Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-02-10 Thread Frederic Weisbecker
On Fri, Feb 09, 2018 at 08:16:12AM +0100, Ingo Molnar wrote: > > * Frederic Weisbecker wrote: > > > When a CPU runs in full dynticks mode, a 1Hz tick remains in order to > > keep the scheduler stats alive. However this residual tick is a burden > > for bare metal tasks that can't stand any inter

Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-02-08 Thread Ingo Molnar
* Frederic Weisbecker wrote: > When a CPU runs in full dynticks mode, a 1Hz tick remains in order to > keep the scheduler stats alive. However this residual tick is a burden > for bare metal tasks that can't stand any interruption at all, or want > to minimize them. > > The usual boot parameter

[PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-02-08 Thread Frederic Weisbecker
When a CPU runs in full dynticks mode, a 1Hz tick remains in order to keep the scheduler stats alive. However this residual tick is a burden for bare metal tasks that can't stand any interruption at all, or want to minimize them. The usual boot parameters "nohz_full=" or "isolcpus=nohz" will now o

Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-01-29 Thread Peter Zijlstra
On Mon, Jan 29, 2018 at 05:48:33PM +0100, Frederic Weisbecker wrote: > On Mon, Jan 29, 2018 at 04:38:39PM +0100, Peter Zijlstra wrote: > > I would very much like a few words on why sched_class::task_tick() is > > safe to call remote -- from a quick look I think it actually is, but it > > would be g

Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-01-29 Thread Frederic Weisbecker
On Mon, Jan 29, 2018 at 04:38:39PM +0100, Peter Zijlstra wrote: 1;4205;0c> On Fri, Jan 19, 2018 at 01:02:18AM +0100, Frederic Weisbecker wrote: > > When a CPU runs in full dynticks mode, a 1Hz tick remains in order to > > keep the scheduler stats alive. However this residual tick is a burden > > fo

Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-01-29 Thread Peter Zijlstra
On Fri, Jan 19, 2018 at 01:02:18AM +0100, Frederic Weisbecker wrote: > + queue_delayed_work(system_unbound_wq, dwork, HZ); Also, why HZ ? No clues were given.

Re: [PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-01-29 Thread Peter Zijlstra
On Fri, Jan 19, 2018 at 01:02:18AM +0100, Frederic Weisbecker wrote: > When a CPU runs in full dynticks mode, a 1Hz tick remains in order to > keep the scheduler stats alive. However this residual tick is a burden > for bare metal tasks that can't stand any interruption at all, or want > to minimiz

[PATCH 4/6] sched/isolation: Residual 1Hz scheduler tick offload

2018-01-18 Thread Frederic Weisbecker
When a CPU runs in full dynticks mode, a 1Hz tick remains in order to keep the scheduler stats alive. However this residual tick is a burden for bare metal tasks that can't stand any interruption at all, or want to minimize them. The usual boot parameters "nohz_full=" or "isolcpus=nohz" will now o