For someone who isn't experienced in how Linux crashes - what currently happens when there's an error that can't be recovered from? Does the device just restart?
A kernel panic handler writes down a piece of stack, a piece of code, contents of registers, and some minimal message.
Normally the machine does not reboot, so that you could photograph the console maybe.
(The screen normally remains white on black. The console can as well be a COM port.)
You can configure it to auto-reboot by putting a timeout value to /proc/sys/kernel/panic.
I don't see why the handler could not go back to textual mode is the console is serial.
Some devices have neither a serial console nor a text-mode screen powered by BIOS; all they have is a framebuffer device. There a QR code could work really well. No need to have any font data, for instance.
"A new component "systemd-bsod" has been added, which can show logged error messages full screen, if they have a log level of LOG_EMERG log level. This component is experimental and its public interface is subject to change."
First step is to log something with LOG_EMERG level. Then maybe make the system crash?
Ahh, I had heard people complaining about systemd having a qr code generator included a few years ago.
Hopefully this one can do something useful, like include details of the actual crash so you don't have to transcribe them. Unlike the windows one, that's hardcoded in.
This is such a pointless feature. I have several daily driver Linux systems. They run for months at a time, sometimes years. The only time I see a kernel panic is if the hardware is bad or I did something wrong (misconfiguration).
Years ago I had a dual boot system with bad RAM and on Windows it would regularly blue screen, on Linux it was fine. Took a long time to figure out what the actual problem was because Linux is just so rock solid.
And no, I don't use any systemd based distros on my daily drivers.
Not everyone runs Linux systems like a Sys Admin. Some folks use it to play games on their steamdeck or to give live to older hardware that can’t run windows/osx anymore. Others don’t even care what they’re running as long as it can open chrome.
BSOD is very clearly an end user feature for folks who don’t have experience debugging kernel panics. If the QR code can generate a copy passable stack trace of the error the it could help make linux more mainstream and easier to debug for non technical folks.
