Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched domain build-up

From: Yuyang Du
Date: Tue May 31 2016 - 21:58:38 EST


On Tue, May 31, 2016 at 12:41:21PM +0200, Peter Zijlstra wrote:
> On Tue, May 31, 2016 at 09:31:32AM +0800, Yuyang Du wrote:
> > On Tue, May 31, 2016 at 11:21:46AM +0200, Peter Zijlstra wrote:
> > > On Tue, May 31, 2016 at 09:11:37AM +0800, Yuyang Du wrote:
> > > > The SD_BALANCE_WAKE is irrelevant in the contexts of these two removals,
> > > > and in addition SD_BALANCE_WAKE is not and should not be set in any
> > > > sched_domain flags so far.
> > >
> > > This Changelog doesn't make any sense...
> >
> > How? SD_BALANCE_WAKE is not in any sched_domain flags (sd->flags), even if
> > it is, it is not used anywhere, no?
>
> It is and it is. See select_task_fair_rq():
>
> if (tmp->flags & sd_flags)
>
> Now, as long as WAKE_AFFINE is also set, its hard to actually get into
> the find_idlest_cpu() balancing, but if you clear all that you will
> still get there.

Well, that is very true, and the next patch (2/2) just makes all this
what this is supposed to be: the SD_BALANCE_WAKE is a meaningful sched_domain
flag.

This particular patch is a pure cleanup, may I amend the changelog to:

According to the comment: "turn off/on idle balance on this domain",
the SD_BALANCE_WAKE has nothing to do with idle balance, so clean them up.