perf hist: Use cached rbtrees

At the cost of an extra pointer, we can avoid the O(logN) cost of
finding the first element in the tree (smallest node), which is
something heavily required for histograms. Specifically, the following
are converted to rb_root_cached, and users accordingly:

hist::entries_in_array
hist::entries_in
hist::entries
hist::entries_collapsed
hist_entry::hroot_in
hist_entry::hroot_out
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/20181206191819.30182-7-dave@stgolabs.net
[ Added some missing conversions to rb_first_cached() ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

tools/perf/builtin-annotate.c

@@ -305,7 +305,7 @@ static void hists__find_annotations(struct hists *hists,
 				    struct perf_evsel *evsel,
 				    struct perf_annotate *ann)
 {
-	struct rb_node *nd = rb_first(&hists->entries), *next;
+	struct rb_node *nd = rb_first_cached(&hists->entries), *next;
 	int key = K_RIGHT;
 
 	while (nd) {

tools/perf/builtin-c2c.c

@@ -2088,7 +2088,7 @@ static int resort_hitm_cb(struct hist_entry *he)
 
 static int hists__iterate_cb(struct hists *hists, hists__resort_cb_t cb)
 {
-	struct rb_node *next = rb_first(&hists->entries);
+	struct rb_node *next = rb_first_cached(&hists->entries);
 	int ret = 0;
 
 	while (next) {
@@ -2215,7 +2215,7 @@ static void print_pareto(FILE *out)
 	if (WARN_ONCE(ret, "failed to setup sort entries\n"))
 		return;
 
-	nd = rb_first(&c2c.hists.hists.entries);
+	nd = rb_first_cached(&c2c.hists.hists.entries);
 
 	for (; nd; nd = rb_next(nd)) {
 		struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node);
@@ -2283,7 +2283,7 @@ static void perf_c2c__hists_fprintf(FILE *out, struct perf_session *session)
 static void c2c_browser__update_nr_entries(struct hist_browser *hb)
 {
 	u64 nr_entries = 0;
-	struct rb_node *nd = rb_first(&hb->hists->entries);
+	struct rb_node *nd = rb_first_cached(&hb->hists->entries);
 
 	while (nd) {
 		struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node);

tools/perf/builtin-diff.c

@@ -429,7 +429,7 @@ get_pair_fmt(struct hist_entry *he, struct diff_hpp_fmt *dfmt)
 
 static void hists__baseline_only(struct hists *hists)
 {
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *next;
 
 	if (hists__has(hists, need_collapse))
@@ -437,13 +437,13 @@ static void hists__baseline_only(struct hists *hists)
 	else
 		root = hists->entries_in;
 
-	next = rb_first(root);
+	next = rb_first_cached(root);
 	while (next != NULL) {
 		struct hist_entry *he = rb_entry(next, struct hist_entry, rb_node_in);
 
 		next = rb_next(&he->rb_node_in);
 		if (!hist_entry__next_pair(he)) {
-			rb_erase(&he->rb_node_in, root);
+			rb_erase_cached(&he->rb_node_in, root);
 			hist_entry__delete(he);
 		}
 	}
@@ -451,7 +451,7 @@ static void hists__baseline_only(struct hists *hists)
 
 static void hists__precompute(struct hists *hists)
 {
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *next;
 
 	if (hists__has(hists, need_collapse))
@@ -459,7 +459,7 @@ static void hists__precompute(struct hists *hists)
 	else
 		root = hists->entries_in;
 
-	next = rb_first(root);
+	next = rb_first_cached(root);
 	while (next != NULL) {
 		struct hist_entry *he, *pair;
 		struct data__file *d;

tools/perf/builtin-top.c

@@ -367,7 +367,7 @@ static void perf_top__prompt_symbol(struct perf_top *top, const char *msg)
 	if (p)
 		*p = 0;
 
-	next = rb_first(&hists->entries);
+	next = rb_first_cached(&hists->entries);
 	while (next) {
 		n = rb_entry(next, struct hist_entry, rb_node);
 		if (n->ms.sym && !strcmp(buf, n->ms.sym->name)) {

tools/perf/tests/hists_common.c

@@ -161,7 +161,7 @@ struct machine *setup_fake_machine(struct machines *machines)
 void print_hists_in(struct hists *hists)
 {
 	int i = 0;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	if (hists__has(hists, need_collapse))
@@ -170,7 +170,7 @@ void print_hists_in(struct hists *hists)
 		root = hists->entries_in;
 
 	pr_info("----- %s --------\n", __func__);
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	while (node) {
 		struct hist_entry *he;
 
@@ -191,13 +191,13 @@ void print_hists_in(struct hists *hists)
 void print_hists_out(struct hists *hists)
 {
 	int i = 0;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	root = &hists->entries;
 
 	pr_info("----- %s --------\n", __func__);
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	while (node) {
 		struct hist_entry *he;
 

tools/perf/tests/hists_cumulate.c

@@ -125,8 +125,8 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
 static void del_hist_entries(struct hists *hists)
 {
 	struct hist_entry *he;
-	struct rb_root *root_in;
-	struct rb_root *root_out;
+	struct rb_root_cached *root_in;
+	struct rb_root_cached *root_out;
 	struct rb_node *node;
 
 	if (hists__has(hists, need_collapse))
@@ -136,12 +136,12 @@ static void del_hist_entries(struct hists *hists)
 
 	root_out = &hists->entries;
 
-	while (!RB_EMPTY_ROOT(root_out)) {
-		node = rb_first(root_out);
+	while (!RB_EMPTY_ROOT(&root_out->rb_root)) {
+		node = rb_first_cached(root_out);
 
 		he = rb_entry(node, struct hist_entry, rb_node);
-		rb_erase(node, root_out);
-		rb_erase(&he->rb_node_in, root_in);
+		rb_erase_cached(node, root_out);
+		rb_erase_cached(&he->rb_node_in, root_in);
 		hist_entry__delete(he);
 	}
 }
@@ -198,7 +198,7 @@ static int do_test(struct hists *hists, struct result *expected, size_t nr_expec
 		print_hists_out(hists);
 	}
 
-	root = &hists->entries;
+	root = &hists->entries.rb_root;
 	for (node = rb_first(root), i = 0;
 	     node && (he = rb_entry(node, struct hist_entry, rb_node));
 	     node = rb_next(node), i++) {

tools/perf/tests/hists_link.c

@@ -142,7 +142,7 @@ static int find_sample(struct sample *samples, size_t nr_samples,
 static int __validate_match(struct hists *hists)
 {
 	size_t count = 0;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	/*
@@ -153,7 +153,7 @@ static int __validate_match(struct hists *hists)
 	else
 		root = hists->entries_in;
 
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	while (node) {
 		struct hist_entry *he;
 
@@ -192,7 +192,7 @@ static int __validate_link(struct hists *hists, int idx)
 	size_t count = 0;
 	size_t count_pair = 0;
 	size_t count_dummy = 0;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	/*
@@ -205,7 +205,7 @@ static int __validate_link(struct hists *hists, int idx)
 	else
 		root = hists->entries_in;
 
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	while (node) {
 		struct hist_entry *he;
 

tools/perf/tests/hists_output.c

@@ -91,8 +91,8 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
 static void del_hist_entries(struct hists *hists)
 {
 	struct hist_entry *he;
-	struct rb_root *root_in;
-	struct rb_root *root_out;
+	struct rb_root_cached *root_in;
+	struct rb_root_cached *root_out;
 	struct rb_node *node;
 
 	if (hists__has(hists, need_collapse))
@@ -102,12 +102,12 @@ static void del_hist_entries(struct hists *hists)
 
 	root_out = &hists->entries;
 
-	while (!RB_EMPTY_ROOT(root_out)) {
-		node = rb_first(root_out);
+	while (!RB_EMPTY_ROOT(&root_out->rb_root)) {
+		node = rb_first_cached(root_out);
 
 		he = rb_entry(node, struct hist_entry, rb_node);
-		rb_erase(node, root_out);
-		rb_erase(&he->rb_node_in, root_in);
+		rb_erase_cached(node, root_out);
+		rb_erase_cached(&he->rb_node_in, root_in);
 		hist_entry__delete(he);
 	}
 }
@@ -126,7 +126,7 @@ static int test1(struct perf_evsel *evsel, struct machine *machine)
 	int err;
 	struct hists *hists = evsel__hists(evsel);
 	struct hist_entry *he;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	field_order = NULL;
@@ -162,7 +162,7 @@ static int test1(struct perf_evsel *evsel, struct machine *machine)
 	}
 
 	root = &hists->entries;
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	he = rb_entry(node, struct hist_entry, rb_node);
 	TEST_ASSERT_VAL("Invalid hist entry",
 			!strcmp(COMM(he), "perf") && !strcmp(DSO(he), "perf") &&
@@ -228,7 +228,7 @@ static int test2(struct perf_evsel *evsel, struct machine *machine)
 	int err;
 	struct hists *hists = evsel__hists(evsel);
 	struct hist_entry *he;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	field_order = "overhead,cpu";
@@ -262,7 +262,7 @@ static int test2(struct perf_evsel *evsel, struct machine *machine)
 	}
 
 	root = &hists->entries;
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	he = rb_entry(node, struct hist_entry, rb_node);
 	TEST_ASSERT_VAL("Invalid hist entry",
 			CPU(he) == 1 && PID(he) == 100 && he->stat.period == 300);
@@ -284,7 +284,7 @@ static int test3(struct perf_evsel *evsel, struct machine *machine)
 	int err;
 	struct hists *hists = evsel__hists(evsel);
 	struct hist_entry *he;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	field_order = "comm,overhead,dso";
@@ -316,7 +316,7 @@ static int test3(struct perf_evsel *evsel, struct machine *machine)
 	}
 
 	root = &hists->entries;
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	he = rb_entry(node, struct hist_entry, rb_node);
 	TEST_ASSERT_VAL("Invalid hist entry",
 			!strcmp(COMM(he), "bash") && !strcmp(DSO(he), "bash") &&
@@ -358,7 +358,7 @@ static int test4(struct perf_evsel *evsel, struct machine *machine)
 	int err;
 	struct hists *hists = evsel__hists(evsel);
 	struct hist_entry *he;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	field_order = "dso,sym,comm,overhead,dso";
@@ -394,7 +394,7 @@ static int test4(struct perf_evsel *evsel, struct machine *machine)
 	}
 
 	root = &hists->entries;
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	he = rb_entry(node, struct hist_entry, rb_node);
 	TEST_ASSERT_VAL("Invalid hist entry",
 			!strcmp(DSO(he), "perf") && !strcmp(SYM(he), "cmd_record") &&
@@ -460,7 +460,7 @@ static int test5(struct perf_evsel *evsel, struct machine *machine)
 	int err;
 	struct hists *hists = evsel__hists(evsel);
 	struct hist_entry *he;
-	struct rb_root *root;
+	struct rb_root_cached *root;
 	struct rb_node *node;
 
 	field_order = "cpu,pid,comm,dso,sym";
@@ -497,7 +497,7 @@ static int test5(struct perf_evsel *evsel, struct machine *machine)
 	}
 
 	root = &hists->entries;
-	node = rb_first(root);
+	node = rb_first_cached(root);
 	he = rb_entry(node, struct hist_entry, rb_node);
 
 	TEST_ASSERT_VAL("Invalid hist entry",

tools/perf/ui/browsers/hists.c

@@ -49,7 +49,7 @@ static int hist_browser__get_folding(struct hist_browser *browser)
 	struct hists *hists = browser->hists;
 	int unfolded_rows = 0;
 
-	for (nd = rb_first(&hists->entries);
+	for (nd = rb_first_cached(&hists->entries);
 	     (nd = hists__filter_entries(nd, browser->min_pcnt)) != NULL;
 	     nd = rb_hierarchy_next(nd)) {
 		struct hist_entry *he =
@@ -267,7 +267,7 @@ static int hierarchy_count_rows(struct hist_browser *hb, struct hist_entry *he,
 	if (he->has_no_entry)
 		return 1;
 
-	node = rb_first(&he->hroot_out);
+	node = rb_first_cached(&he->hroot_out);
 	while (node) {
 		float percent;
 
@@ -372,7 +372,7 @@ static void hist_entry__init_have_children(struct hist_entry *he)
 		he->has_children = !RB_EMPTY_ROOT(&he->sorted_chain);
 		callchain__init_have_children(&he->sorted_chain);
 	} else {
-		he->has_children = !RB_EMPTY_ROOT(&he->hroot_out);
+		he->has_children = !RB_EMPTY_ROOT(&he->hroot_out.rb_root);
 	}
 
 	he->init_have_children = true;
@@ -508,7 +508,7 @@ static int hierarchy_set_folding(struct hist_browser *hb, struct hist_entry *he,
 	struct hist_entry *child;
 	int n = 0;
 
-	for (nd = rb_first(&he->hroot_out); nd; nd = rb_next(nd)) {
+	for (nd = rb_first_cached(&he->hroot_out); nd; nd = rb_next(nd)) {
 		child = rb_entry(nd, struct hist_entry, rb_node);
 		percent = hist_entry__get_percent_limit(child);
 		if (!child->filtered && percent >= hb->min_pcnt)
@@ -566,7 +566,7 @@ __hist_browser__set_folding(struct hist_browser *browser, bool unfold)
 	struct rb_node *nd;
 	struct hist_entry *he;
 
-	nd = rb_first(&browser->hists->entries);
+	nd = rb_first_cached(&browser->hists->entries);
 	while (nd) {
 		he = rb_entry(nd, struct hist_entry, rb_node);
 
@@ -1738,7 +1738,7 @@ static void ui_browser__hists_init_top(struct ui_browser *browser)
 		struct hist_browser *hb;
 
 		hb = container_of(browser, struct hist_browser, b);
-		browser->top = rb_first(&hb->hists->entries);
+		browser->top = rb_first_cached(&hb->hists->entries);
 	}
 }
 
@@ -2649,7 +2649,7 @@ add_socket_opt(struct hist_browser *browser, struct popup_action *act,
 static void hist_browser__update_nr_entries(struct hist_browser *hb)
 {
 	u64 nr_entries = 0;
-	struct rb_node *nd = rb_first(&hb->hists->entries);
+	struct rb_node *nd = rb_first_cached(&hb->hists->entries);
 
 	if (hb->min_pcnt == 0 && !symbol_conf.report_hierarchy) {
 		hb->nr_non_filtered_entries = hb->hists->nr_non_filtered_entries;
@@ -2669,7 +2669,7 @@ static void hist_browser__update_percent_limit(struct hist_browser *hb,
 					       double percent)
 {
 	struct hist_entry *he;
-	struct rb_node *nd = rb_first(&hb->hists->entries);
+	struct rb_node *nd = rb_first_cached(&hb->hists->entries);
 	u64 total = hists__total_period(hb->hists);
 	u64 min_callchain_hits = total * (percent / 100);
 

tools/perf/ui/gtk/hists.c

@@ -353,7 +353,7 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists,
 
 	g_object_unref(GTK_TREE_MODEL(store));
 
-	for (nd = rb_first(&hists->entries); nd; nd = rb_next(nd)) {
+	for (nd = rb_first_cached(&hists->entries); nd; nd = rb_next(nd)) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
 		GtkTreeIter iter;
 		u64 total = hists__total_period(h->hists);
@@ -401,7 +401,7 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists,
 }
 
 static void perf_gtk__add_hierarchy_entries(struct hists *hists,
-					    struct rb_root *root,
+					    struct rb_root_cached *root,
 					    GtkTreeStore *store,
 					    GtkTreeIter *parent,
 					    struct perf_hpp *hpp,
@@ -415,7 +415,7 @@ static void perf_gtk__add_hierarchy_entries(struct hists *hists,
 	u64 total = hists__total_period(hists);
 	int size;
 
-	for (node = rb_first(root); node; node = rb_next(node)) {
+	for (node = rb_first_cached(root); node; node = rb_next(node)) {
 		GtkTreeIter iter;
 		float percent;
 		char *bf;

tools/perf/ui/stdio/hist.c

@@ -788,7 +788,8 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows,
 
 	indent = hists__overhead_width(hists) + 4;
 
-	for (nd = rb_first(&hists->entries); nd; nd = __rb_hierarchy_next(nd, HMD_FORCE_CHILD)) {
+	for (nd = rb_first_cached(&hists->entries); nd;
+	     nd = __rb_hierarchy_next(nd, HMD_FORCE_CHILD)) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
 		float percent;
 

tools/perf/util/hist.h

@@ -70,10 +70,10 @@ struct thread;
 struct dso;
 
 struct hists {
-	struct rb_root		entries_in_array[2];
-	struct rb_root		*entries_in;
-	struct rb_root		entries;
-	struct rb_root		entries_collapsed;
+	struct rb_root_cached	entries_in_array[2];
+	struct rb_root_cached	*entries_in;
+	struct rb_root_cached	entries;
+	struct rb_root_cached	entries_collapsed;
 	u64			nr_entries;
 	u64			nr_non_filtered_entries;
 	u64			callchain_period;
@@ -230,7 +230,7 @@ static __pure inline bool hists__has_callchains(struct hists *hists)
 int hists__init(void);
 int __hists__init(struct hists *hists, struct perf_hpp_list *hpp_list);
 
-struct rb_root *hists__get_rotate_entries_in(struct hists *hists);
+struct rb_root_cached *hists__get_rotate_entries_in(struct hists *hists);
 
 struct perf_hpp {
 	char *buf;

tools/perf/util/sort.h

@@ -145,8 +145,8 @@ struct hist_entry {
 	union {
 		/* this is for hierarchical entry structure */
 		struct {
-			struct rb_root	hroot_in;
-			struct rb_root  hroot_out;
+			struct rb_root_cached	hroot_in;
+			struct rb_root_cached   hroot_out;
 		};				/* non-leaf entries */
 		struct rb_root	sorted_chain;	/* leaf entry has callchains */
 	};