lkml.org 
[lkml]   [2019]   [Jan]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH v6 5/5] cpuset: Rebuild root domain deadline accounting information
Date
When the topology of root domains is modified by CPUset or CPUhotplug
operations information about the current deadline bandwidth held in the
root domain is lost.

This patch addresses the issue by recalculating the lost deadline
bandwidth information by circling through the deadline tasks held in
CPUsets and adding their current load to the root domain they are
associated with.

Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>

---

v5 -> v6: Since the list of tasks that belongs to each cgroup is lazily
built only the first time cgroups are mounted, to be able to use it for
restoring deadline accounting information, we need to make sure we build
it also when cgroups support is enabled, but cgroups are not yet mounted
(early hotplugging of CPUs might verify this condition).
---
include/linux/cgroup.h | 1 +
include/linux/sched.h | 5 +++
include/linux/sched/deadline.h | 8 +++++
kernel/cgroup/cgroup.c | 2 +-
kernel/cgroup/cpuset.c | 64 +++++++++++++++++++++++++++++++++-
kernel/sched/deadline.c | 31 ++++++++++++++++
kernel/sched/sched.h | 3 --
kernel/sched/topology.c | 15 ++++++--
8 files changed, 122 insertions(+), 7 deletions(-)

diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 9d12757a65b0..ffa97eea5dfd 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -141,6 +141,7 @@ struct task_struct *cgroup_taskset_first(struct cgroup_taskset *tset,
struct task_struct *cgroup_taskset_next(struct cgroup_taskset *tset,
struct cgroup_subsys_state **dst_cssp);

+void cgroup_enable_task_cg_lists(void);
void css_task_iter_start(struct cgroup_subsys_state *css, unsigned int flags,
struct css_task_iter *it);
struct task_struct *css_task_iter_next(struct css_task_iter *it);
diff --git a/include/linux/sched.h b/include/linux/sched.h
index a51c13c2b1a0..cb3b14f6069b 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -280,6 +280,11 @@ struct vtime {
u64 gtime;
};

+#ifdef CONFIG_SMP
+extern struct root_domain def_root_domain;
+extern struct mutex sched_domains_mutex;
+#endif
+
struct sched_info {
#ifdef CONFIG_SCHED_INFO
/* Cumulative counters: */
diff --git a/include/linux/sched/deadline.h b/include/linux/sched/deadline.h
index 0cb034331cbb..1aff00b65f3c 100644
--- a/include/linux/sched/deadline.h
+++ b/include/linux/sched/deadline.h
@@ -24,3 +24,11 @@ static inline bool dl_time_before(u64 a, u64 b)
{
return (s64)(a - b) < 0;
}
+
+#ifdef CONFIG_SMP
+
+struct root_domain;
+extern void dl_add_task_root_domain(struct task_struct *p);
+extern void dl_clear_root_domain(struct root_domain *rd);
+
+#endif /* CONFIG_SMP */
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 879c9f191f66..bc51786a4ae6 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -1802,7 +1802,7 @@ static int cgroup_remount(struct kernfs_root *kf_root, int *flags, char *data)
*/
static bool use_task_css_set_links __read_mostly;

-static void cgroup_enable_task_cg_lists(void)
+void cgroup_enable_task_cg_lists(void)
{
struct task_struct *p, *g;

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 8ee1b24b6353..338cd5149436 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -44,6 +44,7 @@
#include <linux/proc_fs.h>
#include <linux/rcupdate.h>
#include <linux/sched.h>
+#include <linux/sched/deadline.h>
#include <linux/sched/mm.h>
#include <linux/sched/task.h>
#include <linux/seq_file.h>
@@ -934,6 +935,67 @@ static int generate_sched_domains(cpumask_var_t **domains,
return ndoms;
}

+static void update_tasks_root_domain(struct cpuset *cs)
+{
+ struct css_task_iter it;
+ struct task_struct *task;
+
+ css_task_iter_start(&cs->css, 0, &it);
+
+ while ((task = css_task_iter_next(&it)))
+ dl_add_task_root_domain(task);
+
+ css_task_iter_end(&it);
+}
+
+static void rebuild_root_domains(void)
+{
+ struct cpuset *cs = NULL;
+ struct cgroup_subsys_state *pos_css;
+
+ lockdep_assert_held(&cpuset_mutex);
+ lockdep_assert_cpus_held();
+ lockdep_assert_held(&sched_domains_mutex);
+
+ cgroup_enable_task_cg_lists();
+
+ rcu_read_lock();
+
+ /*
+ * Clear default root domain DL accounting, it will be computed again
+ * if a task belongs to it.
+ */
+ dl_clear_root_domain(&def_root_domain);
+
+ cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) {
+
+ if (cpumask_empty(cs->effective_cpus)) {
+ pos_css = css_rightmost_descendant(pos_css);
+ continue;
+ }
+
+ css_get(&cs->css);
+
+ rcu_read_unlock();
+
+ update_tasks_root_domain(cs);
+
+ rcu_read_lock();
+ css_put(&cs->css);
+ }
+ rcu_read_unlock();
+}
+
+static void
+partition_and_rebuild_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
+ struct sched_domain_attr *dattr_new)
+{
+ mutex_lock(&sched_domains_mutex);
+ partition_sched_domains_locked(ndoms_new, doms_new, dattr_new);
+ rebuild_root_domains();
+ mutex_unlock(&sched_domains_mutex);
+}
+
/*
* Rebuild scheduler domains.
*
@@ -971,7 +1033,7 @@ static void rebuild_sched_domains_locked(void)
ndoms = generate_sched_domains(&doms, &attr);

/* Have scheduler rebuild the domains */
- partition_sched_domains(ndoms, doms, attr);
+ partition_and_rebuild_sched_domains(ndoms, doms, attr);
out:
put_online_cpus();
}
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 91e4202b0634..63d36958cbba 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -2285,6 +2285,37 @@ void __init init_sched_dl_class(void)
GFP_KERNEL, cpu_to_node(i));
}

+void dl_add_task_root_domain(struct task_struct *p)
+{
+ unsigned long flags;
+ struct rq_flags rf;
+ struct rq *rq;
+ struct dl_bw *dl_b;
+
+ rq = task_rq_lock(p, &rf);
+ if (!dl_task(p))
+ goto unlock;
+
+ dl_b = &rq->rd->dl_bw;
+ raw_spin_lock_irqsave(&dl_b->lock, flags);
+
+ dl_b->total_bw += p->dl.dl_bw;
+
+ raw_spin_unlock_irqrestore(&dl_b->lock, flags);
+
+unlock:
+ task_rq_unlock(rq, p, &rf);
+}
+
+void dl_clear_root_domain(struct root_domain *rd)
+{
+ unsigned long flags;
+
+ raw_spin_lock_irqsave(&rd->dl_bw.lock, flags);
+ rd->dl_bw.total_bw = 0;
+ raw_spin_unlock_irqrestore(&rd->dl_bw.lock, flags);
+}
+
#endif /* CONFIG_SMP */

static void switched_from_dl(struct rq *rq, struct task_struct *p)
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 618577fc9aa8..f34cc648c613 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -757,9 +757,6 @@ struct root_domain {
unsigned long max_cpu_capacity;
};

-extern struct root_domain def_root_domain;
-extern struct mutex sched_domains_mutex;
-
extern void init_defrootdomain(void);
extern int sched_init_domains(const struct cpumask *cpu_map);
extern void rq_attach_root(struct rq *rq, struct root_domain *rd);
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index dd94000cb378..dc3468b4b883 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -1961,9 +1961,20 @@ void partition_sched_domains_locked(int ndoms_new, cpumask_var_t doms_new[],
/* Destroy deleted domains: */
for (i = 0; i < ndoms_cur; i++) {
for (j = 0; j < n && !new_topology; j++) {
- if (cpumask_equal(doms_cur[i], doms_new[j])
- && dattrs_equal(dattr_cur, i, dattr_new, j))
+ if (cpumask_equal(doms_cur[i], doms_new[j]) &&
+ dattrs_equal(dattr_cur, i, dattr_new, j)) {
+ struct root_domain *rd;
+
+ /*
+ * This domain won't be destroyed and as such
+ * its dl_bw->total_bw needs to be cleared. It
+ * will be recomputed in function
+ * update_tasks_root_domain().
+ */
+ rd = cpu_rq(cpumask_any(doms_cur[i]))->rd;
+ dl_clear_root_domain(rd);
goto match1;
+ }
}
/* No match - a current sched domain not in new doms_new[] */
detach_destroy_domains(doms_cur[i]);
--
2.17.2
\
 
 \ /
  Last update: 2019-01-17 09:48    [W:0.080 / U:0.252 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site