On Mon, 22 Mar 2021 11:03:06 +
Mel Gorman wrote:
> Second, select_idle_smt() does not use the cpus mask so consider moving
> the cpus initialisation after select_idle_smt() has been called.
> Specifically this initialisation
>
> cpumask_and(cpus, sched_domain_span(sd), p->cpus_ptr);
>
On Wed, 2021-04-07 at 12:19 +0200, Peter Zijlstra wrote:
> On Wed, Apr 07, 2021 at 11:54:37AM +0200, Peter Zijlstra wrote:
>
> > Let me have another poke at it.
>
> Pretty much what you did, except I also did s/smt/has_idle_core/ and
> fixed that @sd thing.
>
> Like so then?
Looks good to me. T
On Fri, 26 Mar 2021 at 20:19, Rik van Riel wrote:
>
> On Mon, 22 Mar 2021 11:03:06 +
> Mel Gorman wrote:
>
>
> > Second, select_idle_smt() does not use the cpus mask so consider moving
> > the cpus initialisation after select_idle_smt() has been called.
> > Specifically this initialisation
>
On Tue, 2021-04-06 at 17:10 +0200, Vincent Guittot wrote:
> On Fri, 26 Mar 2021 at 20:19, Rik van Riel wrote:
>
> > -static int select_idle_cpu(struct task_struct *p, struct
> > sched_domain *sd, int target)
> > +static int select_idle_cpu(struct task_struct *p, struct
> > sched_domain *sd, int p
On Tue, 6 Apr 2021 at 17:26, Rik van Riel wrote:
>
> On Tue, 2021-04-06 at 17:10 +0200, Vincent Guittot wrote:
> > On Fri, 26 Mar 2021 at 20:19, Rik van Riel wrote:
> >
> > > -static int select_idle_cpu(struct task_struct *p, struct
> > > sched_domain *sd, int target)
> > > +static int select_idl
On Tue, 6 Apr 2021 at 17:31, Vincent Guittot wrote:
>
> On Tue, 6 Apr 2021 at 17:26, Rik van Riel wrote:
> >
> > On Tue, 2021-04-06 at 17:10 +0200, Vincent Guittot wrote:
> > > On Fri, 26 Mar 2021 at 20:19, Rik van Riel wrote:
> > >
> > > > -static int select_idle_cpu(struct task_struct *p, stru
On Tue, 2021-04-06 at 17:31 +0200, Vincent Guittot wrote:
> On Tue, 6 Apr 2021 at 17:26, Rik van Riel wrote:
> > On Tue, 2021-04-06 at 17:10 +0200, Vincent Guittot wrote:
> > > On Fri, 26 Mar 2021 at 20:19, Rik van Riel
> > > wrote:
> > >
> > > > -static int select_idle_cpu(struct task_struct *p
On Tue, 6 Apr 2021 at 17:55, Rik van Riel wrote:
>
> On Tue, 2021-04-06 at 17:31 +0200, Vincent Guittot wrote:
> > On Tue, 6 Apr 2021 at 17:26, Rik van Riel wrote:
> > > On Tue, 2021-04-06 at 17:10 +0200, Vincent Guittot wrote:
> > > > On Fri, 26 Mar 2021 at 20:19, Rik van Riel
> > > > wrote:
>
On Tue, Apr 06, 2021 at 11:26:37AM -0400, Rik van Riel wrote:
> I would be happy to pull the static branch out of select_idle_smt()
> and place it into this if condition, though. You are right that
> would save some overhead on non-smt systems.
>
> Peter, would you prefer a follow-up patch for tha
On Wed, Apr 07, 2021 at 09:17:18AM +0200, Peter Zijlstra wrote:
> Subject: sched/fair: Bring back select_idle_smt(), but differently
> From: Rik van Riel
> Date: Fri, 26 Mar 2021 15:19:32 -0400
>
> From: Rik van Riel
>
> Mel Gorman did some nice work in 9fe1f127b913 ("sched/fair: Merge
> select
Le mercredi 07 avril 2021 à 09:17:18 (+0200), Peter Zijlstra a écrit :
> On Tue, Apr 06, 2021 at 11:26:37AM -0400, Rik van Riel wrote:
> > I would be happy to pull the static branch out of select_idle_smt()
> > and place it into this if condition, though. You are right that
> > would save some over
On Wed, Apr 07, 2021 at 11:42:17AM +0200, Vincent Guittot wrote:
> I would really prefer to keep that out of select_idle_cpu which aims to merge
> in one
> single loop the walk through sd_llc. In the case of select_idle_smt, this is
> done outside
> the loop:
Fair enough.
> @@ -6317,11 +6339,21
On Wed, 7 Apr 2021 at 11:55, Peter Zijlstra wrote:
>
> On Wed, Apr 07, 2021 at 11:42:17AM +0200, Vincent Guittot wrote:
> > I would really prefer to keep that out of select_idle_cpu which aims to
> > merge in one
> > single loop the walk through sd_llc. In the case of select_idle_smt, this
> > i
On Wed, Apr 07, 2021 at 10:41:06AM +0100, Mel Gorman wrote:
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -6112,6 +6112,27 @@ static int select_idle_core(struct task_
> > return -1;
> > }
> >
> > +/*
> > + * Scan the local SMT mask for idle CPUs.
> > + */
> > +static in
On Wed, Apr 07, 2021 at 11:54:37AM +0200, Peter Zijlstra wrote:
> Let me have another poke at it.
Pretty much what you did, except I also did s/smt/has_idle_core/ and
fixed that @sd thing.
Like so then?
---
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6038,11 +6038,9 @@ static inline
On Wed, 7 Apr 2021 at 12:19, Peter Zijlstra wrote:
>
> On Wed, Apr 07, 2021 at 11:54:37AM +0200, Peter Zijlstra wrote:
>
> > Let me have another poke at it.
>
> Pretty much what you did, except I also did s/smt/has_idle_core/ and
> fixed that @sd thing.
>
> Like so then?
Yes. Looks good to me
>
On Wed, Apr 07, 2021 at 12:15:13PM +0200, Peter Zijlstra wrote:
> On Wed, Apr 07, 2021 at 10:41:06AM +0100, Mel Gorman wrote:
>
> > > --- a/kernel/sched/fair.c
> > > +++ b/kernel/sched/fair.c
> > > @@ -6112,6 +6112,27 @@ static int select_idle_core(struct task_
> > > return -1;
> > > }
> > >
On Wed, Apr 07, 2021 at 11:47:17AM +0100, Mel Gorman wrote:
> Ok, cpusets do split domains. I can't imagine the logic of splitting SMT
> siblings across cpusets but if it's possible, it has to be checked and
> protecting that with cpusets_enabled() would be a little overkill and
> possibly miss so
On Fri, Mar 26, 2021 at 03:19:32PM -0400, Rik van Riel wrote:
> ---8<---
> sched,fair: bring back select_idle_smt, but differently
>
> Mel Gorman did some nice work in 9fe1f127b913
> ("sched/fair: Merge select_idle_core/cpu()"), resulting in the kernel
> being more efficient at finding an idle CPU
19 matches
Mail list logo