Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-14 Thread Peter Zijlstra
On Wed, Oct 14, 2015 at 10:05:16PM +0200, Oleg Nesterov wrote: > Yes, because that damn cpu_active() check doesn't look strictly necessary ;) > Or I misunderstood. How about we sit down and have a hard look after Thomas is done revamping hotplug? I don't want to go pour over hotplug code that is g

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-14 Thread Oleg Nesterov
On 10/14, Peter Zijlstra wrote: > > On Mon, Oct 12, 2015 at 07:34:02PM +0200, Oleg Nesterov wrote: > > > > I simply can't understand... To me it looks as if we can simply remove > > the cpu_active() check in select_fallback_rq(). > > > > If we race with cpu_down(), cpu_active() is cleared by sched_

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-14 Thread Peter Zijlstra
On Mon, Oct 12, 2015 at 07:34:02PM +0200, Oleg Nesterov wrote: > On 10/12, Peter Zijlstra wrote: > > > > On Sat, Oct 10, 2015 at 08:53:09PM +0200, Oleg Nesterov wrote: > > > I do not understand the cpu_active() check in select_fallback_rq(). > > > x86 doesn't need it, and the recent commit dd9d3843

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-12 Thread Oleg Nesterov
On 10/12, Peter Zijlstra wrote: > > On Sat, Oct 10, 2015 at 08:53:09PM +0200, Oleg Nesterov wrote: > > I do not understand the cpu_active() check in select_fallback_rq(). > > x86 doesn't need it, and the recent commit dd9d3843755d "sched: Fix > > cpu_active_mask/cpu_online_mask race" documents the

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-12 Thread Peter Zijlstra
On Sat, Oct 10, 2015 at 08:53:09PM +0200, Oleg Nesterov wrote: > I do not understand the cpu_active() check in select_fallback_rq(). > x86 doesn't need it, and the recent commit dd9d3843755d "sched: Fix > cpu_active_mask/cpu_online_mask race" documents the fact that on any > architecture we can ign

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-11 Thread Thomas Gleixner
On Sun, 11 Oct 2015, Oleg Nesterov wrote: > On 10/10, Oleg Nesterov wrote: > > > > I do not understand the cpu_active() check in select_fallback_rq(). > > x86 doesn't need it, and the recent commit dd9d3843755d "sched: Fix > > cpu_active_mask/cpu_online_mask race" documents the fact that on any > >

Re: [PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-11 Thread Oleg Nesterov
On 10/10, Oleg Nesterov wrote: > > I do not understand the cpu_active() check in select_fallback_rq(). > x86 doesn't need it, and the recent commit dd9d3843755d "sched: Fix > cpu_active_mask/cpu_online_mask race" documents the fact that on any > architecture we can ignore !active starting from CPU_

[PATCH 1/3] sched: select_task_rq() should check cpu_active() like select_fallback_rq()

2015-10-10 Thread Oleg Nesterov
I do not understand the cpu_active() check in select_fallback_rq(). x86 doesn't need it, and the recent commit dd9d3843755d "sched: Fix cpu_active_mask/cpu_online_mask race" documents the fact that on any architecture we can ignore !active starting from CPU_ONLINE stage. But any possible reason wh