Currently cgroup_get_from_path() and cgroup_get_from_id() grab
cgroup_mutex before traversing the default hierarchy to find the
kernfs_node corresponding to the path/id and then extract the linked
cgroup. Since cgroup_mutex is still held, it is guaranteed that the
cgroup will be alive and the reference can be taken on it.
However similar guarantee can be provided without depending on the
cgroup_mutex and potentially reducing avenues of cgroup_mutex contentions.
The kernfs_node's priv pointer is RCU protected pointer and with just
rcu read lock we can grab the reference on the cgroup without
cgroup_mutex. So, remove cgroup_mutex from them.
Signed-off-by: Shakeel Butt <[email protected]>
---
kernel/cgroup/cgroup.c | 51 ++++++++++++++++++++++++------------------
1 file changed, 29 insertions(+), 22 deletions(-)
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 02e2ac83a94e..f8605caba7ed 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -5932,17 +5932,20 @@ struct cgroup *cgroup_get_from_id(u64 id)
struct kernfs_node *kn;
struct cgroup *cgrp = NULL;
- mutex_lock(&cgroup_mutex);
kn = kernfs_find_and_get_node_by_id(cgrp_dfl_root.kf_root, id);
if (!kn)
- goto out_unlock;
+ goto out;
+
+ rcu_read_lock();
- cgrp = kn->priv;
- if (cgroup_is_dead(cgrp) || !cgroup_tryget(cgrp))
+ cgrp = rcu_dereference(*(void __rcu __force **)&kn->priv);
+ if (cgrp && !cgroup_tryget(cgrp))
cgrp = NULL;
+
+ rcu_read_unlock();
+
kernfs_put(kn);
-out_unlock:
- mutex_unlock(&cgroup_mutex);
+out:
return cgrp;
}
EXPORT_SYMBOL_GPL(cgroup_get_from_id);
@@ -6495,30 +6498,34 @@ struct cgroup_subsys_state *css_from_id(int id, struct cgroup_subsys *ss)
*
* Find the cgroup at @path on the default hierarchy, increment its
* reference count and return it. Returns pointer to the found cgroup on
- * success, ERR_PTR(-ENOENT) if @path doesn't exist and ERR_PTR(-ENOTDIR)
- * if @path points to a non-directory.
+ * success, ERR_PTR(-ENOENT) if @path doesn't exist or if the cgroup has already
+ * been released and ERR_PTR(-ENOTDIR) if @path points to a non-directory.
*/
struct cgroup *cgroup_get_from_path(const char *path)
{
struct kernfs_node *kn;
- struct cgroup *cgrp;
-
- mutex_lock(&cgroup_mutex);
+ struct cgroup *cgrp = ERR_PTR(-ENOENT);
kn = kernfs_walk_and_get(cgrp_dfl_root.cgrp.kn, path);
- if (kn) {
- if (kernfs_type(kn) == KERNFS_DIR) {
- cgrp = kn->priv;
- cgroup_get_live(cgrp);
- } else {
- cgrp = ERR_PTR(-ENOTDIR);
- }
- kernfs_put(kn);
- } else {
- cgrp = ERR_PTR(-ENOENT);
+ if (!kn)
+ goto out;
+
+ if (kernfs_type(kn) != KERNFS_DIR) {
+ cgrp = ERR_PTR(-ENOTDIR);
+ goto out_kernfs;
}
- mutex_unlock(&cgroup_mutex);
+ rcu_read_lock();
+
+ cgrp = rcu_dereference(*(void __rcu __force **)&kn->priv);
+ if (!cgrp || !cgroup_tryget(cgrp))
+ cgrp = ERR_PTR(-ENOENT);
+
+ rcu_read_unlock();
+
+out_kernfs:
+ kernfs_put(kn);
+out:
return cgrp;
}
EXPORT_SYMBOL_GPL(cgroup_get_from_path);
--
2.33.0.1079.g6e70778dc9-goog
The function cgroupstats_build extracts cgroup from the kernfs_node's
priv pointer which is a RCU pointer. So, there is no need to grab
cgroup_mutex. Just get the reference on the cgroup before using and
remove the cgroup_mutex altogether.
Signed-off-by: Shakeel Butt <[email protected]>
---
kernel/cgroup/cgroup-v1.c | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c
index f6cc5f8484dc..fd14a60379c1 100644
--- a/kernel/cgroup/cgroup-v1.c
+++ b/kernel/cgroup/cgroup-v1.c
@@ -698,8 +698,6 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry)
kernfs_type(kn) != KERNFS_DIR)
return -EINVAL;
- mutex_lock(&cgroup_mutex);
-
/*
* We aren't being called from kernfs and there's no guarantee on
* @kn->priv's validity. For this and css_tryget_online_from_dir(),
@@ -707,9 +705,8 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry)
*/
rcu_read_lock();
cgrp = rcu_dereference(*(void __rcu __force **)&kn->priv);
- if (!cgrp || cgroup_is_dead(cgrp)) {
+ if (!cgrp || !cgroup_tryget(cgrp)) {
rcu_read_unlock();
- mutex_unlock(&cgroup_mutex);
return -ENOENT;
}
rcu_read_unlock();
@@ -737,7 +734,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry)
}
css_task_iter_end(&it);
- mutex_unlock(&cgroup_mutex);
+ cgroup_put(cgrp);
return 0;
}
--
2.33.0.1079.g6e70778dc9-goog