Merge branch 'for-3.9' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup changes from Tejun Heo:
 "Nothing too drastic.

   - Removal of synchronize_rcu() from userland visible paths.

   - Various fixes and cleanups from Li.

   - cgroup_rightmost_descendant() added which will be used by cpuset
     changes (it will be a separate pull request)."

* 'for-3.9' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup: fail if monitored file and event_control are in different cgroup
  cgroup: fix cgroup_rmdir() vs close(eventfd) race
  cpuset: fix cpuset_print_task_mems_allowed() vs rename() race
  cgroup: fix exit() vs rmdir() race
  cgroup: remove bogus comments in cgroup_diput()
  cgroup: remove synchronize_rcu() from cgroup_diput()
  cgroup: remove duplicate RCU free on struct cgroup
  sched: remove redundant NULL cgroup check in task_group_path()
  sched: split out css_online/css_offline from tg creation/destruction
  cgroup: initialize cgrp->dentry before css_alloc()
  cgroup: remove a NULL check in cgroup_exit()
  cgroup: fix bogus kernel warnings when cgroup_create() failed
  cgroup: remove synchronize_rcu() from rebind_subsystems()
  cgroup: remove synchronize_rcu() from cgroup_attach_{task|proc}()
  cgroup: use new hashtable implementation
  cgroups: fix cgroup_event_listener error handling
  cgroups: move cgroup_event_listener.c to tools/cgroup
  cgroup: implement cgroup_rightmost_descendant()
  cgroup: remove unused dummy cgroup_fork_callbacks()

1 files changed, 168 insertions, 120 deletions

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 4855892798fd..b5c64327e712 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -52,7 +52,7 @@
 #include <linux/module.h>
 #include <linux/delayacct.h>
 #include <linux/cgroupstats.h>
-#include <linux/hash.h>
+#include <linux/hashtable.h>
 #include <linux/namei.h>
 #include <linux/pid_namespace.h>
 #include <linux/idr.h>
@@ -376,22 +376,18 @@ static int css_set_count;
  * account cgroups in empty hierarchies.
  */
 #define CSS_SET_HASH_BITS	7
-#define CSS_SET_TABLE_SIZE	(1 << CSS_SET_HASH_BITS)
-static struct hlist_head css_set_table[CSS_SET_TABLE_SIZE];
+static DEFINE_HASHTABLE(css_set_table, CSS_SET_HASH_BITS);
 
-static struct hlist_head *css_set_hash(struct cgroup_subsys_state *css[])
+static unsigned long css_set_hash(struct cgroup_subsys_state *css[])
 {
 	int i;
-	int index;
-	unsigned long tmp = 0UL;
+	unsigned long key = 0UL;
 
 	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++)
-		tmp += (unsigned long)css[i];
-	tmp = (tmp >> 16) ^ tmp;
+		key += (unsigned long)css[i];
+	key = (key >> 16) ^ key;
 
-	index = hash_long(tmp, CSS_SET_HASH_BITS);
-
-	return &css_set_table[index];
+	return key;
 }
 
 /* We don't maintain the lists running through each css_set to its
@@ -418,7 +414,7 @@ static void __put_css_set(struct css_set *cg, int taskexit)
 	}
 
 	/* This css_set is dead. unlink it and release cgroup refcounts */
-	hlist_del(&cg->hlist);
+	hash_del(&cg->hlist);
 	css_set_count--;
 
 	list_for_each_entry_safe(link, saved_link, &cg->cg_links,
@@ -426,12 +422,20 @@ static void __put_css_set(struct css_set *cg, int taskexit)
 		struct cgroup *cgrp = link->cgrp;
 		list_del(&link->cg_link_list);
 		list_del(&link->cgrp_link_list);
+
+		/*
+		 * We may not be holding cgroup_mutex, and if cgrp->count is
+		 * dropped to 0 the cgroup can be destroyed at any time, hence
+		 * rcu_read_lock is used to keep it alive.
+		 */
+		rcu_read_lock();
 		if (atomic_dec_and_test(&cgrp->count) &&
 		    notify_on_release(cgrp)) {
 			if (taskexit)
 				set_bit(CGRP_RELEASABLE, &cgrp->flags);
 			check_for_release(cgrp);
 		}
+		rcu_read_unlock();
 
 		kfree(link);
 	}
@@ -550,9 +554,9 @@ static struct css_set *find_existing_css_set(
 {
 	int i;
 	struct cgroupfs_root *root = cgrp->root;
-	struct hlist_head *hhead;
 	struct hlist_node *node;
 	struct css_set *cg;
+	unsigned long key;
 
 	/*
 	 * Build the set of subsystem state objects that we want to see in the
@@ -572,8 +576,8 @@ static struct css_set *find_existing_css_set(
 		}
 	}
 
-	hhead = css_set_hash(template);
-	hlist_for_each_entry(cg, node, hhead, hlist) {
+	key = css_set_hash(template);
+	hash_for_each_possible(css_set_table, cg, node, hlist, key) {
 		if (!compare_css_sets(cg, oldcg, cgrp, template))
 			continue;
 
@@ -657,8 +661,8 @@ static struct css_set *find_css_set(
 
 	struct list_head tmp_cg_links;
 
-	struct hlist_head *hhead;
 	struct cg_cgroup_link *link;
+	unsigned long key;
 
 	/* First see if we already have a cgroup group that matches
 	 * the desired set */
@@ -704,8 +708,8 @@ static struct css_set *find_css_set(
 	css_set_count++;
 
 	/* Add this cgroup group to the hash table */
-	hhead = css_set_hash(res->subsys);
-	hlist_add_head(&res->hlist, hhead);
+	key = css_set_hash(res->subsys);
+	hash_add(css_set_table, &res->hlist, key);
 
 	write_unlock(&css_set_lock);
 
@@ -856,47 +860,54 @@ static struct inode *cgroup_new_inode(umode_t mode, struct super_block *sb)
 	return inode;
 }
 
-static void cgroup_diput(struct dentry *dentry, struct inode *inode)
+static void cgroup_free_fn(struct work_struct *work)
 {
-	/* is dentry a directory ? if so, kfree() associated cgroup */
-	if (S_ISDIR(inode->i_mode)) {
-		struct cgroup *cgrp = dentry->d_fsdata;
-		struct cgroup_subsys *ss;
-		BUG_ON(!(cgroup_is_removed(cgrp)));
-		/* It's possible for external users to be holding css
-		 * reference counts on a cgroup; css_put() needs to
-		 * be able to access the cgroup after decrementing
-		 * the reference count in order to know if it needs to
-		 * queue the cgroup to be handled by the release
-		 * agent */
-		synchronize_rcu();
+	struct cgroup *cgrp = container_of(work, struct cgroup, free_work);
+	struct cgroup_subsys *ss;
 
-		mutex_lock(&cgroup_mutex);
-		/*
-		 * Release the subsystem state objects.
-		 */
-		for_each_subsys(cgrp->root, ss)
-			ss->css_free(cgrp);
+	mutex_lock(&cgroup_mutex);
+	/*
+	 * Release the subsystem state objects.
+	 */
+	for_each_subsys(cgrp->root, ss)
+		ss->css_free(cgrp);
 
-		cgrp->root->number_of_cgroups--;
-		mutex_unlock(&cgroup_mutex);
+	cgrp->root->number_of_cgroups--;
+	mutex_unlock(&cgroup_mutex);
 
-		/*
-		 * Drop the active superblock reference that we took when we
-		 * created the cgroup
-		 */
-		deactivate_super(cgrp->root->sb);
+	/*
+	 * Drop the active superblock reference that we took when we
+	 * created the cgroup
+	 */
+	deactivate_super(cgrp->root->sb);
 
-		/*
-		 * if we're getting rid of the cgroup, refcount should ensure
-		 * that there are no pidlists left.
-		 */
-		BUG_ON(!list_empty(&cgrp->pidlists));
+	/*
+	 * if we're getting rid of the cgroup, refcount should ensure
+	 * that there are no pidlists left.
+	 */
+	BUG_ON(!list_empty(&cgrp->pidlists));
 
-		simple_xattrs_free(&cgrp->xattrs);
+	simple_xattrs_free(&cgrp->xattrs);
 
-		ida_simple_remove(&cgrp->root->cgroup_ida, cgrp->id);
-		kfree_rcu(cgrp, rcu_head);
+	ida_simple_remove(&cgrp->root->cgroup_ida, cgrp->id);
+	kfree(cgrp);
+}
+
+static void cgroup_free_rcu(struct rcu_head *head)
+{
+	struct cgroup *cgrp = container_of(head, struct cgroup, rcu_head);
+
+	schedule_work(&cgrp->free_work);
+}
+
+static void cgroup_diput(struct dentry *dentry, struct inode *inode)
+{
+	/* is dentry a directory ? if so, kfree() associated cgroup */
+	if (S_ISDIR(inode->i_mode)) {
+		struct cgroup *cgrp = dentry->d_fsdata;
+
+		BUG_ON(!(cgroup_is_removed(cgrp)));
+		call_rcu(&cgrp->rcu_head, cgroup_free_rcu);
 	} else {
 		struct cfent *cfe = __d_cfe(dentry);
 		struct cgroup *cgrp = dentry->d_parent->d_fsdata;
@@ -925,13 +936,17 @@ static void remove_dir(struct dentry *d)
 	dput(parent);
 }
 
-static int cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft)
+static void cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft)
 {
 	struct cfent *cfe;
 
 	lockdep_assert_held(&cgrp->dentry->d_inode->i_mutex);
 	lockdep_assert_held(&cgroup_mutex);
 
+	/*
+	 * If we're doing cleanup due to failure of cgroup_create(),
+	 * the corresponding @cfe may not exist.
+	 */
 	list_for_each_entry(cfe, &cgrp->files, node) {
 		struct dentry *d = cfe->dentry;
 
@@ -944,9 +959,8 @@ static int cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft)
 		list_del_init(&cfe->node);
 		dput(d);
 
-		return 0;
+		break;
 	}
-	return -ENOENT;
 }
 
 /**
@@ -1083,7 +1097,6 @@ static int rebind_subsystems(struct cgroupfs_root *root,
 		}
 	}
 	root->subsys_mask = root->actual_subsys_mask = final_subsys_mask;
-	synchronize_rcu();
 
 	return 0;
 }
@@ -1393,6 +1406,7 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp)
 	INIT_LIST_HEAD(&cgrp->allcg_node);
 	INIT_LIST_HEAD(&cgrp->release_list);
 	INIT_LIST_HEAD(&cgrp->pidlists);
+	INIT_WORK(&cgrp->free_work, cgroup_free_fn);
 	mutex_init(&cgrp->pidlist_mutex);
 	INIT_LIST_HEAD(&cgrp->event_list);
 	spin_lock_init(&cgrp->event_list_lock);
@@ -1597,6 +1611,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type,
 		struct cgroupfs_root *existing_root;
 		const struct cred *cred;
 		int i;
+		struct hlist_node *node;
+		struct css_set *cg;
 
 		BUG_ON(sb->s_root != NULL);
 
@@ -1650,14 +1666,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type,
 		/* Link the top cgroup in this hierarchy into all
 		 * the css_set objects */
 		write_lock(&css_set_lock);
-		for (i = 0; i < CSS_SET_TABLE_SIZE; i++) {
-			struct hlist_head *hhead = &css_set_table[i];
-			struct hlist_node *node;
-			struct css_set *cg;
-
-			hlist_for_each_entry(cg, node, hhead, hlist)
-				link_css_set(&tmp_cg_links, cg, root_cgrp);
-		}
+		hash_for_each(css_set_table, i, node, cg, hlist)
+			link_css_set(&tmp_cg_links, cg, root_cgrp);
 		write_unlock(&css_set_lock);
 
 		free_cg_links(&tmp_cg_links);
@@ -1773,7 +1783,7 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen)
 	rcu_lockdep_assert(rcu_read_lock_held() || cgroup_lock_is_held(),
 			   "cgroup_path() called without proper locking");
 
-	if (!dentry || cgrp == dummytop) {
+	if (cgrp == dummytop) {
 		/*
 		 * Inactive subsystems have no dentry for their root
 		 * cgroup
@@ -1982,7 +1992,6 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk)
 			ss->attach(cgrp, &tset);
 	}
 
-	synchronize_rcu();
 out:
 	if (retval) {
 		for_each_subsys(root, ss) {
@@ -2151,7 +2160,6 @@ static int cgroup_attach_proc(struct cgroup *cgrp, struct task_struct *leader)
 	/*
 	 * step 5: success! and cleanup
 	 */
-	synchronize_rcu();
 	retval = 0;
 out_put_css_set_refs:
 	if (retval) {
@@ -2769,14 +2777,14 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys,
 		if ((cft->flags & CFTYPE_ONLY_ON_ROOT) && cgrp->parent)
 			continue;
 
-		if (is_add)
+		if (is_add) {
 			err = cgroup_add_file(cgrp, subsys, cft);
-		else
-			err = cgroup_rm_file(cgrp, cft);
-		if (err) {
-			pr_warning("cgroup_addrm_files: failed to %s %s, err=%d\n",
-				   is_add ? "add" : "remove", cft->name, err);
+			if (err)
+				pr_warn("cgroup_addrm_files: failed to add %s, err=%d\n",
+					cft->name, err);
 			ret = err;
+		} else {
+			cgroup_rm_file(cgrp, cft);
 		}
 	}
 	return ret;
@@ -3017,6 +3025,32 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos,
 }
 EXPORT_SYMBOL_GPL(cgroup_next_descendant_pre);
 
+/**
+ * cgroup_rightmost_descendant - return the rightmost descendant of a cgroup
+ * @pos: cgroup of interest
+ *
+ * Return the rightmost descendant of @pos.  If there's no descendant,
+ * @pos is returned.  This can be used during pre-order traversal to skip
+ * subtree of @pos.
+ */
+struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos)
+{
+	struct cgroup *last, *tmp;
+
+	WARN_ON_ONCE(!rcu_read_lock_held());
+
+	do {
+		last = pos;
+		/* ->prev isn't RCU safe, walk ->next till the end */
+		pos = NULL;
+		list_for_each_entry_rcu(tmp, &last->children, sibling)
+			pos = tmp;
+	} while (pos);
+
+	return last;
+}
+EXPORT_SYMBOL_GPL(cgroup_rightmost_descendant);
+
 static struct cgroup *cgroup_leftmost_descendant(struct cgroup *pos)
 {
 	struct cgroup *last;
@@ -3752,8 +3786,13 @@ static void cgroup_event_remove(struct work_struct *work)
 			remove);
 	struct cgroup *cgrp = event->cgrp;
 
+	remove_wait_queue(event->wqh, &event->wait);
+
 	event->cft->unregister_event(cgrp, event->cft, event->eventfd);
 
+	/* Notify userspace the event is going away. */
+	eventfd_signal(event->eventfd, 1);
+
 	eventfd_ctx_put(event->eventfd);
 	kfree(event);
 	dput(cgrp->dentry);
@@ -3773,15 +3812,25 @@ static int cgroup_event_wake(wait_queue_t *wait, unsigned mode,
 	unsigned long flags = (unsigned long)key;
 
 	if (flags & POLLHUP) {
-		__remove_wait_queue(event->wqh, &event->wait);
-		spin_lock(&cgrp->event_list_lock);
-		list_del_init(&event->list);
-		spin_unlock(&cgrp->event_list_lock);
 		/*
-		 * We are in atomic context, but cgroup_event_remove() may
-		 * sleep, so we have to call it in workqueue.
+		 * If the event has been detached at cgroup removal, we
+		 * can simply return knowing the other side will cleanup
+		 * for us.
+		 *
+		 * We can't race against event freeing since the other
+		 * side will require wqh->lock via remove_wait_queue(),
+		 * which we hold.
 		 */
-		schedule_work(&event->remove);
+		spin_lock(&cgrp->event_list_lock);
+		if (!list_empty(&event->list)) {
+			list_del_init(&event->list);
+			/*
+			 * We are in atomic context, but cgroup_event_remove()
+			 * may sleep, so we have to call it in workqueue.
+			 */
+			schedule_work(&event->remove);
+		}
+		spin_unlock(&cgrp->event_list_lock);
 	}
 
 	return 0;
@@ -3807,6 +3856,7 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft,
 				      const char *buffer)
 {
 	struct cgroup_event *event = NULL;
+	struct cgroup *cgrp_cfile;
 	unsigned int efd, cfd;
 	struct file *efile = NULL;
 	struct file *cfile = NULL;
@@ -3862,6 +3912,16 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft,
 		goto fail;
 	}
 
+	/*
+	 * The file to be monitored must be in the same cgroup as
+	 * cgroup.event_control is.
+	 */
+	cgrp_cfile = __d_cgrp(cfile->f_dentry->d_parent);
+	if (cgrp_cfile != cgrp) {
+		ret = -EINVAL;
+		goto fail;
+	}
+
 	if (!event->cft->register_event || !event->cft->unregister_event) {
 		ret = -EINVAL;
 		goto fail;
@@ -4135,6 +4195,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 
 	init_cgroup_housekeeping(cgrp);
 
+	dentry->d_fsdata = cgrp;
+	cgrp->dentry = dentry;
+
 	cgrp->parent = parent;
 	cgrp->root = parent->root;
 	cgrp->top_cgroup = parent->top_cgroup;
@@ -4172,8 +4235,6 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry,
 	lockdep_assert_held(&dentry->d_inode->i_mutex);
 
 	/* allocation complete, commit to creation */
-	dentry->d_fsdata = cgrp;
-	cgrp->dentry = dentry;
 	list_add_tail(&cgrp->allcg_node, &root->allcg_list);
 	list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children);
 	root->number_of_cgroups++;
@@ -4340,20 +4401,14 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
 	/*
 	 * Unregister events and notify userspace.
 	 * Notify userspace about cgroup removing only after rmdir of cgroup
-	 * directory to avoid race between userspace and kernelspace. Use
-	 * a temporary list to avoid a deadlock with cgroup_event_wake(). Since
-	 * cgroup_event_wake() is called with the wait queue head locked,
-	 * remove_wait_queue() cannot be called while holding event_list_lock.
+	 * directory to avoid race between userspace and kernelspace.
 	 */
 	spin_lock(&cgrp->event_list_lock);
-	list_splice_init(&cgrp->event_list, &tmp_list);
-	spin_unlock(&cgrp->event_list_lock);
-	list_for_each_entry_safe(event, tmp, &tmp_list, list) {
+	list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) {
 		list_del_init(&event->list);
-		remove_wait_queue(event->wqh, &event->wait);
-		eventfd_signal(event->eventfd, 1);
 		schedule_work(&event->remove);
 	}
+	spin_unlock(&cgrp->event_list_lock);
 
 	return 0;
 }
@@ -4438,6 +4493,9 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss)
 {
 	struct cgroup_subsys_state *css;
 	int i, ret;
+	struct hlist_node *node, *tmp;
+	struct css_set *cg;
+	unsigned long key;
 
 	/* check name and function validity */
 	if (ss->name == NULL || strlen(ss->name) > MAX_CGROUP_TYPE_NAMELEN ||
@@ -4503,23 +4561,17 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss)
 	 * this is all done under the css_set_lock.
 	 */
 	write_lock(&css_set_lock);
-	for (i = 0; i < CSS_SET_TABLE_SIZE; i++) {
-		struct css_set *cg;
-		struct hlist_node *node, *tmp;
-		struct hlist_head *bucket = &css_set_table[i], *new_bucket;
-
-		hlist_for_each_entry_safe(cg, node, tmp, bucket, hlist) {
-			/* skip entries that we already rehashed */
-			if (cg->subsys[ss->subsys_id])
-				continue;
-			/* remove existing entry */
-			hlist_del(&cg->hlist);
-			/* set new value */
-			cg->subsys[ss->subsys_id] = css;
-			/* recompute hash and restore entry */
-			new_bucket = css_set_hash(cg->subsys);
-			hlist_add_head(&cg->hlist, new_bucket);
-		}
+	hash_for_each_safe(css_set_table, i, node, tmp, cg, hlist) {
+		/* skip entries that we already rehashed */
+		if (cg->subsys[ss->subsys_id])
+			continue;
+		/* remove existing entry */
+		hash_del(&cg->hlist);
+		/* set new value */
+		cg->subsys[ss->subsys_id] = css;
+		/* recompute hash and restore entry */
+		key = css_set_hash(cg->subsys);
+		hash_add(css_set_table, node, key);
 	}
 	write_unlock(&css_set_lock);
 
@@ -4551,7 +4603,6 @@ EXPORT_SYMBOL_GPL(cgroup_load_subsys);
 void cgroup_unload_subsys(struct cgroup_subsys *ss)
 {
 	struct cg_cgroup_link *link;
-	struct hlist_head *hhead;
 
 	BUG_ON(ss->module == NULL);
 
@@ -4585,11 +4636,12 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss)
 	write_lock(&css_set_lock);
 	list_for_each_entry(link, &dummytop->css_sets, cgrp_link_list) {
 		struct css_set *cg = link->cg;
+		unsigned long key;
 
-		hlist_del(&cg->hlist);
+		hash_del(&cg->hlist);
 		cg->subsys[ss->subsys_id] = NULL;
-		hhead = css_set_hash(cg->subsys);
-		hlist_add_head(&cg->hlist, hhead);
+		key = css_set_hash(cg->subsys);
+		hash_add(css_set_table, &cg->hlist, key);
 	}
 	write_unlock(&css_set_lock);
 
@@ -4631,9 +4683,6 @@ int __init cgroup_init_early(void)
 	list_add(&init_css_set_link.cg_link_list,
 		 &init_css_set.cg_links);
 
-	for (i = 0; i < CSS_SET_TABLE_SIZE; i++)
-		INIT_HLIST_HEAD(&css_set_table[i]);
-
 	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
 		struct cgroup_subsys *ss = subsys[i];
 
@@ -4667,7 +4716,7 @@ int __init cgroup_init(void)
 {
 	int err;
 	int i;
-	struct hlist_head *hhead;
+	unsigned long key;
 
 	err = bdi_init(&cgroup_backing_dev_info);
 	if (err)
@@ -4686,8 +4735,8 @@ int __init cgroup_init(void)
 	}
 
 	/* Add init_css_set to the hash table */
-	hhead = css_set_hash(init_css_set.subsys);
-	hlist_add_head(&init_css_set.hlist, hhead);
+	key = css_set_hash(init_css_set.subsys);
+	hash_add(css_set_table, &init_css_set.hlist, key);
 	BUG_ON(!init_root_id(&rootnode));
 
 	cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj);
@@ -4982,8 +5031,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks)
 	}
 	task_unlock(tsk);
 
-	if (cg)
-		put_css_set_taskexit(cg);
+	put_css_set_taskexit(cg);
 }
 
 /**