[mirror_ubuntu-jammy-kernel.git] / kernel / cgroup / cgroup-internal.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef __CGROUP_INTERNAL_H
#define __CGROUP_INTERNAL_H

#include <linux/cgroup.h>
#include <linux/kernfs.h>
#include <linux/workqueue.h>
#include <linux/list.h>
#include <linux/refcount.h>
#include <linux/fs_parser.h>

#define TRACE_CGROUP_PATH_LEN 1024
extern spinlock_t trace_cgroup_path_lock;
extern char trace_cgroup_path[TRACE_CGROUP_PATH_LEN];
extern bool cgroup_debug;
extern void __init enable_debug_cgroup(void);

/*
 * cgroup_path() takes a spin lock. It is good practice not to take
 * spin locks within trace point handlers, as they are mostly hidden
 * from normal view. As cgroup_path() can take the kernfs_rename_lock
 * spin lock, it is best to not call that function from the trace event
 * handler.
 *
 * Note: trace_cgroup_##type##_enabled() is a static branch that will only
 *       be set when the trace event is enabled.
 */
#define TRACE_CGROUP_PATH(type, cgrp, ...)				\
	do {								\
		if (trace_cgroup_##type##_enabled()) {			\
			unsigned long flags;				\
			spin_lock_irqsave(&trace_cgroup_path_lock,	\
					  flags);			\
			cgroup_path(cgrp, trace_cgroup_path,		\
				    TRACE_CGROUP_PATH_LEN);		\
			trace_cgroup_##type(cgrp, trace_cgroup_path,	\
					    ##__VA_ARGS__);		\
			spin_unlock_irqrestore(&trace_cgroup_path_lock, \
					       flags);			\
		}							\
	} while (0)

/*
 * The cgroup filesystem superblock creation/mount context.
 */
struct cgroup_fs_context {
	struct kernfs_fs_context kfc;
	struct cgroup_root	*root;
	struct cgroup_namespace	*ns;
	unsigned int	flags;			/* CGRP_ROOT_* flags */

	/* cgroup1 bits */
	bool		cpuset_clone_children;
	bool		none;			/* User explicitly requested empty subsystem */
	bool		all_ss;			/* Seen 'all' option */
	u16		subsys_mask;		/* Selected subsystems */
	char		*name;			/* Hierarchy name */
	char		*release_agent;		/* Path for release notifications */
};

static inline struct cgroup_fs_context *cgroup_fc2context(struct fs_context *fc)
{
	struct kernfs_fs_context *kfc = fc->fs_private;

	return container_of(kfc, struct cgroup_fs_context, kfc);
}

/*
 * A cgroup can be associated with multiple css_sets as different tasks may
 * belong to different cgroups on different hierarchies.  In the other
 * direction, a css_set is naturally associated with multiple cgroups.
 * This M:N relationship is represented by the following link structure
 * which exists for each association and allows traversing the associations
 * from both sides.
 */
struct cgrp_cset_link {
	/* the cgroup and css_set this link associates */
	struct cgroup		*cgrp;
	struct css_set		*cset;

	/* list of cgrp_cset_links anchored at cgrp->cset_links */
	struct list_head	cset_link;

	/* list of cgrp_cset_links anchored at css_set->cgrp_links */
	struct list_head	cgrp_link;
};

/* used to track tasks and csets during migration */
struct cgroup_taskset {
	/* the src and dst cset list running through cset->mg_node */
	struct list_head	src_csets;
	struct list_head	dst_csets;

	/* the number of tasks in the set */
	int			nr_tasks;

	/* the subsys currently being processed */
	int			ssid;

	/*
	 * Fields for cgroup_taskset_*() iteration.
	 *
	 * Before migration is committed, the target migration tasks are on
	 * ->mg_tasks of the csets on ->src_csets.  After, on ->mg_tasks of
	 * the csets on ->dst_csets.  ->csets point to either ->src_csets
	 * or ->dst_csets depending on whether migration is committed.
	 *
	 * ->cur_csets and ->cur_task point to the current task position
	 * during iteration.
	 */
	struct list_head	*csets;
	struct css_set		*cur_cset;
	struct task_struct	*cur_task;
};

/* migration context also tracks preloading */
struct cgroup_mgctx {
	/*
	 * Preloaded source and destination csets.  Used to guarantee
	 * atomic success or failure on actual migration.
	 */
	struct list_head	preloaded_src_csets;
	struct list_head	preloaded_dst_csets;

	/* tasks and csets to migrate */
	struct cgroup_taskset	tset;

	/* subsystems affected by migration */
	u16			ss_mask;
};

#define CGROUP_TASKSET_INIT(tset)						\
{										\
	.src_csets		= LIST_HEAD_INIT(tset.src_csets),		\
	.dst_csets		= LIST_HEAD_INIT(tset.dst_csets),		\
	.csets			= &tset.src_csets,				\
}

#define CGROUP_MGCTX_INIT(name)							\
{										\
	LIST_HEAD_INIT(name.preloaded_src_csets),				\
	LIST_HEAD_INIT(name.preloaded_dst_csets),				\
	CGROUP_TASKSET_INIT(name.tset),						\
}

#define DEFINE_CGROUP_MGCTX(name)						\
	struct cgroup_mgctx name = CGROUP_MGCTX_INIT(name)

extern struct mutex cgroup_mutex;
extern spinlock_t css_set_lock;
extern struct cgroup_subsys *cgroup_subsys[];
extern struct list_head cgroup_roots;
extern struct file_system_type cgroup_fs_type;

/* iterate across the hierarchies */
#define for_each_root(root)						\
	list_for_each_entry((root), &cgroup_roots, root_list)

/**
 * for_each_subsys - iterate all enabled cgroup subsystems
 * @ss: the iteration cursor
 * @ssid: the index of @ss, CGROUP_SUBSYS_COUNT after reaching the end
 */
#define for_each_subsys(ss, ssid)					\
	for ((ssid) = 0; (ssid) < CGROUP_SUBSYS_COUNT &&		\
	     (((ss) = cgroup_subsys[ssid]) || true); (ssid)++)

static inline bool cgroup_is_dead(const struct cgroup *cgrp)
{
	return !(cgrp->self.flags & CSS_ONLINE);
}

static inline bool notify_on_release(const struct cgroup *cgrp)
{
	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
}

void put_css_set_locked(struct css_set *cset);

static inline void put_css_set(struct css_set *cset)
{
	unsigned long flags;

	/*
	 * Ensure that the refcount doesn't hit zero while any readers
	 * can see it. Similar to atomic_dec_and_lock(), but for an
	 * rwlock
	 */
	if (refcount_dec_not_one(&cset->refcount))
		return;

	spin_lock_irqsave(&css_set_lock, flags);
	put_css_set_locked(cset);
	spin_unlock_irqrestore(&css_set_lock, flags);
}

/*
 * refcounted get/put for css_set objects
 */
static inline void get_css_set(struct css_set *cset)
{
	refcount_inc(&cset->refcount);
}

bool cgroup_ssid_enabled(int ssid);
bool cgroup_on_dfl(const struct cgroup *cgrp);
bool cgroup_is_thread_root(struct cgroup *cgrp);
bool cgroup_is_threaded(struct cgroup *cgrp);

struct cgroup_root *cgroup_root_from_kf(struct kernfs_root *kf_root);
struct cgroup *task_cgroup_from_root(struct task_struct *task,
				     struct cgroup_root *root);
struct cgroup *cgroup_kn_lock_live(struct kernfs_node *kn, bool drain_offline);
void cgroup_kn_unlock(struct kernfs_node *kn);
int cgroup_path_ns_locked(struct cgroup *cgrp, char *buf, size_t buflen,
			  struct cgroup_namespace *ns);

void cgroup_free_root(struct cgroup_root *root);
void init_cgroup_root(struct cgroup_fs_context *ctx);
int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask);
int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask);
int cgroup_do_get_tree(struct fs_context *fc);

int cgroup_migrate_vet_dst(struct cgroup *dst_cgrp);
void cgroup_migrate_finish(struct cgroup_mgctx *mgctx);
void cgroup_migrate_add_src(struct css_set *src_cset, struct cgroup *dst_cgrp,
			    struct cgroup_mgctx *mgctx);
int cgroup_migrate_prepare_dst(struct cgroup_mgctx *mgctx);
int cgroup_migrate(struct task_struct *leader, bool threadgroup,
		   struct cgroup_mgctx *mgctx);

int cgroup_attach_task(struct cgroup *dst_cgrp, struct task_struct *leader,
		       bool threadgroup);
struct task_struct *cgroup_procs_write_start(char *buf, bool threadgroup,
					     bool *locked)
	__acquires(&cgroup_threadgroup_rwsem);
void cgroup_procs_write_finish(struct task_struct *task, bool locked)
	__releases(&cgroup_threadgroup_rwsem);

void cgroup_lock_and_drain_offline(struct cgroup *cgrp);

int cgroup_mkdir(struct kernfs_node *parent_kn, const char *name, umode_t mode);
int cgroup_rmdir(struct kernfs_node *kn);
int cgroup_show_path(struct seq_file *sf, struct kernfs_node *kf_node,
		     struct kernfs_root *kf_root);

int __cgroup_task_count(const struct cgroup *cgrp);
int cgroup_task_count(const struct cgroup *cgrp);

/*
 * rstat.c
 */
int cgroup_rstat_init(struct cgroup *cgrp);
void cgroup_rstat_exit(struct cgroup *cgrp);
void cgroup_rstat_boot(void);
void cgroup_base_stat_cputime_show(struct seq_file *seq);

/*
 * namespace.c
 */
extern const struct proc_ns_operations cgroupns_operations;

/*
 * cgroup-v1.c
 */
extern struct cftype cgroup1_base_files[];
extern struct kernfs_syscall_ops cgroup1_kf_syscall_ops;
extern const struct fs_parameter_spec cgroup1_fs_parameters[];

int proc_cgroupstats_show(struct seq_file *m, void *v);
bool cgroup1_ssid_disabled(int ssid);
void cgroup1_pidlist_destroy_all(struct cgroup *cgrp);
void cgroup1_release_agent(struct work_struct *work);
void cgroup1_check_for_release(struct cgroup *cgrp);
int cgroup1_parse_param(struct fs_context *fc, struct fs_parameter *param);
int cgroup1_get_tree(struct fs_context *fc);
int cgroup1_reconfigure(struct fs_context *ctx);

#endif /* __CGROUP_INTERNAL_H */
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
0a268dbd TH	2	#ifndef __CGROUP_INTERNAL_H
	3	#define __CGROUP_INTERNAL_H
	4
	5	#include <linux/cgroup.h>
	6	#include <linux/kernfs.h>
	7	#include <linux/workqueue.h>
	8	#include <linux/list.h>
4b9502e6	9	#include <linux/refcount.h>
d7167b14	10	#include <linux/fs_parser.h>
0a268dbd	11
e4f8d81c SRV	12	#define TRACE_CGROUP_PATH_LEN 1024
	13	extern spinlock_t trace_cgroup_path_lock;
	14	extern char trace_cgroup_path[TRACE_CGROUP_PATH_LEN];
5cf8114d WL	15	extern bool cgroup_debug;
5cf8114d WL	16	extern void __init enable_debug_cgroup(void);
e4f8d81c SRV	17
	18	/*
	19	* cgroup_path() takes a spin lock. It is good practice not to take
	20	* spin locks within trace point handlers, as they are mostly hidden
	21	* from normal view. As cgroup_path() can take the kernfs_rename_lock
	22	* spin lock, it is best to not call that function from the trace event
	23	* handler.
	24	*
	25	* Note: trace_cgroup_##type##_enabled() is a static branch that will only
	26	* be set when the trace event is enabled.
	27	*/
	28	#define TRACE_CGROUP_PATH(type, cgrp, ...) \
	29	do { \
	30	if (trace_cgroup_##type##_enabled()) { \
712e3517 RG	31	unsigned long flags; \
	32	spin_lock_irqsave(&trace_cgroup_path_lock, \
	33	flags); \
e4f8d81c SRV	34	cgroup_path(cgrp, trace_cgroup_path, \
	35	TRACE_CGROUP_PATH_LEN); \
	36	trace_cgroup_##type(cgrp, trace_cgroup_path, \
	37	##__VA_ARGS__); \
712e3517 RG	38	spin_unlock_irqrestore(&trace_cgroup_path_lock, \
712e3517 RG	39	flags); \
e4f8d81c SRV	40	} \
	41	} while (0)
	42
90129625 AV	43	/*
	44	* The cgroup filesystem superblock creation/mount context.
	45	*/
	46	struct cgroup_fs_context {
23bf1b6b	47	struct kernfs_fs_context kfc;
cf6299b1	48	struct cgroup_root *root;
cca8f327	49	struct cgroup_namespace *ns;
f5dfb531 AV	50	unsigned int flags; /* CGRP_ROOT_* flags */
	51
	52	/* cgroup1 bits */
	53	bool cpuset_clone_children;
	54	bool none; /* User explicitly requested empty subsystem */
	55	bool all_ss; /* Seen 'all' option */
	56	u16 subsys_mask; /* Selected subsystems */
	57	char name; / Hierarchy name */
	58	char release_agent; / Path for release notifications */
90129625 AV	59	};
	60
	61	static inline struct cgroup_fs_context cgroup_fc2context(struct fs_context fc)
	62	{
23bf1b6b DH	63	struct kernfs_fs_context *kfc = fc->fs_private;
	64
	65	return container_of(kfc, struct cgroup_fs_context, kfc);
90129625 AV	66	}
90129625 AV	67
0a268dbd TH	68	/*
	69	* A cgroup can be associated with multiple css_sets as different tasks may
	70	* belong to different cgroups on different hierarchies. In the other
	71	* direction, a css_set is naturally associated with multiple cgroups.
	72	* This M:N relationship is represented by the following link structure
	73	* which exists for each association and allows traversing the associations
	74	* from both sides.
	75	*/
	76	struct cgrp_cset_link {
	77	/* the cgroup and css_set this link associates */
	78	struct cgroup *cgrp;
	79	struct css_set *cset;
	80
	81	/* list of cgrp_cset_links anchored at cgrp->cset_links */
	82	struct list_head cset_link;
	83
	84	/* list of cgrp_cset_links anchored at css_set->cgrp_links */
	85	struct list_head cgrp_link;
	86	};
	87
e595cd70 TH	88	/* used to track tasks and csets during migration */
	89	struct cgroup_taskset {
	90	/* the src and dst cset list running through cset->mg_node */
	91	struct list_head src_csets;
	92	struct list_head dst_csets;
	93
61046727 TH	94	/* the number of tasks in the set */
	95	int nr_tasks;
	96
e595cd70 TH	97	/* the subsys currently being processed */
	98	int ssid;
	99
	100	/*
	101	* Fields for cgroup_taskset_*() iteration.
	102	*
	103	* Before migration is committed, the target migration tasks are on
	104	* ->mg_tasks of the csets on ->src_csets. After, on ->mg_tasks of
	105	* the csets on ->dst_csets. ->csets point to either ->src_csets
	106	* or ->dst_csets depending on whether migration is committed.
	107	*
	108	* ->cur_csets and ->cur_task point to the current task position
	109	* during iteration.
	110	*/
	111	struct list_head *csets;
	112	struct css_set *cur_cset;
	113	struct task_struct *cur_task;
	114	};
	115
	116	/* migration context also tracks preloading */
	117	struct cgroup_mgctx {
	118	/*
	119	* Preloaded source and destination csets. Used to guarantee
	120	* atomic success or failure on actual migration.
	121	*/
	122	struct list_head preloaded_src_csets;
	123	struct list_head preloaded_dst_csets;
	124
	125	/* tasks and csets to migrate */
	126	struct cgroup_taskset tset;
bfc2cf6f TH	127
	128	/* subsystems affected by migration */
	129	u16 ss_mask;
e595cd70 TH	130	};
	131
	132	#define CGROUP_TASKSET_INIT(tset) \
	133	{ \
	134	.src_csets = LIST_HEAD_INIT(tset.src_csets), \
	135	.dst_csets = LIST_HEAD_INIT(tset.dst_csets), \
	136	.csets = &tset.src_csets, \
	137	}
	138
	139	#define CGROUP_MGCTX_INIT(name) \
	140	{ \
	141	LIST_HEAD_INIT(name.preloaded_src_csets), \
	142	LIST_HEAD_INIT(name.preloaded_dst_csets), \
	143	CGROUP_TASKSET_INIT(name.tset), \
	144	}
	145
	146	#define DEFINE_CGROUP_MGCTX(name) \
	147	struct cgroup_mgctx name = CGROUP_MGCTX_INIT(name)
	148
0a268dbd TH	149	extern struct mutex cgroup_mutex;
	150	extern spinlock_t css_set_lock;
	151	extern struct cgroup_subsys *cgroup_subsys[];
	152	extern struct list_head cgroup_roots;
	153	extern struct file_system_type cgroup_fs_type;
	154
	155	/* iterate across the hierarchies */
	156	#define for_each_root(root) \
	157	list_for_each_entry((root), &cgroup_roots, root_list)
	158
	159	/**
	160	* for_each_subsys - iterate all enabled cgroup subsystems
	161	* @ss: the iteration cursor
	162	* @ssid: the index of @ss, CGROUP_SUBSYS_COUNT after reaching the end
	163	*/
	164	#define for_each_subsys(ss, ssid) \
	165	for ((ssid) = 0; (ssid) < CGROUP_SUBSYS_COUNT && \
	166	(((ss) = cgroup_subsys[ssid]) \|\| true); (ssid)++)
	167
	168	static inline bool cgroup_is_dead(const struct cgroup *cgrp)
	169	{
	170	return !(cgrp->self.flags & CSS_ONLINE);
	171	}
	172
	173	static inline bool notify_on_release(const struct cgroup *cgrp)
	174	{
	175	return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
	176	}
	177
dcfe149b TH	178	void put_css_set_locked(struct css_set *cset);
	179
	180	static inline void put_css_set(struct css_set *cset)
	181	{
	182	unsigned long flags;
	183
	184	/*
	185	* Ensure that the refcount doesn't hit zero while any readers
	186	* can see it. Similar to atomic_dec_and_lock(), but for an
	187	* rwlock
	188	*/
4b9502e6	189	if (refcount_dec_not_one(&cset->refcount))
dcfe149b TH	190	return;
	191
	192	spin_lock_irqsave(&css_set_lock, flags);
	193	put_css_set_locked(cset);
	194	spin_unlock_irqrestore(&css_set_lock, flags);
	195	}
	196
	197	/*
	198	* refcounted get/put for css_set objects
	199	*/
	200	static inline void get_css_set(struct css_set *cset)
	201	{
4b9502e6	202	refcount_inc(&cset->refcount);
dcfe149b TH	203	}
dcfe149b TH	204
0a268dbd TH	205	bool cgroup_ssid_enabled(int ssid);
0a268dbd TH	206	bool cgroup_on_dfl(const struct cgroup *cgrp);
7a0cf0e7 WL	207	bool cgroup_is_thread_root(struct cgroup *cgrp);
7a0cf0e7 WL	208	bool cgroup_is_threaded(struct cgroup *cgrp);
0a268dbd TH	209
	210	struct cgroup_root cgroup_root_from_kf(struct kernfs_root kf_root);
	211	struct cgroup task_cgroup_from_root(struct task_struct task,
	212	struct cgroup_root *root);
	213	struct cgroup cgroup_kn_lock_live(struct kernfs_node kn, bool drain_offline);
	214	void cgroup_kn_unlock(struct kernfs_node *kn);
	215	int cgroup_path_ns_locked(struct cgroup cgrp, char buf, size_t buflen,
	216	struct cgroup_namespace *ns);
	217
1592c9b2	218	void cgroup_free_root(struct cgroup_root *root);
cf6299b1	219	void init_cgroup_root(struct cgroup_fs_context *ctx);
35ac1184	220	int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask);
0a268dbd	221	int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask);
cca8f327	222	int cgroup_do_get_tree(struct fs_context *fc);
0a268dbd	223
8cfd8147	224	int cgroup_migrate_vet_dst(struct cgroup *dst_cgrp);
e595cd70 TH	225	void cgroup_migrate_finish(struct cgroup_mgctx *mgctx);
	226	void cgroup_migrate_add_src(struct css_set src_cset, struct cgroup dst_cgrp,
	227	struct cgroup_mgctx *mgctx);
	228	int cgroup_migrate_prepare_dst(struct cgroup_mgctx *mgctx);
0a268dbd	229	int cgroup_migrate(struct task_struct *leader, bool threadgroup,
bfc2cf6f	230	struct cgroup_mgctx *mgctx);
0a268dbd TH	231
	232	int cgroup_attach_task(struct cgroup dst_cgrp, struct task_struct leader,
	233	bool threadgroup);
9a3284fa MK	234	struct task_struct cgroup_procs_write_start(char buf, bool threadgroup,
9a3284fa MK	235	bool *locked)
715c809d	236	__acquires(&cgroup_threadgroup_rwsem);
9a3284fa	237	void cgroup_procs_write_finish(struct task_struct *task, bool locked)
715c809d	238	__releases(&cgroup_threadgroup_rwsem);
0a268dbd TH	239
	240	void cgroup_lock_and_drain_offline(struct cgroup *cgrp);
	241
1592c9b2 TH	242	int cgroup_mkdir(struct kernfs_node parent_kn, const char name, umode_t mode);
	243	int cgroup_rmdir(struct kernfs_node *kn);
	244	int cgroup_show_path(struct seq_file sf, struct kernfs_node kf_node,
	245	struct kernfs_root *kf_root);
	246
aade7f9e	247	int __cgroup_task_count(const struct cgroup *cgrp);
a28f8f5e WL	248	int cgroup_task_count(const struct cgroup *cgrp);
a28f8f5e WL	249
041cd640	250	/*
c58632b3	251	* rstat.c
041cd640	252	*/
c58632b3 TH	253	int cgroup_rstat_init(struct cgroup *cgrp);
c58632b3 TH	254	void cgroup_rstat_exit(struct cgroup *cgrp);
c58632b3	255	void cgroup_rstat_boot(void);
a17556f8	256	void cgroup_base_stat_cputime_show(struct seq_file *seq);
041cd640	257
dcfe149b TH	258	/*
	259	* namespace.c
	260	*/
	261	extern const struct proc_ns_operations cgroupns_operations;
	262
0a268dbd TH	263	/*
	264	* cgroup-v1.c
	265	*/
d62beb7f	266	extern struct cftype cgroup1_base_files[];
1592c9b2	267	extern struct kernfs_syscall_ops cgroup1_kf_syscall_ops;
d7167b14	268	extern const struct fs_parameter_spec cgroup1_fs_parameters[];
0a268dbd	269
3f3942ac	270	int proc_cgroupstats_show(struct seq_file m, void v);
d62beb7f TH	271	bool cgroup1_ssid_disabled(int ssid);
	272	void cgroup1_pidlist_destroy_all(struct cgroup *cgrp);
	273	void cgroup1_release_agent(struct work_struct *work);
	274	void cgroup1_check_for_release(struct cgroup *cgrp);
8d2451f4	275	int cgroup1_parse_param(struct fs_context fc, struct fs_parameter param);
7feeef58	276	int cgroup1_get_tree(struct fs_context *fc);
90129625	277	int cgroup1_reconfigure(struct fs_context *ctx);
0a268dbd TH	278
0a268dbd TH	279	#endif /* __CGROUP_INTERNAL_H */