Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Wed, Sep 23, 2015 at 08:43:21AM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 08:28:22 -0700 > "Paul E. McKenney" wrote: > > > On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > > > On Tue, 22 Sep 2015 21:29:14 +0800 > > > Boqun Feng wrote: > > > > > > > On Tue,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Wed, Sep 23, 2015 at 08:43:21AM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 08:28:22 -0700 > "Paul E. McKenney" wrote: > > > On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > > > On Tue, 22 Sep 2015 21:29:14 +0800 > > > Boqun Feng

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 08:28:22 -0700 "Paul E. McKenney" wrote: > On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > > On Tue, 22 Sep 2015 21:29:14 +0800 > > Boqun Feng wrote: > > > > > On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > > > > On Tue, 22 Sep

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 08:28:22 -0700 "Paul E. McKenney" wrote: > On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > > On Tue, 22 Sep 2015 21:29:14 +0800 > > Boqun Feng wrote: > > > > > On Tue, Sep 22, 2015 at 02:51:36PM +0200,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 21:29:14 +0800 > Boqun Feng wrote: > > > On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > > > On Tue, 22 Sep 2015 20:23:26 +0800 > > > Boqun Feng wrote: > > > > > > > Hi Martin, > > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 21:29:14 +0800 Boqun Feng wrote: > On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > > On Tue, 22 Sep 2015 20:23:26 +0800 > > Boqun Feng wrote: > > > > > Hi Martin, > > > > > > On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > > > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 20:23:26 +0800 > Boqun Feng wrote: > > > Hi Martin, > > > > On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > > > On Mon, 21 Sep 2015 11:22:52 +0200 > > > Martin Schwidefsky wrote: > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 20:23:26 +0800 Boqun Feng wrote: > Hi Martin, > > On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > > On Mon, 21 Sep 2015 11:22:52 +0200 > > Martin Schwidefsky wrote: > > > > > On Fri, 18 Sep 2015 14:41:20 -0700 > > > "Paul E. McKenney" wrote: > > > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

Hi Martin, On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > On Mon, 21 Sep 2015 11:22:52 +0200 > Martin Schwidefsky wrote: > > > On Fri, 18 Sep 2015 14:41:20 -0700 > > "Paul E. McKenney" wrote: > > > > > On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, 21 Sep 2015 11:22:52 +0200 Martin Schwidefsky wrote: > On Fri, 18 Sep 2015 14:41:20 -0700 > "Paul E. McKenney" wrote: > > > On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 20:23:26 +0800 > Boqun Feng wrote: > > > Hi Martin, > > > > On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > > > On Mon, 21 Sep 2015 11:22:52 +0200 > > > Martin

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, 21 Sep 2015 11:22:52 +0200 Martin Schwidefsky wrote: > On Fri, 18 Sep 2015 14:41:20 -0700 > "Paul E. McKenney" wrote: > > > On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 06:30:28PM

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 21:29:14 +0800 Boqun Feng wrote: > On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > > On Tue, 22 Sep 2015 20:23:26 +0800 > > Boqun Feng wrote: > > > > > Hi Martin, > > > > > > On Tue, Sep 22, 2015 at

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

Hi Martin, On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > On Mon, 21 Sep 2015 11:22:52 +0200 > Martin Schwidefsky wrote: > > > On Fri, 18 Sep 2015 14:41:20 -0700 > > "Paul E. McKenney" wrote: > > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 22 Sep 2015 20:23:26 +0800 Boqun Feng wrote: > Hi Martin, > > On Tue, Sep 22, 2015 at 12:27:35PM +0200, Martin Schwidefsky wrote: > > On Mon, 21 Sep 2015 11:22:52 +0200 > > Martin Schwidefsky wrote: > > > > > On Fri, 18 Sep 2015 14:41:20

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 22, 2015 at 04:33:07PM +0200, Martin Schwidefsky wrote: > On Tue, 22 Sep 2015 21:29:14 +0800 > Boqun Feng wrote: > > > On Tue, Sep 22, 2015 at 02:51:36PM +0200, Martin Schwidefsky wrote: > > > On Tue, 22 Sep 2015 20:23:26 +0800 > > > Boqun Feng

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Fri, 18 Sep 2015 14:41:20 -0700 "Paul E. McKenney" wrote: > On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Fri, 18 Sep 2015 14:41:20 -0700 "Paul E. McKenney" wrote: > On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote:

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 10:09:41AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 12:49:46PM -0700, Davidlohr Bueso wrote: > On Tue, 15 Sep 2015, Peter Zijlstra wrote: > > >>Secondly, the wake queues are not concurrent, they're in context, so I > >>don't see ordering matter at all. The only reason its a cmpxchg() is > >>because there is the (small)

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 12:49:46PM -0700, Davidlohr Bueso wrote: > On Tue, 15 Sep 2015, Peter Zijlstra wrote: > > >>Secondly, the wake queues are not concurrent, they're in context, so I > >>don't see ordering matter at all. The only reason its a cmpxchg() is > >>because there is the (small)

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 15 Sep 2015, Peter Zijlstra wrote: Secondly, the wake queues are not concurrent, they're in context, so I don't see ordering matter at all. The only reason its a cmpxchg() is because there is the (small) possibility of two contexts wanting to wake the same task, and we use task_struct

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > > > Never mind, the

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > > Never mind, the PPC people will implement this with lwsync and that is > > > > very much

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > Never mind, the PPC people will implement this with lwsync and that is > > > very much not transitive IIRC. > > > > I am probably lost on context, but... > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > Never mind, the PPC people will implement this with lwsync and that is > > very much not transitive IIRC. > > I am probably lost on context, but... > > It turns out that lwsync is transitive in special cases. One of them > is

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 11:55:12AM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 11:49:49AM +0200, Peter Zijlstra wrote: > > On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > > > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > > > > > >On Mon, Sep 14, 2015 at 12:37:23AM

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 11:49:49AM +0200, Peter Zijlstra wrote: > On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > > > >On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > > >> /* > > >>+ * Atomically grab the

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > >On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > >>/* > >>+* Atomically grab the task. If ->wake_q is non-nil (failed cmpxchg) > >>+* then the task is

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 06:30:28PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > >On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > >>/* > >>+* Atomically grab the task. If ->wake_q is non-nil (failed cmpxchg) > >>+* then the task is

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 11:49:49AM +0200, Peter Zijlstra wrote: > On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > > > >On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > > >> /* > > >>+ * Atomically grab the

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 08:34:48AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > > > On Tue, Sep 15,

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 04:14:39PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > > > Never mind, the

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, 15 Sep 2015, Peter Zijlstra wrote: Secondly, the wake queues are not concurrent, they're in context, so I don't see ordering matter at all. The only reason its a cmpxchg() is because there is the (small) possibility of two contexts wanting to wake the same task, and we use task_struct

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 07:09:22AM -0700, Paul E. McKenney wrote: > On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > > Never mind, the PPC people will implement this with lwsync and that is > > > > very much

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 02:48:00PM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > > Never mind, the PPC people will implement this with lwsync and that is > > > very much not transitive IIRC. > > > > I am probably lost on context, but... > >

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 11:55:12AM +0200, Peter Zijlstra wrote: > On Tue, Sep 15, 2015 at 11:49:49AM +0200, Peter Zijlstra wrote: > > On Mon, Sep 14, 2015 at 02:08:06PM -0700, Davidlohr Bueso wrote: > > > On Mon, 14 Sep 2015, Peter Zijlstra wrote: > > > > > > >On Mon, Sep 14, 2015 at 12:37:23AM

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Tue, Sep 15, 2015 at 05:41:42AM -0700, Paul E. McKenney wrote: > > Never mind, the PPC people will implement this with lwsync and that is > > very much not transitive IIRC. > > I am probably lost on context, but... > > It turns out that lwsync is transitive in special cases. One of them > is

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, 14 Sep 2015, Peter Zijlstra wrote: On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: /* +* Atomically grab the task. If ->wake_q is non-nil (failed cmpxchg) +* then the task is already queued (by us or someone else) and will +* get the

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > The barrier parings for wake-queues are very straightforward, and thus > we can ease the barrier requirements, for archs that support it, for > wake_q_add by relying on acquire semantics. As such, (i) we keep the > pairing

[PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

The barrier parings for wake-queues are very straightforward, and thus we can ease the barrier requirements, for archs that support it, for wake_q_add by relying on acquire semantics. As such, (i) we keep the pairing structure/logic and (ii) users, such as mqueues, can continue to rely on a full

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: > The barrier parings for wake-queues are very straightforward, and thus > we can ease the barrier requirements, for archs that support it, for > wake_q_add by relying on acquire semantics. As such, (i) we keep the > pairing

Re: [PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

On Mon, 14 Sep 2015, Peter Zijlstra wrote: On Mon, Sep 14, 2015 at 12:37:23AM -0700, Davidlohr Bueso wrote: /* +* Atomically grab the task. If ->wake_q is non-nil (failed cmpxchg) +* then the task is already queued (by us or someone else) and will +* get the

[PATCH -tip 2/3] sched/wake_q: Relax to acquire semantics

The barrier parings for wake-queues are very straightforward, and thus we can ease the barrier requirements, for archs that support it, for wake_q_add by relying on acquire semantics. As such, (i) we keep the pairing structure/logic and (ii) users, such as mqueues, can continue to rely on a full