Skip to content
Browse files

sched/fair: Fix bandwidth timer clock drift condition

I noticed that cgroup task groups constantly get throttled even
if they have low CPU usage, this causes some jitters on the response
time to some of our business containers when enabling CPU quotas.

It's very simple to reproduce:

  mkdir /sys/fs/cgroup/cpu/test
  cd /sys/fs/cgroup/cpu/test
  echo 100000 > cpu.cfs_quota_us
  echo $$ > tasks

then repeat:

  cat cpu.stat | grep nr_throttled  # nr_throttled will increase steadily

After some analysis, we found that cfs_rq::runtime_remaining will
be cleared by expire_cfs_rq_runtime() due to two equal but stale
"cfs_{b|q}->runtime_expires" after period timer is re-armed.

The current condition to judge clock drift in expire_cfs_rq_runtime()
is wrong, the two runtime_expires are actually the same when clock
drift happens, so this condtion can never hit. The orginal design was
correctly done by this commit:

  a9cf55b ("sched: Expire invalid runtime")

... but was changed to be the current implementation due to its locking bug.

This patch introduces another way, it adds a new field in both structures
cfs_rq and cfs_bandwidth to record the expiration update sequence, and
uses them to figure out if clock drift happens (true if they are equal).

Signed-off-by: Xunlei Pang <>
Signed-off-by: Peter Zijlstra (Intel) <>
Reviewed-by: Ben Segall <>
Cc: Linus Torvalds <>
Cc: Peter Zijlstra <>
Cc: Thomas Gleixner <>
Fixes: 51f2176 ("sched/fair: Fix unlocked reads of some cfs_b->quota/period")
Signed-off-by: Ingo Molnar <>
  • Loading branch information...
Xunlei Pang Ingo Molnar
Xunlei Pang authored and Ingo Molnar committed Jun 20, 2018
1 parent 296b2ff commit 512ac999d2755d2b7109e996a76b6fb8b888631d
Showing with 12 additions and 8 deletions.
  1. +8 −6 kernel/sched/fair.c
  2. +4 −2 kernel/sched/sched.h
@@ -4590,6 +4590,7 @@ void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b)
now = sched_clock_cpu(smp_processor_id());
cfs_b->runtime = cfs_b->quota;
cfs_b->runtime_expires = now + ktime_to_ns(cfs_b->period);

static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg)
@@ -4612,6 +4613,7 @@ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq)
struct task_group *tg = cfs_rq->tg;
struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(tg);
u64 amount = 0, min_amount, expires;
int expires_seq;

/* note: this is a positive sum as runtime_remaining <= 0 */
min_amount = sched_cfs_bandwidth_slice() - cfs_rq->runtime_remaining;
@@ -4628,6 +4630,7 @@ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq)
cfs_b->idle = 0;
expires_seq = cfs_b->expires_seq;
expires = cfs_b->runtime_expires;

@@ -4637,8 +4640,10 @@ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq)
* spread between our sched_clock and the one on which runtime was
* issued.
if ((s64)(expires - cfs_rq->runtime_expires) > 0)
if (cfs_rq->expires_seq != expires_seq) {
cfs_rq->expires_seq = expires_seq;
cfs_rq->runtime_expires = expires;

return cfs_rq->runtime_remaining > 0;
@@ -4664,12 +4669,9 @@ static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq)
* has not truly expired.
* Fortunately we can check determine whether this the case by checking
* whether the global deadline has advanced. It is valid to compare
* cfs_b->runtime_expires without any locks since we only care about
* exact equality, so a partial write will still work.
* whether the global deadline(cfs_b->expires_seq) has advanced.

if (cfs_rq->runtime_expires != cfs_b->runtime_expires) {
if (cfs_rq->expires_seq == cfs_b->expires_seq) {
/* extend local deadline, drift is bounded above by 2 ticks */
cfs_rq->runtime_expires += TICK_NSEC;
} else {
@@ -334,9 +334,10 @@ struct cfs_bandwidth {
u64 runtime;
s64 hierarchical_quota;
u64 runtime_expires;
int expires_seq;

int idle;
int period_active;
short idle;
short period_active;
struct hrtimer period_timer;
struct hrtimer slack_timer;
struct list_head throttled_cfs_rq;
@@ -551,6 +552,7 @@ struct cfs_rq {

int runtime_enabled;
int expires_seq;
u64 runtime_expires;
s64 runtime_remaining;

0 comments on commit 512ac99

Please sign in to comment.
You can’t perform that action at this time.