A few days ago I booted up my Arch Linux install and was greeted with a black screen right after GRUB. No error message, no flicker, nothing.
The weird part? The system was clearly running. I could seemingly “blindly” log in and even access media servers from my phone. So it wasn’t a total crash, just no display.
Naturally, my first thought was: great, another driver issue.
Step 1: Suspecting the update
The problem started right after I did a system upgrade. But since I had already successfully rebooted once after pacman finished its dirty deeds, I was pretty confident the update itself wasn’t the culprit.
I started digging through logs and found this suspicious line (journalctl -k -b -1):
amdgpu 0000:0a:00.0: [drm] Cannot find any crtc or sizesThe monitor worked perfectly in BIOS, GRUB, and Windows, so I was almost convinced that I somehow managed to do something funny with the GPU driver without even touching it.
Step 2: Kernel parameters and downgrades
I tried gpu-related kernel parameters:
nomodeset- system booted fineamdgpu.dc=0- also workedvideo=HDMI-A-1:1920x1080@100- black screenvideo=HDMI-A-1:1920x1080@60- black screen
Okay, let's check which amdgpu-related packages were upgraded:
grep -r "amdgpu\|drm" /var/log/pacman.log | grep -i "upgrade\|update"
> [2025-08-18T17:43:57+0300] [ALPM] upgraded linux-firmware-amdgpu (20250708-1 -> 20250808-1)
> [2025-08-18T17:43:58+0300] [ALPM] upgraded xf86-video-amdgpu (23.0.0-2 -> 25.0.0-1)I installed the LTS kernel, downgraded linux-firmware-amdgpu and xf86-video-amdgpu packages… nope. Same black screen.
Step 3: The paranoia kicks in
At some point I noticed something weird:
host: MacPro7,1 on my Arch install, on a classic desktop PC.
The day before, I had been messing around with macOS 26 installation. It had immediately kernel panicked, but I've still had working macOS 15, so shelved it for later. But now my Arch box thought it was a Mac Pro?
Even worse: macOS was also booting to a black screen. System was alive (keyboard responsive), but no display.
I'm not an OS-expert, so I've immediately made some dumb assumptions - did macOS somehow corrupt NVRAM or did some other black magic to my motherboard?
Step 4: Any issue can be resolved by nuking something
In desperation, I flashed the latest BIOS for my motherboard. At least that fixed the weird host string and I got new cool ROG splash screen on startup.
But the black screen? Still there.
Step 5: The Real Fix
After four hours of troubleshooting, log scraping, parameter testing, BIOS flashing, and general paranoia… I looked at my cat. Then at my HDMI cable.
I tightened the cable on the monitor side.
And suddenly, everything worked.
Conclussion
Debugging often starts at the wrong end of the stack. It’s tempting to dive straight into logs, kernel flags, and firmware updates, but sometimes the issue is sitting right on your desk.
