Re: [GIT PULL] Introduce housekeeping subsystem v4

[Frederic Weisbecker]

2017-10-23 14:06 UTC+02:00, Ingo Molnar <mingo@xxxxxxxxxx>:
>
> * Frederic Weisbecker <frederic@xxxxxxxxxx> wrote:
>
>> In fact, CPU affinity is the only high level concept I found to gather all
>> these
>> housekeeping elements.
>>
>> Perhaps I should use "cpu_isolation" instead of "housekeeping" naming.
>
> The problem with names based on that, like cpu_isolation_map, is that
> there's
> really two concepts here: there's the isolcpus feature where the 'mask' is
> in fact
> the CPUs that are isolated - while the 'housekeeping CPUs' is the mask of
> CPUs
> that _support_ the isolated set of CPUs. The two are different roles but
> easily
> confused if named similarly.

Indeed, housekeeping is in fact the machinery that supports cpu isolation.

> So I guess 'housekeeping CPUs' is as good as it gets for now.

Agreed, And I have no doubt the concept will evolve. We can always
split it into high level concepts such as those we discussed if it
appears necessary later.

> Mind sending a refreshed queue against the latest kernel? There's some new
> conflicts in kernel/watchdog.c for example.

Sure, here is a rebase against -rc6:

git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks.git
    core/isolation-v5

Thanks!