Greg K-H

Not all are bad. The one that affects my code was legit. I just doubled checked to make sure.

Same here, but I fully support a wholesale revert given the behavior of that group. Valid fixes can be re-done separately on a one-by-one basis. What a bunch of...

Or have an opt out. If you see patches that are legit, let Greg know, so he can remove them from the wholesale cleansing. But I agree. The default should be removal unless a maintainer tells him otherwise.

I fully support a full rage revert at this point, better to re-fix after the fact rather than delay.

The problem is, the real fixes may never be fixed again. I know I'll probably forget about it if I don't fix it immediately, and the bug will then reappear and not be fixed for several releases.

At least for me, it was a fix for the Sega Dreamcast cdrom driver… So not what I’d consider high priority, the zero users probably don’t care ;-)

And for me, it was a memory leak fix on an error path. Not high priority, but one I don't want to reintroduce. Is the purpose of this exercise to punish the university, or to make Linux more robust?

A blind revert is fine for punishment, but will weaken the kernel. The way to make the kernel better is for every patch to be scrutinized, and remove the broken ones. Then ban them from submitting more patches. But I'm against the blind revert.