[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1531319025.8759.57.camel@infradead.org>
Date: Wed, 11 Jul 2018 15:23:45 +0100
From: David Woodhouse <dwmw2@...radead.org>
To: paulmck@...ux.vnet.ibm.com
Cc: Peter Zijlstra <peterz@...radead.org>, mhillenb@...zon.de,
linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [RFC] Make need_resched() return true when rcu_urgent_qs
requested
On Mon, 2018-07-09 at 15:08 -0700, Paul E. McKenney wrote:
> index f9c0ca2ccf0c..3350ece366ab 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -2839,6 +2839,15 @@ void rcu_check_callbacks(int user)
> rcu_bh_qs();
> }
> rcu_preempt_check_callbacks();
> + /* The load-acquire pairs with the store-release setting to true. */
> + if (smp_load_acquire(this_cpu_ptr(&rcu_dynticks.rcu_urgent_qs))) {
> + /* Idle and userspace execution already are quiescent states. */
> + if (rcu_is_cpu_rrupt_from_idle() && !user) {
if (idle && !user) seems tautological and... illogical.
If I make it 'if (!rcu_is_cpu_rrput_from_idle() && !user)' it seems to
work better. Ripping out my debugging printks now to check that's still
true...
(Also, isn't userspace execution only a quiescent state if NO_HZ_FULL?)
> + set_tsk_need_resched(current);
> + set_preempt_need_resched();
> + }
> + __this_cpu_write(rcu_dynticks.rcu_urgent_qs, false);
> + }
Download attachment "smime.p7s" of type "application/x-pkcs7-signature" (5213 bytes)
Powered by blists - more mailing lists