[PATCH tip/core/rcu 03/10] rcutorture: Add grace period after CPU offline

From: Paul E. McKenney
Date: Wed Jan 09 2019 - 16:50:26 EST


Beyond a certain point in the CPU-hotplug offline process, timers get
stranded on the outgoing CPU, and won't fire until that CPU comes back
online, which might well be never. This commit therefore adds a hook
in torture_onoff_init() that is invoked from torture_offline(), which
rcutorture uses to occasionally wait for a grace period. This should
result in failures for RCU implementations that rely on stranded timers
eventually firing in the absence of the CPU coming back online.

Reported-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxx>
---
include/linux/torture.h | 3 ++-
kernel/locking/locktorture.c | 2 +-
kernel/rcu/rcutorture.c | 11 ++++++++++-
kernel/torture.c | 6 +++++-
4 files changed, 18 insertions(+), 4 deletions(-)

diff --git a/include/linux/torture.h b/include/linux/torture.h
index 48fad21109fc..f2d3bcbf4337 100644
--- a/include/linux/torture.h
+++ b/include/linux/torture.h
@@ -50,11 +50,12 @@
do { if (verbose) pr_alert("%s" TORTURE_FLAG "!!! %s\n", torture_type, s); } while (0)

/* Definitions for online/offline exerciser. */
+typedef void torture_ofl_func(void);
bool torture_offline(int cpu, long *n_onl_attempts, long *n_onl_successes,
unsigned long *sum_offl, int *min_onl, int *max_onl);
bool torture_online(int cpu, long *n_onl_attempts, long *n_onl_successes,
unsigned long *sum_onl, int *min_onl, int *max_onl);
-int torture_onoff_init(long ooholdoff, long oointerval);
+int torture_onoff_init(long ooholdoff, long oointerval, torture_ofl_func *f);
void torture_onoff_stats(void);
bool torture_onoff_failures(void);

diff --git a/kernel/locking/locktorture.c b/kernel/locking/locktorture.c
index 7d0b0ed74404..c8b348097bb5 100644
--- a/kernel/locking/locktorture.c
+++ b/kernel/locking/locktorture.c
@@ -970,7 +970,7 @@ static int __init lock_torture_init(void)
/* Prepare torture context. */
if (onoff_interval > 0) {
firsterr = torture_onoff_init(onoff_holdoff * HZ,
- onoff_interval * HZ);
+ onoff_interval * HZ, NULL);
if (firsterr)
goto unwind;
}
diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
index 0955f3a20952..9eb9235c1ec9 100644
--- a/kernel/rcu/rcutorture.c
+++ b/kernel/rcu/rcutorture.c
@@ -2243,6 +2243,14 @@ static void rcu_test_debug_objects(void)
#endif /* #else #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */
}

+static void rcutorture_sync(void)
+{
+ static unsigned long n;
+
+ if (cur_ops->sync && !(++n & 0xfff))
+ cur_ops->sync();
+}
+
static int __init
rcu_torture_init(void)
{
@@ -2404,7 +2412,8 @@ rcu_torture_init(void)
firsterr = torture_shutdown_init(shutdown_secs, rcu_torture_cleanup);
if (firsterr)
goto unwind;
- firsterr = torture_onoff_init(onoff_holdoff * HZ, onoff_interval);
+ firsterr = torture_onoff_init(onoff_holdoff * HZ, onoff_interval,
+ rcutorture_sync);
if (firsterr)
goto unwind;
firsterr = rcu_torture_stall_init();
diff --git a/kernel/torture.c b/kernel/torture.c
index bbf6d473e50c..a03ff722352b 100644
--- a/kernel/torture.c
+++ b/kernel/torture.c
@@ -75,6 +75,7 @@ static DEFINE_MUTEX(fullstop_mutex);
static struct task_struct *onoff_task;
static long onoff_holdoff;
static long onoff_interval;
+static torture_ofl_func *onoff_f;
static long n_offline_attempts;
static long n_offline_successes;
static unsigned long sum_offline;
@@ -118,6 +119,8 @@ bool torture_offline(int cpu, long *n_offl_attempts, long *n_offl_successes,
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: offlined %d\n",
torture_type, cpu);
+ if (onoff_f)
+ onoff_f();
(*n_offl_successes)++;
delta = jiffies - starttime;
*sum_offl += delta;
@@ -243,11 +246,12 @@ torture_onoff(void *arg)
/*
* Initiate online-offline handling.
*/
-int torture_onoff_init(long ooholdoff, long oointerval)
+int torture_onoff_init(long ooholdoff, long oointerval, torture_ofl_func *f)
{
#ifdef CONFIG_HOTPLUG_CPU
onoff_holdoff = ooholdoff;
onoff_interval = oointerval;
+ onoff_f = f;
if (onoff_interval <= 0)
return 0;
return torture_create_kthread(torture_onoff, NULL, onoff_task);
--
2.17.1