@@ -344,10 +344,29 @@ struct cgroup_base_stat {
* frequency decreases the cost of each read.
*
* This struct hosts both the fields which implement the above -
- * updated_children and updated_next - and the fields which track basic
- * resource statistics on top of it - bsync, bstat and last_bstat.
+ * updated_children and updated_next.
*/
struct cgroup_rstat_cpu {
+ /*
+ * Child cgroups with stat updates on this cpu since the last read
+ * are linked on the parent's ->updated_children through
+ * ->updated_next.
+ *
+ * In addition to being more compact, singly-linked list pointing
+ * to the cgroup makes it unnecessary for each per-cpu struct to
+ * point back to the associated cgroup.
+ *
+ * Protected by per-cpu cgroup_rstat_cpu_lock.
+ */
+ struct cgroup *updated_children; /* terminated by self cgroup */
+ struct cgroup *updated_next; /* NULL iff not on the list */
+};
+
+/*
+ * This struct hosts the fields which track basic resource statistics on
+ * top of it - bsync, bstat and last_bstat.
+ */
+struct cgroup_rstat_base_cpu {
/*
* ->bsync protects ->bstat. These are the only fields which get
* updated in the hot path.
@@ -374,20 +393,6 @@ struct cgroup_rstat_cpu {
* deltas to propagate to the per-cpu subtree_bstat.
*/
struct cgroup_base_stat last_subtree_bstat;
-
- /*
- * Child cgroups with stat updates on this cpu since the last read
- * are linked on the parent's ->updated_children through
- * ->updated_next.
- *
- * In addition to being more compact, singly-linked list pointing
- * to the cgroup makes it unnecessary for each per-cpu struct to
- * point back to the associated cgroup.
- *
- * Protected by per-cpu cgroup_rstat_cpu_lock.
- */
- struct cgroup *updated_children; /* terminated by self cgroup */
- struct cgroup *updated_next; /* NULL iff not on the list */
};
struct cgroup_freezer_state {
@@ -518,6 +523,7 @@ struct cgroup {
/* per-cpu recursive resource statistics */
struct cgroup_rstat_cpu __percpu *rstat_cpu;
+ struct cgroup_rstat_base_cpu __percpu *rstat_base_cpu;
struct list_head rstat_css_list;
/*
@@ -161,10 +161,14 @@ static struct static_key_true *cgroup_subsys_on_dfl_key[] = {
};
#undef SUBSYS
-static DEFINE_PER_CPU(struct cgroup_rstat_cpu, cgrp_dfl_root_rstat_cpu);
+static DEFINE_PER_CPU(struct cgroup_rstat_cpu, root_rstat_cpu);
+static DEFINE_PER_CPU(struct cgroup_rstat_base_cpu, root_rstat_base_cpu);
/* the default hierarchy */
-struct cgroup_root cgrp_dfl_root = { .cgrp.rstat_cpu = &cgrp_dfl_root_rstat_cpu };
+struct cgroup_root cgrp_dfl_root = {
+ .cgrp.rstat_cpu = &root_rstat_cpu,
+ .cgrp.rstat_base_cpu = &root_rstat_base_cpu,
+};
EXPORT_SYMBOL_GPL(cgrp_dfl_root);
/*
@@ -19,6 +19,12 @@ static struct cgroup_rstat_cpu *cgroup_rstat_cpu(struct cgroup *cgrp, int cpu)
return per_cpu_ptr(cgrp->rstat_cpu, cpu);
}
+static struct cgroup_rstat_base_cpu *cgroup_rstat_base_cpu(
+ struct cgroup *cgrp, int cpu)
+{
+ return per_cpu_ptr(cgrp->rstat_base_cpu, cpu);
+}
+
/*
* Helper functions for rstat per CPU lock (cgroup_rstat_cpu_lock).
*
@@ -351,12 +357,22 @@ int cgroup_rstat_init(struct cgroup *cgrp)
return -ENOMEM;
}
+ if (!cgrp->rstat_base_cpu) {
+ cgrp->rstat_base_cpu = alloc_percpu(struct cgroup_rstat_base_cpu);
+ if (!cgrp->rstat_cpu) {
+ free_percpu(cgrp->rstat_cpu);
+ return -ENOMEM;
+ }
+ }
+
/* ->updated_children list is self terminated */
for_each_possible_cpu(cpu) {
struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu);
+ struct cgroup_rstat_base_cpu *rstatbc =
+ cgroup_rstat_base_cpu(cgrp, cpu);
rstatc->updated_children = cgrp;
- u64_stats_init(&rstatc->bsync);
+ u64_stats_init(&rstatbc->bsync);
}
return 0;
@@ -379,6 +395,8 @@ void cgroup_rstat_exit(struct cgroup *cgrp)
free_percpu(cgrp->rstat_cpu);
cgrp->rstat_cpu = NULL;
+ free_percpu(cgrp->rstat_base_cpu);
+ cgrp->rstat_base_cpu = NULL;
}
void __init cgroup_rstat_boot(void)
@@ -419,9 +437,9 @@ static void cgroup_base_stat_sub(struct cgroup_base_stat *dst_bstat,
static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu)
{
- struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu);
+ struct cgroup_rstat_base_cpu *rstatbc = cgroup_rstat_base_cpu(cgrp, cpu);
struct cgroup *parent = cgroup_parent(cgrp);
- struct cgroup_rstat_cpu *prstatc;
+ struct cgroup_rstat_base_cpu *prstatbc;
struct cgroup_base_stat delta;
unsigned seq;
@@ -431,15 +449,15 @@ static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu)
/* fetch the current per-cpu values */
do {
- seq = __u64_stats_fetch_begin(&rstatc->bsync);
- delta = rstatc->bstat;
- } while (__u64_stats_fetch_retry(&rstatc->bsync, seq));
+ seq = __u64_stats_fetch_begin(&rstatbc->bsync);
+ delta = rstatbc->bstat;
+ } while (__u64_stats_fetch_retry(&rstatbc->bsync, seq));
/* propagate per-cpu delta to cgroup and per-cpu global statistics */
- cgroup_base_stat_sub(&delta, &rstatc->last_bstat);
+ cgroup_base_stat_sub(&delta, &rstatbc->last_bstat);
cgroup_base_stat_add(&cgrp->bstat, &delta);
- cgroup_base_stat_add(&rstatc->last_bstat, &delta);
- cgroup_base_stat_add(&rstatc->subtree_bstat, &delta);
+ cgroup_base_stat_add(&rstatbc->last_bstat, &delta);
+ cgroup_base_stat_add(&rstatbc->subtree_bstat, &delta);
/* propagate cgroup and per-cpu global delta to parent (unless that's root) */
if (cgroup_parent(parent)) {
@@ -448,73 +466,73 @@ static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu)
cgroup_base_stat_add(&parent->bstat, &delta);
cgroup_base_stat_add(&cgrp->last_bstat, &delta);
- delta = rstatc->subtree_bstat;
- prstatc = cgroup_rstat_cpu(parent, cpu);
- cgroup_base_stat_sub(&delta, &rstatc->last_subtree_bstat);
- cgroup_base_stat_add(&prstatc->subtree_bstat, &delta);
- cgroup_base_stat_add(&rstatc->last_subtree_bstat, &delta);
+ delta = rstatbc->subtree_bstat;
+ prstatbc = cgroup_rstat_base_cpu(parent, cpu);
+ cgroup_base_stat_sub(&delta, &rstatbc->last_subtree_bstat);
+ cgroup_base_stat_add(&prstatbc->subtree_bstat, &delta);
+ cgroup_base_stat_add(&rstatbc->last_subtree_bstat, &delta);
}
}
-static struct cgroup_rstat_cpu *
+static struct cgroup_rstat_base_cpu *
cgroup_base_stat_cputime_account_begin(struct cgroup *cgrp, unsigned long *flags)
{
- struct cgroup_rstat_cpu *rstatc;
+ struct cgroup_rstat_base_cpu *rstatbc;
- rstatc = get_cpu_ptr(cgrp->rstat_cpu);
- *flags = u64_stats_update_begin_irqsave(&rstatc->bsync);
- return rstatc;
+ rstatbc = get_cpu_ptr(cgrp->rstat_base_cpu);
+ *flags = u64_stats_update_begin_irqsave(&rstatbc->bsync);
+ return rstatbc;
}
static void cgroup_base_stat_cputime_account_end(struct cgroup *cgrp,
- struct cgroup_rstat_cpu *rstatc,
+ struct cgroup_rstat_base_cpu *rstatbc,
unsigned long flags)
{
- u64_stats_update_end_irqrestore(&rstatc->bsync, flags);
+ u64_stats_update_end_irqrestore(&rstatbc->bsync, flags);
cgroup_rstat_updated(cgrp, smp_processor_id());
- put_cpu_ptr(rstatc);
+ put_cpu_ptr(rstatbc);
}
void __cgroup_account_cputime(struct cgroup *cgrp, u64 delta_exec)
{
- struct cgroup_rstat_cpu *rstatc;
+ struct cgroup_rstat_base_cpu *rstatbc;
unsigned long flags;
- rstatc = cgroup_base_stat_cputime_account_begin(cgrp, &flags);
- rstatc->bstat.cputime.sum_exec_runtime += delta_exec;
- cgroup_base_stat_cputime_account_end(cgrp, rstatc, flags);
+ rstatbc = cgroup_base_stat_cputime_account_begin(cgrp, &flags);
+ rstatbc->bstat.cputime.sum_exec_runtime += delta_exec;
+ cgroup_base_stat_cputime_account_end(cgrp, rstatbc, flags);
}
void __cgroup_account_cputime_field(struct cgroup *cgrp,
enum cpu_usage_stat index, u64 delta_exec)
{
- struct cgroup_rstat_cpu *rstatc;
+ struct cgroup_rstat_base_cpu *rstatbc;
unsigned long flags;
- rstatc = cgroup_base_stat_cputime_account_begin(cgrp, &flags);
+ rstatbc = cgroup_base_stat_cputime_account_begin(cgrp, &flags);
switch (index) {
case CPUTIME_NICE:
- rstatc->bstat.ntime += delta_exec;
+ rstatbc->bstat.ntime += delta_exec;
fallthrough;
case CPUTIME_USER:
- rstatc->bstat.cputime.utime += delta_exec;
+ rstatbc->bstat.cputime.utime += delta_exec;
break;
case CPUTIME_SYSTEM:
case CPUTIME_IRQ:
case CPUTIME_SOFTIRQ:
- rstatc->bstat.cputime.stime += delta_exec;
+ rstatbc->bstat.cputime.stime += delta_exec;
break;
#ifdef CONFIG_SCHED_CORE
case CPUTIME_FORCEIDLE:
- rstatc->bstat.forceidle_sum += delta_exec;
+ rstatbc->bstat.forceidle_sum += delta_exec;
break;
#endif
default:
break;
}
- cgroup_base_stat_cputime_account_end(cgrp, rstatc, flags);
+ cgroup_base_stat_cputime_account_end(cgrp, rstatbc, flags);
}
/*
This non-functional change serves as preparation for moving to subsystem-based rstat trees. The base stats are not an actual subsystem, but in future commits they will have exclusive rstat trees just as other subsystems will. Moving the base stat objects into a new struct allows the cgroup_rstat_cpu struct to become more compact since it now only contains the minimum amount of pointers needed for rstat participation. Subsystems will (in future commits) make use of the compact cgroup_rstat_cpu struct while avoiding the memory overhead of the base stat objects which they will not use. An instance of the new struct cgroup_rstat_base_cpu was placed on the cgroup struct so it can retain ownership of these base stats common to all cgroups. A helper function was added for looking up the cpu-specific base stats of a given cgroup. Finally, initialization and variable names were adjusted where applicable. Signed-off-by: JP Kobryn <inwardvessel@gmail.com> --- include/linux/cgroup-defs.h | 38 ++++++++++------- kernel/cgroup/cgroup.c | 8 +++- kernel/cgroup/rstat.c | 84 ++++++++++++++++++++++--------------- 3 files changed, 79 insertions(+), 51 deletions(-)