You're making a false equivalency where stern is the same as toxic. There are more professional and clear ways to communicate the issues with code quality. No one is disagreeing that those need to be communicated. The Issue is how.

And because you seem to take stock in what Torvalds says, then consider that if he himself admitted these were harmful and inefficient methods of communication then they probably were. If it was leading to fantastic results in the kernel i don't see why he would've stopped. My guess is that he learned something that it seems you may still have yet to: empathy.