Yes, I agree that the atmosphere and the rules enforced differ between communities. It's also true that the Linux kernel community is known for its strong language and personal attacks (the last time I remember it being featured prominently on HN and elsewhere was when Sarah Sharp called out Linus on his behaviour [1]). But maybe LKML feels like it needs drama from time to time?
Most of the particularly vicious flak Poettering gets isn't even really from the Linux kernel community, but just from random internet people who have an opinion. There's traditionally a bit of an expectation of a "meritocracy of code" in LKML where you can't yell too loudly if you don't also contribute (not necessarily literally "code", also documentation, testing, code review, hardware specs, etc.). I think the degree to which it's an idealized meritocracy can easily be overstated, but it does tend to at least raise the bar for ranting, since you can't only rant without ever doing anything else. The general internet doesn't really feel constrained by that, so you have people who know next to nothing about the Linux kernel who nevertheless feel entitled to very strong and intemperately expressed opinions about it.
This is pretty much it: Linus Thorvald telling people to go screw themselves is an argument of "I am widely recognized to have the loudest voice because of many useful contributions, now accept my opinion or go away".
This carries the danger of people shortening it into "I have the loudest voice" in their mind, which begs the question how other communities unite the role of "last instance for contentious questions" (aka benevolent dictator) with the ideal of being nice to everyone.
Linus's flames also tend not to be personal: they're more like "this architecture is shit" than "you're shit". They can be taken personally (and imo are sometimes over the top), but they don't seem intended in a more vicious way as trying to "hit" a particular person. When he does criticize people it ends up being pretty measured, e.g. a widely talked about blog post where he "attacks" RMS [1] is not a flame at all, but just him saying he isn't a huge fan of RMS and thinks RMS is a bit of a single-issue, black-and-white type of person. Which is a personal criticism, but a pretty civil one. And he definitely doesn't stoop to starting campaigns to try to drive people out of the open-source community who he doesn't like; he just works with them or doesn't.
How common are flames like "How did they not die as babies, considering that they were likely too stupid to find a tit to suck on?"? Those seem very personal, even if the target isn't specifically named. Is it 40%, or more like 10%? Since to me 0% seems like the right level.
Nothing like a good old cherry-picked fiery comment out of the thousands of public messages Torvalds posts. Not to mention that you've sizably mischaracterised the 'meat' of his argument - which I find more offensive than hearing someone called 'full of BS' followed by an explanation.
Frankly, Linus is in the right here. There's nothing constructive about the initial tone of the post being made. It is a hostile question from the outset. He goes on to explain in detail the why of the answer. It's not a flame.
To bring this back to Lennart Poettering, I'd argue this is why community issues are hard to fix: there's far too many people who are happy to use efforts to stamp out really harmful stuff (threats of violence/death, flames and spamming/harassment) to instead try and knock down a prominent figure just because. It's a distraction that ultimately makes it irrelevant, and to boot ruins any forward progress to be made.
For those paying attention, that above lkml thread ended with Sarah Sharp stating that she'll bring and distribute pot brownies to everyone on stage at linuxconf. No word on whether GKH or Linus enjoyed them. :-)
1: https://lkml.org/lkml/2013/7/19/634