From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965104AbcDLPCY (ORCPT ); Tue, 12 Apr 2016 11:02:24 -0400 Received: from e37.co.us.ibm.com ([32.97.110.158]:35964 "EHLO e37.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964872AbcDLPCV (ORCPT ); Tue, 12 Apr 2016 11:02:21 -0400 X-IBM-Helo: d03dlp02.boulder.ibm.com X-IBM-MailFrom: paulmck@linux.vnet.ibm.com X-IBM-RcptTo: linux-kernel@vger.kernel.org From: "Paul E. McKenney" To: linux-kernel@vger.kernel.org Cc: mingo@kernel.org, jiangshanlai@gmail.com, dipankar@in.ibm.com, akpm@linux-foundation.org, mathieu.desnoyers@efficios.com, josh@joshtriplett.org, tglx@linutronix.de, peterz@infradead.org, rostedt@goodmis.org, dhowells@redhat.com, edumazet@google.com, dvhart@linux.intel.com, fweisbec@gmail.com, oleg@redhat.com, bobby.prani@gmail.com, "Paul E. McKenney" Subject: [PATCH tip/core/rcu 4/6] rcu: Awaken grace-period kthread if too long since FQS Date: Tue, 12 Apr 2016 08:02:20 -0700 Message-Id: <1460473342-20836-4-git-send-email-paulmck@linux.vnet.ibm.com> X-Mailer: git-send-email 2.5.2 In-Reply-To: <20160412150157.GA19367@linux.vnet.ibm.com> References: <20160412150157.GA19367@linux.vnet.ibm.com> X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 16041215-0025-0000-0000-00002CD80C0B Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Recent kernels can fail to awaken the grace-period kthread for quiescent-state forcing. This commit is a crude hack that does a wakeup if a scheduling-clock interrupt sees that it has been too long since force-quiescent-state (FQS) processing. Signed-off-by: Paul E. McKenney --- kernel/rcu/tree.c | 39 +++++++++++++++++++++++++++++++++++++-- kernel/rcu/tree.h | 2 ++ 2 files changed, 39 insertions(+), 2 deletions(-) diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index 6116cfad18ff..a739292be605 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -385,9 +385,11 @@ module_param(qlowmark, long, 0444); static ulong jiffies_till_first_fqs = ULONG_MAX; static ulong jiffies_till_next_fqs = ULONG_MAX; +static bool rcu_kick_kthreads; module_param(jiffies_till_first_fqs, ulong, 0644); module_param(jiffies_till_next_fqs, ulong, 0644); +module_param(rcu_kick_kthreads, bool, 0644); /* * How long the grace period must be before we start recruiting @@ -1251,6 +1253,24 @@ static void rcu_dump_cpu_stacks(struct rcu_state *rsp) } } +/* + * If too much time has passed in the current grace period, and if + * so configured, go kick the relevant kthreads. + */ +static void rcu_stall_kick_kthreads(struct rcu_state *rsp) +{ + unsigned long j; + + if (!rcu_kick_kthreads) + return; + j = READ_ONCE(rsp->jiffies_kick_kthreads); + if (time_after(jiffies, j) && rsp->gp_kthread) { + WARN_ONCE(1, "Kicking %s grace-period kthread\n", rsp->name); + wake_up_process(rsp->gp_kthread); + WRITE_ONCE(rsp->jiffies_kick_kthreads, j + HZ); + } +} + static void print_other_cpu_stall(struct rcu_state *rsp, unsigned long gpnum) { int cpu; @@ -1262,6 +1282,11 @@ static void print_other_cpu_stall(struct rcu_state *rsp, unsigned long gpnum) struct rcu_node *rnp = rcu_get_root(rsp); long totqlen = 0; + /* Kick and suppress, if so configured. */ + rcu_stall_kick_kthreads(rsp); + if (rcu_cpu_stall_suppress) + return; + /* Only let one CPU complain about others per time interval. */ raw_spin_lock_irqsave_rcu_node(rnp, flags); @@ -1335,6 +1360,11 @@ static void print_cpu_stall(struct rcu_state *rsp) struct rcu_node *rnp = rcu_get_root(rsp); long totqlen = 0; + /* Kick and suppress, if so configured. */ + rcu_stall_kick_kthreads(rsp); + if (rcu_cpu_stall_suppress) + return; + /* * OK, time to rat on ourselves... * See Documentation/RCU/stallwarn.txt for info on how to debug @@ -1379,8 +1409,10 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) unsigned long js; struct rcu_node *rnp; - if (rcu_cpu_stall_suppress || !rcu_gp_in_progress(rsp)) + if ((rcu_cpu_stall_suppress && !rcu_kick_kthreads) || + !rcu_gp_in_progress(rsp)) return; + rcu_stall_kick_kthreads(rsp); j = jiffies; /* @@ -2119,8 +2151,11 @@ static int __noreturn rcu_gp_kthread(void *arg) } ret = 0; for (;;) { - if (!ret) + if (!ret) { rsp->jiffies_force_qs = jiffies + j; + WRITE_ONCE(rsp->jiffies_kick_kthreads, + jiffies + 3 * j); + } trace_rcu_grace_period(rsp->name, READ_ONCE(rsp->gpnum), TPS("fqswait")); diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h index df668c0f9e64..34d3973f7223 100644 --- a/kernel/rcu/tree.h +++ b/kernel/rcu/tree.h @@ -513,6 +513,8 @@ struct rcu_state { unsigned long jiffies_force_qs; /* Time at which to invoke */ /* force_quiescent_state(). */ + unsigned long jiffies_kick_kthreads; /* Time at which to kick */ + /* kthreads, if configured. */ unsigned long n_force_qs; /* Number of calls to */ /* force_quiescent_state(). */ unsigned long n_force_qs_lh; /* ~Number of calls leaving */ -- 2.5.2