Placing timers at enqueue time on a target CPU based on dubious heuristics
does not make any sense:
1) Most timer wheel timers are canceled or rearmed before they expire.
2) The heuristics to predict which CPU will be busy when the timer expires
are wrong by definition.
So placing the timers at enqueue wastes precious cycles.
The proper solution to this problem is to always queue the timers on the
local CPU and allow the non pinned timers to be pulled onto a busy CPU at
expiry time.
Therefore split the timer storage into local pinned and global timers:
Local pinned timers are always expired on the CPU on which they have been
queued. Global timers can be expired on any CPU.
As long as a CPU is busy it expires both local and global timers. When a
CPU goes idle it arms for the first expiring local timer. If the first
expiring pinned (local) timer is before the first expiring movable timer,
then no action is required because the CPU will wake up before the first
movable timer expires. If the first expiring movable timer is before the
first expiring pinned (local) timer, then this timer is queued into a idle
timerqueue and eventually expired by some other active CPU.
To avoid global locking the timerqueues are implemented as a hierarchy. The
lowest level of the hierarchy holds the CPUs. The CPUs are associated to
groups of 8, which are seperated per node. If more than one CPU group
exist, then a second level in the hierarchy collects the groups. Depending
on the size of the system more than 2 levels are required. Each group has a
"migrator" which checks the timerqueue during the tick for remote expirable
timers.
If the last CPU in a group goes idle it reports the first expiring event in
the group up to the next group(s) in the hierarchy. If the last CPU goes
idle it arms its timer for the first system wide expiring timer to ensure
that no timer event is missed.
Testing
~~~~~~~
The impact of wasting cycles during enqueue by using the heuristic in
contrast to always queueing the timer on the local CPU was measured with a
micro benchmark. Therefore a timer is enqueued and dequeued in a loop with
1000 repetitions on a isolated CPU. The time the loop takes is measured. A
quater of the remaining CPUs was kept busy. This measurement was repeated
several times. With the patch queue the average duration was reduced by
approximately 25%.
145ns plain v6
109ns v6 with patch queue
Furthermore the impact of residence in deep idle states of an idle system
was investigated. The patch queue doesn't downgrade this behavior.
During testing on a mostly idle machine a ping pong game could be observed:
a process_timeout timer is expired remotely on a non idle CPU. Then the CPU
where the schedule_timeout() was executed to enqueue the timer comes out of
idle and restarts the timer using schedule_timeout() and goes back to idle
again. This is due to the fair scheduler which tries to keep the task on
the CPU which it previously executed on.
Next Steps
~~~~~~~~~~
Deferrable timers are no longer required as they can be converted to global
timers. If a CPU goes idle, a formerly deferrable timer will not prevent
the CPU to sleep as long as possible. Only the last migrator CPU has to
take care of them. This conversation will be done in a follow up series.
Changes vs. v2: https://lore.kernel.org/r/[email protected]/
- Minimize usage of locks by storing data using atomic_cmpxchg() for
migrator information and information about active cpus.
Thanks,
Anna-Maria
Anna-Maria Behnsen (14):
cpufreq: Prepare timer flags for hierarchical timer pull model
tick-sched: Warn when next tick seems to be in the past
timer: Move store of next event into __next_timer_interrupt()
timer: Split next timer interrupt logic
timer: Keep the pinned timers separate from the others
timer: Retrieve next expiry of pinned/non-pinned timers seperately
timer: Rename get_next_timer_interrupt()
timer: Split out "get next timer interrupt" functionality
timer: Add get next timer interrupt functionality for remote CPUs
timer: Check if timers base is handled already
timer: Implement the hierarchical pull model
timer_migration: Add tracepoints
add_timer_on(): Make sure callers have TIMER_PINNED flag
timer: Always queue timers on the local CPU
Richard Cochran (linutronix GmbH) (2):
timer: Restructure internal locking
tick/sched: Split out jiffies update helper function
Thomas Gleixner (1):
timer: Rework idle logic
arch/x86/kernel/tsc_sync.c | 3 +-
drivers/cpufreq/powernv-cpufreq.c | 15 +-
include/linux/cpuhotplug.h | 1 +
include/trace/events/timer_migration.h | 277 ++++++
kernel/time/Makefile | 3 +
kernel/time/clocksource.c | 2 +-
kernel/time/tick-internal.h | 12 +-
kernel/time/tick-sched.c | 50 +-
kernel/time/timer.c | 360 +++++--
kernel/time/timer_migration.c | 1263 ++++++++++++++++++++++++
kernel/time/timer_migration.h | 121 +++
kernel/workqueue.c | 7 +-
12 files changed, 2011 insertions(+), 103 deletions(-)
create mode 100644 include/trace/events/timer_migration.h
create mode 100644 kernel/time/timer_migration.c
create mode 100644 kernel/time/timer_migration.h
--
2.30.2
The timer pull logic needs proper debugging aids. Add tracepoints so the
hierarchical idle machinery can be diagnosed.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
include/trace/events/timer_migration.h | 277 +++++++++++++++++++++++++
kernel/time/timer_migration.c | 24 +++
2 files changed, 301 insertions(+)
create mode 100644 include/trace/events/timer_migration.h
diff --git a/include/trace/events/timer_migration.h b/include/trace/events/timer_migration.h
new file mode 100644
index 000000000000..0c4824056930
--- /dev/null
+++ b/include/trace/events/timer_migration.h
@@ -0,0 +1,277 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM timer_migration
+
+#if !defined(_TRACE_TIMER_MIGRATION_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_TIMER_MIGRATION_H
+
+#include <linux/tracepoint.h>
+
+/* Group events */
+TRACE_EVENT(tmigr_group_set,
+
+ TP_PROTO(struct tmigr_group *group),
+
+ TP_ARGS(group),
+
+ TP_STRUCT__entry(
+ __field( void *, group )
+ __field( unsigned int, lvl )
+ __field( unsigned int, numa_node )
+ ),
+
+ TP_fast_assign(
+ __entry->group = group;
+ __entry->lvl = group->level;
+ __entry->numa_node = group->numa_node;
+ ),
+
+ TP_printk("group=%p lvl=%d numa=%d",
+ __entry->group, __entry->lvl, __entry->numa_node)
+);
+
+TRACE_EVENT(tmigr_connect_child_parent,
+
+ TP_PROTO(struct tmigr_group *child),
+
+ TP_ARGS(child),
+
+ TP_STRUCT__entry(
+ __field( void *, child )
+ __field( void *, parent )
+ __field( unsigned int, lvl )
+ __field( unsigned int, numa_node )
+ __field( unsigned int, num_childs )
+ __field( u32, childmask )
+ ),
+
+ TP_fast_assign(
+ __entry->child = child;
+ __entry->parent = child->parent;
+ __entry->lvl = child->parent->level;
+ __entry->numa_node = child->parent->numa_node;
+ __entry->numa_node = child->parent->num_childs;
+ __entry->childmask = child->childmask;
+ ),
+
+ TP_printk("group=%p childmask=%0x parent=%p lvl=%d numa=%d num_childs=%d",
+ __entry->child, __entry->childmask, __entry->parent,
+ __entry->lvl, __entry->numa_node, __entry->num_childs)
+);
+
+TRACE_EVENT(tmigr_connect_cpu_parent,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc),
+
+ TP_STRUCT__entry(
+ __field( void *, parent )
+ __field( unsigned int, cpu )
+ __field( unsigned int, lvl )
+ __field( unsigned int, numa_node )
+ __field( unsigned int, num_childs )
+ __field( u32, childmask )
+ ),
+
+ TP_fast_assign(
+ __entry->parent = tmc->tmgroup;
+ __entry->cpu = tmc->cpuevt.cpu;
+ __entry->lvl = tmc->tmgroup->level;
+ __entry->numa_node = tmc->tmgroup->numa_node;
+ __entry->numa_node = tmc->tmgroup->num_childs;
+ __entry->childmask = tmc->childmask;
+ ),
+
+ TP_printk("cpu=%d childmask=%0x parent=%p lvl=%d numa=%d num_childs=%d",
+ __entry->cpu, __entry->childmask, __entry->parent,
+ __entry->lvl, __entry->numa_node, __entry->num_childs)
+);
+
+DECLARE_EVENT_CLASS(tmigr_group_and_cpu,
+
+ TP_PROTO(struct tmigr_group *group, union tmigr_state state, u32 childmask),
+
+ TP_ARGS(group, state, childmask),
+
+ TP_STRUCT__entry(
+ __field( void *, group )
+ __field( void *, parent )
+ __field( unsigned int, lvl )
+ __field( unsigned int, numa_node )
+ __field( u8, active )
+ __field( u8, migrator )
+ __field( u32, childmask )
+ ),
+
+ TP_fast_assign(
+ __entry->group = group;
+ __entry->parent = group->parent;
+ __entry->lvl = group->level;
+ __entry->numa_node = group->numa_node;
+ __entry->active = state.active;
+ __entry->migrator = state.migrator;
+ __entry->childmask = childmask;
+ ),
+
+ TP_printk("group=%p lvl=%d numa=%d active=%0x migrator=%0x "
+ "parent=%p childmask=%0x",
+ __entry->group, __entry->lvl, __entry->numa_node,
+ __entry->active, __entry->migrator,
+ __entry->parent, __entry->childmask)
+);
+
+DEFINE_EVENT(tmigr_group_and_cpu, tmigr_group_set_cpu_inactive,
+
+ TP_PROTO(struct tmigr_group *group, union tmigr_state state, u32 childmask),
+
+ TP_ARGS(group, state, childmask)
+);
+
+DEFINE_EVENT(tmigr_group_and_cpu, tmigr_group_set_cpu_active,
+
+ TP_PROTO(struct tmigr_group *group, union tmigr_state state, u32 childmask),
+
+ TP_ARGS(group, state, childmask)
+);
+
+/* CPU events*/
+DECLARE_EVENT_CLASS(tmigr_cpugroup,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc),
+
+ TP_STRUCT__entry(
+ __field( void *, parent)
+ __field( unsigned int, cpu)
+ ),
+
+ TP_fast_assign(
+ __entry->cpu = tmc->cpuevt.cpu;
+ __entry->parent = tmc->tmgroup;
+ ),
+
+ TP_printk("cpu=%d parent=%p", __entry->cpu, __entry->parent)
+);
+
+DEFINE_EVENT(tmigr_cpugroup, tmigr_cpu_new_timer,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc)
+);
+
+DEFINE_EVENT(tmigr_cpugroup, tmigr_cpu_active,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc)
+);
+
+DEFINE_EVENT(tmigr_cpugroup, tmigr_cpu_online,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc)
+);
+
+DEFINE_EVENT(tmigr_cpugroup, tmigr_cpu_offline,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc)
+);
+
+DEFINE_EVENT(tmigr_cpugroup, tmigr_handle_remote_cpu,
+
+ TP_PROTO(struct tmigr_cpu *tmc),
+
+ TP_ARGS(tmc)
+);
+
+TRACE_EVENT(tmigr_cpu_idle,
+
+ TP_PROTO(struct tmigr_cpu *tmc, u64 nextevt),
+
+ TP_ARGS(tmc, nextevt),
+
+ TP_STRUCT__entry(
+ __field( void *, parent)
+ __field( unsigned int, cpu)
+ __field( u64, nextevt)
+ ),
+
+ TP_fast_assign(
+ __entry->cpu = tmc->cpuevt.cpu;
+ __entry->parent = tmc->tmgroup;
+ __entry->nextevt = nextevt;
+ ),
+
+ TP_printk("cpu=%d parent=%p nextevt=%llu",
+ __entry->cpu, __entry->parent, __entry->nextevt)
+);
+
+TRACE_EVENT(tmigr_update_events,
+
+ TP_PROTO(struct tmigr_group *child, struct tmigr_group *group,
+ union tmigr_state childstate, union tmigr_state groupstate,
+ u64 nextevt),
+
+ TP_ARGS(child, group, childstate, groupstate, nextevt),
+
+ TP_STRUCT__entry(
+ __field( void *, child )
+ __field( void *, group )
+ __field( u64, nextevt )
+ __field( u64, group_next_expiry )
+ __field( unsigned int, group_lvl )
+ __field( u8, child_active )
+ __field( u8, group_active )
+ __field( unsigned int, child_evtcpu )
+ __field( u64, child_evt_expiry )
+ ),
+
+ TP_fast_assign(
+ __entry->child = child;
+ __entry->group = group;
+ __entry->nextevt = nextevt;
+ __entry->group_next_expiry = group->next_expiry;
+ __entry->group_lvl = group->level;
+ __entry->child_active = childstate.active;
+ __entry->group_active = groupstate.active;
+ __entry->child_evtcpu = child ? child->groupevt.cpu : 0;
+ __entry->child_evt_expiry = child ? child->groupevt.nextevt.expires : 0;
+ ),
+
+ TP_printk("child=%p group=%p group_lvl=%d child_active=%0x group_active=%0x "
+ "nextevt=%llu next_expiry=%llu child_evt_expiry=%llu child_evtcpu=%d",
+ __entry->child, __entry->group, __entry->group_lvl, __entry->child_active,
+ __entry->group_active,
+ __entry->nextevt, __entry->group_next_expiry, __entry->child_evt_expiry,
+ __entry->child_evtcpu)
+);
+
+TRACE_EVENT(tmigr_handle_remote,
+
+ TP_PROTO(struct tmigr_group *group),
+
+ TP_ARGS(group),
+
+ TP_STRUCT__entry(
+ __field( void * , group )
+ __field( unsigned int , lvl )
+ ),
+
+ TP_fast_assign(
+ __entry->group = group;
+ __entry->lvl = group->level;
+ ),
+
+ TP_printk("group=%p lvl=%d",
+ __entry->group, __entry->lvl)
+);
+
+#endif /* _TRACE_TIMER_MIGRATION_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
diff --git a/kernel/time/timer_migration.c b/kernel/time/timer_migration.c
index d68b8fd06a97..6b954e034959 100644
--- a/kernel/time/timer_migration.c
+++ b/kernel/time/timer_migration.c
@@ -13,6 +13,9 @@
#include "timer_migration.h"
#include "tick-internal.h"
+#define CREATE_TRACE_POINTS
+#include <trace/events/timer_migration.h>
+
/*
* The timer migration mechanism is built on a hierarchy of groups. The
* lowest level group contains CPUs, the next level groups of CPU groups
@@ -317,6 +320,8 @@ static bool tmigr_active_up(struct tmigr_group *group,
*/
set_bit(0, &group->groupevt.ignore);
+ trace_tmigr_group_set_cpu_active(group, newstate, childmask);
+
return walk_done;
}
@@ -341,6 +346,7 @@ void tmigr_cpu_activate(void)
raw_spin_lock(&tmc->lock);
tmc->idle = 0;
tmc->wakeup = KTIME_MAX;
+ trace_tmigr_cpu_active(tmc);
__tmigr_cpu_activate(tmc);
raw_spin_unlock(&tmc->lock);
}
@@ -435,6 +441,9 @@ static bool tmigr_update_events(struct tmigr_group *group,
data->nextexp = tmigr_next_groupevt_expires(group);
}
+ trace_tmigr_update_events(child, group, data->childstate,
+ data->groupstate, nextexp);
+
unlock:
raw_spin_unlock(&group->lock);
@@ -478,6 +487,8 @@ static u64 tmigr_new_timer(struct tmigr_cpu *tmc, u64 nextexp)
if (tmc->remote)
return KTIME_MAX;
+ trace_tmigr_cpu_new_timer(tmc);
+
clear_bit(0, &tmc->cpuevt.ignore);
data.groupstate.state = atomic_read(tmc->tmgroup->migr_state);
@@ -576,6 +587,8 @@ static bool tmigr_inactive_up(struct tmigr_group *group,
}
}
+ trace_tmigr_group_set_cpu_inactive(group, data->childstate, childmask);
+
return walk_done;
}
@@ -652,6 +665,7 @@ u64 tmigr_cpu_deactivate(u64 nextexp)
tmc->idle = 1;
unlock:
+ trace_tmigr_cpu_idle(tmc, ret);
raw_spin_unlock(&tmc->lock);
return ret;
}
@@ -678,6 +692,8 @@ static u64 tmigr_handle_remote_cpu(unsigned int cpu, u64 now,
return next;
}
+ trace_tmigr_handle_remote_cpu(tmc);
+
tmc->remote = 1;
/* Drop the lock to allow the remote CPU to exit idle */
@@ -728,6 +744,7 @@ static bool tmigr_handle_remote_up(struct tmigr_group *group,
childmask = data->childmask;
+ trace_tmigr_handle_remote(group);
again:
/*
* Handle the group only if @childmask is the migrator or if the
@@ -963,6 +980,7 @@ static struct tmigr_group *tmigr_get_group(unsigned int cpu, unsigned int node,
tmigr_init_group(group, lvl, node, migr_state);
/* Setup successful. Add it to the hierarchy */
list_add(&group->list, &tmigr_level_list[lvl]);
+ trace_tmigr_group_set(group);
return group;
}
@@ -981,6 +999,8 @@ static void tmigr_connect_child_parent(struct tmigr_group *child,
raw_spin_unlock(&parent->lock);
raw_spin_unlock_irqrestore(&child->lock, flags);
+ trace_tmigr_connect_child_parent(child);
+
/*
* To prevent inconsistent states, active childs needs to be active
* in new parent as well. Inactive childs are already marked
@@ -1067,6 +1087,8 @@ static int tmigr_setup_groups(unsigned int cpu, unsigned int node)
raw_spin_unlock_irqrestore(&group->lock, flags);
+ trace_tmigr_connect_cpu_parent(tmc);
+
/* There are no childs that needs to be connected */
continue;
} else {
@@ -1135,6 +1157,7 @@ static int tmigr_cpu_online(unsigned int cpu)
tmc->wakeup = KTIME_MAX;
}
raw_spin_lock_irqsave(&tmc->lock, flags);
+ trace_tmigr_cpu_online(tmc);
__tmigr_cpu_activate(tmc);
tmc->online = 1;
raw_spin_unlock_irqrestore(&tmc->lock, flags);
@@ -1148,6 +1171,7 @@ static int tmigr_cpu_offline(unsigned int cpu)
raw_spin_lock_irq(&tmc->lock);
tmc->online = 0;
__tmigr_cpu_deactivate(tmc, KTIME_MAX);
+ trace_tmigr_cpu_offline(tmc);
raw_spin_unlock_irq(&tmc->lock);
return 0;
--
2.30.2
For the conversion of the NOHZ timer placement to a pull at expiry time
model it's required to have seperate expiry times for the pinned and the
non-pinned (movable) timers. Therefore struct timer_events is introduced.
No functional change
Originally-by: Richard Cochran (linutronix GmbH) <[email protected]>
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/tick-internal.h | 8 ++++-
kernel/time/tick-sched.c | 20 ++++++++----
kernel/time/timer.c | 65 ++++++++++++++++++++++++++++---------
3 files changed, 70 insertions(+), 23 deletions(-)
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index 649f2b48e8f0..fcb2d45c2934 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -8,6 +8,11 @@
#include "timekeeping.h"
#include "tick-sched.h"
+struct timer_events {
+ u64 local;
+ u64 global;
+};
+
#ifdef CONFIG_GENERIC_CLOCKEVENTS
# define TICK_DO_TIMER_NONE -1
@@ -163,7 +168,8 @@ static inline void timers_update_nohz(void) { }
DECLARE_PER_CPU(struct hrtimer_cpu_base, hrtimer_bases);
-extern u64 get_next_timer_interrupt(unsigned long basej, u64 basem);
+extern void get_next_timer_interrupt(unsigned long basej, u64 basem,
+ struct timer_events *tevt);
void timer_clear_idle(void);
#define CLOCK_SET_WALL \
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 7ffdc7ba19b4..78f172d1f3d2 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -784,7 +784,8 @@ static inline bool local_timer_softirq_pending(void)
static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
{
- u64 basemono, next_tick, delta, expires;
+ struct timer_events tevt = { .local = KTIME_MAX, .global = KTIME_MAX };
+ u64 basemono, delta, expires;
unsigned long basejiff;
unsigned int seq;
@@ -809,7 +810,11 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
*/
if (rcu_needs_cpu() || arch_needs_cpu() ||
irq_work_needs_cpu() || local_timer_softirq_pending()) {
- next_tick = basemono + TICK_NSEC;
+ /*
+ * If anyone needs the CPU, treat this as a local
+ * timer expiring in a jiffy.
+ */
+ tevt.local = basemono + TICK_NSEC;
} else {
/*
* Get the next pending timer. If high resolution
@@ -818,17 +823,18 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
* disabled this also looks at the next expiring
* hrtimer.
*/
- next_tick = get_next_timer_interrupt(basejiff, basemono);
- ts->next_timer = next_tick;
+ get_next_timer_interrupt(basejiff, basemono, &tevt);
+ tevt.local = min_t(u64, tevt.local, tevt.global);
+ ts->next_timer = tevt.local;
}
/*
* If the tick is due in the next period, keep it ticking or
* force prod the timer.
*/
- WARN_ON_ONCE(basemono > next_tick);
+ WARN_ON_ONCE(basemono > tevt.local);
- delta = next_tick - basemono;
+ delta = tevt.local - basemono;
if (delta <= (u64)TICK_NSEC) {
/*
* Tell the timer code that the base is not idle, i.e. undo
@@ -861,7 +867,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
else
expires = KTIME_MAX;
- ts->timer_expires = min_t(u64, expires, next_tick);
+ ts->timer_expires = min_t(u64, expires, tevt.local);
out:
return ts->timer_expires;
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index b3eea90cb212..069478e65230 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1664,7 +1664,7 @@ static void __next_timer_interrupt(struct timer_base *base)
* Check, if the next hrtimer event is before the next timer wheel
* event:
*/
-static u64 cmp_next_hrtimer_event(u64 basem, u64 expires)
+static void cmp_next_hrtimer_event(u64 basem, struct timer_events *tevt)
{
u64 nextevt = hrtimer_get_next_event();
@@ -1672,15 +1672,17 @@ static u64 cmp_next_hrtimer_event(u64 basem, u64 expires)
* If high resolution timers are enabled
* hrtimer_get_next_event() returns KTIME_MAX.
*/
- if (expires <= nextevt)
- return expires;
+ if (tevt->local <= nextevt)
+ return;
/*
* If the next timer is already expired, return the tick base
* time so the tick is fired immediately.
*/
- if (nextevt <= basem)
- return basem;
+ if (nextevt <= basem) {
+ tevt->local = basem;
+ return;
+ }
/*
* Round up to the next jiffie. High resolution timers are
@@ -1690,7 +1692,7 @@ static u64 cmp_next_hrtimer_event(u64 basem, u64 expires)
*
* Use DIV_ROUND_UP_ULL to prevent gcc calling __divdi3
*/
- return DIV_ROUND_UP_ULL(nextevt, TICK_NSEC) * TICK_NSEC;
+ tevt->local = DIV_ROUND_UP_ULL(nextevt, TICK_NSEC) * TICK_NSEC;
}
@@ -1703,26 +1705,31 @@ static unsigned long next_timer_interrupt(struct timer_base *base)
}
/**
- * get_next_timer_interrupt - return the time (clock mono) of the next timer
+ * get_next_timer_interrupt
* @basej: base time jiffies
* @basem: base time clock monotonic
+ * @tevt: Pointer to the storage for the expiry values
*
- * Returns the tick aligned clock monotonic time of the next pending
- * timer or KTIME_MAX if no timer is pending.
+ * Stores the next pending local and global timer expiry values in the
+ * struct pointed to by @tevt. If a queue is empty the corresponding field
+ * is set to KTIME_MAX.
*/
-u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
+void get_next_timer_interrupt(unsigned long basej, u64 basem,
+ struct timer_events *tevt)
{
unsigned long nextevt, nextevt_local, nextevt_global;
struct timer_base *base_local, *base_global;
bool local_first, is_idle;
- u64 expires = KTIME_MAX;
+
+ /* Preset local / global events */
+ tevt->local = tevt->global = KTIME_MAX;
/*
* Pretend that there is no timer pending if the cpu is offline.
* Possible pending timers will be migrated later to an active cpu.
*/
if (cpu_is_offline(smp_processor_id()))
- return expires;
+ return;
base_local = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
base_global = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
@@ -1779,16 +1786,44 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
/* We need to mark both bases in sync */
base_local->is_idle = base_global->is_idle = is_idle;
- if (base_local->timers_pending || base_global->timers_pending) {
+ /*
+ * If the bases are not marked idle, i.e one of the events is at
+ * max. one tick away, then the CPU can't go into a NOHZ idle
+ * sleep. Use the earlier event of both and store it in the local
+ * expiry value. The next global event is irrelevant in this case
+ * and can be left as KTIME_MAX. CPU will wakeup on time.
+ */
+ if (!is_idle) {
/* If we missed a tick already, force 0 delta */
if (time_before_eq(nextevt, basej))
nextevt = basej;
- expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
+ tevt->local = basem + (u64)(nextevt - basej) * TICK_NSEC;
+ goto unlock;
}
+
+ /*
+ * If the bases are marked idle, i.e. the next event on both the
+ * local and the global queue are farther away than a tick,
+ * evaluate both bases. No need to check whether one of the bases
+ * has an already expired timer as this is caught by the !is_idle
+ * condition above.
+ */
+ if (base_local->timers_pending)
+ tevt->local = basem + (u64)(nextevt_local - basej) * TICK_NSEC;
+
+ /*
+ * If the local queue expires first, then the global event can be
+ * ignored. The CPU wakes up before that. If the global queue is
+ * empty, nothing to do either.
+ */
+ if (!local_first && base_global->timers_pending)
+ tevt->global = basem + (u64)(nextevt_global - basej) * TICK_NSEC;
+
+unlock:
raw_spin_unlock(&base_global->lock);
raw_spin_unlock(&base_local->lock);
- return cmp_next_hrtimer_event(basem, expires);
+ cmp_next_hrtimer_event(basem, tevt);
}
/**
--
2.30.2
From: Thomas Gleixner <[email protected]>
To improve readability of the code, split base->idle calculation and
expires calculation into separate parts.
Signed-off-by: Thomas Gleixner <[email protected]>
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/timer.c | 29 ++++++++++++++---------------
1 file changed, 14 insertions(+), 15 deletions(-)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index f34bc75ff848..cb4194ecca60 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1727,21 +1727,20 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
base->clk = nextevt;
}
- if (time_before_eq(nextevt, basej)) {
- expires = basem;
- base->is_idle = false;
- } else {
- if (base->timers_pending)
- expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
- /*
- * If we expect to sleep more than a tick, mark the base idle.
- * Also the tick is stopped so any added timer must forward
- * the base clk itself to keep granularity small. This idle
- * logic is only maintained for the BASE_STD base, deferrable
- * timers may still see large granularity skew (by design).
- */
- if ((expires - basem) > TICK_NSEC)
- base->is_idle = true;
+ /*
+ * Base is idle if the next event is more than a tick away. Also
+ * the tick is stopped so any added timer must forward the base clk
+ * itself to keep granularity small. This idle logic is only
+ * maintained for the BASE_STD base, deferrable timers may still
+ * see large granularity skew (by design).
+ */
+ base->is_idle = time_after(nextevt, basej + 1);
+
+ if (base->timers_pending) {
+ /* If we missed a tick already, force 0 delta */
+ if (time_before_eq(nextevt, basej))
+ nextevt = basej;
+ expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
}
raw_spin_unlock(&base->lock);
--
2.30.2
When the next tick is in the past, the delta between basemono and the next
tick gets negativ. But the next tick should never be in the past. The
negative effect of a wrong next tick might be a stop of the tick and timers
might expire late.
To prevent expensive debugging when changing underlying code, add a
WARN_ON_ONCE into this code path.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
---
kernel/time/tick-sched.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index b0e3c9205946..7ffdc7ba19b4 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -826,6 +826,8 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
* If the tick is due in the next period, keep it ticking or
* force prod the timer.
*/
+ WARN_ON_ONCE(basemono > next_tick);
+
delta = next_tick - basemono;
if (delta <= (u64)TICK_NSEC) {
/*
--
2.30.2
Seperate the storage space for pinned timers.
This is preparatory work for changing the NOHZ timer placement from a push
at enqueue time to a pull at expiry time model.
No functional change.
Originally-by: Richard Cochran (linutronix GmbH) <[email protected]>
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/timer.c | 109 ++++++++++++++++++++++++++++++++------------
1 file changed, 81 insertions(+), 28 deletions(-)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index cb4194ecca60..b3eea90cb212 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -187,12 +187,18 @@ EXPORT_SYMBOL(jiffies_64);
#define WHEEL_SIZE (LVL_SIZE * LVL_DEPTH)
#ifdef CONFIG_NO_HZ_COMMON
-# define NR_BASES 2
-# define BASE_STD 0
-# define BASE_DEF 1
+/*
+ * If multiple bases need to be locked, use the base ordering for lock
+ * nesting, i.e. lowest number first.
+ */
+# define NR_BASES 3
+# define BASE_LOCAL 0
+# define BASE_GLOBAL 1
+# define BASE_DEF 2
#else
# define NR_BASES 1
-# define BASE_STD 0
+# define BASE_LOCAL 0
+# define BASE_GLOBAL 0
# define BASE_DEF 0
#endif
@@ -902,7 +908,10 @@ static int detach_if_pending(struct timer_list *timer, struct timer_base *base,
static inline struct timer_base *get_timer_cpu_base(u32 tflags, u32 cpu)
{
- struct timer_base *base = per_cpu_ptr(&timer_bases[BASE_STD], cpu);
+ int index = tflags & TIMER_PINNED ? BASE_LOCAL : BASE_GLOBAL;
+ struct timer_base *base;
+
+ base = per_cpu_ptr(&timer_bases[index], cpu);
/*
* If the timer is deferrable and NO_HZ_COMMON is set then we need
@@ -915,7 +924,10 @@ static inline struct timer_base *get_timer_cpu_base(u32 tflags, u32 cpu)
static inline struct timer_base *get_timer_this_cpu_base(u32 tflags)
{
- struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
+ int index = tflags & TIMER_PINNED ? BASE_LOCAL : BASE_GLOBAL;
+ struct timer_base *base;
+
+ base = this_cpu_ptr(&timer_bases[index]);
/*
* If the timer is deferrable and NO_HZ_COMMON is set then we need
@@ -1700,9 +1712,10 @@ static unsigned long next_timer_interrupt(struct timer_base *base)
*/
u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
{
- struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
+ unsigned long nextevt, nextevt_local, nextevt_global;
+ struct timer_base *base_local, *base_global;
+ bool local_first, is_idle;
u64 expires = KTIME_MAX;
- unsigned long nextevt;
/*
* Pretend that there is no timer pending if the cpu is offline.
@@ -1711,38 +1724,69 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
if (cpu_is_offline(smp_processor_id()))
return expires;
- raw_spin_lock(&base->lock);
+ base_local = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
+ base_global = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
- nextevt = next_timer_interrupt(base);
+ raw_spin_lock(&base_local->lock);
+ raw_spin_lock_nested(&base_global->lock, SINGLE_DEPTH_NESTING);
+
+ nextevt_local = next_timer_interrupt(base_local);
+ nextevt_global = next_timer_interrupt(base_global);
/*
* We have a fresh next event. Check whether we can forward the
* base. We can only do that when @basej is past base->clk
* otherwise we might rewind base->clk.
*/
- if (time_after(basej, base->clk)) {
- if (time_after(nextevt, basej))
- base->clk = basej;
- else if (time_after(nextevt, base->clk))
- base->clk = nextevt;
+ if (time_after(basej, base_local->clk)) {
+ if (time_after(nextevt_local, basej))
+ base_local->clk = basej;
+ else if (time_after(nextevt_local, base_local->clk))
+ base_local->clk = nextevt_local;
+ }
+
+ if (time_after(basej, base_global->clk)) {
+ if (time_after(nextevt_global, basej))
+ base_global->clk = basej;
+ else if (time_after(nextevt_global, base_global->clk))
+ base_global->clk = nextevt_global;
}
/*
- * Base is idle if the next event is more than a tick away. Also
+ * Check whether the local event is expiring before or at the same
+ * time as the global event.
+ *
+ * Note, that nextevt_global and nextevt_local might be based on
+ * different base->clk values. So it's not guaranteed that
+ * comparing with empty bases results in a correct local_first.
+ */
+ if (base_local->timers_pending && base_global->timers_pending)
+ local_first = time_before_eq(nextevt_local, nextevt_global);
+ else
+ local_first = base_local->timers_pending;
+
+ nextevt = local_first ? nextevt_local : nextevt_global;
+
+ /*
+ * Bases are idle if the next event is more than a tick away. Also
* the tick is stopped so any added timer must forward the base clk
* itself to keep granularity small. This idle logic is only
- * maintained for the BASE_STD base, deferrable timers may still
- * see large granularity skew (by design).
+ * maintained for the BASE_LOCAL and BASE_GLOBAL base, deferrable
+ * timers may still see large granularity skew (by design).
*/
- base->is_idle = time_after(nextevt, basej + 1);
+ is_idle = time_after(nextevt, basej + 1);
+
+ /* We need to mark both bases in sync */
+ base_local->is_idle = base_global->is_idle = is_idle;
- if (base->timers_pending) {
+ if (base_local->timers_pending || base_global->timers_pending) {
/* If we missed a tick already, force 0 delta */
if (time_before_eq(nextevt, basej))
nextevt = basej;
expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
}
- raw_spin_unlock(&base->lock);
+ raw_spin_unlock(&base_global->lock);
+ raw_spin_unlock(&base_local->lock);
return cmp_next_hrtimer_event(basem, expires);
}
@@ -1754,7 +1798,7 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
*/
void timer_clear_idle(void)
{
- struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
+ struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
/*
* We do this unlocked. The worst outcome is a remote enqueue sending
@@ -1763,6 +1807,9 @@ void timer_clear_idle(void)
* the lock in the exit from idle path.
*/
base->is_idle = false;
+
+ base = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
+ base->is_idle = false;
}
#endif
@@ -1808,11 +1855,13 @@ static inline void __run_timers(struct timer_base *base)
*/
static __latent_entropy void run_timer_softirq(struct softirq_action *h)
{
- struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
+ struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
__run_timers(base);
- if (IS_ENABLED(CONFIG_NO_HZ_COMMON))
+ if (IS_ENABLED(CONFIG_NO_HZ_COMMON)) {
+ __run_timers(this_cpu_ptr(&timer_bases[BASE_GLOBAL]));
__run_timers(this_cpu_ptr(&timer_bases[BASE_DEF]));
+ }
}
/*
@@ -1820,17 +1869,21 @@ static __latent_entropy void run_timer_softirq(struct softirq_action *h)
*/
static void run_local_timers(void)
{
- struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
+ struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
hrtimer_run_queues();
/* Raise the softirq only if required. */
if (time_before(jiffies, base->next_expiry)) {
if (!IS_ENABLED(CONFIG_NO_HZ_COMMON))
return;
- /* CPU is awake, so check the deferrable base. */
+ /* CPU is awake, so check for the global base. */
base++;
- if (time_before(jiffies, base->next_expiry))
- return;
+ if (time_before(jiffies, base->next_expiry)) {
+ /* CPU is awake, so check the deferrable base. */
+ base++;
+ if (time_before(jiffies, base->next_expiry))
+ return;
+ }
}
raise_softirq(TIMER_SOFTIRQ);
}
--
2.30.2
The hierachical timer pull model at expiry time is now in place. Timers
which should expire on a dedicated CPU needs the TIMER_PINNED
flag. Otherwise they will be queued on the dedicated CPU but in global
timer base and those timers could also expire on other CPUs. Only timers
with TIMER_PINNED flag will end up in local timer base
Therefore add the missing TIMER_PINNED flag for those who use
add_timer_on() without the flag.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
arch/x86/kernel/tsc_sync.c | 3 ++-
kernel/time/clocksource.c | 2 +-
kernel/workqueue.c | 7 +++++--
3 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c
index 9452dc9664b5..eab827288e0f 100644
--- a/arch/x86/kernel/tsc_sync.c
+++ b/arch/x86/kernel/tsc_sync.c
@@ -110,7 +110,8 @@ static int __init start_sync_check_timer(void)
if (!cpu_feature_enabled(X86_FEATURE_TSC_ADJUST) || tsc_clocksource_reliable)
return 0;
- timer_setup(&tsc_sync_check_timer, tsc_sync_check_timer_fn, 0);
+ timer_setup(&tsc_sync_check_timer, tsc_sync_check_timer_fn,
+ TIMER_PINNED);
tsc_sync_check_timer.expires = jiffies + SYNC_CHECK_INTERVAL;
add_timer(&tsc_sync_check_timer);
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index 8058bec87ace..f8c310e62758 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -523,7 +523,7 @@ static inline void clocksource_start_watchdog(void)
{
if (watchdog_running || !watchdog || list_empty(&watchdog_list))
return;
- timer_setup(&watchdog_timer, clocksource_watchdog, 0);
+ timer_setup(&watchdog_timer, clocksource_watchdog, TIMER_PINNED);
watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask));
watchdog_running = 1;
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 7cd5f5e7e0a1..a0f7bf7be6f2 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -1670,10 +1670,13 @@ static void __queue_delayed_work(int cpu, struct workqueue_struct *wq,
dwork->cpu = cpu;
timer->expires = jiffies + delay;
- if (unlikely(cpu != WORK_CPU_UNBOUND))
+ if (unlikely(cpu != WORK_CPU_UNBOUND)) {
+ timer->flags |= TIMER_PINNED;
add_timer_on(timer, cpu);
- else
+ } else {
+ timer->flags &= ~TIMER_PINNED;
add_timer(timer);
+ }
}
/**
--
2.30.2
To prepare for the conversion of the NOHZ timer placement to a pull at
expiry time model it's required to have functionality available getting the
next timer interrupt on a remote CPU.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/tick-internal.h | 3 +++
kernel/time/timer.c | 41 +++++++++++++++++++++++++++++++++++++
2 files changed, 44 insertions(+)
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index 6f5f164506d5..7b65abc9d803 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -170,6 +170,9 @@ DECLARE_PER_CPU(struct hrtimer_cpu_base, hrtimer_bases);
extern void forward_and_idle_timer_bases(unsigned long basej, u64 basem,
struct timer_events *tevt);
+extern void get_next_timer_interrupt_remote(unsigned long basej, u64 basem,
+ struct timer_events *tevt,
+ unsigned int cpu);
void timer_clear_idle(void);
#define CLOCK_SET_WALL \
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 167ae89f0f46..54631e86763f 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1744,6 +1744,47 @@ static unsigned long get_next_timer_interrupt(struct timer_base *base_local,
return local_first ? nextevt_local : nextevt_global;
}
+/**
+ * get_next_timer_interrupt_remote
+ * @basej: base time jiffies
+ * @basem: base time clock monotonic
+ * @tevt: Pointer to the storage for the expiry values
+ * @cpu: Remote CPU
+ *
+ * Stores the next pending local and global timer expiry values in the
+ * struct pointed to by @tevt. If a queue is empty the corresponding
+ * field is set to KTIME_MAX. If local event expires before global
+ * event, global event is set to KTIME_MAX as well.
+ */
+void get_next_timer_interrupt_remote(unsigned long basej, u64 basem,
+ struct timer_events *tevt,
+ unsigned int cpu)
+{
+ struct timer_base *base_local, *base_global;
+ unsigned long flags;
+
+ /* Preset local / global events */
+ tevt->local = tevt->global = KTIME_MAX;
+
+ /*
+ * Pretend that there is no timer pending if the cpu is offline.
+ * Possible pending timers will be migrated later to an active cpu.
+ */
+ if (cpu_is_offline(cpu))
+ return;
+
+ base_local = per_cpu_ptr(&timer_bases[BASE_LOCAL], cpu);
+ base_global = per_cpu_ptr(&timer_bases[BASE_GLOBAL], cpu);
+
+ raw_spin_lock_irqsave(&base_local->lock, flags);
+ raw_spin_lock_nested(&base_global->lock, SINGLE_DEPTH_NESTING);
+
+ get_next_timer_interrupt(base_local, base_global, basej, basem, tevt);
+
+ raw_spin_unlock(&base_global->lock);
+ raw_spin_unlock_irqrestore(&base_local->lock, flags);
+}
+
/**
* forward_and_idle_timer_bases
* @basej: base time jiffies
--
2.30.2
From: "Richard Cochran (linutronix GmbH)" <[email protected]>
The logic to get the time of the last jiffies update will be needed by
the timer pull model as well.
Move the code into a global funtion in anticipation of the new caller.
No functional change.
Signed-off-by: Richard Cochran (linutronix GmbH) <[email protected]>
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/tick-internal.h | 1 +
kernel/time/tick-sched.c | 20 ++++++++++++++++----
2 files changed, 17 insertions(+), 4 deletions(-)
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index 7b65abc9d803..8f99e35e75b4 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -158,6 +158,7 @@ static inline void tick_nohz_init(void) { }
#ifdef CONFIG_NO_HZ_COMMON
extern unsigned long tick_nohz_active;
extern void timers_update_nohz(void);
+extern u64 get_jiffies_update(unsigned long *basej);
# ifdef CONFIG_SMP
extern struct static_key_false timers_migration_enabled;
# endif
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 7f7bfe8b498d..868679ff3421 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -782,19 +782,31 @@ static inline bool local_timer_softirq_pending(void)
return local_softirq_pending() & BIT(TIMER_SOFTIRQ);
}
-static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
+/*
+ * Read jiffies and the time when jiffies were updated last
+ */
+u64 get_jiffies_update(unsigned long *basej)
{
- struct timer_events tevt = { .local = KTIME_MAX, .global = KTIME_MAX };
- u64 basemono, delta, expires;
unsigned long basejiff;
unsigned int seq;
+ u64 basemono;
- /* Read jiffies and the time when jiffies were updated last */
do {
seq = read_seqcount_begin(&jiffies_seq);
basemono = last_jiffies_update;
basejiff = jiffies;
} while (read_seqcount_retry(&jiffies_seq, seq));
+ *basej = basejiff;
+ return basemono;
+}
+
+static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
+{
+ struct timer_events tevt = { .local = KTIME_MAX, .global = KTIME_MAX };
+ u64 basemono, delta, expires;
+ unsigned long basejiff;
+
+ basemono = get_jiffies_update(&basejiff);
ts->last_jiffies = basejiff;
ts->timer_expires_base = basemono;
--
2.30.2
Placing timers at enqueue time on a target CPU based on dubious heuristics
does not make any sense:
1) Most timer wheel timers are canceled or rearmed before they expire.
2) The heuristics to predict which CPU will be busy when the timer expires
are wrong by definition.
So placing the timers at enqueue wastes precious cycles.
The proper solution to this problem is to always queue the timers on the
local CPU and allow the non pinned timers to be pulled onto a busy CPU at
expiry time.
Therefore split the timer storage into local pinned and global timers:
Local pinned timers are always expired on the CPU on which they have been
queued. Global timers can be expired on any CPU.
As long as a CPU is busy it expires both local and global timers. When a
CPU goes idle it arms for the first expiring local timer. If the first
expiring pinned (local) timer is before the first expiring movable timer,
then no action is required because the CPU will wake up before the first
movable timer expires. If the first expiring movable timer is before the
first expiring pinned (local) timer, then this timer is queued into a idle
timerqueue and eventually expired by some other active CPU.
To avoid global locking the timerqueues are implemented as a hierarchy. The
lowest level of the hierarchy holds the CPUs. The CPUs are associated to
groups of 8, which are seperated per node. If more than one CPU group
exist, then a second level in the hierarchy collects the groups. Depending
on the size of the system more than 2 levels are required. Each group has a
"migrator" which checks the timerqueue during the tick for remote expirable
timers.
If the last CPU in a group goes idle it reports the first expiring event in
the group up to the next group(s) in the hierarchy. If the last CPU goes
idle it arms its timer for the first system wide expiring timer to ensure
that no timer event is missed.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
include/linux/cpuhotplug.h | 1 +
kernel/time/Makefile | 3 +
kernel/time/timer.c | 61 +-
kernel/time/timer_migration.c | 1239 +++++++++++++++++++++++++++++++++
kernel/time/timer_migration.h | 121 ++++
5 files changed, 1420 insertions(+), 5 deletions(-)
create mode 100644 kernel/time/timer_migration.c
create mode 100644 kernel/time/timer_migration.h
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index f61447913db9..95e89495251a 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -244,6 +244,7 @@ enum cpuhp_state {
CPUHP_AP_PERF_POWERPC_HV_24x7_ONLINE,
CPUHP_AP_PERF_POWERPC_HV_GPCI_ONLINE,
CPUHP_AP_PERF_CSKY_ONLINE,
+ CPUHP_AP_TMIGR_ONLINE,
CPUHP_AP_WATCHDOG_ONLINE,
CPUHP_AP_WORKQUEUE_ONLINE,
CPUHP_AP_RANDOM_ONLINE,
diff --git a/kernel/time/Makefile b/kernel/time/Makefile
index 7e875e63ff3b..4af2a264a160 100644
--- a/kernel/time/Makefile
+++ b/kernel/time/Makefile
@@ -17,6 +17,9 @@ endif
obj-$(CONFIG_GENERIC_SCHED_CLOCK) += sched_clock.o
obj-$(CONFIG_TICK_ONESHOT) += tick-oneshot.o tick-sched.o
obj-$(CONFIG_LEGACY_TIMER_TICK) += tick-legacy.o
+ifeq ($(CONFIG_SMP),y)
+ obj-$(CONFIG_NO_HZ_COMMON) += timer_migration.o
+endif
obj-$(CONFIG_HAVE_GENERIC_VDSO) += vsyscall.o
obj-$(CONFIG_DEBUG_FS) += timekeeping_debug.o
obj-$(CONFIG_TEST_UDELAY) += test_udelay.o
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 0790ec8efe82..67f9cc077ff7 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -53,6 +53,7 @@
#include <asm/io.h>
#include "tick-internal.h"
+#include "timer_migration.h"
#define CREATE_TRACE_POINTS
#include <trace/events/timer.h>
@@ -592,10 +593,13 @@ trigger_dyntick_cpu(struct timer_base *base, struct timer_list *timer)
/*
* We might have to IPI the remote CPU if the base is idle and the
- * timer is not deferrable. If the other CPU is on the way to idle
- * then it can't set base->is_idle as we hold the base lock:
+ * timer is pinned. If it is a non pinned timer, it is only queued
+ * on the remote CPU, when timer was running during queueing. Then
+ * everything is handled by remote CPU anyway. If the other CPU is
+ * on the way to idle then it can't set base->is_idle as we hold
+ * the base lock:
*/
- if (base->is_idle)
+ if (base->is_idle && timer->flags & TIMER_PINNED)
wake_up_nohz_cpu(base->cpu);
}
@@ -1719,7 +1723,7 @@ static unsigned long get_next_timer_interrupt(struct timer_base *base_local,
* Check whether the local event is expiring before or at the same
* time as the global event.
*
- * Note, that nextevt_global and next_evtlocal might be based on
+ * Note, that nextevt_global and nextevt_local might be based on
* different base->clk values. So it's not guaranteed that
* comparing with empty bases results in a correct local_first.
*/
@@ -1856,6 +1860,36 @@ void forward_and_idle_timer_bases(unsigned long basej, u64 basem,
*/
is_idle = time_after(nextevt, basej + 1);
+ if (is_idle) {
+ u64 next_tmigr;
+
+ next_tmigr = tmigr_cpu_deactivate(tevt->global);
+
+ tevt->global = KTIME_MAX;
+
+ /*
+ * If CPU is the last going idle in timer migration
+ * hierarchy, make sure CPU will wake up in time to handle
+ * remote timers. next_tmigr == KTIME_MAX if other CPUs are
+ * still active.
+ */
+ if (next_tmigr < tevt->local) {
+ u64 tmp;
+
+ /* If we missed a tick already, force 0 delta */
+ if (next_tmigr < basem)
+ next_tmigr = basem;
+
+ tmp = div_u64(next_tmigr - basem, TICK_NSEC);
+
+ nextevt = basej + (unsigned long)tmp;
+ tevt->local = next_tmigr;
+ is_idle = time_after(nextevt, basej + 1);
+ } else {
+ nextevt = nextevt_local;
+ }
+ }
+
/* We need to mark both bases in sync */
base_local->is_idle = base_global->is_idle = is_idle;
@@ -1899,6 +1933,9 @@ void timer_clear_idle(void)
base = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
base->is_idle = false;
+
+ /* Activate without holding the timer_base->lock */
+ tmigr_cpu_activate();
}
#endif
@@ -1955,6 +1992,15 @@ static void run_timer_base(int index)
__run_timer_base(base);
}
+#ifdef CONFIG_SMP
+void timer_expire_remote(unsigned int cpu)
+{
+ struct timer_base *base = per_cpu_ptr(&timer_bases[BASE_GLOBAL], cpu);
+
+ __run_timer_base(base);
+}
+#endif
+
/*
* This function runs timers and the timer-tq in bottom half context.
*/
@@ -1964,6 +2010,9 @@ static __latent_entropy void run_timer_softirq(struct softirq_action *h)
if (IS_ENABLED(CONFIG_NO_HZ_COMMON)) {
run_timer_base(BASE_GLOBAL);
run_timer_base(BASE_DEF);
+
+ if (is_timers_nohz_active())
+ tmigr_handle_remote();
}
}
@@ -1985,7 +2034,9 @@ static void run_local_timers(void)
/* CPU is awake, so check the deferrable base. */
base++;
if (time_before(jiffies, base->next_expiry))
- return;
+ /* CPU is awake, so check tmigr hierarchy. */
+ if (!tmigr_requires_handle_remote())
+ return;
}
}
raise_softirq(TIMER_SOFTIRQ);
diff --git a/kernel/time/timer_migration.c b/kernel/time/timer_migration.c
new file mode 100644
index 000000000000..d68b8fd06a97
--- /dev/null
+++ b/kernel/time/timer_migration.c
@@ -0,0 +1,1239 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Infrastructure for migrateable timers
+ *
+ * Copyright(C) 2022 linutronix GmbH
+ */
+#include <linux/cpuhotplug.h>
+#include <linux/slab.h>
+#include <linux/smp.h>
+#include <linux/spinlock.h>
+#include <linux/timerqueue.h>
+
+#include "timer_migration.h"
+#include "tick-internal.h"
+
+/*
+ * The timer migration mechanism is built on a hierarchy of groups. The
+ * lowest level group contains CPUs, the next level groups of CPU groups
+ * and so forth. The CPU groups are kept per node so for the normal case
+ * lock contention won't happen accross nodes. Depending on the number of
+ * CPUs per node even the next level might be kept as groups of CPU groups
+ * per node and only the levels above cross the node topology.
+ *
+ * Example topology for a two node system with 24 CPUs each.
+ *
+ * LVL 2 [GRP2:0]
+ * GRP1:0 = GRP1:M
+ *
+ * LVL 1 [GRP1:0] [GRP1:1]
+ * GRP0:0 - GRP0:2 GRP0:3 - GRP0:5
+ *
+ * LVL 0 [GRP0:0] [GRP0:1] [GRP0:2] [GRP0:3] [GRP0:4] [GRP0:5]
+ * CPUS 0-7 8-15 16-23 24-31 32-39 40-47
+ *
+ * The groups hold a timer queue of events sorted by expiry time. These
+ * queues are updated when CPUs go in idle. When they come out of idle
+ * ignore bit of events is set.
+ *
+ * Each group has a designated migrator CPU/group as long as a CPU/group is
+ * active in the group. This designated role is necessary to avoid that all
+ * active CPUs in a group try to migrate expired timers from other cpus,
+ * which would result in massive lock bouncing.
+ *
+ * When a CPU is awake, it checks in it's own timer tick the group
+ * hierarchy up to the point where it is assigned the migrator role or if
+ * no CPU is active, it also checks the groups where no migrator is set
+ * (TMIGR_NONE).
+ *
+ * If it finds expired timers in one of the group queues it pulls them over
+ * from the idle CPU and runs the timer function. After that it updates the
+ * group and the parent groups if required.
+ *
+ * CPUs which go idle arm their CPU local timer hardware for the next local
+ * (pinned) timer event. If the next migrateable timer expires after the
+ * next local timer or the CPU has no migrateable timer pending then the
+ * CPU does not queue an event in the LVL0 group. If the next migrateable
+ * timer expires before the next local timer then the CPU queues that timer
+ * in the LVL0 group. In both cases the CPU marks itself idle in the LVL0
+ * group.
+ *
+ * If the CPU is the migrator of the group then it delegates that role to
+ * the next active CPU in the group or sets migrator to TMIGR_NONE when
+ * there is no active CPU in the group. This delegation needs to be
+ * propagated up the hierarchy so hand over from other leaves can happen at
+ * all hierarchy levels w/o doing a search.
+ *
+ * When the last CPU in the system goes idle, then it drops all migrator
+ * duties up to the top level of the hierarchy (LVL2 in the example). It
+ * then has to make sure, that it arms it's own local hardware timer for
+ * the earliest event in the system.
+ *
+ * Lifetime rules:
+ *
+ * The groups are built up at init time or when CPUs come online. They are
+ * not destroyed when a group becomes empty due to offlining. The group
+ * just won't participate in the hierachary management anymore. Destroying
+ * groups would result in interesting race conditions which would just make
+ * the whole mechanism slow and complex.
+ *
+ * Locking rules:
+ *
+ * For setting up new groups and handling events it's required to lock both
+ * child and parent group. The lock odering is always bottom up. This also
+ * includes the per CPU locks in struct tmigr_cpu. For updating migrator
+ * and active CPU/group information atomic_cmpxchg() is used instead and
+ * only per CPU tmigr_cpu->lock is held.
+ *
+ * During setup of groups tmigr_level_list is required. It is protected by
+ * tmigr_mutex.
+ */
+
+#ifdef DEBUG
+# define DBG_BUG_ON(x) BUG_ON(x)
+#else
+# define DBG_BUG_ON(x)
+#endif
+
+static DEFINE_MUTEX(tmigr_mutex);
+static struct list_head *tmigr_level_list __read_mostly;
+
+static unsigned int tmigr_cores_per_group __read_mostly;
+static unsigned int tmigr_hierarchy_levels __read_mostly;
+static unsigned int tmigr_crossnode_level __read_mostly;
+
+static DEFINE_PER_CPU(struct tmigr_cpu, tmigr_cpu);
+
+#define TMIGR_NONE 0xFF
+#define BIT_CNT 8
+
+static DEFINE_STATIC_KEY_FALSE(tmigr_enabled);
+
+static inline bool is_tmigr_enabled(void)
+{
+ return static_branch_unlikely(&tmigr_enabled);
+}
+
+/*
+ * Returns true, when @childmask corresponds to group migrator or when group
+ * is not active - so no migrator is set.
+ */
+static bool tmigr_check_migrator(struct tmigr_group *group, u32 childmask)
+{
+ union tmigr_state s;
+
+ s.state = atomic_read(group->migr_state);
+
+ if ((s.migrator != (u8)childmask) && (s.migrator != TMIGR_NONE))
+ return false;
+
+ return true;
+}
+
+typedef bool (*up_f)(struct tmigr_group *, struct tmigr_group *, void *);
+
+static void __walk_groups(up_f up, void *data,
+ struct tmigr_cpu *tmc)
+{
+ struct tmigr_group *child = NULL, *group = tmc->tmgroup;
+
+ do {
+ DBG_BUG_ON(group->level >= tmigr_hierarchy_levels);
+
+ if (up(group, child, data))
+ break;
+
+ child = group;
+ group = group->parent;
+ } while (group);
+}
+
+static void walk_groups(up_f up, void *data, struct tmigr_cpu *tmc)
+{
+ lockdep_assert_held(&tmc->lock);
+
+ __walk_groups(up, data, tmc);
+}
+
+/**
+ * struct tmigr_walk - data required for walking the hierarchy
+ * @evt: Pointer to tmigr_event which needs to be queued (of idle
+ * child group)
+ * @childmask: childmask of child group
+ * @nextexp: Next CPU event expiry information which is handed
+ * into tmigr code by timer code
+ * (forward_and_idle_timer_bases()); it is furthermore
+ * used for next event which is queued if timer
+ * migration hierarchy is completely idle
+ * @childstate: tmigr_group->migr_state of child - will be only reread
+ * when cmpxchg in group fails (is required for deactive
+ * path and new timer path)
+ * @groupstate: tmigr_group->migr_state of group - will be only reread
+ * when cmpxchg in group fails (is required for active,
+ * deactive and new timer path)
+ */
+struct tmigr_walk {
+ struct tmigr_event *evt;
+ u32 childmask;
+ u64 nextexp;
+ union tmigr_state childstate;
+ union tmigr_state groupstate;
+};
+
+/**
+ * struct tmigr_remote_data - data required for (check) remote expiry
+ * hierarchy walk
+ * @basej: timer base in jiffies
+ * @now: timer base monotonic
+ * @childmask: childmask of child group
+ * @check: is set to 1 if there is the need to handle remote timers;
+ * required in tmigr_check_handle_remote() only
+ * @wakeup: returns expiry of first timer in idle timer migration hierarchy
+ * to make sure timer is handled in time; it is stored in per CPU
+ * tmigr_cpu struct of CPU which expires remote timers
+ */
+struct tmigr_remote_data {
+ unsigned long basej;
+ u64 now;
+ u32 childmask;
+ int check;
+ u64 wakeup;
+};
+
+/*
+ * Returns next event of timerqueue @group->events
+ *
+ * Removes timers with ignore bits and update next_expiry and event cpu
+ * value in group. Expiry value of group event is updated in
+ * tmigr_update_events() only.
+ */
+static struct tmigr_event *tmigr_next_groupevt(struct tmigr_group *group)
+{
+ struct timerqueue_node *node = NULL;
+ struct tmigr_event *evt = NULL;
+
+ lockdep_assert_held(&group->lock);
+
+ group->next_expiry = KTIME_MAX;
+
+ while ((node = timerqueue_getnext(&group->events))) {
+ evt = container_of(node, struct tmigr_event, nextevt);
+
+ if (!test_bit(0, &evt->ignore)) {
+ group->next_expiry = evt->nextevt.expires;
+ return evt;
+ }
+
+ /*
+ * Remove next timers with ignore bits, because group lock
+ * is held anyway
+ */
+ if (!timerqueue_del(&group->events, node))
+ break;
+ }
+
+ return NULL;
+}
+
+/*
+ * Return next event which is already expired of group timerqueue
+ *
+ * Event is also removed from queue.
+ */
+static struct tmigr_event *tmigr_next_expired_groupevt(struct tmigr_group *group,
+ u64 now)
+{
+ struct tmigr_event *evt = tmigr_next_groupevt(group);
+
+ if (!evt || now < evt->nextevt.expires)
+ return NULL;
+
+ /*
+ * Event is already expired. Remove it. If it's not the last event,
+ * then update all group event related information.
+ */
+ if (timerqueue_del(&group->events, &evt->nextevt))
+ tmigr_next_groupevt(group);
+ else
+ group->next_expiry = KTIME_MAX;
+
+ return evt;
+}
+
+static u64 tmigr_next_groupevt_expires(struct tmigr_group *group)
+{
+ struct tmigr_event *evt;
+
+ evt = tmigr_next_groupevt(group);
+
+ if (!evt)
+ return KTIME_MAX;
+ else
+ return evt->nextevt.expires;
+}
+
+static bool tmigr_active_up(struct tmigr_group *group,
+ struct tmigr_group *child,
+ void *ptr)
+{
+ union tmigr_state curstate, newstate;
+ struct tmigr_walk *data = ptr;
+ bool walk_done;
+ u32 childmask;
+
+ childmask = data->childmask;
+ newstate = curstate = data->groupstate;
+
+retry:
+ walk_done = true;
+
+ if (newstate.migrator == TMIGR_NONE) {
+ newstate.migrator = (u8)childmask;
+
+ /* Changes need to be propagated */
+ walk_done = false;
+ }
+
+ newstate.active |= (u8)childmask;
+
+ newstate.seq++;
+
+ if (atomic_cmpxchg(group->migr_state, curstate.state, newstate.state) != curstate.state) {
+ newstate.state = curstate.state = atomic_read(group->migr_state);
+ goto retry;
+ }
+
+ if (group->parent && (walk_done == false)) {
+ data->groupstate.state = atomic_read(group->parent->migr_state);
+ data->childmask = group->childmask;
+ }
+
+ /*
+ * Group is active, event will be ignored - bit is updated without
+ * holding the lock. In case bit is set while another CPU already
+ * handles remote events, nothing happens, because it is clear that
+ * CPU became active just in this moment, or in worst case event is
+ * handled remote. Nothing to worry about.
+ */
+ set_bit(0, &group->groupevt.ignore);
+
+ return walk_done;
+}
+
+static void __tmigr_cpu_activate(struct tmigr_cpu *tmc)
+{
+ struct tmigr_walk data;
+ data.childmask = tmc->childmask;
+ data.groupstate.state = atomic_read(tmc->tmgroup->migr_state);
+
+ set_bit(0, &tmc->cpuevt.ignore);
+
+ walk_groups(&tmigr_active_up, &data, tmc);
+}
+
+void tmigr_cpu_activate(void)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+
+ if (!is_tmigr_enabled() || !tmc->tmgroup || !tmc->online || !tmc->idle)
+ return;
+
+ raw_spin_lock(&tmc->lock);
+ tmc->idle = 0;
+ tmc->wakeup = KTIME_MAX;
+ __tmigr_cpu_activate(tmc);
+ raw_spin_unlock(&tmc->lock);
+}
+
+/*
+ * Returns true, if there is nothing to be propagated to the next level
+ *
+ * @data->nextexp is reset to KTIME_MAX; it is reused for first global
+ * event which needs to be handled by migrator (in toplevel group)
+ *
+ * This is the only place where group event expiry value is set.
+ */
+static bool tmigr_update_events(struct tmigr_group *group,
+ struct tmigr_group *child,
+ struct tmigr_walk *data)
+{
+ struct tmigr_event *evt, *first_childevt;
+ bool walk_done;
+ u64 nextexp;
+
+ if (child) {
+ if (data->childstate.active)
+ return true;
+
+ raw_spin_lock(&child->lock);
+ raw_spin_lock_nested(&group->lock, SINGLE_DEPTH_NESTING);
+
+ first_childevt = tmigr_next_groupevt(child);
+ nextexp = child->next_expiry;
+ evt = &child->groupevt;
+ } else {
+ nextexp = data->nextexp;
+
+ /*
+ * Set @data->nextexp to KTIME_MAX; it is reused for first
+ * global event which needs to be handled by migrator (in
+ * toplevel group)
+ */
+ data->nextexp = KTIME_MAX;
+
+ first_childevt = evt = data->evt;
+ if (test_bit(0, &evt->ignore))
+ return true;
+
+ raw_spin_lock(&group->lock);
+ }
+
+ if (nextexp == KTIME_MAX) {
+ set_bit(0, &evt->ignore);
+ walk_done = true;
+ goto unlock;
+ }
+
+ walk_done = !group->parent;
+
+ /*
+ * Update of event cpu and ignore bit is required only when @child
+ * is set (child is equal or higher than lvl0), but it doesn't
+ * matter if it is written once more to per cpu event; make the
+ * update unconditional.
+ */
+ evt->cpu = first_childevt->cpu;
+ clear_bit(0, &evt->ignore);
+
+ /*
+ * If child event is already queued in group, remove it from queue
+ * when expiry time changed only
+ */
+ if (timerqueue_node_queued(&evt->nextevt)) {
+ if (evt->nextevt.expires == nextexp)
+ goto check_toplvl;
+
+ if (evt->nextevt.expires != nextexp &&
+ !timerqueue_del(&group->events, &evt->nextevt))
+ group->next_expiry = KTIME_MAX;
+ }
+
+ evt->nextevt.expires = nextexp;
+
+ if (timerqueue_add(&group->events, &evt->nextevt))
+ group->next_expiry = nextexp;
+
+check_toplvl:
+ if (walk_done && (data->groupstate.migrator == TMIGR_NONE)) {
+ /*
+ * Toplevel group is idle and it has to be ensured global
+ * timers are handled in time. (This could be optimized by
+ * keeping track of the last global scheduled event and
+ * only arming it on CPU if the new event is earlier. Not
+ * sure if its worth the complexity.)
+ */
+ data->nextexp = tmigr_next_groupevt_expires(group);
+ }
+
+unlock:
+ raw_spin_unlock(&group->lock);
+
+ if (child)
+ raw_spin_unlock(&child->lock);
+
+ return walk_done;
+}
+
+static bool tmigr_new_timer_up(struct tmigr_group *group,
+ struct tmigr_group *child,
+ void *ptr)
+{
+ struct tmigr_walk *data = ptr;
+ bool walk_done;
+
+ walk_done = tmigr_update_events(group, child, data);
+
+ if (!walk_done) {
+ /* Update state information for next iteration */
+ data->childstate.state = atomic_read(group->migr_state);
+ if (group->parent)
+ data->groupstate.state = atomic_read(group->parent->migr_state);
+ }
+
+ return walk_done;
+}
+
+/*
+ * Returns expiry of next timer that needs to be handled. KTIME_MAX is
+ * returned, when an active CPU will handle all timer migration hierarchy
+ * timers.
+ */
+static u64 tmigr_new_timer(struct tmigr_cpu *tmc, u64 nextexp)
+{
+ struct tmigr_walk data = { .evt = &tmc->cpuevt,
+ .nextexp = nextexp };
+
+ lockdep_assert_held(&tmc->lock);
+
+ if (tmc->remote)
+ return KTIME_MAX;
+
+ clear_bit(0, &tmc->cpuevt.ignore);
+
+ data.groupstate.state = atomic_read(tmc->tmgroup->migr_state);
+
+ walk_groups(&tmigr_new_timer_up, &data, tmc);
+
+ /* If there is a new first global event, make sure it is handled */
+ return data.nextexp;
+}
+
+static bool tmigr_inactive_up(struct tmigr_group *group,
+ struct tmigr_group *child,
+ void *ptr)
+{
+ union tmigr_state curstate, newstate;
+ struct tmigr_walk *data = ptr;
+ bool walk_done;
+ u32 childmask;
+
+ childmask = data->childmask;
+ newstate = curstate = data->groupstate;
+
+retry:
+ walk_done = true;
+
+ /* Reset active bit when child is no longer active */
+ if (!data->childstate.active)
+ newstate.active &= ~(u8)childmask;
+
+ if (newstate.migrator == (u8)childmask) {
+ /*
+ * Find a new migrator for the group, because child group
+ * is idle!
+ */
+ if (!data->childstate.active) {
+ unsigned long new_migr_bit;
+
+ new_migr_bit = find_first_bit((const long *)&newstate.active, BIT_CNT);
+
+ /* Changes need to be propagated */
+ walk_done = false;
+ data->childmask = group->childmask;
+
+ if (new_migr_bit != BIT_CNT)
+ newstate.migrator = BIT(new_migr_bit);
+ else
+ newstate.migrator = TMIGR_NONE;
+ }
+ }
+
+ newstate.seq++;
+
+ DBG_BUG_ON((newstate.migrator != TMIGR_NONE) && !(newstate.active));
+
+ if (atomic_cmpxchg(group->migr_state, curstate.state, newstate.state) != curstate.state) {
+ /*
+ * Something changed in child/parent group in the meantime,
+ * reread the state of child and parent; Update of
+ * data->childstate is required for event handling;
+ */
+ if (child)
+ data->childstate.state = atomic_read(child->migr_state);
+ newstate.state = curstate.state = atomic_read(group->migr_state);
+
+ goto retry;
+ }
+
+ data->groupstate = newstate;
+
+ /* Event Handling */
+ tmigr_update_events(group, child, data);
+
+ if (group->parent && (walk_done == false)) {
+ data->childstate = newstate;
+ data->groupstate.state = atomic_read(group->parent->migr_state);
+ }
+
+ /*
+ * data->nextexp was set by tmigr_update_events() and contains the
+ * expiry of first global event which needs to be handled
+ */
+ if (data->nextexp != KTIME_MAX) {
+ DBG_BUG_ON(group->parent);
+ /*
+ * Toplevel path - If this cpu is about going offline wake
+ * up some random other cpu so it will take over the
+ * migrator duty and program its timer properly. Ideally
+ * wake the cpu with the closest expiry time, but that's
+ * overkill to figure out.
+ */
+ if (!(this_cpu_ptr(&tmigr_cpu)->online)) {
+ unsigned int cpu = smp_processor_id();
+
+ cpu = cpumask_any_but(cpu_online_mask, cpu);
+ smp_send_reschedule(cpu);
+ }
+ }
+
+ return walk_done;
+}
+
+static u64 __tmigr_cpu_deactivate(struct tmigr_cpu *tmc, u64 nextexp)
+{
+ struct tmigr_walk data = { .childmask = tmc->childmask,
+ .evt = &tmc->cpuevt,
+ .nextexp = nextexp,
+ .childstate.state = 0 };
+
+ data.groupstate.state = atomic_read(tmc->tmgroup->migr_state);
+
+ /*
+ * If nextexp is KTIME_MAX, CPU event will be ignored because,
+ * local timer expires before global timer, no global timer is set
+ * or CPU goes offline.
+ */
+ if (nextexp != KTIME_MAX)
+ clear_bit(0, &tmc->cpuevt.ignore);
+
+ walk_groups(&tmigr_inactive_up, &data, tmc);
+ return data.nextexp;
+}
+
+/**
+ * tmigr_cpu_deactivate - Put current CPU into inactive state
+ * @nextexp: The next timer event expiry set in the current CPU
+ *
+ * Must be called with interrupts disabled.
+ *
+ * Return: next event of the current CPU or next event from the hierarchy
+ * if this CPU is the top level migrator or hierarchy is completely idle.
+ */
+u64 tmigr_cpu_deactivate(u64 nextexp)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+ u64 ret = KTIME_MAX;
+
+ if (!is_tmigr_enabled() || !tmc->tmgroup || !tmc->online)
+ return ret;
+
+ raw_spin_lock(&tmc->lock);
+
+ /*
+ * CPU is already deactivated in timer migration
+ * hierarchy. tick_nohz_get_sleep_length() calls
+ * tick_nohz_next_event() and thereby timer idle path is
+ * executed once more. tmc->wakeup holds the first timer, when
+ * timer migration hierarchy is completely idle and remote
+ * expiry was done. If there is no new next expiry value
+ * handed in which should be inserted into the timer migration
+ * hierarchy, wakeup value is returned.
+ */
+ if (tmc->idle) {
+ u64 ret = tmc->wakeup;
+
+ tmc->wakeup = KTIME_MAX;
+
+ if (nextexp != KTIME_MAX) {
+ if (nextexp != tmc->cpuevt.nextevt.expires ||
+ test_bit(0, &tmc->cpuevt.ignore))
+ ret = tmigr_new_timer(tmc, nextexp);
+ }
+
+ goto unlock;
+ }
+
+ /*
+ * When tmigr_remote is active, set cpu inactive path and queuing of
+ * nextexp is done by handle remote path.
+ */
+ ret = __tmigr_cpu_deactivate(tmc, nextexp);
+
+ tmc->idle = 1;
+
+unlock:
+ raw_spin_unlock(&tmc->lock);
+ return ret;
+}
+
+static u64 tmigr_handle_remote_cpu(unsigned int cpu, u64 now,
+ unsigned long jif)
+{
+ struct timer_events tevt;
+ struct tmigr_walk data;
+ struct tmigr_cpu *tmc;
+ u64 next = KTIME_MAX;
+ unsigned long flags;
+
+ tmc = per_cpu_ptr(&tmigr_cpu, cpu);
+
+ raw_spin_lock_irqsave(&tmc->lock, flags);
+ /*
+ * Remote CPU is offline or no longer idle or other cpu handles cpu
+ * timers already or next event was already expired - return!
+ */
+ if (!tmc->online || tmc->remote || test_bit(0, &tmc->cpuevt.ignore) ||
+ now < tmc->cpuevt.nextevt.expires) {
+ raw_spin_unlock_irqrestore(&tmc->lock, flags);
+ return next;
+ }
+
+ tmc->remote = 1;
+
+ /* Drop the lock to allow the remote CPU to exit idle */
+ raw_spin_unlock_irqrestore(&tmc->lock, flags);
+
+ if (cpu != smp_processor_id())
+ timer_expire_remote(cpu);
+
+ /* next event of cpu */
+ get_next_timer_interrupt_remote(jif, now, &tevt, cpu);
+
+ raw_spin_lock_irqsave(&tmc->lock, flags);
+ /*
+ * Nothing more to do when CPU came out of idle in the meantime - needs
+ * to be checked when holding the base lock to prevent race.
+ */
+ if (!tmc->idle)
+ goto unlock;
+
+ data.evt = &tmc->cpuevt;
+ data.nextexp = tevt.global;
+ data.groupstate.state = atomic_read(tmc->tmgroup->migr_state);
+ clear_bit(0, &tmc->cpuevt.ignore);
+
+ walk_groups(&tmigr_new_timer_up, &data, tmc);
+
+ next = data.nextexp;
+
+unlock:
+ tmc->remote = 0;
+ raw_spin_unlock_irqrestore(&tmc->lock, flags);
+
+ return next;
+}
+
+static bool tmigr_handle_remote_up(struct tmigr_group *group,
+ struct tmigr_group *child,
+ void *ptr)
+{
+ struct tmigr_remote_data *data = ptr;
+ u64 now, next = KTIME_MAX;
+ unsigned long flags, jif;
+ struct tmigr_event *evt;
+ u32 childmask;
+
+ jif = data->basej;
+ now = data->now;
+
+ childmask = data->childmask;
+
+again:
+ /*
+ * Handle the group only if @childmask is the migrator or if the
+ * group has no migrator. Otherwise the group is active and is
+ * handled by its own migrator.
+ */
+ if (!tmigr_check_migrator(group, childmask))
+ return true;
+
+ raw_spin_lock_irqsave(&group->lock, flags);
+
+ evt = tmigr_next_expired_groupevt(group, now);
+
+ if (evt) {
+ unsigned int remote_cpu;
+
+ remote_cpu = READ_ONCE(evt->cpu);
+
+ raw_spin_unlock_irqrestore(&group->lock, flags);
+
+ next = tmigr_handle_remote_cpu(remote_cpu, now, jif);
+
+ /* check if there is another event, that needs to be handled */
+ goto again;
+ } else {
+ raw_spin_unlock_irqrestore(&group->lock, flags);
+ }
+
+ /* Update of childmask for next level */
+ data->childmask = group->childmask;
+ data->wakeup = next;
+
+ return false;
+}
+
+/**
+ * tmigr_handle_remote - Handle migratable timers on remote idle CPUs
+ *
+ * Called from the timer soft interrupt with interrupts enabled.
+ */
+void tmigr_handle_remote(void)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+ struct tmigr_remote_data data;
+ unsigned long flags;
+
+ if (!is_tmigr_enabled() || !tmc->tmgroup || !tmc->online)
+ return;
+
+ /*
+ * NOTE: This is a doubled check because migrator test will be done
+ * in tmigr_handle_remote_up() anyway. Keep this check to fasten
+ * the return when nothing has to be done.
+ */
+ if (!tmigr_check_migrator(tmc->tmgroup, tmc->childmask))
+ return;
+
+ data.now = get_jiffies_update(&data.basej);
+ data.childmask = tmc->childmask;
+ data.wakeup = KTIME_MAX;
+
+ __walk_groups(&tmigr_handle_remote_up, &data, tmc);
+
+ raw_spin_lock_irqsave(&tmc->lock, flags);
+ if (tmc->idle)
+ tmc->wakeup = data.wakeup;
+
+ raw_spin_unlock_irqrestore(&tmc->lock, flags);
+
+ return;
+}
+
+static bool tmigr_requires_handle_remote_up(struct tmigr_group *group,
+ struct tmigr_group *child,
+ void *ptr)
+{
+ struct tmigr_remote_data *data = ptr;
+ u32 childmask;
+
+ childmask = data->childmask;
+
+ /*
+ * Handle the group only if child is the migrator or if the group
+ * has no migrator. Otherwise the group is active and is handled by
+ * its own migrator.
+ */
+ if (!tmigr_check_migrator(group, childmask))
+ return true;
+
+ /*
+ * Racy lockless check for next_expiry
+ */
+ if (data->now >= group->next_expiry) {
+ data->check = 1;
+ return true;
+ }
+
+ /* Update of childmask for next level */
+ data->childmask = group->childmask;
+ return false;
+}
+
+int tmigr_requires_handle_remote(void)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+ struct tmigr_remote_data data;
+
+ if (!is_tmigr_enabled() || !tmc->tmgroup || !tmc->online)
+ return 0;
+
+ if (!tmigr_check_migrator(tmc->tmgroup, tmc->childmask))
+ return 0;
+
+ data.now = get_jiffies_update(&data.basej);
+ data.childmask = tmc->childmask;
+
+ __walk_groups(&tmigr_requires_handle_remote_up, &data, tmc);
+
+ return data.check;
+}
+
+static void tmigr_init_group(struct tmigr_group *group, unsigned int lvl,
+ unsigned int node, atomic_t *migr_state)
+{
+ union tmigr_state s;
+
+ raw_spin_lock_init(&group->lock);
+
+ group->level = lvl;
+ group->numa_node = lvl < tmigr_crossnode_level ? node : NUMA_NO_NODE;
+
+ group->num_childs = 0;
+
+ /*
+ * num_cores is required for level=0 groups only during setup and
+ * when siblings exists but it doesn't matter if this value is set
+ * in other groups as well
+ */
+ group->num_cores = 1;
+
+ s.migrator = TMIGR_NONE;
+ s.active = 0;
+ s.seq = 0;
+ atomic_set(migr_state, s.state);
+
+ group->migr_state = migr_state;
+
+ timerqueue_init_head(&group->events);
+ timerqueue_init(&group->groupevt.nextevt);
+ group->groupevt.nextevt.expires = KTIME_MAX;
+ group->next_expiry = KTIME_MAX;
+ set_bit(0, &group->groupevt.ignore);
+}
+
+static bool sibling_in_group(int newcpu, struct tmigr_group *group)
+{
+ int i, cpu;
+
+ /* Find a sibling of newcpu in group members */
+ for (i = 0; i < group->num_childs; i++) {
+ cpu = group->cpus[i];
+
+ if (cpumask_test_cpu(newcpu, topology_sibling_cpumask(cpu)))
+ return true;
+ }
+ return false;
+}
+
+static struct tmigr_group *tmigr_get_group(unsigned int cpu, unsigned int node,
+ unsigned int lvl)
+{
+ struct tmigr_group *tmp, *group = NULL;
+ bool first_loop = true;
+ atomic_t *migr_state;
+
+reloop:
+ /* Try to attach to an exisiting group first */
+ list_for_each_entry(tmp, &tmigr_level_list[lvl], list) {
+ /*
+ * If @lvl is below the cross numa node level, check whether
+ * this group belongs to the same numa node.
+ */
+ if (lvl < tmigr_crossnode_level && tmp->numa_node != node)
+ continue;
+
+ /* Capacity left? */
+ if (tmp->num_childs >= TMIGR_CHILDS_PER_GROUP)
+ continue;
+
+ /*
+ * If this is the lowest level of the hierarchy, make sure
+ * that thread siblings share a group. It is only executed
+ * when siblings exist. ALL groups of lowest level needs to
+ * be checked for thread sibling, before thread cpu is
+ * added to a random group with capacity. When all groups
+ * are checked and no thread sibling was found, reloop of
+ * level zero groups is required to get a group with
+ * capacity.
+ */
+ if (!lvl && (tmigr_cores_per_group != TMIGR_CHILDS_PER_GROUP)) {
+ if (first_loop == true && !sibling_in_group(cpu, tmp)) {
+ continue;
+ } else if (first_loop == false) {
+ if (tmp->num_cores >= tmigr_cores_per_group)
+ continue;
+ else
+ tmp->num_cores++;
+ }
+ }
+
+ group = tmp;
+ break;
+ }
+
+ if (group) {
+ return group;
+ } else if (first_loop == true) {
+ first_loop = false;
+ goto reloop;
+ }
+
+ /* Allocate and set up a new group with corresponding migr_state */
+ group = kzalloc_node(sizeof(*group), GFP_KERNEL, node);
+ if (!group)
+ return ERR_PTR(-ENOMEM);
+
+ migr_state = kzalloc_node(sizeof(atomic_t), GFP_KERNEL, node);
+ if (!migr_state) {
+ kfree(group);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ tmigr_init_group(group, lvl, node, migr_state);
+ /* Setup successful. Add it to the hierarchy */
+ list_add(&group->list, &tmigr_level_list[lvl]);
+ return group;
+}
+
+static void tmigr_connect_child_parent(struct tmigr_group *child,
+ struct tmigr_group *parent)
+{
+ union tmigr_state childstate;
+ unsigned long flags;
+
+ raw_spin_lock_irqsave(&child->lock, flags);
+ raw_spin_lock_nested(&parent->lock, SINGLE_DEPTH_NESTING);
+
+ child->parent = parent;
+ child->childmask = BIT(parent->num_childs++);
+
+ raw_spin_unlock(&parent->lock);
+ raw_spin_unlock_irqrestore(&child->lock, flags);
+
+ /*
+ * To prevent inconsistent states, active childs needs to be active
+ * in new parent as well. Inactive childs are already marked
+ * inactive in parent group.
+ */
+ childstate.state = atomic_read(child->migr_state);
+ if (childstate.migrator != TMIGR_NONE) {
+ struct tmigr_walk data;
+
+ data.childmask = child->childmask;
+ data.groupstate.state = atomic_read(parent->migr_state);
+
+ /*
+ * There is only one new level per time. When connecting
+ * child and parent and set child active when parent is
+ * inactive, parent needs to be the upperst
+ * level. Otherwise there went something wrong!
+ */
+ WARN_ON(!tmigr_active_up(parent, child, &data) && parent->parent);
+ }
+}
+
+static int tmigr_setup_groups(unsigned int cpu, unsigned int node)
+{
+ struct tmigr_group *group, *child, **stack;
+ int top = 0, err = 0, i = 0;
+ struct list_head *lvllist;
+ size_t sz;
+
+ sz = sizeof(struct tmigr_group *) * tmigr_hierarchy_levels;
+ stack = kzalloc(sz, GFP_KERNEL);
+ if (!stack)
+ return -ENOMEM;
+
+ do {
+ group = tmigr_get_group(cpu, node, i);
+ if (IS_ERR(group)) {
+ err = IS_ERR(group);
+ break;
+ }
+
+ top = i;
+ stack[i++] = group;
+
+ /*
+ * When booting only less CPUs of a system than CPUs are
+ * available, not all calculated hierarchy levels are required.
+ *
+ * The loop is aborted as soon as the highest level, which might
+ * be different from tmigr_hierarchy_levels, contains only a
+ * single group.
+ */
+ if (group->parent || i == tmigr_hierarchy_levels ||
+ (list_empty(&tmigr_level_list[i]) &&
+ list_is_singular(&tmigr_level_list[i - 1])))
+ break;
+
+ } while (i < tmigr_hierarchy_levels);
+
+ do {
+ group = stack[--i];
+
+ if (err < 0) {
+ list_del(&group->list);
+ kfree(group);
+ continue;
+ }
+
+ DBG_BUG_ON(i != group->level);
+
+ /*
+ * Update tmc -> group / child -> group connection
+ */
+ if (i == 0) {
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+ unsigned long flags;
+
+ raw_spin_lock_irqsave(&group->lock, flags);
+
+ tmc->tmgroup = group;
+ tmc->childmask = BIT(group->num_childs);
+
+ group->cpus[group->num_childs++] = cpu;
+
+ raw_spin_unlock_irqrestore(&group->lock, flags);
+
+ /* There are no childs that needs to be connected */
+ continue;
+ } else {
+ child = stack[i - 1];
+ tmigr_connect_child_parent(child, group);
+ }
+
+ /* check if upperst level was newly created */
+ if (top != i)
+ continue;
+
+ DBG_BUG_ON(top == 0);
+
+ lvllist = &tmigr_level_list[top];
+ if (group->num_childs == 1 && list_is_singular(lvllist)) {
+ lvllist = &tmigr_level_list[top - 1];
+ list_for_each_entry(child, lvllist, list) {
+ if (child->parent)
+ continue;
+
+ tmigr_connect_child_parent(child, group);
+ }
+ }
+ } while (i > 0);
+
+ kfree(stack);
+
+ return err;
+}
+
+static int tmigr_add_cpu(unsigned int cpu)
+{
+ unsigned int node = cpu_to_node(cpu);
+ int ret;
+ mutex_lock(&tmigr_mutex);
+ ret = tmigr_setup_groups(cpu, node);
+ mutex_unlock(&tmigr_mutex);
+
+ return ret;
+}
+
+static int tmigr_cpu_online(unsigned int cpu)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+ unsigned long flags;
+ unsigned int ret;
+
+ /* First online attempt? Initialize CPU data */
+ if (!tmc->tmgroup) {
+ raw_spin_lock_init(&tmc->lock);
+
+ ret = tmigr_add_cpu(cpu);
+ if (ret < 0)
+ return ret;
+
+ if (tmc->childmask == 0)
+ return -EINVAL;
+
+ timerqueue_init(&tmc->cpuevt.nextevt);
+ tmc->cpuevt.nextevt.expires = KTIME_MAX;
+ set_bit(0, &tmc->cpuevt.ignore);
+ tmc->cpuevt.cpu = cpu;
+
+ tmc->remote = 0;
+ tmc->idle = 0;
+ tmc->wakeup = KTIME_MAX;
+ }
+ raw_spin_lock_irqsave(&tmc->lock, flags);
+ __tmigr_cpu_activate(tmc);
+ tmc->online = 1;
+ raw_spin_unlock_irqrestore(&tmc->lock, flags);
+ return 0;
+}
+
+static int tmigr_cpu_offline(unsigned int cpu)
+{
+ struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
+
+ raw_spin_lock_irq(&tmc->lock);
+ tmc->online = 0;
+ __tmigr_cpu_deactivate(tmc, KTIME_MAX);
+ raw_spin_unlock_irq(&tmc->lock);
+
+ return 0;
+}
+
+static int __init tmigr_init(void)
+{
+ unsigned int cpulvl, nodelvl, cpus_per_node, i, ns;
+ unsigned int nnodes = num_possible_nodes();
+ unsigned int ncpus = num_possible_cpus();
+ int ret = -ENOMEM;
+ size_t sz;
+
+ /* Nothing to do if running on UP */
+ if (ncpus == 1)
+ return 0;
+
+ /*
+ * Unfortunately there is no reliable way to determine the number of SMT
+ * siblings in a generic way. tmigr_init() is called after SMP bringup,
+ * so for the normal boot case it can be assumed that all siblings have
+ * been brought up and the number of siblings of the current cpu can be
+ * used. If someone booted with 'maxcpus=N/2' on the kernel command line
+ * and (at least x86) bring up the siblings later then the siblings will
+ * end up in different groups. Bad luck.
+ */
+ ns = cpumask_weight(topology_sibling_cpumask(raw_smp_processor_id()));
+ tmigr_cores_per_group = TMIGR_CHILDS_PER_GROUP;
+ if (ns >= 2 && ns < TMIGR_CHILDS_PER_GROUP)
+ tmigr_cores_per_group /= ns;
+
+ /*
+ * Calculate the required hierarchy levels. Unfortunately there is no
+ * reliable information available, unless all possible CPUs have been
+ * brought up and all numa nodes are populated.
+ *
+ * Estimate the number of levels with the number of possible nodes and
+ * the number of possible cpus. Assume CPUs are spread evenly accross
+ * nodes. We cannot rely on cpumask_of_node() because there only already
+ * online CPUs are considered.
+ */
+ cpus_per_node = DIV_ROUND_UP(ncpus, nnodes);
+
+ /* Calc the hierarchy levels required to hold the CPUs of a node */
+ cpulvl = DIV_ROUND_UP(order_base_2(cpus_per_node),
+ ilog2(TMIGR_CHILDS_PER_GROUP));
+
+ /* Calculate the extra levels to connect all nodes */
+ nodelvl = DIV_ROUND_UP(order_base_2(nnodes),
+ ilog2(TMIGR_CHILDS_PER_GROUP));
+
+ tmigr_hierarchy_levels = cpulvl + nodelvl;
+
+ /*
+ * If a numa node spawns more than one CPU level group then the next
+ * level(s) of the hierarchy contains groups which handle all CPU groups
+ * of the same numa node. The level above goes accross numa nodes. Store
+ * this information for the setup code to decide when node matching is
+ * not longer required.
+ */
+ tmigr_crossnode_level = cpulvl;
+
+ sz = sizeof(struct list_head) * tmigr_hierarchy_levels;
+ tmigr_level_list = kzalloc(sz, GFP_KERNEL);
+ if (!tmigr_level_list)
+ goto err;
+
+ for (i = 0; i < tmigr_hierarchy_levels; i++)
+ INIT_LIST_HEAD(&tmigr_level_list[i]);
+
+ pr_info("Timer migration: %d hierarchy levels; %d childs per group;"
+ " %d cores_per_group; %d crossnode level\n",
+ tmigr_hierarchy_levels, TMIGR_CHILDS_PER_GROUP,
+ tmigr_cores_per_group, tmigr_crossnode_level);
+
+ ret = cpuhp_setup_state(CPUHP_AP_TMIGR_ONLINE, "tmigr:online",
+ tmigr_cpu_online, tmigr_cpu_offline);
+ if (ret)
+ goto err;
+
+ static_branch_enable(&tmigr_enabled);
+
+ return 0;
+
+err:
+ pr_err("Timer migration setup failed\n");
+ return ret;
+}
+late_initcall(tmigr_init);
diff --git a/kernel/time/timer_migration.h b/kernel/time/timer_migration.h
new file mode 100644
index 000000000000..42d7550ea231
--- /dev/null
+++ b/kernel/time/timer_migration.h
@@ -0,0 +1,121 @@
+#ifndef _KERNEL_TIME_MIGRATION_H
+#define _KERNEL_TIME_MIGRATION_H
+
+/* Per group capacity. Must be a power of 2! */
+#define TMIGR_CHILDS_PER_GROUP 8
+
+/**
+ * struct tmigr_event - a timer event associated to a CPU
+ * @nextevt: The node to enqueue an event in the parent group queue
+ * @cpu: The CPU to which this event belongs
+ * @ignore: Hint whether the event could be ignored; it is set when
+ * CPU or group is active;
+ */
+struct tmigr_event {
+ struct timerqueue_node nextevt;
+ unsigned int cpu;
+ unsigned long ignore;
+};
+
+/**
+ * struct tmigr_group - timer migration hierarchy group
+ * @lock: Lock protecting the event information
+ * @cpus: Array with CPUs which are member of group; required for
+ * sibling CPUs; used only when level == 0
+ * @parent: Pointer to parent group
+ * @list: List head that is added to per level tmigr_level_list
+ * @level: Hierarchy level of group
+ * @numa_node: Is set to numa node when level < tmigr_crossnode_level;
+ * otherwise it is set to NUMA_NO_NODE; Required for setup
+ * only
+ * @num_childs: Counter of group childs; Required for setup only
+ * @num_cores: Counter of cores per group; Required for setup only when
+ * level == 0 and siblings exist
+ * @migr_state: State of group (see struct tmigr_state)
+ * @childmask: childmask of group in parent group; is set during setup
+ * never changed; could be read lockless
+ * @events: Timer queue for child events queued in the group
+ * @groupevt: Next event of group; it is only reliable when group is
+ * !active (ignore bit is set when group is active)
+ * @next_expiry: Base monotonic expiry time of next event of group;
+ * Used for racy lockless check whether remote expiry is
+ * required; it is always reliable
+ */
+struct tmigr_group {
+ raw_spinlock_t lock;
+ unsigned int cpus[TMIGR_CHILDS_PER_GROUP];
+ struct tmigr_group *parent;
+ struct list_head list;
+ unsigned int level;
+ unsigned int numa_node;
+ unsigned int num_childs;
+ unsigned int num_cores;
+ atomic_t *migr_state;
+ u32 childmask;
+ struct timerqueue_head events;
+ struct tmigr_event groupevt;
+ u64 next_expiry;
+};
+
+/**
+ * struct tmigr_cpu - timer migration per CPU group
+ * @lock: Lock protecting tmigr_cpu group information
+ * @online: Indicates wheter CPU is online
+ * @idle: Indicates wheter CPU is idle in timer migration hierarchy
+ * @remote: Is set when timers of CPU are expired remote
+ * @tmgroup: Pointer to parent group
+ * @childmask: childmask of tmigr_cpu in parent group
+ * @cpuevt: CPU event which could be queued into parent group
+ * @wakeup: Stores the first timer when the timer migration hierarchy is
+ * completely idle and remote expiry was done; is returned to
+ * timer code when tmigr_cpu_deactive() is called and group is
+ * idle; afterwards a reset to KTIME_MAX is required;
+ */
+struct tmigr_cpu {
+ raw_spinlock_t lock;
+ int online;
+ int idle;
+ int remote;
+ struct tmigr_group *tmgroup;
+ u32 childmask;
+ struct tmigr_event cpuevt;
+ u64 wakeup;
+};
+
+/**
+ * union tmigr_state - state of tmigr_group
+ * @complete: Combined version of the state
+ * @split: Splitted version of the state
+ */
+union tmigr_state {
+ u32 state;
+ /**
+ * struct - splitted state of tmigr_group
+ * @active: Contains each childmask bit of active childs
+ * @migrator: Contains childmask of child which is migrator
+ * @seq: Seqence number to prevent race when update in child
+ * group are propagated in wrong order (especially when
+ * migrator changes are involved)
+ */
+ struct {
+ u8 active;
+ u8 migrator;
+ u16 seq;
+ } __packed;
+};
+
+#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
+extern void tmigr_handle_remote(void);
+extern int tmigr_requires_handle_remote(void);
+extern void tmigr_cpu_activate(void);
+extern u64 tmigr_cpu_deactivate(u64 nextevt);
+extern void timer_expire_remote(unsigned int cpu);
+#else
+static inline void tmigr_handle_remote(void) { }
+extern inline int tmigr_requires_handle_remote(void) { return 0; }
+static inline void tmigr_cpu_activate(void) { }
+static inline u64 tmigr_cpu_deactivate(u64 nextevt) { return KTIME_MAX; }
+extern inline void timer_expire_remote(unsigned int cpu) { }
+#endif
+
+#endif
--
2.30.2
get_next_timer_interrupt() does more than simply getting the next timer
interrupt. The timer bases are forwarded and also marked as idle when
possible and the next timer interrupt information is required for this.
To get not confused, rename function to a more descriptive name. No
functional change.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/tick-internal.h | 4 ++--
kernel/time/tick-sched.c | 16 ++++++++--------
kernel/time/timer.c | 10 +++++++---
3 files changed, 17 insertions(+), 13 deletions(-)
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index fcb2d45c2934..6f5f164506d5 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -168,8 +168,8 @@ static inline void timers_update_nohz(void) { }
DECLARE_PER_CPU(struct hrtimer_cpu_base, hrtimer_bases);
-extern void get_next_timer_interrupt(unsigned long basej, u64 basem,
- struct timer_events *tevt);
+extern void forward_and_idle_timer_bases(unsigned long basej, u64 basem,
+ struct timer_events *tevt);
void timer_clear_idle(void);
#define CLOCK_SET_WALL \
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 78f172d1f3d2..7f7bfe8b498d 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -802,11 +802,11 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
* Keep the periodic tick, when RCU, architecture or irq_work
* requests it.
* Aside of that check whether the local timer softirq is
- * pending. If so its a bad idea to call get_next_timer_interrupt()
- * because there is an already expired timer, so it will request
- * immediate expiry, which rearms the hardware timer with a
- * minimal delta which brings us back to this place
- * immediately. Lather, rinse and repeat...
+ * pending. If so its a bad idea to call
+ * forward_and_idle_timer_bases() because there is an already
+ * expired timer, so it will request immeditate expiry, which
+ * rearms the hardware timer with a minimal delta which brings us
+ * back to this place immediately. Lather, rinse and repeat...
*/
if (rcu_needs_cpu() || arch_needs_cpu() ||
irq_work_needs_cpu() || local_timer_softirq_pending()) {
@@ -823,7 +823,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
* disabled this also looks at the next expiring
* hrtimer.
*/
- get_next_timer_interrupt(basejiff, basemono, &tevt);
+ forward_and_idle_timer_bases(basejiff, basemono, &tevt);
tevt.local = min_t(u64, tevt.local, tevt.global);
ts->next_timer = tevt.local;
}
@@ -838,7 +838,7 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
if (delta <= (u64)TICK_NSEC) {
/*
* Tell the timer code that the base is not idle, i.e. undo
- * the effect of get_next_timer_interrupt():
+ * the effect of forward_and_idle_timer_bases():
*/
timer_clear_idle();
/*
@@ -1141,7 +1141,7 @@ void tick_nohz_idle_retain_tick(void)
{
tick_nohz_retain_tick(this_cpu_ptr(&tick_cpu_sched));
/*
- * Undo the effect of get_next_timer_interrupt() called from
+ * Undo the effect of forward_and_idle_timer_bases() called from
* tick_nohz_next_event().
*/
timer_clear_idle();
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 069478e65230..1990eb1b2cc9 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1705,7 +1705,7 @@ static unsigned long next_timer_interrupt(struct timer_base *base)
}
/**
- * get_next_timer_interrupt
+ * forward_and_idle_timer_bases
* @basej: base time jiffies
* @basem: base time clock monotonic
* @tevt: Pointer to the storage for the expiry values
@@ -1713,9 +1713,13 @@ static unsigned long next_timer_interrupt(struct timer_base *base)
* Stores the next pending local and global timer expiry values in the
* struct pointed to by @tevt. If a queue is empty the corresponding field
* is set to KTIME_MAX.
+ *
+ * If required, base->clk is forwarded and base is also marked as
+ * idle. Idle handling of timer bases is allowed only to be done by
+ * CPU itself.
*/
-void get_next_timer_interrupt(unsigned long basej, u64 basem,
- struct timer_events *tevt)
+void forward_and_idle_timer_bases(unsigned long basej, u64 basem,
+ struct timer_events *tevt)
{
unsigned long nextevt, nextevt_local, nextevt_global;
struct timer_base *base_local, *base_global;
--
2.30.2
Both call sides of __next_timer_interrupt() store the return value directly
in base->next_expiry. Move the store into __next_timer_interrupt().
No functional change.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
Reviewed-by: Thomas Gleixner <[email protected]>
---
kernel/time/timer.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 717fcb9fb14a..7695c733dfa5 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1571,8 +1571,10 @@ static int next_pending_bucket(struct timer_base *base, unsigned offset,
/*
* Search the first expiring timer in the various clock levels. Caller must
* hold base->lock.
+ *
+ * Store next expiry time in base->next_expiry.
*/
-static unsigned long __next_timer_interrupt(struct timer_base *base)
+static void __next_timer_interrupt(struct timer_base *base)
{
unsigned long clk, next, adj;
unsigned lvl, offset = 0;
@@ -1638,10 +1640,11 @@ static unsigned long __next_timer_interrupt(struct timer_base *base)
clk += adj;
}
+ base->next_expiry = next;
base->next_expiry_recalc = false;
base->timers_pending = !(next == base->clk + NEXT_TIMER_MAX_DELTA);
- return next;
+ return;
}
#ifdef CONFIG_NO_HZ_COMMON
@@ -1701,7 +1704,7 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
raw_spin_lock(&base->lock);
if (base->next_expiry_recalc)
- base->next_expiry = __next_timer_interrupt(base);
+ __next_timer_interrupt(base);
nextevt = base->next_expiry;
/*
@@ -1784,7 +1787,7 @@ static inline void __run_timers(struct timer_base *base)
WARN_ON_ONCE(!levels && !base->next_expiry_recalc
&& base->timers_pending);
base->clk++;
- base->next_expiry = __next_timer_interrupt(base);
+ __next_timer_interrupt(base);
while (levels--)
expire_timers(base, heads + levels);
--
2.30.2
Due to the conversion of the NOHZ timer placement to a pull at expiry
time model, the per CPU timer bases with non pinned timers are no
longer handled only by the local CPU. In case a remote CPU already
expires the non pinned timers base of the local cpu, nothing more
needs to be done by the local CPU. A check at the begin of the expire
timers routine is required, because timer base lock is dropped before
executing the timer callback function.
This is a preparatory work, but has no functional impact right now.
Signed-off-by: Anna-Maria Behnsen <[email protected]>
---
kernel/time/timer.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index bce2f87d5e70..0790ec8efe82 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1913,6 +1913,9 @@ static inline void __run_timers(struct timer_base *base)
lockdep_assert_held(&base->lock);
+ if (!!base->running_timer)
+ return;
+
while (time_after_eq(jiffies, base->clk) &&
time_after_eq(jiffies, base->next_expiry)) {
levels = collect_expired_timers(base, heads);
--
2.30.2
Hi Anna-Maria,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on tip/timers/core]
[also build test WARNING on powerpc/next tip/x86/core linus/master v6.1-rc2]
[cannot apply to tip/timers/nohz next-20221025]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Anna-Maria-Behnsen/timer-Move-from-a-push-remote-at-enqueue-to-a-pull-at-expiry-model/20221025-220106
patch link: https://lore.kernel.org/r/20221025135850.51044-15-anna-maria%40linutronix.de
patch subject: [PATCH v3 14/17] timer: Implement the hierarchical pull model
config: loongarch-allyesconfig (attached as .config)
compiler: loongarch64-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# https://github.com/intel-lab-lkp/linux/commit/b57766ae36b5e3dd11225f0259f9fd7d39a79e94
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Anna-Maria-Behnsen/timer-Move-from-a-push-remote-at-enqueue-to-a-pull-at-expiry-model/20221025-220106
git checkout b57766ae36b5e3dd11225f0259f9fd7d39a79e94
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=loongarch SHELL=/bin/bash kernel/time/
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
All warnings (new ones prefixed by >>):
kernel/time/timer_migration.c: In function 'tmigr_cpu_deactivate':
>> kernel/time/timer_migration.c:633:21: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
633 | u64 ret = tmc->wakeup;
| ^~~
vim +/ret +633 kernel/time/timer_migration.c
602
603 /**
604 * tmigr_cpu_deactivate - Put current CPU into inactive state
605 * @nextexp: The next timer event expiry set in the current CPU
606 *
607 * Must be called with interrupts disabled.
608 *
609 * Return: next event of the current CPU or next event from the hierarchy
610 * if this CPU is the top level migrator or hierarchy is completely idle.
611 */
612 u64 tmigr_cpu_deactivate(u64 nextexp)
613 {
614 struct tmigr_cpu *tmc = this_cpu_ptr(&tmigr_cpu);
615 u64 ret = KTIME_MAX;
616
617 if (!is_tmigr_enabled() || !tmc->tmgroup || !tmc->online)
618 return ret;
619
620 raw_spin_lock(&tmc->lock);
621
622 /*
623 * CPU is already deactivated in timer migration
624 * hierarchy. tick_nohz_get_sleep_length() calls
625 * tick_nohz_next_event() and thereby timer idle path is
626 * executed once more. tmc->wakeup holds the first timer, when
627 * timer migration hierarchy is completely idle and remote
628 * expiry was done. If there is no new next expiry value
629 * handed in which should be inserted into the timer migration
630 * hierarchy, wakeup value is returned.
631 */
632 if (tmc->idle) {
> 633 u64 ret = tmc->wakeup;
634
635 tmc->wakeup = KTIME_MAX;
636
637 if (nextexp != KTIME_MAX) {
638 if (nextexp != tmc->cpuevt.nextevt.expires ||
639 test_bit(0, &tmc->cpuevt.ignore))
640 ret = tmigr_new_timer(tmc, nextexp);
641 }
642
643 goto unlock;
644 }
645
646 /*
647 * When tmigr_remote is active, set cpu inactive path and queuing of
648 * nextexp is done by handle remote path.
649 */
650 ret = __tmigr_cpu_deactivate(tmc, nextexp);
651
652 tmc->idle = 1;
653
654 unlock:
655 raw_spin_unlock(&tmc->lock);
656 return ret;
657 }
658
--
0-DAY CI Kernel Test Service
https://01.org/lkp
On Tue, Oct 25, 2022 at 03:58:35PM +0200, Anna-Maria Behnsen wrote:
> When the next tick is in the past, the delta between basemono and the next
> tick gets negativ. But the next tick should never be in the past. The
> negative effect of a wrong next tick might be a stop of the tick and timers
> might expire late.
>
> To prevent expensive debugging when changing underlying code, add a
> WARN_ON_ONCE into this code path.
>
> Signed-off-by: Anna-Maria Behnsen <[email protected]>
> Reviewed-by: Thomas Gleixner <[email protected]>
Reviewed-by: Frederic Weisbecker <[email protected]>
> ---
> kernel/time/tick-sched.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> index b0e3c9205946..7ffdc7ba19b4 100644
> --- a/kernel/time/tick-sched.c
> +++ b/kernel/time/tick-sched.c
> @@ -826,6 +826,8 @@ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu)
> * If the tick is due in the next period, keep it ticking or
> * force prod the timer.
> */
> + WARN_ON_ONCE(basemono > next_tick);
> +
> delta = next_tick - basemono;
> if (delta <= (u64)TICK_NSEC) {
> /*
> --
> 2.30.2
>
Hi Anna-Maria,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on tip/timers/core]
[also build test WARNING on powerpc/next tip/x86/core linus/master tj-wq/for-next v6.1-rc2]
[cannot apply to tip/timers/nohz next-20221025]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Anna-Maria-Behnsen/timer-Move-from-a-push-remote-at-enqueue-to-a-pull-at-expiry-model/20221025-220106
patch link: https://lore.kernel.org/r/20221025135850.51044-15-anna-maria%40linutronix.de
patch subject: [PATCH v3 14/17] timer: Implement the hierarchical pull model
config: x86_64-rhel-8.3-func (attached as .config)
compiler: gcc-11 (Debian 11.3.0-8) 11.3.0
reproduce (this is a W=1 build):
# https://github.com/intel-lab-lkp/linux/commit/b57766ae36b5e3dd11225f0259f9fd7d39a79e94
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Anna-Maria-Behnsen/timer-Move-from-a-push-remote-at-enqueue-to-a-pull-at-expiry-model/20221025-220106
git checkout b57766ae36b5e3dd11225f0259f9fd7d39a79e94
# save the config file
mkdir build_dir && cp config build_dir/.config
make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash kernel/time/
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
All warnings (new ones prefixed by >>):
kernel/time/timer_migration.c: In function 'tmigr_cpu_deactivate':
kernel/time/timer_migration.c:633:21: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
633 | u64 ret = tmc->wakeup;
| ^~~
In file included from include/linux/bitmap.h:9,
from include/linux/cpumask.h:12,
from arch/x86/include/asm/cpumask.h:5,
from arch/x86/include/asm/msr.h:11,
from arch/x86/include/asm/processor.h:22,
from arch/x86/include/asm/cpufeature.h:5,
from arch/x86/include/asm/thread_info.h:53,
from include/linux/thread_info.h:60,
from arch/x86/include/asm/preempt.h:7,
from include/linux/preempt.h:78,
from include/linux/spinlock.h:56,
from include/linux/mmzone.h:8,
from include/linux/gfp.h:7,
from include/linux/slab.h:15,
from kernel/time/timer_migration.c:8:
kernel/time/timer_migration.c: In function 'tmigr_inactive_up':
>> include/linux/find.h:168:37: warning: array subscript 'long unsigned int[0]' is partly outside array bounds of 'union tmigr_state[1]' [-Warray-bounds]
168 | unsigned long val = *addr & GENMASK(size - 1, 0);
| ^~~~~
kernel/time/timer_migration.c:495:37: note: while referencing 'newstate'
495 | union tmigr_state curstate, newstate;
| ^~~~~~~~
vim +168 include/linux/find.h
c7f612cdf091de include/asm-generic/bitops/find.h Akinobu Mita 2006-03-26 154
b7ec62d7ee0f0b include/asm-generic/bitops/find.h Yury Norov 2021-08-14 155 #ifndef find_first_bit
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 156 /**
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 157 * find_first_bit - find the first set bit in a memory region
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 158 * @addr: The address to start the search at
ec778edf97dcaa include/asm-generic/bitops/find.h Cody P Schafer 2013-11-12 159 * @size: The maximum number of bits to search
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 160 *
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 161 * Returns the bit number of the first set bit.
ec778edf97dcaa include/asm-generic/bitops/find.h Cody P Schafer 2013-11-12 162 * If no bits are set, returns @size.
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 163 */
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 164 static inline
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 165 unsigned long find_first_bit(const unsigned long *addr, unsigned long size)
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 166 {
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 167 if (small_const_nbits(size)) {
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 @168 unsigned long val = *addr & GENMASK(size - 1, 0);
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 169
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 170 return val ? __ffs(val) : size;
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 171 }
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 172
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 173 return _find_first_bit(addr, size);
2cc7b6a44ac21d include/asm-generic/bitops/find.h Yury Norov 2021-05-06 174 }
b7ec62d7ee0f0b include/asm-generic/bitops/find.h Yury Norov 2021-08-14 175 #endif
708ff2a0097b02 include/asm-generic/bitops/find.h Akinobu Mita 2010-09-29 176
--
0-DAY CI Kernel Test Service
https://01.org/lkp
On Tue, Oct 25, 2022 at 03:58:36PM +0200, Anna-Maria Behnsen wrote:
> Both call sides of __next_timer_interrupt() store the return value directly
> in base->next_expiry. Move the store into __next_timer_interrupt().
>
> No functional change.
>
> Signed-off-by: Anna-Maria Behnsen <[email protected]>
> Reviewed-by: Thomas Gleixner <[email protected]>
> ---
> kernel/time/timer.c | 11 +++++++----
> 1 file changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/time/timer.c b/kernel/time/timer.c
> index 717fcb9fb14a..7695c733dfa5 100644
> --- a/kernel/time/timer.c
> +++ b/kernel/time/timer.c
> @@ -1571,8 +1571,10 @@ static int next_pending_bucket(struct timer_base *base, unsigned offset,
> /*
> * Search the first expiring timer in the various clock levels. Caller must
> * hold base->lock.
> + *
> + * Store next expiry time in base->next_expiry.
> */
> -static unsigned long __next_timer_interrupt(struct timer_base *base)
> +static void __next_timer_interrupt(struct timer_base *base)
> {
> unsigned long clk, next, adj;
> unsigned lvl, offset = 0;
> @@ -1638,10 +1640,11 @@ static unsigned long __next_timer_interrupt(struct timer_base *base)
> clk += adj;
> }
>
> + base->next_expiry = next;
> base->next_expiry_recalc = false;
In that case, maybe rename that function as next_expiry_recalc() to make its
purpose clearer?
Thanks!
> base->timers_pending = !(next == base->clk + NEXT_TIMER_MAX_DELTA);
>
> - return next;
> + return;
> }
>
> #ifdef CONFIG_NO_HZ_COMMON
> @@ -1701,7 +1704,7 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
>
> raw_spin_lock(&base->lock);
> if (base->next_expiry_recalc)
> - base->next_expiry = __next_timer_interrupt(base);
> + __next_timer_interrupt(base);
> nextevt = base->next_expiry;
>
> /*
> @@ -1784,7 +1787,7 @@ static inline void __run_timers(struct timer_base *base)
> WARN_ON_ONCE(!levels && !base->next_expiry_recalc
> && base->timers_pending);
> base->clk++;
> - base->next_expiry = __next_timer_interrupt(base);
> + __next_timer_interrupt(base);
>
> while (levels--)
> expire_timers(base, heads + levels);
> --
> 2.30.2
>
On Tue, Oct 25, 2022 at 03:58:38PM +0200, Anna-Maria Behnsen wrote:
> From: Thomas Gleixner <[email protected]>
>
> To improve readability of the code, split base->idle calculation and
> expires calculation into separate parts.
>
> Signed-off-by: Thomas Gleixner <[email protected]>
> Signed-off-by: Anna-Maria Behnsen <[email protected]>
> ---
> kernel/time/timer.c | 29 ++++++++++++++---------------
> 1 file changed, 14 insertions(+), 15 deletions(-)
>
> diff --git a/kernel/time/timer.c b/kernel/time/timer.c
> index f34bc75ff848..cb4194ecca60 100644
> --- a/kernel/time/timer.c
> +++ b/kernel/time/timer.c
> @@ -1727,21 +1727,20 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
> base->clk = nextevt;
> }
>
> - if (time_before_eq(nextevt, basej)) {
> - expires = basem;
> - base->is_idle = false;
> - } else {
> - if (base->timers_pending)
> - expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
> - /*
> - * If we expect to sleep more than a tick, mark the base idle.
> - * Also the tick is stopped so any added timer must forward
> - * the base clk itself to keep granularity small. This idle
> - * logic is only maintained for the BASE_STD base, deferrable
> - * timers may still see large granularity skew (by design).
> - */
> - if ((expires - basem) > TICK_NSEC)
> - base->is_idle = true;
> + /*
> + * Base is idle if the next event is more than a tick away. Also
> + * the tick is stopped so any added timer must forward the base clk
> + * itself to keep granularity small. This idle logic is only
> + * maintained for the BASE_STD base, deferrable timers may still
> + * see large granularity skew (by design).
> + */
> + base->is_idle = time_after(nextevt, basej + 1);
A subtle change, yet welcome, is introduced here. If the next event is just
one jiffy ahead, base->is_idle used to be left untouched. So the behaviour
was:
1) If the tick was already stopped (so we must be after an idle IRQ),
base->is_idle remains true
2) If the tick was not yet stopped, base->is_idle remains false
After this patch, 1) becomes:
1) If the tick was already stopped, turn base->is_idle to false
As a result, we may spare an IPI if we remotely enqueue a timer to
an idle CPU that is going to tick on the next jiffy.
Whether it's worth mentioning in the changelog, I leave it to you.
> +
> + if (base->timers_pending) {
> + /* If we missed a tick already, force 0 delta */
> + if (time_before_eq(nextevt, basej))
> + nextevt = basej;
Can be time_before() then, right?
Reviewed-by: Frederic Weisbecker <[email protected]>
> + expires = basem + (u64)(nextevt - basej) * TICK_NSEC;
> }
> raw_spin_unlock(&base->lock);
>
> --
> 2.30.2
>
On Tue, Oct 25, 2022 at 03:58:39PM +0200, Anna-Maria Behnsen wrote:
> @@ -1711,38 +1724,69 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
> if (cpu_is_offline(smp_processor_id()))
> return expires;
>
> - raw_spin_lock(&base->lock);
> + base_local = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
> + base_global = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
>
> - nextevt = next_timer_interrupt(base);
> + raw_spin_lock(&base_local->lock);
> + raw_spin_lock_nested(&base_global->lock, SINGLE_DEPTH_NESTING);
> +
> + nextevt_local = next_timer_interrupt(base_local);
> + nextevt_global = next_timer_interrupt(base_global);
>
> /*
> * We have a fresh next event. Check whether we can forward the
> * base. We can only do that when @basej is past base->clk
> * otherwise we might rewind base->clk.
> */
> - if (time_after(basej, base->clk)) {
> - if (time_after(nextevt, basej))
> - base->clk = basej;
> - else if (time_after(nextevt, base->clk))
> - base->clk = nextevt;
> + if (time_after(basej, base_local->clk)) {
> + if (time_after(nextevt_local, basej))
> + base_local->clk = basej;
> + else if (time_after(nextevt_local, base_local->clk))
> + base_local->clk = nextevt_local;
> + }
> +
> + if (time_after(basej, base_global->clk)) {
> + if (time_after(nextevt_global, basej))
> + base_global->clk = basej;
> + else if (time_after(nextevt_global, base_global->clk))
> + base_global->clk = nextevt_global;
Perhaps make a helper for the two above?
> }
>
> @@ -1763,6 +1807,9 @@ void timer_clear_idle(void)
> * the lock in the exit from idle path.
> */
> base->is_idle = false;
> +
> + base = this_cpu_ptr(&timer_bases[BASE_GLOBAL]);
> + base->is_idle = false;
May be just:
__this_cpu_write(timer_bases[BASE_LOCAL].is_idle, false)
__this_cpu_write(timer_bases[BASE_GLOBAL].is_idle, false)
> }
> #endif
>
> @@ -1820,17 +1869,21 @@ static __latent_entropy void run_timer_softirq(struct softirq_action *h)
> */
> static void run_local_timers(void)
> {
> - struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_STD]);
> + struct timer_base *base = this_cpu_ptr(&timer_bases[BASE_LOCAL]);
>
> hrtimer_run_queues();
> /* Raise the softirq only if required. */
> if (time_before(jiffies, base->next_expiry)) {
> if (!IS_ENABLED(CONFIG_NO_HZ_COMMON))
> return;
> - /* CPU is awake, so check the deferrable base. */
> + /* CPU is awake, so check for the global base. */
> base++;
> - if (time_before(jiffies, base->next_expiry))
> - return;
> + if (time_before(jiffies, base->next_expiry)) {
> + /* CPU is awake, so check the deferrable base. */
> + base++;
> + if (time_before(jiffies, base->next_expiry))
> + return;
> + }
Could be:
for (i = 0; i < NR_BASES; i++) {
struct timer_base *base = this_cpu_ptr(&timer_bases[i]);
if (time_after_eq(jiffies, base->next_expiry)) {
raise_softirq(TIMER_SOFTIRQ);
return;
}
}
Reviewed-by: Frederic Weisbecker <[email protected]>
Thanks!
> }
> raise_softirq(TIMER_SOFTIRQ);
> }
> --
> 2.30.2
>
Greeting,
FYI, we noticed BUG:KASAN:stack-out-of-bounds_in_tmigr_inactive_up due to commit (built with gcc-11):
commit: b57766ae36b5e3dd11225f0259f9fd7d39a79e94 ("[PATCH v3 14/17] timer: Implement the hierarchical pull model")
url: https://github.com/intel-lab-lkp/linux/commits/Anna-Maria-Behnsen/timer-Move-from-a-push-remote-at-enqueue-to-a-pull-at-expiry-model/20221025-220106
base: https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git 8be3f96ceddb911539a53d87a66da84a04502366
patch link: https://lore.kernel.org/lkml/[email protected]
patch subject: [PATCH v3 14/17] timer: Implement the hierarchical pull model
in testcase: boot
on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
[ 387.110466][ T0] BUG: KASAN: stack-out-of-bounds in tmigr_inactive_up (include/linux/find.h:168 kernel/time/timer_migration.c:518)
[ 387.112931][ T0] Read of size 8 at addr ffffffff84a07b30 by task swapper/0/0
[ 387.115269][ T0]
[ 387.116037][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.1.0-rc1-00015-gb57766ae36b5 #1
[ 387.118685][ T0] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
[ 387.121748][ T0] Call Trace:
[ 387.122894][ T0] <TASK>
[ 387.123933][ T0] dump_stack_lvl (lib/dump_stack.c:107 (discriminator 1))
[ 387.125423][ T0] print_address_description+0x87/0x2a5
[ 387.127400][ T0] print_report (mm/kasan/report.c:396)
[ 387.128880][ T0] ? kasan_addr_to_slab (mm/kasan/common.c:35)
[ 387.130367][ T0] ? tmigr_inactive_up (include/linux/find.h:168 kernel/time/timer_migration.c:518)
[ 387.131888][ T0] kasan_report (mm/kasan/report.c:497)
[ 387.133264][ T0] ? tmigr_inactive_up (include/linux/find.h:168 kernel/time/timer_migration.c:518)
[ 387.134852][ T0] tmigr_inactive_up (include/linux/find.h:168 kernel/time/timer_migration.c:518)
[ 387.136406][ T0] ? kasan_save_stack (mm/kasan/common.c:47)
[ 387.137843][ T0] ? tmigr_handle_remote_up (kernel/time/timer_migration.c:494)
[ 387.139488][ T0] ? secondary_startup_64_no_verify (arch/x86/kernel/head_64.S:358)
[ 387.141281][ T0] __tmigr_cpu_deactivate (kernel/time/timer_migration.c:143 kernel/time/timer_migration.c:155 kernel/time/timer_migration.c:599)
[ 387.142987][ T0] ? tmigr_inactive_up (kernel/time/timer_migration.c:583)
[ 387.144609][ T0] ? _raw_write_lock_irq (kernel/locking/spinlock.c:153)
[ 387.146255][ T0] ? sched_clock_local (kernel/sched/clock.c:290)
[ 387.147849][ T0] tmigr_cpu_deactivate (kernel/time/timer_migration.c:652)
[ 387.149436][ T0] ? tmigr_cpu_activate (kernel/time/timer_migration.c:613)
[ 387.150980][ T0] ? _raw_spin_lock (arch/x86/include/asm/atomic.h:202 include/linux/atomic/atomic-instrumented.h:543 include/asm-generic/qspinlock.h:111 include/linux/spinlock.h:186 include/linux/spinlock_api_smp.h:134 kernel/locking/spinlock.c:154)
[ 387.152436][ T0] ? _raw_write_lock_irq (kernel/locking/spinlock.c:153)
[ 387.154068][ T0] ? update_load_avg (kernel/sched/fair.c:3851 kernel/sched/fair.c:4186)
[ 387.155707][ T0] forward_and_idle_timer_bases (kernel/time/timer.c:1868)
[ 387.157540][ T0] tick_nohz_next_event (kernel/time/tick-sched.c:839)
[ 387.159056][ T0] ? tick_nohz_full_kick (kernel/time/tick-sched.c:804)
[ 387.160631][ T0] ? __switch_to (arch/x86/include/asm/bitops.h:55 include/asm-generic/bitops/instrumented-atomic.h:29 include/linux/thread_info.h:89 arch/x86/include/asm/fpu/sched.h:65 arch/x86/kernel/process_64.c:623)
[ 387.162049][ T0] tick_nohz_idle_stop_tick (kernel/time/tick-sched.c:1119 kernel/time/tick-sched.c:1149)
[ 387.163559][ T0] cpuidle_idle_call (kernel/sched/idle.c:191)
[ 387.164948][ T0] ? arch_cpu_idle_exit+0xc0/0xc0
[ 387.166569][ T0] do_idle (kernel/sched/idle.c:303)
[ 387.167905][ T0] cpu_startup_entry (kernel/sched/idle.c:399 (discriminator 1))
[ 387.169470][ T0] rest_init (init/main.c:702)
[ 387.170873][ T0] arch_call_rest_init+0xf/0x19
[ 387.172468][ T0] start_kernel (init/main.c:1147)
[ 387.173851][ T0] secondary_startup_64_no_verify (arch/x86/kernel/head_64.S:358)
[ 387.175657][ T0] </TASK>
[ 387.176668][ T0]
[ 387.177433][ T0] The buggy address belongs to stack of task swapper/0/0
[ 387.179259][ T0] and is located at offset 32 in frame:
[ 387.180879][ T0] tmigr_inactive_up (kernel/time/timer_migration.c:494)
[ 387.182424][ T0]
[ 387.183293][ T0] This frame has 1 object:
[ 387.184728][ T0] [32, 36) 'newstate'
[ 387.184736][ T0]
[ 387.186953][ T0] The buggy address belongs to the physical page:
[ 387.188951][ T0] page:(____ptrval____) refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x4a07
[ 387.192040][ T0] flags: 0xfffffc0001000(reserved|node=0|zone=1|lastcpupid=0x1fffff)
[ 387.194464][ T0] raw: 000fffffc0001000 ffffea00001281c8 ffffea00001281c8 0000000000000000
[ 387.196942][ T0] raw: 0000000000000000 0000000000000000 00000001ffffffff 0000000000000000
[ 387.199575][ T0] page dumped because: kasan: bad access detected
[ 387.201611][ T0] page_owner info is not present (never set?)
[ 387.206375][ T0]
[ 387.207215][ T0] Memory state around the buggy address:
[ 387.208801][ T0] ffffffff84a07a00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[ 387.211366][ T0] ffffffff84a07a80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[ 387.213864][ T0] >ffffffff84a07b00: 00 00 f1 f1 f1 f1 04 f3 f3 f3 00 00 00 00 00 00
[ 387.216353][ T0] ^
[ 387.218190][ T0] ffffffff84a07b80: 00 00 00 00 00 f1 f1 f1 f1 00 00 00 00 f3 f3 f3
[ 387.220848][ T0] ffffffff84a07c00: f3 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1 00
[ 387.223223][ T0] ==================================================================
[ 387.225589][ T0] Disabling lock debugging due to kernel taint
If you fix the issue, kindly add following tag
| Reported-by: kernel test robot <[email protected]>
| Link: https://lore.kernel.org/oe-lkp/[email protected]
To reproduce:
# build kernel
cd linux
cp config-6.1.0-rc1-00015-gb57766ae36b5 .config
make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 olddefconfig prepare modules_prepare bzImage modules
make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 INSTALL_MOD_PATH=<mod-install-dir> modules_install
cd <mod-install-dir>
find lib/ | cpio -o -H newc --quiet | gzip > modules.cgz
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
bin/lkp qemu -k <bzImage> -m modules.cgz job-script # job-script is attached in this email
# if come across any failure that blocks the test,
# please remove ~/.lkp and /lkp dir to run from a clean state.
--
0-DAY CI Kernel Test Service
https://01.org/lkp