800W)
System config snapshot from dpkg -l, lsb_release (filtered)
Ubuntu 18.04.5 LTS
If it matters, these are the kernel related packages
linux-generic-hwe-18.04 5.4.0.74.83~18.04.67
linux-headers-generic-hwe-18.04 5.4.0.74.83~18.04.67
linux-hwe-5.4-headers-5.4.0-73 5.4.0-73.82~18.04.1
linux-hwe-5.4-headers-5.4.0-74 5.4.0-74.83~18.04.1
linux-image-generic-hwe-18.04 5.4.0.74.83~18.04.67
linux-signed-generic-hwe-18.04 5.3.0.62.115
These relate to XOrg
xorg 1:7.7+19ubuntu7.1
xorg-docs-core 1:1.7.1-1.1
xorg-sgml-doctools 1:1.11-1
xserver-xorg-core-hwe-18.04 2:1.20.8-2ubuntu2.2~18.04.5
xserver-xorg-hwe-18.04 1:7.7+19ubuntu8~18.04.3
xserver-xorg-input-all-hwe-18.04 1:7.7+19ubuntu8~18.04.3
xserver-xorg-input-libinput-hwe-18.04 0.28.1-1~18.04.1
xserver-xorg-input-wacom-hwe-18.04 1:0.36.1-0ubuntu1~18.04.1
xserver-xorg-legacy-hwe-18.04 2:1.20.8-2ubuntu2.2~18.04.5
xserver-xorg-video-all-hwe-18.04 1:7.7+19ubuntu8~18.04.3
xserver-xorg-video-amdgpu-hwe-18.04 19.1.0-1~18.04.1
xserver-xorg-video-ati-hwe-18.04 1:19.1.0-1~18.04.1
xserver-xorg-video-fbdev-hwe-18.04 1:0.5.0-1ubuntu1~18.04.1
xserver-xorg-video-intel-hwe-18.04 2:2.99.917+git20171229-1ubuntu1~18.04.1
xserver-xorg-video-nouveau-hwe-18.04 1:1.0.16-1~18.04.1
xserver-xorg-video-qxl-hwe-18.04 0.1.5-2build2~18.04.1
xserver-xorg-video-radeon-hwe-18.04 1:19.1.0-1~18.04.1
xserver-xorg-video-vesa-hwe-18.04 1:2.4.0-1~18.04.1
xserver-xorg-video-vmware-hwe-18.04 1:13.3.0-2build1~18.04.1
These relate to the GPU / DRM
libdrm-radeon1:amd64 2.4.101-2~18.04.1
libdrm-radeon1:i386 2.4.101-2~18.04.1
libva-drm2:amd64 2.1.0-3
libva-drm2:i386 2.1.0-3
libva-wayland2:amd64 2.1.0-3
libva-x11-2:amd64 2.1.0-3
libva-x11-2:i386 2.1.0-3
libva2:amd64 2.1.0-3
libva2:i386 2.1.0-3
libvdpau1:amd64 1.1.1-3ubuntu1
libvidstab1.1:amd64 1.1.0-2~18.04.york0
Specifically the machine would decrypt hard drive, get as far as login screen, but freeze before displaying users. In recovery mode it could boot all the way through to login screen, login and work, but in 800x600 or 1024x768 mode.
Text mode booted, All EFI Firmware checks passed (HDD, RAM, self-test).
Graphical mode booted with modesetting disabled with and without recovery mode.
I had noticed the day prior that a firefox window coming through the second older GPU on one specific website would cause a screen to stop receiving signal, or stop displaying. I'd put that down to a linux quirk as it didn't freeze the machine and I was able to relocate the window to the main display, which didn't deactivate, but did cause the first display to reactivate.
Recent entries from apt log
Start-Date: 2021-06-03 19:44:00
Commandline: /usr/bin/unattended-upgrade
Install: linux-image-5.4.0-74-generic:amd64 (5.4.0-74.83~18.04.1, automatic), linux-modules-extra-5.4.0-74-generic:amd64 (5.4.0-74.83~18.04.1, automatic), linux-headers-5.4.0-74-generic:amd64 (5.4.0-74.83~18.04.1, automatic), linux-hwe-5.4-headers-5.4.0-74:amd64 (5.4.0-74.83~18.04.1, automatic), linux-modules-5.4.0-74-generic:amd64 (5.4.0-74.83~18.04.1, automatic)
Upgrade: linux-headers-generic-hwe-18.04:amd64 (5.4.0.73.82~18.04.66, 5.4.0.74.83~18.04.67), linux-image-generic-hwe-18.04:amd64 (5.4.0.73.82~18.04.66, 5.4.0.74.83~18.04.67), linux-generic-hwe-18.04:amd64 (5.4.0.73.82~18.04.66, 5.4.0.74.83~18.04.67)
End-Date: 2021-06-03 19:45:13
Start-Date: 2021-06-04 06:21:24
Commandline: /usr/bin/unattended-upgrade
Remove: linux-headers-5.4.0-54-generic:amd64 (5.4.0-54.60~18.04.1)
End-Date: 2021-06-04 06:21:26
Start-Date: 2021-06-04 06:21:31
Commandline: /usr/bin/unattended-upgrade
Remove: linux-image-5.4.0-54-generic:amd64 (5.4.0-54.60~18.04.1), linux-modules-extra-5.4.0-54-generic:amd64 (5.4.0-54.60~18.04.1), linux-modules-5.4.0-54-generic:amd64 (5.4.0-54.60~18.04.1)
End-Date: 2021-06-04 06:21:40
These are the only kernel or xorg changes in the past 3 months. The other card works fine on a different machine (which is where it shall live forever more now)
