cpuset: replace cgroup_mutex locking with cpuset internal locking

		if (is_cpuset_online(((child_cs) = cgroup_cs((pos_cgrp)))))

		if (is_cpuset_online(((child_cs) = cgroup_cs((pos_cgrp)))))

/*

/*

 * There are two global mutexes guarding cpuset structures.  The first

 * There are two global mutexes guarding cpuset structures - cpuset_mutex

 * is the main control groups cgroup_mutex, accessed via

 * and callback_mutex.  The latter may nest inside the former.  We also

 * cgroup_lock()/cgroup_unlock().  The second is the cpuset-specific

 * require taking task_lock() when dereferencing a task's cpuset pointer.

 * callback_mutex, below. They can nest.  It is ok to first take

 * See "The task_lock() exception", at the end of this comment.

 * cgroup_mutex, then nest callback_mutex.  We also require taking

 *

 * task_lock() when dereferencing a task's cpuset pointer.  See "The

 * A task must hold both mutexes to modify cpusets.  If a task holds

 * task_lock() exception", at the end of this comment.

 * cpuset_mutex, then it blocks others wanting that mutex, ensuring that it

 *

 * is the only task able to also acquire callback_mutex and be able to

 * A task must hold both mutexes to modify cpusets.  If a task

 * modify cpusets.  It can perform various checks on the cpuset structure

 * holds cgroup_mutex, then it blocks others wanting that mutex,

 * first, knowing nothing will change.  It can also allocate memory while

 * ensuring that it is the only task able to also acquire callback_mutex

 * just holding cpuset_mutex.  While it is performing these checks, various

 * and be able to modify cpusets.  It can perform various checks on

 * callback routines can briefly acquire callback_mutex to query cpusets.

 * the cpuset structure first, knowing nothing will change.  It can

 * Once it is ready to make the changes, it takes callback_mutex, blocking

 * also allocate memory while just holding cgroup_mutex.  While it is

 * everyone else.

 * performing these checks, various callback routines can briefly

 * acquire callback_mutex to query cpusets.  Once it is ready to make

 * the changes, it takes callback_mutex, blocking everyone else.

 *

 *

 * Calls to the kernel memory allocator can not be made while holding

 * Calls to the kernel memory allocator can not be made while holding

 * callback_mutex, as that would risk double tripping on callback_mutex

 * callback_mutex, as that would risk double tripping on callback_mutex

 * guidelines for accessing subsystem state in kernel/cgroup.c

 * guidelines for accessing subsystem state in kernel/cgroup.c

 */

 */

static DEFINE_MUTEX(cpuset_mutex);

static DEFINE_MUTEX(callback_mutex);

static DEFINE_MUTEX(callback_mutex);

/*

/*

/*

/*

 * update task's spread flag if cpuset's page/slab spread flag is set

 * update task's spread flag if cpuset's page/slab spread flag is set

 *

 *

 * Called with callback_mutex/cgroup_mutex held

 * Called with callback_mutex/cpuset_mutex held

 */

 */

static void cpuset_update_task_spread_flag(struct cpuset *cs,

static void cpuset_update_task_spread_flag(struct cpuset *cs,

					struct task_struct *tsk)

					struct task_struct *tsk)

 *

 *

 * One cpuset is a subset of another if all its allowed CPUs and

 * One cpuset is a subset of another if all its allowed CPUs and

 * Memory Nodes are a subset of the other, and its exclusive flags

 * Memory Nodes are a subset of the other, and its exclusive flags

 * are only set if the other's are set.  Call holding cgroup_mutex.

 * are only set if the other's are set.  Call holding cpuset_mutex.

 */

 */

static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q)

static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q)

 * If we replaced the flag and mask values of the current cpuset

 * If we replaced the flag and mask values of the current cpuset

 * (cur) with those values in the trial cpuset (trial), would

 * (cur) with those values in the trial cpuset (trial), would

 * our various subset and exclusive rules still be valid?  Presumes

 * our various subset and exclusive rules still be valid?  Presumes

 * cgroup_mutex held.

 * cpuset_mutex held.

 *

 *

 * 'cur' is the address of an actual, in-use cpuset.  Operations

 * 'cur' is the address of an actual, in-use cpuset.  Operations

 * such as list traversal that depend on the actual address of the

 * such as list traversal that depend on the actual address of the

 * domains when operating in the severe memory shortage situations

 * domains when operating in the severe memory shortage situations

 * that could cause allocation failures below.

 * that could cause allocation failures below.

 *

 *

 * Must be called with cgroup_lock held.

 * Must be called with cpuset_mutex held.

 *

 *

 * The three key local variables below are:

 * The three key local variables below are:

 *    q  - a linked-list queue of cpuset pointers, used to implement a

 *    q  - a linked-list queue of cpuset pointers, used to implement a

 * 'cpus' is removed, then call this routine to rebuild the

 * 'cpus' is removed, then call this routine to rebuild the

 * scheduler's dynamic sched domains.

 * scheduler's dynamic sched domains.

 *

 *

 * Call with cgroup_mutex held.  Takes get_online_cpus().

 * Call with cpuset_mutex held.  Takes get_online_cpus().

 */

 */

static void rebuild_sched_domains_locked(void)

static void rebuild_sched_domains_locked(void)

{

{

	cpumask_var_t *doms;

	cpumask_var_t *doms;

	int ndoms;

	int ndoms;

	WARN_ON_ONCE(!cgroup_lock_is_held());

	lockdep_assert_held(&cpuset_mutex);

	get_online_cpus();

	get_online_cpus();

	/* Generate domain masks and attrs */

	/* Generate domain masks and attrs */

void rebuild_sched_domains(void)

void rebuild_sched_domains(void)

{

{

	cgroup_lock();

	mutex_lock(&cpuset_mutex);

	rebuild_sched_domains_locked();

	rebuild_sched_domains_locked();

	cgroup_unlock();

	mutex_unlock(&cpuset_mutex);

}

}

/**

/**

 * @tsk: task to test

 * @tsk: task to test

 * @scan: struct cgroup_scanner contained in its struct cpuset_hotplug_scanner

 * @scan: struct cgroup_scanner contained in its struct cpuset_hotplug_scanner

 *

 *

 * Call with cgroup_mutex held.  May take callback_mutex during call.

 * Call with cpuset_mutex held.  May take callback_mutex during call.

 * Called for each task in a cgroup by cgroup_scan_tasks().

 * Called for each task in a cgroup by cgroup_scan_tasks().

 * Return nonzero if this tasks's cpus_allowed mask should be changed (in other

 * Return nonzero if this tasks's cpus_allowed mask should be changed (in other

 * words, if its mask is not equal to its cpuset's mask).

 * words, if its mask is not equal to its cpuset's mask).

 * cpus_allowed mask needs to be changed.

 * cpus_allowed mask needs to be changed.

 *

 *

 * We don't need to re-check for the cgroup/cpuset membership, since we're

 * We don't need to re-check for the cgroup/cpuset membership, since we're

 * holding cgroup_lock() at this point.

 * holding cpuset_mutex at this point.

 */

 */

static void cpuset_change_cpumask(struct task_struct *tsk,

static void cpuset_change_cpumask(struct task_struct *tsk,

				  struct cgroup_scanner *scan)

				  struct cgroup_scanner *scan)

 * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed

 * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 *

 *

 * Called with cgroup_mutex held

 * Called with cpuset_mutex held

 *

 *

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * calling callback functions for each.

 * calling callback functions for each.

 *    Temporarilly set tasks mems_allowed to target nodes of migration,

 *    Temporarilly set tasks mems_allowed to target nodes of migration,

 *    so that the migration code can allocate pages on these nodes.

 *    so that the migration code can allocate pages on these nodes.

 *

 *

 *    Call holding cgroup_mutex, so current's cpuset won't change

 *    Call holding cpuset_mutex, so current's cpuset won't change

 *    during this call, as manage_mutex holds off any cpuset_attach()

 *    during this call, as manage_mutex holds off any cpuset_attach()

 *    calls.  Therefore we don't need to take task_lock around the

 *    calls.  Therefore we don't need to take task_lock around the

 *    call to guarantee_online_mems(), as we know no one is changing

 *    call to guarantee_online_mems(), as we know no one is changing

/*

/*

 * Update task's mems_allowed and rebind its mempolicy and vmas' mempolicy

 * Update task's mems_allowed and rebind its mempolicy and vmas' mempolicy

 * of it to cpuset's new mems_allowed, and migrate pages to new nodes if

 * of it to cpuset's new mems_allowed, and migrate pages to new nodes if

 * memory_migrate flag is set. Called with cgroup_mutex held.

 * memory_migrate flag is set. Called with cpuset_mutex held.

 */

 */

static void cpuset_change_nodemask(struct task_struct *p,

static void cpuset_change_nodemask(struct task_struct *p,

				   struct cgroup_scanner *scan)

				   struct cgroup_scanner *scan)

	struct cpuset *cs;

	struct cpuset *cs;

	int migrate;

	int migrate;

	const nodemask_t *oldmem = scan->data;

	const nodemask_t *oldmem = scan->data;

	static nodemask_t newmems;	/* protected by cgroup_mutex */

	static nodemask_t newmems;	/* protected by cpuset_mutex */

	cs = cgroup_cs(scan->cg);

	cs = cgroup_cs(scan->cg);

	guarantee_online_mems(cs, &newmems);

	guarantee_online_mems(cs, &newmems);

 * @oldmem: old mems_allowed of cpuset cs

 * @oldmem: old mems_allowed of cpuset cs

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 *

 *

 * Called with cgroup_mutex held

 * Called with cpuset_mutex held

 * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0

 * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0

 * if @heap != NULL.

 * if @heap != NULL.

 */

 */

	 * take while holding tasklist_lock.  Forks can happen - the

	 * take while holding tasklist_lock.  Forks can happen - the

	 * mpol_dup() cpuset_being_rebound check will catch such forks,

	 * mpol_dup() cpuset_being_rebound check will catch such forks,

	 * and rebind their vma mempolicies too.  Because we still hold

	 * and rebind their vma mempolicies too.  Because we still hold

	 * the global cgroup_mutex, we know that no other rebind effort

	 * the global cpuset_mutex, we know that no other rebind effort

	 * will be contending for the global variable cpuset_being_rebound.

	 * will be contending for the global variable cpuset_being_rebound.

	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()

	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()

	 * is idempotent.  Also migrate pages in each mm to new nodes.

	 * is idempotent.  Also migrate pages in each mm to new nodes.

 * mempolicies and if the cpuset is marked 'memory_migrate',

 * mempolicies and if the cpuset is marked 'memory_migrate',

 * migrate the tasks pages to the new memory.

 * migrate the tasks pages to the new memory.

 *

 *

 * Call with cgroup_mutex held.  May take callback_mutex during call.

 * Call with cpuset_mutex held.  May take callback_mutex during call.

 * Will take tasklist_lock, scan tasklist for tasks in cpuset cs,

 * Will take tasklist_lock, scan tasklist for tasks in cpuset cs,

 * lock each such tasks mm->mmap_sem, scan its vma's and rebind

 * lock each such tasks mm->mmap_sem, scan its vma's and rebind

 * their mempolicies to the cpusets new mems_allowed.

 * their mempolicies to the cpusets new mems_allowed.

 * Called by cgroup_scan_tasks() for each task in a cgroup.

 * Called by cgroup_scan_tasks() for each task in a cgroup.

 *

 *

 * We don't need to re-check for the cgroup/cpuset membership, since we're

 * We don't need to re-check for the cgroup/cpuset membership, since we're

 * holding cgroup_lock() at this point.

 * holding cpuset_mutex at this point.

 */

 */

static void cpuset_change_flag(struct task_struct *tsk,

static void cpuset_change_flag(struct task_struct *tsk,

				struct cgroup_scanner *scan)

				struct cgroup_scanner *scan)

 * @cs: the cpuset in which each task's spread flags needs to be changed

 * @cs: the cpuset in which each task's spread flags needs to be changed

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks()

 *

 *

 * Called with cgroup_mutex held

 * Called with cpuset_mutex held

 *

 *

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * calling callback functions for each.

 * calling callback functions for each.

 * cs:		the cpuset to update

 * cs:		the cpuset to update

 * turning_on: 	whether the flag is being set or cleared

 * turning_on: 	whether the flag is being set or cleared

 *

 *

 * Call with cgroup_mutex held.

 * Call with cpuset_mutex held.

 */

 */

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs,

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs,

	return val;

	return val;

}

}

/* Called by cgroups to determine if a cpuset is usable; cgroup_mutex held */

/* Called by cgroups to determine if a cpuset is usable; cpuset_mutex held */

static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)

static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)

{

{

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	struct task_struct *task;

	struct task_struct *task;

	int ret;

	int ret;

	mutex_lock(&cpuset_mutex);

	ret = -ENOSPC;

	if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))

	if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))

		return -ENOSPC;

		goto out_unlock;

	cgroup_taskset_for_each(task, cgrp, tset) {

	cgroup_taskset_for_each(task, cgrp, tset) {

		/*

		/*

		 * set_cpus_allowed_ptr() on all attached tasks before

		 * set_cpus_allowed_ptr() on all attached tasks before

		 * cpus_allowed may be changed.

		 * cpus_allowed may be changed.

		 */

		 */

		ret = -EINVAL;

		if (task->flags & PF_THREAD_BOUND)

		if (task->flags & PF_THREAD_BOUND)

			return -EINVAL;

			goto out_unlock;

		if ((ret = security_task_setscheduler(task)))

		ret = security_task_setscheduler(task);

			return ret;

		if (ret)

			goto out_unlock;

	}

	}

	/*

	/*

	 * changes which zero cpus/mems_allowed.

	 * changes which zero cpus/mems_allowed.

	 */

	 */

	cs->attach_in_progress++;

	cs->attach_in_progress++;

	ret = 0;

	return 0;

out_unlock:

	mutex_unlock(&cpuset_mutex);

	return ret;

}

}

static void cpuset_cancel_attach(struct cgroup *cgrp,

static void cpuset_cancel_attach(struct cgroup *cgrp,

				 struct cgroup_taskset *tset)

				 struct cgroup_taskset *tset)

{

{

	mutex_lock(&cpuset_mutex);

	cgroup_cs(cgrp)->attach_in_progress--;

	cgroup_cs(cgrp)->attach_in_progress--;

	mutex_unlock(&cpuset_mutex);

}

}

/*

/*

 * Protected by cgroup_mutex.  cpus_attach is used only by cpuset_attach()

 * Protected by cpuset_mutex.  cpus_attach is used only by cpuset_attach()

 * but we can't allocate it dynamically there.  Define it global and

 * but we can't allocate it dynamically there.  Define it global and

 * allocate from cpuset_init().

 * allocate from cpuset_init().

 */

 */

static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)

static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)

{

{

	/* static bufs protected by cgroup_mutex */

	/* static bufs protected by cpuset_mutex */

	static nodemask_t cpuset_attach_nodemask_from;

	static nodemask_t cpuset_attach_nodemask_from;

	static nodemask_t cpuset_attach_nodemask_to;

	static nodemask_t cpuset_attach_nodemask_to;

	struct mm_struct *mm;

	struct mm_struct *mm;

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *oldcs = cgroup_cs(oldcgrp);

	struct cpuset *oldcs = cgroup_cs(oldcgrp);

	mutex_lock(&cpuset_mutex);

	/* prepare for attach */

	/* prepare for attach */

	if (cs == &top_cpuset)

	if (cs == &top_cpuset)

		cpumask_copy(cpus_attach, cpu_possible_mask);

		cpumask_copy(cpus_attach, cpu_possible_mask);

	 */

	 */

	if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))

	if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))

		schedule_cpuset_propagate_hotplug(cs);

		schedule_cpuset_propagate_hotplug(cs);

	mutex_unlock(&cpuset_mutex);

}

}

/* The various types of files and directories in a cpuset file system */

/* The various types of files and directories in a cpuset file system */

static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val)

static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val)

{

{

	int retval = 0;

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	cpuset_filetype_t type = cft->private;

	cpuset_filetype_t type = cft->private;

	int retval = -ENODEV;

	if (!cgroup_lock_live_group(cgrp))

	mutex_lock(&cpuset_mutex);

		return -ENODEV;

	if (!is_cpuset_online(cs))

		goto out_unlock;

	switch (type) {

	switch (type) {

	case FILE_CPU_EXCLUSIVE:

	case FILE_CPU_EXCLUSIVE:

		retval = -EINVAL;

		retval = -EINVAL;

		break;

		break;

	}

	}

	cgroup_unlock();

out_unlock:

	mutex_unlock(&cpuset_mutex);

	return retval;

	return retval;

}

}

static int cpuset_write_s64(struct cgroup *cgrp, struct cftype *cft, s64 val)

static int cpuset_write_s64(struct cgroup *cgrp, struct cftype *cft, s64 val)

{

{

	int retval = 0;

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	cpuset_filetype_t type = cft->private;

	cpuset_filetype_t type = cft->private;

	int retval = -ENODEV;

	if (!cgroup_lock_live_group(cgrp))

	mutex_lock(&cpuset_mutex);

		return -ENODEV;

	if (!is_cpuset_online(cs))

		goto out_unlock;

	switch (type) {

	switch (type) {

	case FILE_SCHED_RELAX_DOMAIN_LEVEL:

	case FILE_SCHED_RELAX_DOMAIN_LEVEL:

		retval = -EINVAL;

		retval = -EINVAL;

		break;

		break;

	}

	}

	cgroup_unlock();

out_unlock:

	mutex_unlock(&cpuset_mutex);

	return retval;

	return retval;

}

}

static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft,

static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft,

				const char *buf)

				const char *buf)

{

{

	int retval = 0;

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *trialcs;

	struct cpuset *trialcs;

	int retval = -ENODEV;

	/*

	/*

	 * CPU or memory hotunplug may leave @cs w/o any execution

	 * CPU or memory hotunplug may leave @cs w/o any execution

	flush_work(&cpuset_hotplug_work);

	flush_work(&cpuset_hotplug_work);

	flush_workqueue(cpuset_propagate_hotplug_wq);

	flush_workqueue(cpuset_propagate_hotplug_wq);

	if (!cgroup_lock_live_group(cgrp))

	mutex_lock(&cpuset_mutex);

		return -ENODEV;

	if (!is_cpuset_online(cs))

		goto out_unlock;

	trialcs = alloc_trial_cpuset(cs);

	trialcs = alloc_trial_cpuset(cs);

	if (!trialcs) {

	if (!trialcs) {

		retval = -ENOMEM;

		retval = -ENOMEM;

		goto out;

		goto out_unlock;

	}

	}

	switch (cft->private) {

	switch (cft->private) {

	}

	}

	free_trial_cpuset(trialcs);

	free_trial_cpuset(trialcs);

out:

out_unlock:

	cgroup_unlock();

	mutex_unlock(&cpuset_mutex);

	return retval;

	return retval;

}

}

	if (!parent)

	if (!parent)

		return 0;

		return 0;

	mutex_lock(&cpuset_mutex);

	set_bit(CS_ONLINE, &cs->flags);

	set_bit(CS_ONLINE, &cs->flags);

	if (is_spread_page(parent))

	if (is_spread_page(parent))

		set_bit(CS_SPREAD_PAGE, &cs->flags);

		set_bit(CS_SPREAD_PAGE, &cs->flags);

	number_of_cpusets++;

	number_of_cpusets++;

	if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags))

	if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags))

		return 0;

		goto out_unlock;

	/*

	/*

	 * Clone @parent's configuration if CGRP_CPUSET_CLONE_CHILDREN is

	 * Clone @parent's configuration if CGRP_CPUSET_CLONE_CHILDREN is

	cpuset_for_each_child(tmp_cs, pos_cg, parent) {

	cpuset_for_each_child(tmp_cs, pos_cg, parent) {

		if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs)) {

		if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs)) {

			rcu_read_unlock();

			rcu_read_unlock();

			return 0;

			goto out_unlock;

		}

		}

	}

	}

	rcu_read_unlock();

	rcu_read_unlock();

	cs->mems_allowed = parent->mems_allowed;

	cs->mems_allowed = parent->mems_allowed;

	cpumask_copy(cs->cpus_allowed, parent->cpus_allowed);

	cpumask_copy(cs->cpus_allowed, parent->cpus_allowed);

	mutex_unlock(&callback_mutex);

	mutex_unlock(&callback_mutex);

out_unlock:

	mutex_unlock(&cpuset_mutex);

	return 0;

	return 0;

}

}

{

{

	struct cpuset *cs = cgroup_cs(cgrp);

	struct cpuset *cs = cgroup_cs(cgrp);

	/* css_offline is called w/o cgroup_mutex, grab it */

	mutex_lock(&cpuset_mutex);

	cgroup_lock();

	if (is_sched_load_balance(cs))

	if (is_sched_load_balance(cs))

		update_flag(CS_SCHED_LOAD_BALANCE, cs, 0);

		update_flag(CS_SCHED_LOAD_BALANCE, cs, 0);

	number_of_cpusets--;

	number_of_cpusets--;

	clear_bit(CS_ONLINE, &cs->flags);

	clear_bit(CS_ONLINE, &cs->flags);

	cgroup_unlock();

	mutex_unlock(&cpuset_mutex);

}

}

/*

/*

{

{

	struct cgroup *new_cgroup = scan->data;

	struct cgroup *new_cgroup = scan->data;

	cgroup_lock();

	cgroup_attach_task(new_cgroup, tsk);

	cgroup_attach_task(new_cgroup, tsk);

	cgroup_unlock();

}

}

/**

/**

 * @from: cpuset in which the tasks currently reside

 * @from: cpuset in which the tasks currently reside

 * @to: cpuset to which the tasks will be moved

 * @to: cpuset to which the tasks will be moved

 *

 *

 * Called with cgroup_mutex held

 * Called with cpuset_mutex held

 * callback_mutex must not be held, as cpuset_attach() will take it.

 * callback_mutex must not be held, as cpuset_attach() will take it.

 *

 *

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * The cgroup_scan_tasks() function will scan all the tasks in a cgroup,

 * removing that CPU or node from all cpusets.  If this removes the

 * removing that CPU or node from all cpusets.  If this removes the

 * last CPU or node from a cpuset, then move the tasks in the empty

 * last CPU or node from a cpuset, then move the tasks in the empty

 * cpuset to its next-highest non-empty parent.

 * cpuset to its next-highest non-empty parent.

 *

 * Called with cgroup_mutex held

 * callback_mutex must not be held, as cpuset_attach() will take it.

 */

 */

static void remove_tasks_in_empty_cpuset(struct cpuset *cs)

static void remove_tasks_in_empty_cpuset(struct cpuset *cs)

{

{

	static cpumask_t off_cpus;

	static cpumask_t off_cpus;

	static nodemask_t off_mems, tmp_mems;

	static nodemask_t off_mems, tmp_mems;

	struct cpuset *cs = container_of(work, struct cpuset, hotplug_work);

	struct cpuset *cs = container_of(work, struct cpuset, hotplug_work);

	bool is_empty;

	cgroup_lock();

	mutex_lock(&cpuset_mutex);

	cpumask_andnot(&off_cpus, cs->cpus_allowed, top_cpuset.cpus_allowed);

	cpumask_andnot(&off_cpus, cs->cpus_allowed, top_cpuset.cpus_allowed);

	nodes_andnot(off_mems, cs->mems_allowed, top_cpuset.mems_allowed);

	nodes_andnot(off_mems, cs->mems_allowed, top_cpuset.mems_allowed);

		update_tasks_nodemask(cs, &tmp_mems, NULL);

		update_tasks_nodemask(cs, &tmp_mems, NULL);

	}

	}

	if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))

	is_empty = cpumask_empty(cs->cpus_allowed) ||

		remove_tasks_in_empty_cpuset(cs);

		nodes_empty(cs->mems_allowed);

	cgroup_unlock();

	mutex_unlock(&cpuset_mutex);

	/*

	 * If @cs became empty, move tasks to the nearest ancestor with

	 * execution resources.  This is full cgroup operation which will

	 * also call back into cpuset.  Should be done outside any lock.

	 */

	if (is_empty)

		remove_tasks_in_empty_cpuset(cs);

	/* the following may free @cs, should be the last operation */

	/* the following may free @cs, should be the last operation */

	css_put(&cs->css);

	css_put(&cs->css);

	bool cpus_updated, mems_updated;

	bool cpus_updated, mems_updated;

	bool cpus_offlined, mems_offlined;

	bool cpus_offlined, mems_offlined;

	cgroup_lock();

	mutex_lock(&cpuset_mutex);

	/* fetch the available cpus/mems and find out which changed how */

	/* fetch the available cpus/mems and find out which changed how */

	cpumask_copy(&new_cpus, cpu_active_mask);

	cpumask_copy(&new_cpus, cpu_active_mask);

				schedule_cpuset_propagate_hotplug(cs);

				schedule_cpuset_propagate_hotplug(cs);

	}

	}

	cgroup_unlock();

	mutex_unlock(&cpuset_mutex);

	/* wait for propagations to finish */

	/* wait for propagations to finish */

	flush_workqueue(cpuset_propagate_hotplug_wq);

	flush_workqueue(cpuset_propagate_hotplug_wq);

		cpumask_var_t *doms;

		cpumask_var_t *doms;

		int ndoms;

		int ndoms;

		cgroup_lock();

		mutex_lock(&cpuset_mutex);

		ndoms = generate_sched_domains(&doms, &attr);

		ndoms = generate_sched_domains(&doms, &attr);

		cgroup_unlock();

		mutex_unlock(&cpuset_mutex);

		partition_sched_domains(ndoms, doms, attr);

		partition_sched_domains(ndoms, doms, attr);

	}

	}

 *  - Used for /proc/<pid>/cpuset.

 *  - Used for /proc/<pid>/cpuset.

 *  - No need to task_lock(tsk) on this tsk->cpuset reference, as it

 *  - No need to task_lock(tsk) on this tsk->cpuset reference, as it

 *    doesn't really matter if tsk->cpuset changes after we read it,

 *    doesn't really matter if tsk->cpuset changes after we read it,

 *    and we take cgroup_mutex, keeping cpuset_attach() from changing it

 *    and we take cpuset_mutex, keeping cpuset_attach() from changing it

 *    anyway.

 *    anyway.

 */

 */

static int proc_cpuset_show(struct seq_file *m, void *unused_v)

static int proc_cpuset_show(struct seq_file *m, void *unused_v)