From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Rafael J. Wysocki" Subject: [RFC/RFT][PATCH v2 3/6] sched: idle: Do not stop the tick before cpuidle_idle_call() Date: Tue, 06 Mar 2018 10:03:16 +0100 Message-ID: <2037426.jsGIfSuYk6@aspire.rjw.lan> References: <2067762.1uWBf5RSRc@aspire.rjw.lan> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Return-path: In-Reply-To: <2067762.1uWBf5RSRc@aspire.rjw.lan> Sender: linux-kernel-owner@vger.kernel.org To: Peter Zijlstra , Linux PM Cc: Thomas Gleixner , Frederic Weisbecker , Paul McKenney , Thomas Ilsche , Doug Smythies , Rik van Riel , Aubrey Li , Mike Galbraith , LKML List-Id: linux-pm@vger.kernel.org From: Rafael J. Wysocki Make cpuidle_idle_call() decide whether or not to stop the tick. First, there are code paths in cpuidle_idle_call() that don't need the tick to be stopped. In particular, the cpuidle_enter_s2idle() path deals with the tick (and with the entire timekeeping for that matter) by itself and it doesn't need the tick to be stopped beforehand. Second, to address the issue with short idle duration predictions by the idle governor after the tick has been stopped, it will be necessary to change the ordering of some governor code with respect to some code in tick_nohz_idle_go_idle(). For this purpose, call tick_nohz_idle_go_idle() in the same branch as cpuidle_select(). Signed-off-by: Rafael J. Wysocki --- -> v2: No changes. --- kernel/sched/idle.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) Index: linux-pm/kernel/sched/idle.c =================================================================== --- linux-pm.orig/kernel/sched/idle.c +++ linux-pm/kernel/sched/idle.c @@ -146,13 +146,14 @@ static void cpuidle_idle_call(void) } /* - * Tell the RCU framework we are entering an idle section, - * so no more rcu read side critical sections and one more + * The RCU framework needs to be told that we are entering an idle + * section, so no more rcu read side critical sections and one more * step to the grace period */ - rcu_idle_enter(); if (cpuidle_not_available(drv, dev)) { + tick_nohz_idle_go_idle(false); + rcu_idle_enter(); default_idle_call(); goto exit_idle; } @@ -169,6 +170,9 @@ static void cpuidle_idle_call(void) if (idle_should_enter_s2idle() || dev->use_deepest_state) { if (idle_should_enter_s2idle()) { + tick_nohz_idle_go_idle(false); + rcu_idle_enter(); + entered_state = cpuidle_enter_s2idle(drv, dev); if (entered_state > 0) { local_irq_enable(); @@ -176,9 +180,15 @@ static void cpuidle_idle_call(void) } } + tick_nohz_idle_go_idle(true); + rcu_idle_enter(); + next_state = cpuidle_find_deepest_state(drv, dev); call_cpuidle(drv, dev, next_state); } else { + tick_nohz_idle_go_idle(true); + rcu_idle_enter(); + /* * Ask the cpuidle framework to choose a convenient idle state. */ @@ -248,7 +258,6 @@ static void do_idle(void) tick_nohz_idle_go_idle(false); cpu_idle_poll(); } else { - tick_nohz_idle_go_idle(true); cpuidle_idle_call(); } arch_cpu_idle_exit();