Re: [PATCH v5 7/7] sched/fair: Add latency list

From: Youssef Esmat
Date: Thu Oct 13 2022 - 13:19:58 EST

Next message: David Matlack: "Re: [PATCH v4 00/11] KVM: x86/mmu: Make tdp_mmu a read-only parameter"
Previous message: Linus Torvalds: "Re: [GIT PULL] virtio: fixes, features"
In reply to: Vincent Guittot: "Re: [PATCH v5 7/7] sched/fair: Add latency list"
Next in thread: Vincent Guittot: "Re: [PATCH v5 7/7] sched/fair: Add latency list"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Oct 12, 2022 at 10:21 AM Vincent Guittot
<vincent.guittot@xxxxxxxxxx> wrote:
>
> Le mardi 11 oct. 2022 à 18:54:27 (-0500), Youssef Esmat a écrit :
> > Hi Vincent,
> >
> > On Tue, Oct 11, 2022 at 12:10 PM Vincent Guittot
> > <vincent.guittot@xxxxxxxxxx> wrote:
>
> ...
>
> > > > > latency 0 latency -20
> > > > > Min Latencies: 60 61
> > > > > Avg Latencies: 1077 86
> > > > > Max Latencies: 87311 444
> > > > > 50% latencies: 92 85
> > > > > 75% latencies: 554 90
> > > > > 85% latencies: 1019 93
> > > > > 90% latencies: 1346 96
> > > > > 95% latencies: 5400 100
> > > > > 99% latencies: 19044 110
> > > > >
> > > > > Signed-off-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
> > > > > ---
> > > >
> > > > The ability to boost the latency sensitivity of a task seems very
> > > > interesting. I have been playing around with these changes and have
> > > > some observations.
> > > >
> > > > I tried 2 bursty tasks affinitized to the same CPU. The tasks sleep
> > > > for 1ms and run for 10ms in a loop. I first tried it without adjusting
> > > > the latency_nice value and took perf sched traces:
> > >
> > > The CPU is overloaded almost all the time as it can't run the 2 tasks
> > > (2*10ms every 11ms)
> > >
> > > >
> > > > latency_test:7040 | 2447.137 ms | 8 | avg: 6.546 ms |
> > > > max: 10.674 ms | max start: 353.809487 s | max end: 353.820161 s
> > > > latency_test:7028 | 2454.777 ms | 7 | avg: 4.494 ms |
> > > > max: 10.609 ms | max start: 354.804386 s | max end: 354.814995 s
> > > >
> > > > Everything looked as expected, for a 5s run they had similar runtime
> > > > and latency.
> > > >
> > > > I then adjusted one task to have a latency_nice of -20 (pid 8614
> > > > below) and took another set of traces:
> > > >
> > > > latency_test:8618 | 1845.534 ms | 131 | avg: 9.764 ms |
> > > > max: 10.686 ms | max start: 1405.737905 s | max end: 1405.748592 s
> > > > latency_test:8614 | 3033.635 ms | 16 | avg: 3.559 ms |
> > > > max: 10.467 ms | max start: 1407.594751 s | max end: 1407.605218 s
> > > >
> > > > The task with -20 latency_nice had significantly more runtime. The
> > > > average latency was improved but the max roughly stayed the same. As
> > > > expected the one with latency_nice value of 0 experienced more
> > > > switches, but so did the one with latency_nice of -20.
> > >
> > > Your results look unexpected because the vruntime of the tasks is not
> > > modified. So I can imagine that the thread with the low latency runs
> > > first up to the offset at the beg of the test but then they should
> > > switch regularly. I have tried a similar test with a modified rt-app
> > > and the result seems ok. I have a small difference but not the
> > > difference that you see.
> > >
> > > Could you share more details about your setup ? I'm going to try to
> > > reproduce your sequence
> >
> > I was using an intel core i7 with this frequency details:
> > CPU MHz: 4200.000
> > CPU max MHz: 4800.0000
> > CPU min MHz: 400.0000
> >
> > This is a snippet of the test I was using:
> >
> > struct sched_attr attr;
> > memset(&attr, 0, sizeof(struct sched_attr));
> > attr.size = sizeof(struct sched_attr);
> > attr.sched_latency_nice = nice_latency;
> > attr.sched_flags = SCHED_FLAG_LATENCY_NICE;
> >
> > // set nice latency value
> > int res = syscall(__NR_sched_setattr, 0, &attr, 0);
> >
> > while(1){
> > // wake up every ms
> > usleep(1000);
> > for(int i = 0; i < 40000000; i++){}
> > }
>
> Between v2 and v3, the sched_feat(GENTLE_FAIR_SLEEPERS) has diseappeared when
> I moved the computation of latency_offset at the setting of the latency prio
> instead of runtime. As a result, the latency nice task can preempt up to
> (sysctl_sched_latency - sysctl_sched_wakeup_granularity) but other threads are
> cap to half of sysctl_sched_latency at wakeup.
>
> Could you try the patch below ?

Thanks Vincent. That improved the runtime for the tests above.

Running two bursty tasks (1ms sleep and 10ms running loop) one of
which had a latency_nice of -20 (pid 9209) and the other 0:

latency_test:9205 | 2429.467 ms | 14 | avg: 6.293 ms |
max: 11.340 ms | max start: 209.543482 s | max end: 209.554821 s
latency_test:9209 | 2467.423 ms | 17 | avg: 3.346 ms |
max: 11.475 ms | max start: 208.374517 s | max end: 208.385992 s

The task with -20 latency_nice had an improved average latency, but
the max roughly stayed the same.

Running the same test with 3 bursty tasks, one of which had a
latency_nice of -20 and the other two with latency_nice = 0. The
results for this case roughly stayed the same after the changes:

latency_test:26088 | 1641.458 ms | 158 | avg: 19.613 ms |
max: 33.000 ms | max start: 871.707231 s | max end: 871.740231 s
latency_test:26295 | 1639.766 ms | 238 | avg: 10.259 ms |
max: 24.289 ms | max start: 873.917231 s | max end: 873.941519 s
latency_test:26401 | 1643.580 ms | 241 | avg: 10.200 ms |
max: 22.124 ms | max start: 876.289233 s | max end: 876.311357 s

The task with latency_nice -20 seemed to have the highest average and
max latencies.

I also tried an additional test where we had a task that was cpu bound
(while(1){}) and a task that ran for 1ms and slept for 1ms. The cpu
bound task had the latency_nice value of -20 and the bursty had
latency_nice of 0.

latency_test:17353 | 4557.699 ms | 356 | avg: 1.058 ms |
max: 1.181 ms | max start: 4324.806123 s | max end: 4324.807304 s
latency_test:17452 | 377.804 ms | 1 | avg: 0.000 ms |
max: 0.000 ms | max start: 0.000000 s | max end: 0.000000 s

The cpu bound task (pid 17353 above) had significantly more runtime.
If I reran the test with latency_nice = 0 for both:

latency_test:20748 | 2478.014 ms | 2367 | avg: 1.037 ms |
max: 1.182 ms | max start: 4460.769240 s | max end: 4460.770423 s
latency_test:20972 | 2455.888 ms | 3 | avg: 0.001 ms |
max: 0.002 ms | max start: 4460.628756 s | max end: 4460.628758 s

Would decreasing the latency_offset as the task runs and increasing it
as it sleeps help here?

>
> ---
> kernel/sched/core.c | 19 ++++++++++++++++---
> 1 file changed, 16 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 8b44685ae247..68f9a83d7089 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -1287,11 +1287,15 @@ static void set_load_weight(struct task_struct *p, bool update_load)
> static void set_latency_offset(struct task_struct *p)
> {
> long weight = sched_latency_to_weight[p->latency_prio];
> + unsigned long period = sysctl_sched_latency;
> s64 offset;
>
> - offset = sysctl_sched_latency * weight;
> + if (sched_feat(GENTLE_FAIR_SLEEPERS))
> + period >>= 1;
> + offset = period * weight;
> offset = div_s64(offset, NICE_LATENCY_WEIGHT_MAX);
> p->se.latency_offset = (long)offset;
> + trace_printk("set_latency_offset pid %d offset %ld", p->pid, offset);
> }
>
> #ifdef CONFIG_UCLAMP_TASK
> @@ -10894,12 +10898,17 @@ static int cpu_idle_write_s64(struct cgroup_subsys_state *css,
> static s64 cpu_latency_nice_read_s64(struct cgroup_subsys_state *css,
> struct cftype *cft)
> {
> + unsigned long period = sysctl_sched_latency;
> int last_delta = INT_MAX;
> int prio, delta;
> s64 weight;
>
> + if (sched_feat(GENTLE_FAIR_SLEEPERS))
> + period >>= 1;
> +
> weight = css_tg(css)->latency_offset * NICE_LATENCY_WEIGHT_MAX;
> - weight = div_s64(weight, sysctl_sched_latency);
> + period = sysctl_sched_latency;
> + weight = div_s64(weight, period);
>
> /* Find the closest nice value to the current weight */
> for (prio = 0; prio < ARRAY_SIZE(sched_latency_to_weight); prio++) {
> @@ -10915,6 +10924,7 @@ static s64 cpu_latency_nice_read_s64(struct cgroup_subsys_state *css,
> static int cpu_latency_nice_write_s64(struct cgroup_subsys_state *css,
> struct cftype *cft, s64 nice)
> {
> + unsigned long period;
> s64 latency_offset;
> long weight;
> int idx;
> @@ -10926,7 +10936,10 @@ static int cpu_latency_nice_write_s64(struct cgroup_subsys_state *css,
> idx = array_index_nospec(idx, LATENCY_NICE_WIDTH);
> weight = sched_latency_to_weight[idx];
>
> - latency_offset = sysctl_sched_latency * weight;
> + period = sysctl_sched_latency;
> + if (sched_feat(GENTLE_FAIR_SLEEPERS))
> + period >>= 1;
> + latency_offset = period * weight;
> latency_offset = div_s64(latency_offset, NICE_LATENCY_WEIGHT_MAX);
>
> return sched_group_set_latency(css_tg(css), latency_offset);
> --
> 2.17.1
>
> >
> > >
> > > >
> > > > Also tried running the same test but instead of using latency nice I
> > > > adjusted the nice value as a comparison. In that case one task had a
> > > > nice of -5 and the other was 0.
> > > >
> > > > nice_test:25219 | 1216.839 ms | 242 | avg: 10.295 ms |
> > > > max: 11.927 ms | max start: 5877.881279 s | max end: 5877.893206 s
> > > > nice_test:25235 | 3711.788 ms | 6 | avg: 1.026 ms |
> > > > max: 6.143 ms | max start: 5875.603741 s | max end: 5875.609883 s
> > > >
> > > > As expected the one with a nice value of -5 had more runtime but also
> > > > had better latency numbers than in the previous case of using
> > > > latency_nice.
> > > >
> > > > I also tried a similar test with 3 bursty tasks instead of two. In
> > > > this case all tasks had a latency_nice of 0:
> > > >
> > > > latency_test:11467 | 1641.131 ms | 161 | avg: 17.489 ms |
> > > > max: 21.011 ms | max start: 1542.656275 s | max end: 1542.677286 s
> > > > latency_test:11463 | 1644.809 ms | 161 | avg: 11.994 ms |
> > > > max: 25.012 ms | max start: 1545.657776 s | max end: 1545.682788 s
> > > > latency_test:11478 | 1643.211 ms | 160 | avg: 11.465 ms |
> > > > max: 21.012 ms | max start: 1546.159026 s | max end: 1546.180038 s
> > > >
> > > > Next I tried two tasks with a latency_nice of 0 and a third one had a
> > > > latency_nice of -20 (pid 11763 below):
> > > >
> > > > latency_test:11763 | 1645.482 ms | 159 | avg: 19.634 ms |
> > > > max: 31.016 ms | max start: 1623.834862 s | max end: 1623.865877 s
> > > > latency_test:11750 | 1644.276 ms | 259 | avg: 9.985 ms |
> > > > max: 21.012 ms | max start: 1623.953921 s | max end: 1623.974933 s
> > > > latency_test:11747 | 1642.745 ms | 262 | avg: 9.079 ms |
> > > > max: 25.013 ms | max start: 1620.980435 s | max end: 1621.005447 s
> > > >
> > > > In this case it seemed like the runtime was not affected by the
> > > > latency_nice value, but strangely the task with the latency nice of
> > > > -20 had a worse average and max latency than the other two. The
> > > > context switch times are also increased from the previous case.
> > > >
> > > > Have we considered an approach where the task that is marked as
> > > > latency sensitive gets a boosted nice value when it sleeps and is
> > > > either scaled down exponentially as it runs, or immediately reset to
> > > > its default when it runs for one tick?
> > > >
> > > > Thanks,
> > > > Youssef
> > > >
> > > >
> > > > > include/linux/sched.h | 2 +
> > > > > kernel/sched/fair.c | 96 +++++++++++++++++++++++++++++++++++++++++--
> > > > > kernel/sched/sched.h | 1 +
> > > > > 3 files changed, 96 insertions(+), 3 deletions(-)
> > > > >
> > > > > diff --git a/include/linux/sched.h b/include/linux/sched.h
> > > > > index a74cad08e91e..0b92674e3664 100644
> > > > > --- a/include/linux/sched.h
> > > > > +++ b/include/linux/sched.h
> > > > > @@ -547,6 +547,8 @@ struct sched_entity {
> > > > > /* For load-balancing: */
> > > > > struct load_weight load;
> > > > > struct rb_node run_node;
> > > > > + struct rb_node latency_node;
> > > > > + unsigned int on_latency;
> > > > > struct list_head group_node;
> > > > > unsigned int on_rq;
> > > > >
> > > > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > > > > index e524e892d118..1a72f34136d8 100644
> > > > > --- a/kernel/sched/fair.c
> > > > > +++ b/kernel/sched/fair.c
> > > > > @@ -664,7 +664,77 @@ struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq)
> > > > >
> > > > > return __node_2_se(last);
> > > > > }
> > > > > +#endif
> > > > >
> > > > > +/**************************************************************
> > > > > + * Scheduling class tree data structure manipulation methods:
> > > > > + * for latency
> > > > > + */
> > > > > +
> > > > > +static inline bool latency_before(struct sched_entity *a,
> > > > > + struct sched_entity *b)
> > > > > +{
> > > > > + return (s64)(a->vruntime + a->latency_offset - b->vruntime - b->latency_offset) < 0;
> > > > > +}
> > > > > +
> > > > > +#define __latency_node_2_se(node) \
> > > > > + rb_entry((node), struct sched_entity, latency_node)
> > > > > +
> > > > > +static inline bool __latency_less(struct rb_node *a, const struct rb_node *b)
> > > > > +{
> > > > > + return latency_before(__latency_node_2_se(a), __latency_node_2_se(b));
> > > > > +}
> > > > > +
> > > > > +/*
> > > > > + * Enqueue an entity into the latency rb-tree:
> > > > > + */
> > > > > +static void __enqueue_latency(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> > > > > +{
> > > > > +
> > > > > + /* Only latency sensitive entity can be added to the list */
> > > > > + if (se->latency_offset >= 0)
> > > > > + return;
> > > > > +
> > > > > + if (se->on_latency)
> > > > > + return;
> > > > > +
> > > > > + /*
> > > > > + * An execution time less than sysctl_sched_min_granularity means that
> > > > > + * the entity has been preempted by a higher sched class or an entity
> > > > > + * with higher latency constraint.
> > > > > + * Put it back in the list so it gets a chance to run 1st during the
> > > > > + * next slice.
> > > > > + */
> > > > > + if (!(flags & ENQUEUE_WAKEUP)) {
> > > > > + u64 delta_exec = se->sum_exec_runtime - se->prev_sum_exec_runtime;
> > > > > +
> > > > > + if (delta_exec >= sysctl_sched_min_granularity)
> > > > > + return;
> > > > > + }
> > > > > +
> > > > > + rb_add_cached(&se->latency_node, &cfs_rq->latency_timeline, __latency_less);
> > > > > + se->on_latency = 1;
> > > > > +}
> > > > > +
> > > > > +static void __dequeue_latency(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > > > > +{
> > > > > + if (se->on_latency) {
> > > > > + rb_erase_cached(&se->latency_node, &cfs_rq->latency_timeline);
> > > > > + se->on_latency = 0;
> > > > > + }
> > > > > +}
> > > > > +
> > > > > +static struct sched_entity *__pick_first_latency(struct cfs_rq *cfs_rq)
> > > > > +{
> > > > > + struct rb_node *left = rb_first_cached(&cfs_rq->latency_timeline);
> > > > > +
> > > > > + if (!left)
> > > > > + return NULL;
> > > > > +
> > > > > + return __latency_node_2_se(left);
> > > > > +}
> > > > > +
> > > > > +#ifdef CONFIG_SCHED_DEBUG
> > > > > /**************************************************************
> > > > > * Scheduling class statistics methods:
> > > > > */
> > > > > @@ -4455,8 +4525,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> > > > > check_schedstat_required();
> > > > > update_stats_enqueue_fair(cfs_rq, se, flags);
> > > > > check_spread(cfs_rq, se);
> > > > > - if (!curr)
> > > > > + if (!curr) {
> > > > > __enqueue_entity(cfs_rq, se);
> > > > > + __enqueue_latency(cfs_rq, se, flags);
> > > > > + }
> > > > > se->on_rq = 1;
> > > > >
> > > > > if (cfs_rq->nr_running == 1) {
> > > > > @@ -4542,8 +4614,10 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> > > > >
> > > > > clear_buddies(cfs_rq, se);
> > > > >
> > > > > - if (se != cfs_rq->curr)
> > > > > + if (se != cfs_rq->curr) {
> > > > > __dequeue_entity(cfs_rq, se);
> > > > > + __dequeue_latency(cfs_rq, se);
> > > > > + }
> > > > > se->on_rq = 0;
> > > > > account_entity_dequeue(cfs_rq, se);
> > > > >
> > > > > @@ -4631,6 +4705,7 @@ set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > > > > */
> > > > > update_stats_wait_end_fair(cfs_rq, se);
> > > > > __dequeue_entity(cfs_rq, se);
> > > > > + __dequeue_latency(cfs_rq, se);
> > > > > update_load_avg(cfs_rq, se, UPDATE_TG);
> > > > > }
> > > > >
> > > > > @@ -4669,7 +4744,7 @@ static struct sched_entity *
> > > > > pick_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *curr)
> > > > > {
> > > > > struct sched_entity *left = __pick_first_entity(cfs_rq);
> > > > > - struct sched_entity *se;
> > > > > + struct sched_entity *latency, *se;
> > > > >
> > > > > /*
> > > > > * If curr is set we have to see if its left of the leftmost entity
> > > > > @@ -4711,6 +4786,12 @@ pick_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *curr)
> > > > > se = cfs_rq->last;
> > > > > }
> > > > >
> > > > > + /* Check for latency sensitive entity waiting for running */
> > > > > + latency = __pick_first_latency(cfs_rq);
> > > > > + if (latency && (latency != se) &&
> > > > > + wakeup_preempt_entity(latency, se) < 1)
> > > > > + se = latency;
> > > > > +
> > > > > return se;
> > > > > }
> > > > >
> > > > > @@ -4734,6 +4815,7 @@ static void put_prev_entity(struct cfs_rq *cfs_rq, struct sched_entity *prev)
> > > > > update_stats_wait_start_fair(cfs_rq, prev);
> > > > > /* Put 'current' back into the tree. */
> > > > > __enqueue_entity(cfs_rq, prev);
> > > > > + __enqueue_latency(cfs_rq, prev, 0);
> > > > > /* in !on_rq case, update occurred at dequeue */
> > > > > update_load_avg(cfs_rq, prev, 0);
> > > > > }
> > > > > @@ -11717,6 +11799,7 @@ static void set_next_task_fair(struct rq *rq, struct task_struct *p, bool first)
> > > > > void init_cfs_rq(struct cfs_rq *cfs_rq)
> > > > > {
> > > > > cfs_rq->tasks_timeline = RB_ROOT_CACHED;
> > > > > + cfs_rq->latency_timeline = RB_ROOT_CACHED;
> > > > > u64_u32_store(cfs_rq->min_vruntime, (u64)(-(1LL << 20)));
> > > > > #ifdef CONFIG_SMP
> > > > > raw_spin_lock_init(&cfs_rq->removed.lock);
> > > > > @@ -12025,8 +12108,15 @@ int sched_group_set_latency(struct task_group *tg, s64 latency)
> > > > >
> > > > > for_each_possible_cpu(i) {
> > > > > struct sched_entity *se = tg->se[i];
> > > > > + struct rq *rq = cpu_rq(i);
> > > > > + struct rq_flags rf;
> > > > > +
> > > > > + rq_lock_irqsave(rq, &rf);
> > > > >
> > > > > + __dequeue_latency(se->cfs_rq, se);
> > > > > WRITE_ONCE(se->latency_offset, latency);
> > > > > +
> > > > > + rq_unlock_irqrestore(rq, &rf);
> > > > > }
> > > > >
> > > > > mutex_unlock(&shares_mutex);
> > > > > diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> > > > > index a15fb955092c..76bca172585c 100644
> > > > > --- a/kernel/sched/sched.h
> > > > > +++ b/kernel/sched/sched.h
> > > > > @@ -599,6 +599,7 @@ struct cfs_rq {
> > > > > #endif
> > > > >
> > > > > struct rb_root_cached tasks_timeline;
> > > > > + struct rb_root_cached latency_timeline;
> > > > >
> > > > > /*
> > > > > * 'curr' points to currently running entity on this cfs_rq.
> > > > > --
> > > > > 2.17.1
> > > > >
> > > > >
> > > > >

Next message: David Matlack: "Re: [PATCH v4 00/11] KVM: x86/mmu: Make tdp_mmu a read-only parameter"
Previous message: Linus Torvalds: "Re: [GIT PULL] virtio: fixes, features"
In reply to: Vincent Guittot: "Re: [PATCH v5 7/7] sched/fair: Add latency list"
Next in thread: Vincent Guittot: "Re: [PATCH v5 7/7] sched/fair: Add latency list"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]