Removing Some Old Arm Drivers & Board/Machine Code To Lighten The Kernel By 154k Lines

[sinepgib]

As far as I'm aware, the vast majority of those security fixes only apply to newer architectures. As far as I'm concerned, the kernel shouldn't be responsible for most userland security, and most tasks associated with the internet are userland.

Buffer overflows are architecture agnostic. UB is (mostly) architecture agnostic. Etc. I don't mean "oh no spectre", but it's just not the only kind of vulnerability you patch. I'm not sure big changes like mitigations for those get backported anyway.
Re: userspace, whole different story, but in any case those are mostly web browsing related AFAIK. I don't expect these to be loading JS ridden websites but to just use a TCP connection or at most call some HTTP API.

[willmore]

I would say it is a bit more complicated. To be honest Harvard architecture is kinda misleading because what makes data a "data". Technically speaking yes you have chip trying to predict what instructions will be used and what data will be used so it is loaded into memory.

And that's why I say they only differ in where they collapse back to a single memory space along the data and instruction paths:

everything merges memory accesses by the time they get to the main memory, though

Of course everything is more complicated that what can be summarized in a short post. There's a tradeoff between being thorough and being concise.