1 files changed, 541 insertions, 88 deletions

diff --git a/tools/perf/util/cgroup.c b/tools/perf/util/cgroup.c
index 96bbda1ddb83..040eb75f0804 100644
--- a/tools/perf/util/cgroup.c
+++ b/tools/perf/util/cgroup.c
@@ -1,152 +1,373 @@
-#include "util.h"
-#include "../perf.h"
-#include "parse-options.h"
+// SPDX-License-Identifier: GPL-2.0
+#include <subcmd/parse-options.h>
 #include "evsel.h"
 #include "cgroup.h"
 #include "evlist.h"
+#include "rblist.h"
+#include "metricgroup.h"
+#include "stat.h"
+#include <linux/zalloc.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/statfs.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <string.h>
+#include <api/fs/fs.h>
+#include <ftw.h>
+#include <regex.h>
 
 int nr_cgroups;
+bool cgrp_event_expanded;
 
-static int
-cgroupfs_find_mountpoint(char *buf, size_t maxlen)
+/* used to match cgroup name with patterns */
+struct cgroup_name {
+	struct list_head list;
+	bool used;
+	char name[];
+};
+static LIST_HEAD(cgroup_list);
+
+static int open_cgroup(const char *name)
 {
-	FILE *fp;
-	char mountpoint[PATH_MAX + 1], tokens[PATH_MAX + 1], type[PATH_MAX + 1];
-	char *token, *saved_ptr = NULL;
-	int found = 0;
+	char path[PATH_MAX + 1];
+	char mnt[PATH_MAX + 1];
+	int fd;
 
-	fp = fopen("/proc/mounts", "r");
-	if (!fp)
+
+	if (cgroupfs_find_mountpoint(mnt, PATH_MAX + 1, "perf_event"))
 		return -1;
 
-	/*
-	 * in order to handle split hierarchy, we need to scan /proc/mounts
-	 * and inspect every cgroupfs mount point to find one that has
-	 * perf_event subsystem
-	 */
-	while (fscanf(fp, "%*s %"STR(PATH_MAX)"s %"STR(PATH_MAX)"s %"
-				STR(PATH_MAX)"s %*d %*d\n",
-				mountpoint, type, tokens) == 3) {
+	scnprintf(path, PATH_MAX, "%s/%s", mnt, name);
 
-		if (!strcmp(type, "cgroup")) {
+	fd = open(path, O_RDONLY);
+	if (fd == -1)
+		fprintf(stderr, "no access to cgroup %s\n", path);
 
-			token = strtok_r(tokens, ",", &saved_ptr);
+	return fd;
+}
 
-			while (token != NULL) {
-				if (!strcmp(token, "perf_event")) {
-					found = 1;
-					break;
-				}
-				token = strtok_r(NULL, ",", &saved_ptr);
-			}
-		}
-		if (found)
-			break;
-	}
-	fclose(fp);
-	if (!found)
-		return -1;
+#ifdef HAVE_FILE_HANDLE
+static u64 __read_cgroup_id(const char *path)
+{
+	struct {
+		struct file_handle fh;
+		uint64_t cgroup_id;
+	} handle;
+	int mount_id;
 
-	if (strlen(mountpoint) < maxlen) {
-		strcpy(buf, mountpoint);
-		return 0;
-	}
-	return -1;
+	handle.fh.handle_bytes = sizeof(handle.cgroup_id);
+	if (name_to_handle_at(AT_FDCWD, path, &handle.fh, &mount_id, 0) < 0)
+		return -1ULL;
+
+	return handle.cgroup_id;
 }
 
-static int open_cgroup(char *name)
+int read_cgroup_id(struct cgroup *cgrp)
 {
 	char path[PATH_MAX + 1];
 	char mnt[PATH_MAX + 1];
-	int fd;
 
-
-	if (cgroupfs_find_mountpoint(mnt, PATH_MAX + 1))
+	if (cgroupfs_find_mountpoint(mnt, PATH_MAX + 1, "perf_event"))
 		return -1;
 
-	snprintf(path, PATH_MAX, "%s/%s", mnt, name);
+	scnprintf(path, PATH_MAX, "%s/%s", mnt, cgrp->name);
 
-	fd = open(path, O_RDONLY);
-	if (fd == -1)
-		fprintf(stderr, "no access to cgroup %s\n", path);
+	cgrp->id = __read_cgroup_id(path);
+	return 0;
+}
+#else
+static inline u64 __read_cgroup_id(const char *path __maybe_unused) { return -1ULL; }
+#endif  /* HAVE_FILE_HANDLE */
 
-	return fd;
+#ifndef CGROUP2_SUPER_MAGIC
+#define CGROUP2_SUPER_MAGIC  0x63677270
+#endif
+
+int cgroup_is_v2(const char *subsys)
+{
+	char mnt[PATH_MAX + 1];
+	struct statfs stbuf;
+
+	if (cgroupfs_find_mountpoint(mnt, PATH_MAX + 1, subsys))
+		return -1;
+
+	if (statfs(mnt, &stbuf) < 0)
+		return -1;
+
+	return (stbuf.f_type == CGROUP2_SUPER_MAGIC);
 }
 
-static int add_cgroup(struct perf_evlist *evlist, char *str)
+static struct cgroup *evlist__find_cgroup(struct evlist *evlist, const char *str)
 {
-	struct perf_evsel *counter;
-	struct cgroup_sel *cgrp = NULL;
-	int n;
+	struct evsel *counter;
 	/*
 	 * check if cgrp is already defined, if so we reuse it
 	 */
-	list_for_each_entry(counter, &evlist->entries, node) {
-		cgrp = counter->cgrp;
-		if (!cgrp)
+	evlist__for_each_entry(evlist, counter) {
+		if (!counter->cgrp)
 			continue;
-		if (!strcmp(cgrp->name, str))
-			break;
-
-		cgrp = NULL;
+		if (!strcmp(counter->cgrp->name, str))
+			return cgroup__get(counter->cgrp);
 	}
 
-	if (!cgrp) {
-		cgrp = zalloc(sizeof(*cgrp));
-		if (!cgrp)
-			return -1;
+	return NULL;
+}
+
+struct cgroup *cgroup__new(const char *name, bool do_open)
+{
+	struct cgroup *cgroup = zalloc(sizeof(*cgroup));
+
+	if (cgroup != NULL) {
+		refcount_set(&cgroup->refcnt, 1);
 
-		cgrp->name = str;
+		cgroup->name = strdup(name);
+		if (!cgroup->name)
+			goto out_err;
 
-		cgrp->fd = open_cgroup(str);
-		if (cgrp->fd == -1) {
-			free(cgrp);
-			return -1;
+		if (do_open) {
+			cgroup->fd = open_cgroup(name);
+			if (cgroup->fd == -1)
+				goto out_free_name;
+		} else {
+			cgroup->fd = -1;
 		}
 	}
 
+	return cgroup;
+
+out_free_name:
+	zfree(&cgroup->name);
+out_err:
+	free(cgroup);
+	return NULL;
+}
+
+struct cgroup *evlist__findnew_cgroup(struct evlist *evlist, const char *name)
+{
+	struct cgroup *cgroup = evlist__find_cgroup(evlist, name);
+
+	return cgroup ?: cgroup__new(name, true);
+}
+
+static int add_cgroup(struct evlist *evlist, const char *str)
+{
+	struct evsel *counter;
+	struct cgroup *cgrp = evlist__findnew_cgroup(evlist, str);
+	int n;
+
+	if (!cgrp)
+		return -1;
 	/*
 	 * find corresponding event
 	 * if add cgroup N, then need to find event N
 	 */
 	n = 0;
-	list_for_each_entry(counter, &evlist->entries, node) {
+	evlist__for_each_entry(evlist, counter) {
 		if (n == nr_cgroups)
 			goto found;
 		n++;
 	}
-	if (cgrp->refcnt == 0)
-		free(cgrp);
 
+	cgroup__put(cgrp);
 	return -1;
 found:
-	cgrp->refcnt++;
 	counter->cgrp = cgrp;
 	return 0;
 }
 
-void close_cgroup(struct cgroup_sel *cgrp)
+static void cgroup__delete(struct cgroup *cgroup)
 {
-	if (!cgrp)
-		return;
+	if (cgroup->fd >= 0)
+		close(cgroup->fd);
+	zfree(&cgroup->name);
+	free(cgroup);
+}
 
-	/* XXX: not reentrant */
-	if (--cgrp->refcnt == 0) {
-		close(cgrp->fd);
-		free(cgrp->name);
-		free(cgrp);
+void cgroup__put(struct cgroup *cgrp)
+{
+	if (cgrp && refcount_dec_and_test(&cgrp->refcnt)) {
+		cgroup__delete(cgrp);
+	}
+}
+
+struct cgroup *cgroup__get(struct cgroup *cgroup)
+{
+       if (cgroup)
+		refcount_inc(&cgroup->refcnt);
+       return cgroup;
+}
+
+static void evsel__set_default_cgroup(struct evsel *evsel, struct cgroup *cgroup)
+{
+	if (evsel->cgrp == NULL)
+		evsel->cgrp = cgroup__get(cgroup);
+}
+
+void evlist__set_default_cgroup(struct evlist *evlist, struct cgroup *cgroup)
+{
+	struct evsel *evsel;
+
+	evlist__for_each_entry(evlist, evsel)
+		evsel__set_default_cgroup(evsel, cgroup);
+}
+
+/* helper function for ftw() in match_cgroups and list_cgroups */
+static int add_cgroup_name(const char *fpath, const struct stat *sb __maybe_unused,
+			   int typeflag, struct FTW *ftwbuf __maybe_unused)
+{
+	struct cgroup_name *cn;
+
+	if (typeflag != FTW_D)
+		return 0;
+
+	cn = malloc(sizeof(*cn) + strlen(fpath) + 1);
+	if (cn == NULL)
+		return -1;
+
+	cn->used = false;
+	strcpy(cn->name, fpath);
+
+	list_add_tail(&cn->list, &cgroup_list);
+	return 0;
+}
+
+static int check_and_add_cgroup_name(const char *fpath)
+{
+	struct cgroup_name *cn;
+
+	list_for_each_entry(cn, &cgroup_list, list) {
+		if (!strcmp(cn->name, fpath))
+			return 0;
+	}
+
+	/* pretend if it's added by ftw() */
+	return add_cgroup_name(fpath, NULL, FTW_D, NULL);
+}
+
+static void release_cgroup_list(void)
+{
+	struct cgroup_name *cn;
+
+	while (!list_empty(&cgroup_list)) {
+		cn = list_first_entry(&cgroup_list, struct cgroup_name, list);
+		list_del(&cn->list);
+		free(cn);
 	}
 }
 
-int parse_cgroups(const struct option *opt __maybe_unused, const char *str,
+/* collect given cgroups only */
+static int list_cgroups(const char *str)
+{
+	const char *p, *e, *eos = str + strlen(str);
+	struct cgroup_name *cn;
+	char *s;
+
+	/* use given name as is when no regex is given */
+	for (;;) {
+		p = strchr(str, ',');
+		e = p ? p : eos;
+
+		if (e - str) {
+			int ret;
+
+			s = strndup(str, e - str);
+			if (!s)
+				return -1;
+
+			ret = check_and_add_cgroup_name(s);
+			free(s);
+			if (ret < 0)
+				return -1;
+		} else {
+			if (check_and_add_cgroup_name("/") < 0)
+				return -1;
+		}
+
+		if (!p)
+			break;
+		str = p+1;
+	}
+
+	/* these groups will be used */
+	list_for_each_entry(cn, &cgroup_list, list)
+		cn->used = true;
+
+	return 0;
+}
+
+/* collect all cgroups first and then match with the pattern */
+static int match_cgroups(const char *str)
+{
+	char mnt[PATH_MAX];
+	const char *p, *e, *eos = str + strlen(str);
+	struct cgroup_name *cn;
+	regex_t reg;
+	int prefix_len;
+	char *s;
+
+	if (cgroupfs_find_mountpoint(mnt, sizeof(mnt), "perf_event"))
+		return -1;
+
+	/* cgroup_name will have a full path, skip the root directory */
+	prefix_len = strlen(mnt);
+
+	/* collect all cgroups in the cgroup_list */
+	if (nftw(mnt, add_cgroup_name, 20, 0) < 0)
+		return -1;
+
+	for (;;) {
+		p = strchr(str, ',');
+		e = p ? p : eos;
+
+		/* allow empty cgroups, i.e., skip */
+		if (e - str) {
+			/* termination added */
+			s = strndup(str, e - str);
+			if (!s)
+				return -1;
+			if (regcomp(&reg, s, REG_NOSUB)) {
+				free(s);
+				return -1;
+			}
+
+			/* check cgroup name with the pattern */
+			list_for_each_entry(cn, &cgroup_list, list) {
+				char *name = cn->name + prefix_len;
+
+				if (name[0] == '/' && name[1])
+					name++;
+				if (!regexec(&reg, name, 0, NULL, 0))
+					cn->used = true;
+			}
+			regfree(&reg);
+			free(s);
+		} else {
+			/* first entry to root cgroup */
+			cn = list_first_entry(&cgroup_list, struct cgroup_name,
+					      list);
+			cn->used = true;
+		}
+
+		if (!p)
+			break;
+		str = p+1;
+	}
+	return prefix_len;
+}
+
+int parse_cgroups(const struct option *opt, const char *str,
 		  int unset __maybe_unused)
 {
-	struct perf_evlist *evlist = *(struct perf_evlist **)opt->value;
+	struct evlist *evlist = *(struct evlist **)opt->value;
+	struct evsel *counter;
+	struct cgroup *cgrp = NULL;
 	const char *p, *e, *eos = str + strlen(str);
 	char *s;
-	int ret;
+	int ret, i;
 
-	if (list_empty(&evlist->entries)) {
+	if (list_empty(&evlist->core.entries)) {
 		fprintf(stderr, "must define events before cgroups\n");
 		return -1;
 	}
@@ -162,10 +383,9 @@ int parse_cgroups(const struct option *opt __maybe_unused, const char *str,
 			if (!s)
 				return -1;
 			ret = add_cgroup(evlist, s);
-			if (ret) {
-				free(s);
+			free(s);
+			if (ret)
 				return -1;
-			}
 		}
 		/* nr_cgroups is increased een for empty cgroups */
 		nr_cgroups++;
@@ -173,5 +393,238 @@ int parse_cgroups(const struct option *opt __maybe_unused, const char *str,
 			break;
 		str = p+1;
 	}
+	/* for the case one cgroup combine to multiple events */
+	i = 0;
+	if (nr_cgroups == 1) {
+		evlist__for_each_entry(evlist, counter) {
+			if (i == 0)
+				cgrp = counter->cgrp;
+			else {
+				counter->cgrp = cgrp;
+				refcount_inc(&cgrp->refcnt);
+			}
+			i++;
+		}
+	}
 	return 0;
 }
+
+static bool has_pattern_string(const char *str)
+{
+	return !!strpbrk(str, "{}[]()|*+?^$");
+}
+
+int evlist__expand_cgroup(struct evlist *evlist, const char *str, bool open_cgroup)
+{
+	struct evlist *orig_list, *tmp_list;
+	struct evsel *pos, *evsel, *leader;
+	struct rblist orig_metric_events;
+	struct cgroup *cgrp = NULL;
+	struct cgroup_name *cn;
+	int ret = -1;
+	int prefix_len;
+
+	if (evlist->core.nr_entries == 0) {
+		fprintf(stderr, "must define events before cgroups\n");
+		return -EINVAL;
+	}
+
+	orig_list = evlist__new();
+	tmp_list = evlist__new();
+	if (orig_list == NULL || tmp_list == NULL) {
+		fprintf(stderr, "memory allocation failed\n");
+		return -ENOMEM;
+	}
+
+	/* save original events and init evlist */
+	evlist__splice_list_tail(orig_list, &evlist->core.entries);
+	evlist->core.nr_entries = 0;
+
+	orig_metric_events = evlist->metric_events;
+	metricgroup__rblist_init(&evlist->metric_events);
+
+	if (has_pattern_string(str))
+		prefix_len = match_cgroups(str);
+	else
+		prefix_len = list_cgroups(str);
+
+	if (prefix_len < 0)
+		goto out_err;
+
+	list_for_each_entry(cn, &cgroup_list, list) {
+		char *name;
+
+		if (!cn->used)
+			continue;
+
+		/* cgroup_name might have a full path, skip the prefix */
+		name = cn->name + prefix_len;
+		if (name[0] == '/' && name[1])
+			name++;
+
+		/* the cgroup can go away in the meantime */
+		cgrp = cgroup__new(name, open_cgroup);
+		if (cgrp == NULL)
+			continue;
+
+		leader = NULL;
+		evlist__for_each_entry(orig_list, pos) {
+			evsel = evsel__clone(/*dest=*/NULL, pos);
+			if (evsel == NULL)
+				goto out_err;
+
+			cgroup__put(evsel->cgrp);
+			evsel->cgrp = cgroup__get(cgrp);
+
+			if (evsel__is_group_leader(pos))
+				leader = evsel;
+			evsel__set_leader(evsel, leader);
+
+			evlist__add(tmp_list, evsel);
+		}
+		/* cgroup__new() has a refcount, release it here */
+		cgroup__put(cgrp);
+		nr_cgroups++;
+
+		if (metricgroup__copy_metric_events(tmp_list, cgrp,
+						    &evlist->metric_events,
+						    &orig_metric_events) < 0)
+			goto out_err;
+
+		evlist__splice_list_tail(evlist, &tmp_list->core.entries);
+		tmp_list->core.nr_entries = 0;
+	}
+
+	if (list_empty(&evlist->core.entries)) {
+		fprintf(stderr, "no cgroup matched: %s\n", str);
+		goto out_err;
+	}
+
+	ret = 0;
+	cgrp_event_expanded = true;
+
+out_err:
+	evlist__delete(orig_list);
+	evlist__delete(tmp_list);
+	metricgroup__rblist_exit(&orig_metric_events);
+	release_cgroup_list();
+
+	return ret;
+}
+
+static struct cgroup *__cgroup__findnew(struct rb_root *root, uint64_t id,
+					bool create, const char *path)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct cgroup *cgrp;
+
+	while (*p != NULL) {
+		parent = *p;
+		cgrp = rb_entry(parent, struct cgroup, node);
+
+		if (cgrp->id == id)
+			return cgrp;
+
+		if (cgrp->id < id)
+			p = &(*p)->rb_left;
+		else
+			p = &(*p)->rb_right;
+	}
+
+	if (!create)
+		return NULL;
+
+	cgrp = malloc(sizeof(*cgrp));
+	if (cgrp == NULL)
+		return NULL;
+
+	cgrp->name = strdup(path);
+	if (cgrp->name == NULL) {
+		free(cgrp);
+		return NULL;
+	}
+
+	cgrp->fd = -1;
+	cgrp->id = id;
+	refcount_set(&cgrp->refcnt, 1);
+
+	rb_link_node(&cgrp->node, parent, p);
+	rb_insert_color(&cgrp->node, root);
+
+	return cgrp;
+}
+
+struct cgroup *cgroup__findnew(struct perf_env *env, uint64_t id,
+			       const char *path)
+{
+	struct cgroup *cgrp;
+
+	down_write(&env->cgroups.lock);
+	cgrp = __cgroup__findnew(&env->cgroups.tree, id, true, path);
+	up_write(&env->cgroups.lock);
+	return cgrp;
+}
+
+struct cgroup *__cgroup__find(struct rb_root *root, uint64_t id)
+{
+	return __cgroup__findnew(root, id, /*create=*/false, /*path=*/NULL);
+}
+
+struct cgroup *cgroup__find(struct perf_env *env, uint64_t id)
+{
+	struct cgroup *cgrp;
+
+	down_read(&env->cgroups.lock);
+	cgrp = __cgroup__findnew(&env->cgroups.tree, id, false, NULL);
+	up_read(&env->cgroups.lock);
+	return cgrp;
+}
+
+void perf_env__purge_cgroups(struct perf_env *env)
+{
+	struct rb_node *node;
+	struct cgroup *cgrp;
+
+	down_write(&env->cgroups.lock);
+	while (!RB_EMPTY_ROOT(&env->cgroups.tree)) {
+		node = rb_first(&env->cgroups.tree);
+		cgrp = rb_entry(node, struct cgroup, node);
+
+		rb_erase(node, &env->cgroups.tree);
+		cgroup__put(cgrp);
+	}
+	up_write(&env->cgroups.lock);
+}
+
+void read_all_cgroups(struct rb_root *root)
+{
+	char mnt[PATH_MAX];
+	struct cgroup_name *cn;
+	int prefix_len;
+
+	if (cgroupfs_find_mountpoint(mnt, sizeof(mnt), "perf_event"))
+		return;
+
+	/* cgroup_name will have a full path, skip the root directory */
+	prefix_len = strlen(mnt);
+
+	/* collect all cgroups in the cgroup_list */
+	if (nftw(mnt, add_cgroup_name, 20, 0) < 0)
+		return;
+
+	list_for_each_entry(cn, &cgroup_list, list) {
+		const char *name;
+		u64 cgrp_id;
+
+		/* cgroup_name might have a full path, skip the prefix */
+		name = cn->name + prefix_len;
+		if (name[0] == '\0')
+			name = "/";
+
+		cgrp_id = __read_cgroup_id(cn->name);
+		__cgroup__findnew(root, cgrp_id, /*create=*/true, name);
+	}
+
+	release_cgroup_list();
+}