On Sat, Mar 4, 2023 at 1:01 PM Linus Torvalds
[off-list ref] wrote:

It's still completely untested. Treat this very much as a "Let's make
the common cases faster, at least for !MAXSMP".

Ok, so I started "testing" it in the sense that I actually looked at
the code it generated, and went all "why didn't it make any
difference".

And that's because the patch had the

  #ifdef CONFIG_CPUMASK_OFFSTACK

condition exactly the wrong way around.

So if somebody actually wants to play with that patch, you need to
change that to be

  #ifndef CONFIG_CPUMASK_OFFSTACK

(and then you obviously need to have a kernel config that does *not*
have MAXSMP set).

That at least simplifies some of the code generation when I look at
it. Whether the end result _works_ or not, I still haven't checked.
That patch is still very much meant as a "look, something like this
should make our cpumask handling much more efficient"

              Linus