[PATCH v5 08/10] perf: cache perf_event_groups_first for cgroups

[Ian Rogers]

Add a per-CPU cache of the pinned and flexible perf_event_groups_first
value for a cgroup avoiding an O(log(#perf events)) searches during
sched_in.

Based-on-work-by: Kan Liang <kan.liang@xxxxxxxxxxxxxxx>
Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
---
 include/linux/perf_event.h |  7 ++++
 kernel/events/core.c       | 84 ++++++++++++++++++++++++++++++++++----
 2 files changed, 82 insertions(+), 9 deletions(-)

diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index cd7d3b624655..a29a38df909e 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -892,6 +892,13 @@ struct perf_cgroup_info {
 struct perf_cgroup {
 	struct cgroup_subsys_state	css;
 	struct perf_cgroup_info	__percpu *info;
+	/*
+	 * A cache of the first event with the perf_cpu_context's
+	 * perf_event_context for the first event in pinned_groups or
+	 * flexible_groups. Avoids an rbtree search during sched_in.
+	 */
+	struct perf_event * __percpu    *pinned_event;
+	struct perf_event * __percpu    *flexible_event;
 };
 
 /*
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 3da9cc1ebc2d..5935d2474050 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -1638,6 +1638,27 @@ perf_event_groups_insert(struct perf_event_groups *groups,
 
 	rb_link_node(&event->group_node, parent, node);
 	rb_insert_color(&event->group_node, &groups->tree);
+#ifdef CONFIG_CGROUP_PERF
+	if (is_cgroup_event(event)) {
+		struct perf_event **cgrp_event;
+
+		if (event->attr.pinned) {
+			cgrp_event = per_cpu_ptr(event->cgrp->pinned_event,
+						event->cpu);
+		} else {
+			cgrp_event = per_cpu_ptr(event->cgrp->flexible_event,
+						event->cpu);
+		}
+		/*
+		 * Remember smallest, left-most, group index event. The
+		 * less-than condition is only possible if the group_index
+		 * overflows.
+		 */
+		if (!*cgrp_event ||
+			event->group_index < (*cgrp_event)->group_index)
+			*cgrp_event = event;
+	}
+#endif
 }
 
 /*
@@ -1652,6 +1673,9 @@ add_event_to_groups(struct perf_event *event, struct perf_event_context *ctx)
 	perf_event_groups_insert(groups, event);
 }
 
+static struct perf_event *
+perf_event_groups_next(struct perf_event *event);
+
 /*
  * Delete a group from a tree.
  */
@@ -1662,6 +1686,22 @@ perf_event_groups_delete(struct perf_event_groups *groups,
 	WARN_ON_ONCE(RB_EMPTY_NODE(&event->group_node) ||
 		     RB_EMPTY_ROOT(&groups->tree));
 
+#ifdef CONFIG_CGROUP_PERF
+	if (is_cgroup_event(event)) {
+		struct perf_event **cgrp_event;
+
+		if (event->attr.pinned) {
+			cgrp_event = per_cpu_ptr(event->cgrp->pinned_event,
+						event->cpu);
+		} else {
+			cgrp_event = per_cpu_ptr(event->cgrp->flexible_event,
+						event->cpu);
+		}
+		if (*cgrp_event == event)
+			*cgrp_event = perf_event_groups_next(event);
+	}
+#endif
+
 	rb_erase(&event->group_node, &groups->tree);
 	init_event_group(event);
 }
@@ -1679,7 +1719,8 @@ del_event_from_groups(struct perf_event *event, struct perf_event_context *ctx)
 }
 
 /*
- * Get the leftmost event in the cpu/cgroup subtree.
+ * Get the leftmost event in the cpu subtree without a cgroup (ie task or
+ * system-wide).
  */
 static struct perf_event *
 perf_event_groups_first(struct perf_event_groups *groups, int cpu,
@@ -3596,8 +3637,8 @@ static int ctx_groups_sched_in(struct perf_event_context *ctx,
 			.cap = ARRAY_SIZE(itrs),
 		};
 		/* Events not within a CPU context may be on any CPU. */
-		__heap_add(&event_heap, perf_event_groups_first(groups, -1,
-									NULL));
+		__heap_add(&event_heap,
+			perf_event_groups_first(groups, -1, NULL));
 	}
 	evt = event_heap.data;
 
@@ -3605,8 +3646,16 @@ static int ctx_groups_sched_in(struct perf_event_context *ctx,
 
 #ifdef CONFIG_CGROUP_PERF
 	for (; css; css = css->parent) {
-		__heap_add(&event_heap, perf_event_groups_first(groups, cpu,
-								css->cgroup));
+		struct perf_cgroup *cgrp;
+
+		/* root cgroup doesn't have events */
+		if (css->id == 1)
+			break;
+
+		cgrp = container_of(css, struct perf_cgroup, css);
+		__heap_add(&event_heap, is_pinned
+			? *per_cpu_ptr(cgrp->pinned_event, cpu)
+			: *per_cpu_ptr(cgrp->flexible_event, cpu));
 	}
 #endif
 
@@ -12672,18 +12721,35 @@ perf_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
 		return ERR_PTR(-ENOMEM);
 
 	jc->info = alloc_percpu(struct perf_cgroup_info);
-	if (!jc->info) {
-		kfree(jc);
-		return ERR_PTR(-ENOMEM);
-	}
+	if (!jc->info)
+		goto free_jc;
+
+	jc->pinned_event = alloc_percpu(struct perf_event *);
+	if (!jc->pinned_event)
+		goto free_jc_info;
+
+	jc->flexible_event = alloc_percpu(struct perf_event *);
+	if (!jc->flexible_event)
+		goto free_jc_pinned;
 
 	return &jc->css;
+
+free_jc_pinned:
+	free_percpu(jc->pinned_event);
+free_jc_info:
+	free_percpu(jc->info);
+free_jc:
+	kfree(jc);
+
+	return ERR_PTR(-ENOMEM);
 }
 
 static void perf_cgroup_css_free(struct cgroup_subsys_state *css)
 {
 	struct perf_cgroup *jc = container_of(css, struct perf_cgroup, css);
 
+	free_percpu(jc->pinned_event);
+	free_percpu(jc->flexible_event);
 	free_percpu(jc->info);
 	kfree(jc);
 }
-- 
2.24.0.393.g34dc348eaf-goog