In cgroup2, we plan to combine cpuacct and cpu subsystem.
This patch implement cpu subsystem in unified hierarchy
and implement a cpu.usage in cpu subsystem.
Signed-off-by: Dongsheng Yang <yangds.fnst@xxxxxxxxxxxxxx>
---
kernel/sched/core.c | 187 ++++++++++++++++++++++++++++++++++++++++++++++-
kernel/sched/deadline.c | 2 +-
kernel/sched/fair.c | 2 +-
kernel/sched/rt.c | 2 +-
kernel/sched/sched.h | 10 +++
kernel/sched/stop_task.c | 2 +-
6 files changed, 199 insertions(+), 6 deletions(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 732e993..1333b6c 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7349,11 +7349,14 @@ int in_sched_functions(unsigned long addr)
}
#ifdef CONFIG_CGROUP_SCHED
+DEFINE_PER_CPU(u64, root_cpuusage);
/*
* Default task group.
* Every task in system belongs to this group at bootup.
*/
-struct task_group root_task_group;
+struct task_group root_task_group = {
+ .cpuusage = &root_cpuusage,
+};
LIST_HEAD(task_groups);
#endif
@@ -7692,10 +7695,26 @@ void set_curr_task(int cpu, struct task_struct *p)
/* task_group_lock serializes the addition/removal of task groups */
static DEFINE_SPINLOCK(task_group_lock);
+static int alloc_cpuusage(struct task_group *tg)
+{
+ tg->cpuusage = alloc_percpu(u64);
+ if (!tg->cpuusage)
+ goto err;
+ return 0;
+err:
+ return -ENOMEM;
+}
+
+static void free_cpuusage(struct task_group *tg)
+{
+ free_percpu(tg->cpuusage);
+}
+
static void free_sched_group(struct task_group *tg)
{
free_fair_sched_group(tg);
free_rt_sched_group(tg);
+ free_cpuusage(tg);
autogroup_free(tg);
kfree(tg);
}
@@ -7715,6 +7734,9 @@ struct task_group *sched_create_group(struct task_group *parent)
if (!alloc_rt_sched_group(tg, parent))
goto err;
+ if (alloc_cpuusage(tg))
+ goto err;
+
return tg;
err:
@@ -8194,6 +8216,35 @@ static inline struct task_group *css_tg(struct cgroup_subsys_state *css)
return css ? container_of(css, struct task_group, css) : NULL;
}
+static inline struct task_group *parent_tg(struct task_group *tg)
+{
+ return css_tg(tg->css.parent);
+}
+
+void cpu_usage_charge(struct task_struct *tsk, u64 cputime)
+{
+ struct task_group *tg;
+ int cpu;
+
+ cpu = task_cpu(tsk);
+
+ rcu_read_lock();
+
+ tg = task_group(tsk);
+
+ while (true) {
+ u64 *cpuusage = per_cpu_ptr(tg->cpuusage, cpu);
+ *cpuusage += cputime;
+
+ tg = parent_tg(tg);
+ if (!tg)
+ break;
+ }
+ rcu_read_unlock();
+
+ cpuacct_charge(tsk, cputime);
+}
+
static struct cgroup_subsys_state *
cpu_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
{
@@ -8552,6 +8603,128 @@ static u64 cpu_rt_period_read_uint(struct cgroup_subsys_state *css,
}
#endif /* CONFIG_RT_GROUP_SCHED */
+static u64 cpu_usage_percpu_read(struct task_group *tg, int cpu)
+{
+ u64 *cpuusage = per_cpu_ptr(tg->cpuusage, cpu);
+ u64 data;
+
+#ifndef CONFIG_64BIT
+ /*
+ * Take rq->lock to make 64-bit read safe on 32-bit platforms.
+ */
+ raw_spin_lock_irq(&cpu_rq(cpu)->lock);
+ data = *cpuusage;
+ raw_spin_unlock_irq(&cpu_rq(cpu)->lock);
+#else
+ data = *cpuusage;
+#endif
+
+ return data;
+}
+
+static void cpu_usage_percpu_write(struct task_group *tg, int cpu, u64 val)
+{
+ u64 *cpuusage = per_cpu_ptr(tg->cpuusage, cpu);
+
+#ifndef CONFIG_64BIT
+ /*
+ * Take rq->lock to make 64-bit write safe on 32-bit platforms.
+ */
+ raw_spin_lock_irq(&cpu_rq(cpu)->lock);
+ *cpuusage = val;
+ raw_spin_unlock_irq(&cpu_rq(cpu)->lock);
+#else
+ *cpuusage = val;
+#endif
+}
+
+static u64 cpu_usage_read(struct cgroup_subsys_state *css, struct cftype *cft)
+{
+ struct task_group *tg = css_tg(css);
+ u64 totalcpuusage = 0;
+ int i;
+
+ for_each_present_cpu(i)
+ totalcpuusage += cpu_usage_percpu_read(tg, i);
+
+ return totalcpuusage;
+}
+
+static int cpu_usage_write(struct cgroup_subsys_state *css, struct cftype *cft,
+ u64 val)
+{
+ struct task_group *tg = css_tg(css);
+ int err = 0;
+ int i;
+
+ /*
+ * Only allow '0' here to do a reset.
+ */
+ if (val) {
+ err = -EINVAL;
+ goto out;
+ }
+
+ for_each_present_cpu(i)
+ cpu_usage_percpu_write(tg, i, 0);
+
+out:
+ return err;
+}
+
+static int cpu_usage_percpu_seq_show(struct seq_file *m, void *V)
+{
+ struct task_group *tg = css_tg(seq_css(m));
+ u64 percpu;
+ int i;
+
+ for_each_present_cpu(i) {
+ percpu = cpu_usage_percpu_read(tg, i);
+ seq_printf(m, "%llu ", (unsigned long long) percpu);
+ }
+ seq_printf(m, "\n");
+ return 0;
+}
+
+static struct cftype cpu_legacy_files[] = {
+#ifdef CONFIG_FAIR_GROUP_SCHED
+ {
+ .name = "shares",
+ .read_u64 = cpu_shares_read_u64,
+ .write_u64 = cpu_shares_write_u64,
+ },
+#endif
+#ifdef CONFIG_CFS_BANDWIDTH
+ {
+ .name = "cfs_quota_us",
+ .read_s64 = cpu_cfs_quota_read_s64,
+ .write_s64 = cpu_cfs_quota_write_s64,
+ },
+ {
+ .name = "cfs_period_us",
+ .read_u64 = cpu_cfs_period_read_u64,
+ .write_u64 = cpu_cfs_period_write_u64,
+ },
+ {
+ .name = "stat",
+ .seq_show = cpu_stats_show,
+ },
+#endif
+#ifdef CONFIG_RT_GROUP_SCHED
+ {
+ .name = "rt_runtime_us",
+ .read_s64 = cpu_rt_runtime_read,
+ .write_s64 = cpu_rt_runtime_write,
+ },
+ {
+ .name = "rt_period_us",
+ .read_u64 = cpu_rt_period_read_uint,
+ .write_u64 = cpu_rt_period_write_uint,
+ },
+#endif
+ { } /* terminate */
+};
+
static struct cftype cpu_files[] = {
#ifdef CONFIG_FAIR_GROUP_SCHED
{
@@ -8588,6 +8761,15 @@ static struct cftype cpu_files[] = {
.write_u64 = cpu_rt_period_write_uint,
},
#endif
+ {
+ .name = "usage",
+ .read_u64 = cpu_usage_read,
+ .write_u64 = cpu_usage_write,
+ },
+ {
+ .name = "usage_percpu",
+ .seq_show = cpu_usage_percpu_seq_show,
+ },
{ } /* terminate */
};
@@ -8599,7 +8781,8 @@ struct cgroup_subsys cpu_cgrp_subsys = {
.fork = cpu_cgroup_fork,
.can_attach = cpu_cgroup_can_attach,
.attach = cpu_cgroup_attach,
- .legacy_cftypes = cpu_files,
+ .dfl_cftypes = cpu_files,
+ .legacy_cftypes = cpu_legacy_files,
.early_init = 1,
};
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 8b0a15e..efae061 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -739,7 +739,7 @@ static void update_curr_dl(struct rq *rq)
account_group_exec_runtime(curr, delta_exec);
curr->se.exec_start = rq_clock_task(rq);
- cpuacct_charge(curr, delta_exec);
+ cpu_usage_charge(curr, delta_exec);
sched_rt_avg_update(rq, delta_exec);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index cfdc0e6..ee2fb15 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -726,7 +726,7 @@ static void update_curr(struct cfs_rq *cfs_rq)
struct task_struct *curtask = task_of(curr);
trace_sched_stat_runtime(curtask, delta_exec, curr->vruntime);
- cpuacct_charge(curtask, delta_exec);
+ cpu_usage_charge(curtask, delta_exec);
account_group_exec_runtime(curtask, delta_exec);
}
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 8ec86ab..efd666c 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -956,7 +956,7 @@ static void update_curr_rt(struct rq *rq)
account_group_exec_runtime(curr, delta_exec);
curr->se.exec_start = rq_clock_task(rq);
- cpuacct_charge(curr, delta_exec);
+ cpu_usage_charge(curr, delta_exec);
sched_rt_avg_update(rq, delta_exec);
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 141a16c..fc8db08 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -266,6 +266,9 @@ struct task_group {
struct list_head siblings;
struct list_head children;
+ /* cpuusage holds pointer to a u64-type object on every cpu */
+ u64 __percpu *cpuusage;
+
#ifdef CONFIG_SCHED_AUTOGROUP
struct autogroup *autogroup;
#endif
@@ -336,10 +339,17 @@ extern void sched_move_task(struct task_struct *tsk);
extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
#endif
+extern void cpu_usage_charge(struct task_struct *tsk, u64 cputime);
+
#else /* CONFIG_CGROUP_SCHED */
struct cfs_bandwidth { };
+static inline void cpu_usage_charge(struct task_struct *tsk, u64 cputime)
+{
+ cpuacct_charge(tsk, cputime);
+}
+
#endif /* CONFIG_CGROUP_SCHED */
/* CFS-related fields in a runqueue */
diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c
index cbc67da..132d056 100644
--- a/kernel/sched/stop_task.c
+++ b/kernel/sched/stop_task.c
@@ -71,7 +71,7 @@ static void put_prev_task_stop(struct rq *rq, struct task_struct *prev)
account_group_exec_runtime(curr, delta_exec);
curr->se.exec_start = rq_clock_task(rq);
- cpuacct_charge(curr, delta_exec);
+ cpu_usage_charge(curr, delta_exec);
}
static void task_tick_stop(struct rq *rq, struct task_struct *curr, int queued)
--
1.8.4.2
--
To unsubscribe from this list: send the line "unsubscribe cgroups" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
