lkml.org 
[lkml]   [2011]   [Mar]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    Date
    From
    Subject[patch 12/15] sched: maintain throttled rqs as a list
    Maintain a list of throttle runqueues instead of having to test every runqueue
    to determine whether we need to unthrottle.

    Signed-off-by: Paul Turner <pjt@google.com>

    ---
    kernel/sched.c | 7 +++++++
    kernel/sched_fair.c | 25 +++++++++++++------------
    2 files changed, 20 insertions(+), 12 deletions(-)

    Index: tip/kernel/sched.c
    ===================================================================
    --- tip.orig/kernel/sched.c
    +++ tip/kernel/sched.c
    @@ -256,6 +256,7 @@ struct cfs_bandwidth {
    s64 hierarchal_quota; /* used for validating consistency */
    struct hrtimer period_timer;

    + struct list_head throttled_cfs_rq;
    /* throttle statistics */
    u64 nr_periods;
    u64 nr_throttled;
    @@ -394,6 +395,8 @@ struct cfs_rq {
    int quota_enabled, throttled, throttle_count;
    s64 quota_remaining;
    u64 throttled_timestamp;
    +
    + struct list_head throttled_list;
    #endif
    #endif
    };
    @@ -433,6 +436,7 @@ void init_cfs_bandwidth(struct cfs_bandw
    raw_spin_lock_init(&cfs_b->lock);
    cfs_b->quota = cfs_b->runtime = quota;
    cfs_b->period = ns_to_ktime(period);
    + INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);

    hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
    cfs_b->period_timer.function = sched_cfs_period_timer;
    @@ -8141,6 +8145,9 @@ static void init_cfs_rq(struct cfs_rq *c
    #ifdef CONFIG_FAIR_GROUP_SCHED
    cfs_rq->rq = rq;
    /* allow initial update_cfs_load() to truncate */
    +#ifdef CONFIG_CFS_BANDWIDTH
    + INIT_LIST_HEAD(&cfs_rq->throttled_list);
    +#endif
    #ifdef CONFIG_SMP
    cfs_rq->load_stamp = 1;
    #endif
    Index: tip/kernel/sched_fair.c
    ===================================================================
    --- tip.orig/kernel/sched_fair.c
    +++ tip/kernel/sched_fair.c
    @@ -1405,6 +1405,7 @@ static int tg_throttle_down(struct task_
    static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
    {
    struct rq *rq = rq_of(cfs_rq);
    + struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg);
    struct sched_entity *se;
    long task_delta, dequeue = 1;

    @@ -1434,7 +1435,11 @@ static void throttle_cfs_rq(struct cfs_r
    if (!se)
    rq->nr_running += task_delta;

    + raw_spin_lock(&cfs_b->lock);
    cfs_rq->throttled = 1;
    + list_add_tail_rcu(&cfs_rq->throttled_list, &cfs_b->throttled_cfs_rq);
    + raw_spin_unlock(&cfs_b->lock);
    +
    cfs_rq->throttled_timestamp = rq_of(cfs_rq)->clock;
    }

    @@ -1450,6 +1455,10 @@ static void unthrottle_cfs_rq(struct cfs
    update_rq_clock(rq);
    raw_spin_lock(&cfs_b->lock);
    cfs_b->throttled_time += (rq->clock - cfs_rq->throttled_timestamp);
    +
    + cfs_rq->throttled = 0;
    + list_del_rcu(&cfs_rq->throttled_list);
    +
    raw_spin_unlock(&cfs_b->lock);
    cfs_rq->throttled_timestamp = 0;

    @@ -1459,7 +1468,6 @@ static void unthrottle_cfs_rq(struct cfs
    walk_tg_tree_from(cfs_rq->tg, tg_unthrottle_down, tg_nop,
    (void*)&udd);

    - cfs_rq->throttled = 0;
    if (!cfs_rq->load.weight)
    return;

    @@ -1484,22 +1492,15 @@ static void unthrottle_cfs_rq(struct cfs
    resched_task(rq->curr);
    }

    -static inline struct task_group *cfs_bandwidth_tg(struct cfs_bandwidth *cfs_b)
    -{
    - return container_of(cfs_b, struct task_group, cfs_bandwidth);
    -}
    -
    static u64 distribute_cfs_bandwidth(struct cfs_bandwidth *cfs_b, u64 runtime)
    {
    - int i;
    + struct cfs_rq *cfs_rq;
    u64 quota, remaining = runtime;
    - const struct cpumask *span;

    rcu_read_lock();
    - span = sched_bw_period_mask();
    - for_each_cpu(i, span) {
    - struct rq *rq = cpu_rq(i);
    - struct cfs_rq *cfs_rq = cfs_bandwidth_tg(cfs_b)->cfs_rq[i];
    + list_for_each_entry_rcu(cfs_rq, &cfs_b->throttled_cfs_rq,
    + throttled_list) {
    + struct rq *rq = rq_of(cfs_rq);

    raw_spin_lock(&rq->lock);
    if (within_bandwidth(cfs_rq))



    \
     
     \ /
      Last update: 2011-03-23 04:15    [W:3.521 / U:0.056 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site