Revert "cgroup: Fix threadgroup_rwsem <-> cpus_read_lock() deadlock"

hulk inclusion
category: bugfix
bugzilla: https://gitee.com/openeuler/kernel/issues/I6TI3Y
CVE: NA

--------------------------------

This reverts commit 4924308a.
Signed-off-by: NCai Xinchen <caixinchen1@huawei.com>
Reviewed-by: NWang Weiyang <wangweiyang2@huawei.com>
Signed-off-by: NYongqiang Liu <liuyongqiang13@huawei.com>

kernel/cgroup/cgroup.c

@@ -56,7 +56,6 @@
 #include <linux/file.h>
 #include <linux/sched/cputime.h>
 #include <net/sock.h>
-#include <linux/cpu.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/cgroup.h>
@@ -2213,45 +2212,6 @@ int task_cgroup_path(struct task_struct *task, char *buf, size_t buflen)
 }
 EXPORT_SYMBOL_GPL(task_cgroup_path);
 
-/**
- * cgroup_attach_lock - Lock for ->attach()
- * @lock_threadgroup: whether to down_write cgroup_threadgroup_rwsem
- *
- * cgroup migration sometimes needs to stabilize threadgroups against forks and
- * exits by write-locking cgroup_threadgroup_rwsem. However, some ->attach()
- * implementations (e.g. cpuset), also need to disable CPU hotplug.
- * Unfortunately, letting ->attach() operations acquire cpus_read_lock() can
- * lead to deadlocks.
- *
- * Bringing up a CPU may involve creating and destroying tasks which requires
- * read-locking threadgroup_rwsem, so threadgroup_rwsem nests inside
- * cpus_read_lock(). If we call an ->attach() which acquires the cpus lock while
- * write-locking threadgroup_rwsem, the locking order is reversed and we end up
- * waiting for an on-going CPU hotplug operation which in turn is waiting for
- * the threadgroup_rwsem to be released to create new tasks. For more details:
- *
- *   http://lkml.kernel.org/r/20220711174629.uehfmqegcwn2lqzu@wubuntu
- *
- * Resolve the situation by always acquiring cpus_read_lock() before optionally
- * write-locking cgroup_threadgroup_rwsem. This allows ->attach() to assume that
- * CPU hotplug is disabled on entry.
- */
-static void cgroup_attach_lock(void)
-{
-	cpus_read_lock();
-	percpu_down_write(&cgroup_threadgroup_rwsem);
-}
-
-/**
- * cgroup_attach_unlock - Undo cgroup_attach_lock()
- * @lock_threadgroup: whether to up_write cgroup_threadgroup_rwsem
- */
-static void cgroup_attach_unlock(void)
-{
-	percpu_up_write(&cgroup_threadgroup_rwsem);
-	cpus_read_unlock();
-}
-
 /**
  * cgroup_migrate_add_task - add a migration target task to a migration context
  * @task: target task
@@ -2731,7 +2691,7 @@ struct task_struct *cgroup_procs_write_start(char *buf, bool threadgroup)
 	if (kstrtoint(strstrip(buf), 0, &pid) || pid < 0)
 		return ERR_PTR(-EINVAL);
 
-	cgroup_attach_lock();
+	percpu_down_write(&cgroup_threadgroup_rwsem);
 
 	rcu_read_lock();
 	if (pid) {
@@ -2762,7 +2722,7 @@ struct task_struct *cgroup_procs_write_start(char *buf, bool threadgroup)
 	goto out_unlock_rcu;
 
 out_unlock_threadgroup:
-	cgroup_attach_unlock();
+	percpu_up_write(&cgroup_threadgroup_rwsem);
 out_unlock_rcu:
 	rcu_read_unlock();
 	return tsk;
@@ -2777,7 +2737,7 @@ void cgroup_procs_write_finish(struct task_struct *task)
 	/* release reference from cgroup_procs_write_start() */
 	put_task_struct(task);
 
-	cgroup_attach_unlock();
+	percpu_up_write(&cgroup_threadgroup_rwsem);
 	for_each_subsys(ss, ssid)
 		if (ss->post_attach)
 			ss->post_attach();
@@ -2858,7 +2818,7 @@ static int cgroup_update_dfl_csses(struct cgroup *cgrp)
 
 	lockdep_assert_held(&cgroup_mutex);
 
-	cgroup_attach_lock();
+	percpu_down_write(&cgroup_threadgroup_rwsem);
 
 	/* look up all csses currently attached to @cgrp's subtree */
 	spin_lock_irq(&css_set_lock);
@@ -2888,7 +2848,7 @@ static int cgroup_update_dfl_csses(struct cgroup *cgrp)
 	ret = cgroup_migrate_execute(&mgctx);
 out_finish:
 	cgroup_migrate_finish(&mgctx);
-	cgroup_attach_unlock();
+	percpu_up_write(&cgroup_threadgroup_rwsem);
 	return ret;
 }
 

kernel/cgroup/cpuset.c

@@ -1612,7 +1612,11 @@ static void cpuset_attach(struct cgroup_taskset *tset)
 	cgroup_taskset_first(tset, &css);
 	cs = css_cs(css);
 
-	lockdep_assert_cpus_held();     /* see cgroup_attach_lock() */
+	/*
+	 * It should hold cpus lock because a cpu offline event can
+	 * cause set_cpus_allowed_ptr() failed.
+	 */
+	get_online_cpus();
 	mutex_lock(&cpuset_mutex);
 
 	/* prepare for attach */
@@ -1675,6 +1679,7 @@ static void cpuset_attach(struct cgroup_taskset *tset)
 		wake_up(&cpuset_attach_wq);
 
 	mutex_unlock(&cpuset_mutex);
+	put_online_cpus();
 }
 
 /* The various types of files and directories in a cpuset file system */