Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1411802

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

Show all headers | View raw

On Wed, Jun 01, 2016 at 11:36:39AM +0200, Mike Galbraith wrote:
> > Yup. Up to this point, we don't have any disagreement. And I don't think we
> > have any disagreement conceptually. What the next patch really does is:
> > 
> > (1) we don't remove SD_BALANCE_WAKE as an important sched_domain flag, on
> >     the contrary, we strengthen it.
> > 
> > (2) the semantic of SD_BALANCE_WAKE is currently represented by SD_WAKE_AFFINE,
> >     we actually remove this representation.
> 
> Nope, those two have different meanings.  We pass SD_BALANCE_WAKE to
> identify a ttwu() wakeup, just as we pass SD_BALANCE_FORK to say we're
> waking a child.  SD_WAKE_AFFINE means exactly what it says, but is only
> applicable to ttwu() wakeups.
 
I don't disagree, but want to add that, SD_WAKE_AFFINE has no meaning that is so
special and so important for anyone to use the flag to tune anything. If you want
to do any SD_BALANCE_*, waker CPU is a valid candidate if allowed, that is it.

IIUC your XXX mark and your comment "Prefer wake_affine over balance flags", you
said the same thing: SD_WAKE_AFFINE should be part of SD_BALANCE_WAKE, and should
be part of all SD_BALANCE_* flags,

> > (3) regarding the semantic of SD_WAKE_AFFINE, it is really not about selecting
> >     waker CPU or about the fast path. Conceptually, it is just saying the waker
> >     CPU is a valid and important candidate if SD_BALANCE_WAKE, which is just so
> >     obvious, so I don't think it deserves to be a separate sched_domain flag.
> 
> SD_WAKE_AFFINE being a separate domain flag, the user can turn it
> on/off... separately :)
 
Sure, that is very true, :) But turning it off for what, that is a big question mark.
We don't want a flag unless the flag is for goodness, and not a flag with big question
mark.

> > (4) the outcome is, if SD_BALANCE_WAKE, we definitely will/should try waker CPU,
> >     and if !SD_BALANCE_WAKE, we don't try waker CPU. So nothing functional is
> >     changed.
> 
> If wake_wide() says we do not want an affine wakeup, but you apply
> SD_WAKE_AFFINE meaning to SD_BALANCE_WAKE and turn it on in ->flags,
> we'll give the user a free sample of full balance cost, no?
 
Yes, and otherwise we don't select anything? That is just bad engough whether worse
or not. So the whole fuss I made is really that this is a right thing to start with. :)

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[RFC PATCH 0/2] Remove and replace SD_WAKE_AFFINE with SD_BALANCE_WAKE Yuyang Du <yuyang.du@intel.com> - 2016-05-31 11:10 +0200
  [RFC PATCH 2/2] sched: Remove SD_WAKE_AFFINE flag and replace it with SD_BALANCE_WAKE Yuyang Du <yuyang.du@intel.com> - 2016-05-31 11:10 +0200
    Re: [RFC PATCH 2/2] sched: Remove SD_WAKE_AFFINE flag and replace it  with SD_BALANCE_WAKE Peter Zijlstra <peterz@infradead.org> - 2016-05-31 11:30 +0200
      Re: [RFC PATCH 2/2] sched: Remove SD_WAKE_AFFINE flag and replace it  with SD_BALANCE_WAKE Yuyang Du <yuyang.du@intel.com> - 2016-05-31 11:40 +0200
  [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-05-31 11:10 +0200
    Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-05-31 11:30 +0200
      Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Peter Zijlstra <peterz@infradead.org> - 2016-05-31 12:50 +0200
        Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-01 04:00 +0200
      Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-06-01 07:10 +0200
        Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-01 10:00 +0200
          Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Vincent Guittot <vincent.guittot@linaro.org> - 2016-06-01 10:40 +0200
            Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-01 11:10 +0200
              Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Vincent Guittot <vincent.guittot@linaro.org> - 2016-06-01 11:30 +0200
                Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-02 05:40 +0200
                Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Vincent Guittot <vincent.guittot@linaro.org> - 2016-06-02 09:00 +0200
                Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-02 09:20 +0200
          Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-06-01 11:40 +0200
            Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-02 06:10 +0200
              Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-06-02 08:00 +0200
                Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Yuyang Du <yuyang.du@intel.com> - 2016-06-02 08:40 +0200
                Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-06-02 08:50 +0200
    Re: [RFC PATCH 1/2] sched: Clean up SD_BALANCE_WAKE flags in sched  domain build-up Peter Zijlstra <peterz@infradead.org> - 2016-05-31 11:30 +0200

csiph-web