There was an email posted to the PLUG mailing lists recently about a user that had run rm -rf / and bricked their laptop. It was a fascinating story and I lost a good hour or so reading through that thread, the GitHub issues and responses etc...
The exact specifics might be different but it's an issue I've seen crop up a few times before: Should we (that is the Linux community and more specifically systemd) change our software to work around manufacturers crappy firmware (by making /sys/firmware/efi/efivars/ read only by default) or should we stick to our guns and demand that manufactures ship hardware that doesn't bork when we do something that is allowed by the (UEFI) spec.
I personally lean towards saying this is a manufacturer fault not a Linux issue and that motherboards that can't post after their EFI settings have been wiped should be RMAed.
But both sides have merit and it's not really fair to say to users that have just bricked their device "sure, it's a known issue and we could have done something to prevent that. But we decided it's not our problem to fix, go speak to the manufacturer."
It looks like in this case the user got their laptop replaced under warranty without issues but we can't count on that happening every time. I expect there would be plenty of devices out there that have broken EFI implementations and are out of warranty1 and they won't be able to get their devices replaced.
Our software shouldn't be able to break users hardware. I like to encourage users to explore their system and learn for themselves and I often say things like "Don't worry, try things out, we have backups you can't really break anything" to timid users, I want that to remain true.
-
or vendors who would argue that doing anything advanced link installing Linux voids the warranty, but that's a different
rantdebate. ↩
