Dear RT folks!
I'm pleased to announce the v4.9.9-rt6 patch set.
Changes since v4.9.9-rt5:
- The timer softirq was woken up under certain circumstances where it
could have been avoided. Patch by Haris Okanovic.
- Alex Goins noticed that a GPL only symbol will be forced on -RT
which is not the case on !RT configurations. This has been resolved.
Known issues
- CPU hotplug got a little better but can deadlock.
The delta patch against v4.9.9-rt6 is appended below and can be found here:
https://cdn.kernel.org/pub/linux/kernel/projects/rt/4.9/incr/patch-4.9.9-rt5-rt6.patch.xz
You can get this release via the git tree at:
git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git v4.9.9-rt6
The RT patch against v4.9.9 can be found here:
https://cdn.kernel.org/pub/linux/kernel/projects/rt/4.9/older/patch-4.9.9-rt6.patch.xz
The split quilt queue is available at:
https://cdn.kernel.org/pub/linux/kernel/projects/rt/4.9/older/patches-4.9.9-rt6.tar.xz
Sebastian
diff --git a/include/linux/mutex_rt.h b/include/linux/mutex_rt.h
--- a/include/linux/mutex_rt.h
+++ b/include/linux/mutex_rt.h
@@ -43,7 +43,12 @@ extern void __lockfunc _mutex_unlock(struct mutex *lock);
#define mutex_lock_killable(l) _mutex_lock_killable(l)
#define mutex_trylock(l) _mutex_trylock(l)
#define mutex_unlock(l) _mutex_unlock(l)
+
+#ifdef CONFIG_DEBUG_MUTEXES
#define mutex_destroy(l) rt_mutex_destroy(&(l)->lock)
+#else
+static inline void mutex_destroy(struct mutex *lock) {}
+#endif
#ifdef CONFIG_DEBUG_LOCK_ALLOC
# define mutex_lock_nested(l, s) _mutex_lock_nested(l, s)
diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -2211,8 +2211,7 @@ void rt_mutex_destroy(struct rt_mutex *lock)
lock->magic = NULL;
#endif
}
-
-EXPORT_SYMBOL_GPL(rt_mutex_destroy);
+EXPORT_SYMBOL(rt_mutex_destroy);
/**
* __rt_mutex_init - initialize the rt lock
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -206,6 +206,8 @@ struct timer_base {
bool is_idle;
DECLARE_BITMAP(pending_map, WHEEL_SIZE);
struct hlist_head vectors[WHEEL_SIZE];
+ struct hlist_head expired_lists[LVL_DEPTH];
+ int expired_count;
} ____cacheline_aligned;
static DEFINE_PER_CPU(struct timer_base, timer_bases[NR_BASES]);
@@ -1353,7 +1355,8 @@ static void call_timer_fn(struct timer_list *timer, void (*fn)(unsigned long),
}
}
-static void expire_timers(struct timer_base *base, struct hlist_head *head)
+static inline void __expire_timers(struct timer_base *base,
+ struct hlist_head *head)
{
while (!hlist_empty(head)) {
struct timer_list *timer;
@@ -1384,21 +1387,38 @@ static void expire_timers(struct timer_base *base, struct hlist_head *head)
}
}
-static int __collect_expired_timers(struct timer_base *base,
- struct hlist_head *heads)
+static void expire_timers(struct timer_base *base)
+{
+ struct hlist_head *head;
+
+ while (base->expired_count--) {
+ head = base->expired_lists + base->expired_count;
+ __expire_timers(base, head);
+ }
+ base->expired_count = 0;
+}
+
+static void __collect_expired_timers(struct timer_base *base)
{
unsigned long clk = base->clk;
struct hlist_head *vec;
- int i, levels = 0;
+ int i;
unsigned int idx;
+ /*
+ * expire_timers() must be called at least once before we can
+ * collect more timers
+ */
+ if (WARN_ON(base->expired_count))
+ return;
+
for (i = 0; i < LVL_DEPTH; i++) {
idx = (clk & LVL_MASK) + i * LVL_SIZE;
if (__test_and_clear_bit(idx, base->pending_map)) {
vec = base->vectors + idx;
- hlist_move_list(vec, heads++);
- levels++;
+ hlist_move_list(vec,
+ &base->expired_lists[base->expired_count++]);
}
/* Is it time to look at the next level? */
if (clk & LVL_CLK_MASK)
@@ -1406,7 +1426,6 @@ static int __collect_expired_timers(struct timer_base *base,
/* Shift clock for the next level granularity */
clk >>= LVL_CLK_SHIFT;
}
- return levels;
}
#ifdef CONFIG_NO_HZ_COMMON
@@ -1599,8 +1618,7 @@ void timer_clear_idle(void)
base->is_idle = false;
}
-static int collect_expired_timers(struct timer_base *base,
- struct hlist_head *heads)
+static void collect_expired_timers(struct timer_base *base)
{
/*
* NOHZ optimization. After a long idle sleep we need to forward the
@@ -1617,20 +1635,49 @@ static int collect_expired_timers(struct timer_base *base,
if (time_after(next, jiffies)) {
/* The call site will increment clock! */
base->clk = jiffies - 1;
- return 0;
+ return;
}
base->clk = next;
}
- return __collect_expired_timers(base, heads);
+ __collect_expired_timers(base);
}
#else
-static inline int collect_expired_timers(struct timer_base *base,
- struct hlist_head *heads)
+static inline void collect_expired_timers(struct timer_base *base)
{
- return __collect_expired_timers(base, heads);
+ __collect_expired_timers(base);
}
#endif
+static int find_expired_timers(struct timer_base *base)
+{
+ const unsigned long int end_clk = jiffies;
+
+ while (!base->expired_count && time_after_eq(end_clk, base->clk)) {
+ collect_expired_timers(base);
+ base->clk++;
+ }
+
+ return base->expired_count;
+}
+
+/* Called from CPU tick routine to quickly collect expired timers */
+static int tick_find_expired(struct timer_base *base)
+{
+ int count;
+
+ raw_spin_lock(&base->lock);
+
+ if (unlikely(time_after(jiffies, base->clk + HZ))) {
+ /* defer to ktimersoftd; don't spend too long in irq context */
+ count = -1;
+ } else
+ count = find_expired_timers(base);
+
+ raw_spin_unlock(&base->lock);
+
+ return count;
+}
+
/*
* Called from the timer interrupt handler to charge one tick to the current
* process. user_tick is 1 if the tick is user time, 0 for system.
@@ -1657,22 +1704,11 @@ void update_process_times(int user_tick)
*/
static inline void __run_timers(struct timer_base *base)
{
- struct hlist_head heads[LVL_DEPTH];
- int levels;
-
- if (!time_after_eq(jiffies, base->clk))
- return;
-
raw_spin_lock_irq(&base->lock);
- while (time_after_eq(jiffies, base->clk)) {
+ while (find_expired_timers(base))
+ expire_timers(base);
- levels = collect_expired_timers(base, heads);
- base->clk++;
-
- while (levels--)
- expire_timers(base, heads + levels);
- }
raw_spin_unlock_irq(&base->lock);
wakeup_timer_waiters(base);
}
@@ -1700,12 +1736,12 @@ void run_local_timers(void)
hrtimer_run_queues();
/* Raise the softirq only if required. */
- if (time_before(jiffies, base->clk)) {
+ if (time_before(jiffies, base->clk) || !tick_find_expired(base)) {
if (!IS_ENABLED(CONFIG_NO_HZ_COMMON) || !base->nohz_active)
return;
/* CPU is awake, so check the deferrable base. */
base++;
- if (time_before(jiffies, base->clk))
+ if (time_before(jiffies, base->clk) || !tick_find_expired(base))
return;
}
raise_softirq(TIMER_SOFTIRQ);
@@ -1875,6 +1911,7 @@ int timers_dead_cpu(unsigned int cpu)
raw_spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
BUG_ON(old_base->running_timer);
+ BUG_ON(old_base->expired_count);
for (i = 0; i < WHEEL_SIZE; i++)
migrate_timer_list(new_base, old_base->vectors + i);
@@ -1901,6 +1938,7 @@ static void __init init_timer_cpu(int cpu)
#ifdef CONFIG_PREEMPT_RT_FULL
init_swait_queue_head(&base->wait_for_running_timer);
#endif
+ base->expired_count = 0;
}
}
diff --git a/localversion-rt b/localversion-rt
--- a/localversion-rt
+++ b/localversion-rt
@@ -1 +1 @@
--rt5
+-rt6
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
