Re: [RFC] Make need_resched() return true when rcu_urgent_qs requested

[Paul E. McKenney]

On Mon, Jul 09, 2018 at 04:43:38PM +0200, Peter Zijlstra wrote:
> On Mon, Jul 09, 2018 at 07:29:32AM -0700, Paul E. McKenney wrote:
> 
> > OK, so here are our options:
> > 
> > 1.	Add the RCU conditional to need_resched(), as David suggests.
> > 	Peter has concerns about overhead.
> 
> Not only overhead, its plain broken, because:
> 
>  1) we keep preemption state in other places too
>  2) we check need_resched() in asm

Fair points.  And yes, the various asm checks and transitions have
made their presence known to me in other contexts of late.  :-/

So what would you suggest?

							Thanx, Paul