| Message ID | 1408351052-25075-1-git-send-email-chuansheng.liu@intel.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On 08/18/2014 10:37 AM, Chuansheng Liu wrote: > Implementing one new API wake_up_if_idle(), which is used to > wake up the idle CPU. Is this patchset tested ? Did you check it solves the issue you were facing ? > Suggested-by: Andy Lutomirski <luto@amacapital.net> > Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> > --- > include/linux/sched.h | 1 + > kernel/sched/core.c | 16 ++++++++++++++++ > 2 files changed, 17 insertions(+) > > diff --git a/include/linux/sched.h b/include/linux/sched.h > index 857ba40..3f89ac1 100644 > --- a/include/linux/sched.h > +++ b/include/linux/sched.h > @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { > extern struct sched_domain_topology_level *sched_domain_topology; > > extern void set_sched_topology(struct sched_domain_topology_level *tl); > +extern void wake_up_if_idle(int cpu); > > #ifdef CONFIG_SCHED_DEBUG > # define SD_INIT_NAME(type) .name = #type > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > index 1211575..adf104f 100644 > --- a/kernel/sched/core.c > +++ b/kernel/sched/core.c > @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu) > } > } > > +void wake_up_if_idle(int cpu) > +{ > + struct rq *rq = cpu_rq(cpu); > + unsigned long flags; > + > + if (set_nr_if_polling(rq->idle)) { > + trace_sched_wake_idle_without_ipi(cpu); > + } else { > + raw_spin_lock_irqsave(&rq->lock, flags); > + if (rq->curr == rq->idle) > + smp_send_reschedule(cpu); > + /* Else cpu is not in idle, do nothing here */ > + raw_spin_unlock_irqrestore(&rq->lock, flags); > + } > +} > + > bool cpus_share_cache(int this_cpu, int that_cpu) > { > return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu); >
Hello Daniel, > -----Original Message----- > From: Daniel Lezcano [mailto:daniel.lezcano@linaro.org] > Sent: Thursday, August 21, 2014 9:54 AM > To: Liu, Chuansheng; luto@amacapital.net; peterz@infradead.org; > rjw@rjwysocki.net; mingo@redhat.com > Cc: linux-pm@vger.kernel.org; linux-kernel@vger.kernel.org; Liu, Changcheng; > Wang, Xiaoming; Chakravarty, Souvik K > Subject: Re: [PATCH 1/3] sched: Add new API wake_up_if_idle() to wake up the > idle cpu > > On 08/18/2014 10:37 AM, Chuansheng Liu wrote: > > Implementing one new API wake_up_if_idle(), which is used to > > wake up the idle CPU. > > Is this patchset tested ? Did you check it solves the issue you were > facing ? We have done the basic test, and found the cores can exit C0 quickly with this patchset. Basically once the _TIF_NEED_RESCHED is set, then the poll_idle() can be broken. Please correct me if something is wrong, thanks. > > > Suggested-by: Andy Lutomirski <luto@amacapital.net> > > Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> > > --- > > include/linux/sched.h | 1 + > > kernel/sched/core.c | 16 ++++++++++++++++ > > 2 files changed, 17 insertions(+) > > > > diff --git a/include/linux/sched.h b/include/linux/sched.h > > index 857ba40..3f89ac1 100644 > > --- a/include/linux/sched.h > > +++ b/include/linux/sched.h > > @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { > > extern struct sched_domain_topology_level *sched_domain_topology; > > > > extern void set_sched_topology(struct sched_domain_topology_level *tl); > > +extern void wake_up_if_idle(int cpu); > > > > #ifdef CONFIG_SCHED_DEBUG > > # define SD_INIT_NAME(type) .name = #type > > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > > index 1211575..adf104f 100644 > > --- a/kernel/sched/core.c > > +++ b/kernel/sched/core.c > > @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct > task_struct *p, int cpu) > > } > > } > > > > +void wake_up_if_idle(int cpu) > > +{ > > + struct rq *rq = cpu_rq(cpu); > > + unsigned long flags; > > + > > + if (set_nr_if_polling(rq->idle)) { > > + trace_sched_wake_idle_without_ipi(cpu); > > + } else { > > + raw_spin_lock_irqsave(&rq->lock, flags); > > + if (rq->curr == rq->idle) > > + smp_send_reschedule(cpu); > > + /* Else cpu is not in idle, do nothing here */ > > + raw_spin_unlock_irqrestore(&rq->lock, flags); > > + } > > +} > > + > > bool cpus_share_cache(int this_cpu, int that_cpu) > > { > > return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu); > > > > > -- > <http://www.linaro.org/> Linaro.org ? Open source software for ARM > SoCs > > Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | > <http://twitter.com/#!/linaroorg> Twitter | > <http://www.linaro.org/linaro-blog/> Blog
On 08/21/2014 04:12 AM, Liu, Chuansheng wrote: > Hello Daniel, > >> -----Original Message----- >> From: Daniel Lezcano [mailto:daniel.lezcano@linaro.org] >> Sent: Thursday, August 21, 2014 9:54 AM >> To: Liu, Chuansheng; luto@amacapital.net; peterz@infradead.org; >> rjw@rjwysocki.net; mingo@redhat.com >> Cc: linux-pm@vger.kernel.org; linux-kernel@vger.kernel.org; Liu, Changcheng; >> Wang, Xiaoming; Chakravarty, Souvik K >> Subject: Re: [PATCH 1/3] sched: Add new API wake_up_if_idle() to wake up the >> idle cpu >> >> On 08/18/2014 10:37 AM, Chuansheng Liu wrote: >>> Implementing one new API wake_up_if_idle(), which is used to >>> wake up the idle CPU. >> >> Is this patchset tested ? Did you check it solves the issue you were >> facing ? > We have done the basic test, and found the cores can exit C0 quickly with this patchset. > Basically once the _TIF_NEED_RESCHED is set, then the poll_idle() can be broken. > > Please correct me if something is wrong, thanks. Actually, it was not clear for me if this patch was a proposal or not. I will review it. Thanks -- Daniel >>> Suggested-by: Andy Lutomirski <luto@amacapital.net> >>> Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> >>> --- >>> include/linux/sched.h | 1 + >>> kernel/sched/core.c | 16 ++++++++++++++++ >>> 2 files changed, 17 insertions(+) >>> >>> diff --git a/include/linux/sched.h b/include/linux/sched.h >>> index 857ba40..3f89ac1 100644 >>> --- a/include/linux/sched.h >>> +++ b/include/linux/sched.h >>> @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { >>> extern struct sched_domain_topology_level *sched_domain_topology; >>> >>> extern void set_sched_topology(struct sched_domain_topology_level *tl); >>> +extern void wake_up_if_idle(int cpu); >>> >>> #ifdef CONFIG_SCHED_DEBUG >>> # define SD_INIT_NAME(type) .name = #type >>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c >>> index 1211575..adf104f 100644 >>> --- a/kernel/sched/core.c >>> +++ b/kernel/sched/core.c >>> @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct >> task_struct *p, int cpu) >>> } >>> } >>> >>> +void wake_up_if_idle(int cpu) >>> +{ >>> + struct rq *rq = cpu_rq(cpu); >>> + unsigned long flags; >>> + >>> + if (set_nr_if_polling(rq->idle)) { >>> + trace_sched_wake_idle_without_ipi(cpu); >>> + } else { >>> + raw_spin_lock_irqsave(&rq->lock, flags); >>> + if (rq->curr == rq->idle) >>> + smp_send_reschedule(cpu); >>> + /* Else cpu is not in idle, do nothing here */ >>> + raw_spin_unlock_irqrestore(&rq->lock, flags); >>> + } >>> +} >>> + >>> bool cpus_share_cache(int this_cpu, int that_cpu) >>> { >>> return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu); >>> >> >> >> -- >> <http://www.linaro.org/> Linaro.org ? Open source software for ARM >> SoCs >> >> Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | >> <http://twitter.com/#!/linaroorg> Twitter | >> <http://www.linaro.org/linaro-blog/> Blog >
On Aug 18, 2014 3:52 AM, "Chuansheng Liu" <chuansheng.liu@intel.com> wrote: > > Implementing one new API wake_up_if_idle(), which is used to > wake up the idle CPU. > > Suggested-by: Andy Lutomirski <luto@amacapital.net> > Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> > --- > include/linux/sched.h | 1 + > kernel/sched/core.c | 16 ++++++++++++++++ > 2 files changed, 17 insertions(+) > > diff --git a/include/linux/sched.h b/include/linux/sched.h > index 857ba40..3f89ac1 100644 > --- a/include/linux/sched.h > +++ b/include/linux/sched.h > @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { > extern struct sched_domain_topology_level *sched_domain_topology; > > extern void set_sched_topology(struct sched_domain_topology_level *tl); > +extern void wake_up_if_idle(int cpu); > > #ifdef CONFIG_SCHED_DEBUG > # define SD_INIT_NAME(type) .name = #type > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > index 1211575..adf104f 100644 > --- a/kernel/sched/core.c > +++ b/kernel/sched/core.c > @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu) > } > } > > +void wake_up_if_idle(int cpu) > +{ > + struct rq *rq = cpu_rq(cpu); > + unsigned long flags; > + > + if (set_nr_if_polling(rq->idle)) { > + trace_sched_wake_idle_without_ipi(cpu); > + } else { > + FWIW, adding: if (rq->curr != rq->idle) return; Right here could improve performance on large, mostly non-idle systems. It would skip the spinlock in most cases. I don't know whether this code is relevant on systems like that, though. raw_spin_lock_irqsave(&rq->lock, flags); > + if (rq->curr == rq->idle) > + smp_send_reschedule(cpu); > + /* Else cpu is not in idle, do nothing here */ > + raw_spin_unlock_irqrestore(&rq->lock, flags); > + } > + --Andy -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Aug 18, 2014 at 04:37:30PM +0800, Chuansheng Liu wrote: > Implementing one new API wake_up_if_idle(), which is used to > wake up the idle CPU. > > Suggested-by: Andy Lutomirski <luto@amacapital.net> > Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> > --- > include/linux/sched.h | 1 + > kernel/sched/core.c | 16 ++++++++++++++++ > 2 files changed, 17 insertions(+) > > diff --git a/include/linux/sched.h b/include/linux/sched.h > index 857ba40..3f89ac1 100644 > --- a/include/linux/sched.h > +++ b/include/linux/sched.h > @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { > extern struct sched_domain_topology_level *sched_domain_topology; > > extern void set_sched_topology(struct sched_domain_topology_level *tl); > +extern void wake_up_if_idle(int cpu); > > #ifdef CONFIG_SCHED_DEBUG > # define SD_INIT_NAME(type) .name = #type > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > index 1211575..adf104f 100644 > --- a/kernel/sched/core.c > +++ b/kernel/sched/core.c > @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu) > } > } > > +void wake_up_if_idle(int cpu) > +{ > + struct rq *rq = cpu_rq(cpu); > + unsigned long flags; > + > + if (set_nr_if_polling(rq->idle)) { > + trace_sched_wake_idle_without_ipi(cpu); > + } else { > + raw_spin_lock_irqsave(&rq->lock, flags); > + if (rq->curr == rq->idle) That needs to be is_idle_task(), there might be other 'idle' tasks than rq->idle :/ > + smp_send_reschedule(cpu); > + /* Else cpu is not in idle, do nothing here */ > + raw_spin_unlock_irqrestore(&rq->lock, flags); > + } > +} > + > bool cpus_share_cache(int this_cpu, int that_cpu) > { > return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu); > -- > 1.7.9.5 > -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Aug 21, 2014 at 02:21:51PM -0700, Andy Lutomirski wrote: > On Aug 18, 2014 3:52 AM, "Chuansheng Liu" <chuansheng.liu@intel.com> wrote: > > +void wake_up_if_idle(int cpu) > > +{ > > + struct rq *rq = cpu_rq(cpu); > > + unsigned long flags; > > + > > + if (set_nr_if_polling(rq->idle)) { > > + trace_sched_wake_idle_without_ipi(cpu); > > + } else { > > + > > FWIW, adding: > > if (rq->curr != rq->idle) > return; > > Right here could improve performance on large, mostly non-idle > systems. It would skip the spinlock in most cases. > !is_idle_task() :-) -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
> -----Original Message----- > From: Peter Zijlstra [mailto:peterz@infradead.org] > Sent: Wednesday, September 03, 2014 5:54 PM > To: Andy Lutomirski > Cc: Liu, Chuansheng; linux-pm@vger.kernel.org; Rafael J. Wysocki; Daniel > Lezcano; Wang, Xiaoming; Ingo Molnar; Chakravarty, Souvik K; Liu, Changcheng; > linux-kernel@vger.kernel.org > Subject: Re: [PATCH 1/3] sched: Add new API wake_up_if_idle() to wake up the > idle cpu > > On Thu, Aug 21, 2014 at 02:21:51PM -0700, Andy Lutomirski wrote: > > On Aug 18, 2014 3:52 AM, "Chuansheng Liu" <chuansheng.liu@intel.com> > wrote: > > > > +void wake_up_if_idle(int cpu) > > > +{ > > > + struct rq *rq = cpu_rq(cpu); > > > + unsigned long flags; > > > + > > > + if (set_nr_if_polling(rq->idle)) { > > > + trace_sched_wake_idle_without_ipi(cpu); > > > + } else { > > > + > > > > FWIW, adding: > > > > if (rq->curr != rq->idle) > > return; > > > > Right here could improve performance on large, mostly non-idle > > systems. It would skip the spinlock in most cases. > > > > !is_idle_task() :-) Thanks Peter and Andy, will refine the patches again:) -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/include/linux/sched.h b/include/linux/sched.h index 857ba40..3f89ac1 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1024,6 +1024,7 @@ struct sched_domain_topology_level { extern struct sched_domain_topology_level *sched_domain_topology; extern void set_sched_topology(struct sched_domain_topology_level *tl); +extern void wake_up_if_idle(int cpu); #ifdef CONFIG_SCHED_DEBUG # define SD_INIT_NAME(type) .name = #type diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 1211575..adf104f 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1620,6 +1620,22 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu) } } +void wake_up_if_idle(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long flags; + + if (set_nr_if_polling(rq->idle)) { + trace_sched_wake_idle_without_ipi(cpu); + } else { + raw_spin_lock_irqsave(&rq->lock, flags); + if (rq->curr == rq->idle) + smp_send_reschedule(cpu); + /* Else cpu is not in idle, do nothing here */ + raw_spin_unlock_irqrestore(&rq->lock, flags); + } +} + bool cpus_share_cache(int this_cpu, int that_cpu) { return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu);
Implementing one new API wake_up_if_idle(), which is used to wake up the idle CPU. Suggested-by: Andy Lutomirski <luto@amacapital.net> Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com> --- include/linux/sched.h | 1 + kernel/sched/core.c | 16 ++++++++++++++++ 2 files changed, 17 insertions(+)