[deliverable/linux.git] / kernel / sched_autogroup.c

#ifdef CONFIG_SCHED_AUTOGROUP

#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/kallsyms.h>
#include <linux/utsname.h>

unsigned int __read_mostly sysctl_sched_autogroup_enabled = 1;
static struct autogroup autogroup_default;
static atomic_t autogroup_seq_nr;

static void __init autogroup_init(struct task_struct *init_task)
{
	autogroup_default.tg = &root_task_group;
	root_task_group.autogroup = &autogroup_default;
	kref_init(&autogroup_default.kref);
	init_rwsem(&autogroup_default.lock);
	init_task->signal->autogroup = &autogroup_default;
}

static inline void autogroup_free(struct task_group *tg)
{
	kfree(tg->autogroup);
}

static inline void autogroup_destroy(struct kref *kref)
{
	struct autogroup *ag = container_of(kref, struct autogroup, kref);

#ifdef CONFIG_RT_GROUP_SCHED
	/* We've redirected RT tasks to the root task group... */
	ag->tg->rt_se = NULL;
	ag->tg->rt_rq = NULL;
#endif
	sched_destroy_group(ag->tg);
}

static inline void autogroup_kref_put(struct autogroup *ag)
{
	kref_put(&ag->kref, autogroup_destroy);
}

static inline struct autogroup *autogroup_kref_get(struct autogroup *ag)
{
	kref_get(&ag->kref);
	return ag;
}

static inline struct autogroup *autogroup_task_get(struct task_struct *p)
{
	struct autogroup *ag;
	unsigned long flags;

	if (!lock_task_sighand(p, &flags))
		return autogroup_kref_get(&autogroup_default);

	ag = autogroup_kref_get(p->signal->autogroup);
	unlock_task_sighand(p, &flags);

	return ag;
}

#ifdef CONFIG_RT_GROUP_SCHED
static void free_rt_sched_group(struct task_group *tg);
#endif

static inline struct autogroup *autogroup_create(void)
{
	struct autogroup *ag = kzalloc(sizeof(*ag), GFP_KERNEL);
	struct task_group *tg;

	if (!ag)
		goto out_fail;

	tg = sched_create_group(&root_task_group);

	if (IS_ERR(tg))
		goto out_free;

	kref_init(&ag->kref);
	init_rwsem(&ag->lock);
	ag->id = atomic_inc_return(&autogroup_seq_nr);
	ag->tg = tg;
#ifdef CONFIG_RT_GROUP_SCHED
	/*
	 * Autogroup RT tasks are redirected to the root task group
	 * so we don't have to move tasks around upon policy change,
	 * or flail around trying to allocate bandwidth on the fly.
	 * A bandwidth exception in __sched_setscheduler() allows
	 * the policy change to proceed.  Thereafter, task_group()
	 * returns &root_task_group, so zero bandwidth is required.
	 */
	free_rt_sched_group(tg);
	tg->rt_se = root_task_group.rt_se;
	tg->rt_rq = root_task_group.rt_rq;
#endif
	tg->autogroup = ag;

	return ag;

out_free:
	kfree(ag);
out_fail:
	if (printk_ratelimit()) {
		printk(KERN_WARNING "autogroup_create: %s failure.\n",
			ag ? "sched_create_group()" : "kmalloc()");
	}

	return autogroup_kref_get(&autogroup_default);
}

static inline bool
task_wants_autogroup(struct task_struct *p, struct task_group *tg)
{
	if (tg != &root_task_group)
		return false;

	if (p->sched_class != &fair_sched_class)
		return false;

	/*
	 * We can only assume the task group can't go away on us if
	 * autogroup_move_group() can see us on ->thread_group list.
	 */
	if (p->flags & PF_EXITING)
		return false;

	return true;
}

static inline bool task_group_is_autogroup(struct task_group *tg)
{
	return tg != &root_task_group && tg->autogroup;
}

static inline struct task_group *
autogroup_task_group(struct task_struct *p, struct task_group *tg)
{
	int enabled = ACCESS_ONCE(sysctl_sched_autogroup_enabled);

	if (enabled && task_wants_autogroup(p, tg))
		return p->signal->autogroup->tg;

	return tg;
}

static void
autogroup_move_group(struct task_struct *p, struct autogroup *ag)
{
	struct autogroup *prev;
	struct task_struct *t;
	unsigned long flags;

	BUG_ON(!lock_task_sighand(p, &flags));

	prev = p->signal->autogroup;
	if (prev == ag) {
		unlock_task_sighand(p, &flags);
		return;
	}

	p->signal->autogroup = autogroup_kref_get(ag);

	if (!ACCESS_ONCE(sysctl_sched_autogroup_enabled))
		goto out;

	t = p;
	do {
		sched_move_task(t);
	} while_each_thread(p, t);

out:
	unlock_task_sighand(p, &flags);
	autogroup_kref_put(prev);
}

/* Allocates GFP_KERNEL, cannot be called under any spinlock */
void sched_autogroup_create_attach(struct task_struct *p)
{
	struct autogroup *ag = autogroup_create();

	autogroup_move_group(p, ag);
	/* drop extra refrence added by autogroup_create() */
	autogroup_kref_put(ag);
}
EXPORT_SYMBOL(sched_autogroup_create_attach);

/* Cannot be called under siglock.  Currently has no users */
void sched_autogroup_detach(struct task_struct *p)
{
	autogroup_move_group(p, &autogroup_default);
}
EXPORT_SYMBOL(sched_autogroup_detach);

void sched_autogroup_fork(struct signal_struct *sig)
{
	sig->autogroup = autogroup_task_get(current);
}

void sched_autogroup_exit(struct signal_struct *sig)
{
	autogroup_kref_put(sig->autogroup);
}

static int __init setup_autogroup(char *str)
{
	sysctl_sched_autogroup_enabled = 0;

	return 1;
}

__setup("noautogroup", setup_autogroup);

#ifdef CONFIG_PROC_FS

int proc_sched_autogroup_set_nice(struct task_struct *p, int *nice)
{
	static unsigned long next = INITIAL_JIFFIES;
	struct autogroup *ag;
	int err;

	if (*nice < -20 || *nice > 19)
		return -EINVAL;

	err = security_task_setnice(current, *nice);
	if (err)
		return err;

	if (*nice < 0 && !can_nice(current, *nice))
		return -EPERM;

	/* this is a heavy operation taking global locks.. */
	if (!capable(CAP_SYS_ADMIN) && time_before(jiffies, next))
		return -EAGAIN;

	next = HZ / 10 + jiffies;
	ag = autogroup_task_get(p);

	down_write(&ag->lock);
	err = sched_group_set_shares(ag->tg, prio_to_weight[*nice + 20]);
	if (!err)
		ag->nice = *nice;
	up_write(&ag->lock);

	autogroup_kref_put(ag);

	return err;
}

void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m)
{
	struct autogroup *ag = autogroup_task_get(p);

	down_read(&ag->lock);
	seq_printf(m, "/autogroup-%ld nice %d\n", ag->id, ag->nice);
	up_read(&ag->lock);

	autogroup_kref_put(ag);
}
#endif /* CONFIG_PROC_FS */

#ifdef CONFIG_SCHED_DEBUG
static inline int autogroup_path(struct task_group *tg, char *buf, int buflen)
{
	int enabled = ACCESS_ONCE(sysctl_sched_autogroup_enabled);

	if (!enabled || !tg->autogroup)
		return 0;

	return snprintf(buf, buflen, "%s-%ld", "/autogroup", tg->autogroup->id);
}
#endif /* CONFIG_SCHED_DEBUG */

#endif /* CONFIG_SCHED_AUTOGROUP */
Commit	Line	Data
5091faa4 MG	1	#ifdef CONFIG_SCHED_AUTOGROUP
	2
	3	#include <linux/proc_fs.h>
	4	#include <linux/seq_file.h>
	5	#include <linux/kallsyms.h>
	6	#include <linux/utsname.h>
	7
	8	unsigned int __read_mostly sysctl_sched_autogroup_enabled = 1;
	9	static struct autogroup autogroup_default;
	10	static atomic_t autogroup_seq_nr;
	11
0ca08735	12	static void __init autogroup_init(struct task_struct *init_task)
5091faa4	13	{
07e06b01 YZ	14	autogroup_default.tg = &root_task_group;
07e06b01 YZ	15	root_task_group.autogroup = &autogroup_default;
5091faa4 MG	16	kref_init(&autogroup_default.kref);
	17	init_rwsem(&autogroup_default.lock);
	18	init_task->signal->autogroup = &autogroup_default;
	19	}
	20
	21	static inline void autogroup_free(struct task_group *tg)
	22	{
	23	kfree(tg->autogroup);
	24	}
	25
	26	static inline void autogroup_destroy(struct kref *kref)
	27	{
	28	struct autogroup *ag = container_of(kref, struct autogroup, kref);
	29
f4493771 MG	30	#ifdef CONFIG_RT_GROUP_SCHED
	31	/* We've redirected RT tasks to the root task group... */
	32	ag->tg->rt_se = NULL;
	33	ag->tg->rt_rq = NULL;
	34	#endif
5091faa4 MG	35	sched_destroy_group(ag->tg);
	36	}
	37
	38	static inline void autogroup_kref_put(struct autogroup *ag)
	39	{
	40	kref_put(&ag->kref, autogroup_destroy);
	41	}
	42
	43	static inline struct autogroup autogroup_kref_get(struct autogroup ag)
	44	{
	45	kref_get(&ag->kref);
	46	return ag;
	47	}
	48
4f821987 MG	49	static inline struct autogroup autogroup_task_get(struct task_struct p)
	50	{
	51	struct autogroup *ag;
	52	unsigned long flags;
	53
	54	if (!lock_task_sighand(p, &flags))
	55	return autogroup_kref_get(&autogroup_default);
	56
	57	ag = autogroup_kref_get(p->signal->autogroup);
	58	unlock_task_sighand(p, &flags);
	59
	60	return ag;
	61	}
	62
f4493771 MG	63	#ifdef CONFIG_RT_GROUP_SCHED
	64	static void free_rt_sched_group(struct task_group *tg);
	65	#endif
	66
5091faa4 MG	67	static inline struct autogroup *autogroup_create(void)
	68	{
	69	struct autogroup ag = kzalloc(sizeof(ag), GFP_KERNEL);
	70	struct task_group *tg;
	71
	72	if (!ag)
	73	goto out_fail;
	74
07e06b01	75	tg = sched_create_group(&root_task_group);
5091faa4 MG	76
	77	if (IS_ERR(tg))
	78	goto out_free;
	79
	80	kref_init(&ag->kref);
	81	init_rwsem(&ag->lock);
	82	ag->id = atomic_inc_return(&autogroup_seq_nr);
	83	ag->tg = tg;
f4493771 MG	84	#ifdef CONFIG_RT_GROUP_SCHED
	85	/*
	86	* Autogroup RT tasks are redirected to the root task group
	87	* so we don't have to move tasks around upon policy change,
	88	* or flail around trying to allocate bandwidth on the fly.
	89	* A bandwidth exception in __sched_setscheduler() allows
	90	* the policy change to proceed. Thereafter, task_group()
	91	* returns &root_task_group, so zero bandwidth is required.
	92	*/
	93	free_rt_sched_group(tg);
	94	tg->rt_se = root_task_group.rt_se;
	95	tg->rt_rq = root_task_group.rt_rq;
	96	#endif
5091faa4 MG	97	tg->autogroup = ag;
	98
	99	return ag;
	100
	101	out_free:
	102	kfree(ag);
	103	out_fail:
	104	if (printk_ratelimit()) {
	105	printk(KERN_WARNING "autogroup_create: %s failure.\n",
	106	ag ? "sched_create_group()" : "kmalloc()");
	107	}
	108
	109	return autogroup_kref_get(&autogroup_default);
	110	}
	111
	112	static inline bool
	113	task_wants_autogroup(struct task_struct p, struct task_group tg)
	114	{
	115	if (tg != &root_task_group)
	116	return false;
	117
	118	if (p->sched_class != &fair_sched_class)
	119	return false;
	120
	121	/*
	122	* We can only assume the task group can't go away on us if
	123	* autogroup_move_group() can see us on ->thread_group list.
	124	*/
	125	if (p->flags & PF_EXITING)
	126	return false;
	127
	128	return true;
	129	}
	130
f4493771 MG	131	static inline bool task_group_is_autogroup(struct task_group *tg)
	132	{
	133	return tg != &root_task_group && tg->autogroup;
	134	}
	135
5091faa4 MG	136	static inline struct task_group *
	137	autogroup_task_group(struct task_struct p, struct task_group tg)
	138	{
	139	int enabled = ACCESS_ONCE(sysctl_sched_autogroup_enabled);
	140
	141	if (enabled && task_wants_autogroup(p, tg))
	142	return p->signal->autogroup->tg;
	143
	144	return tg;
	145	}
	146
	147	static void
	148	autogroup_move_group(struct task_struct p, struct autogroup ag)
	149	{
	150	struct autogroup *prev;
	151	struct task_struct *t;
	152	unsigned long flags;
	153
	154	BUG_ON(!lock_task_sighand(p, &flags));
	155
	156	prev = p->signal->autogroup;
	157	if (prev == ag) {
	158	unlock_task_sighand(p, &flags);
	159	return;
	160	}
	161
	162	p->signal->autogroup = autogroup_kref_get(ag);
	163
800d4d30 YZ	164	if (!ACCESS_ONCE(sysctl_sched_autogroup_enabled))
	165	goto out;
	166
5091faa4 MG	167	t = p;
	168	do {
	169	sched_move_task(t);
	170	} while_each_thread(p, t);
	171
800d4d30	172	out:
5091faa4 MG	173	unlock_task_sighand(p, &flags);
	174	autogroup_kref_put(prev);
	175	}
	176
	177	/* Allocates GFP_KERNEL, cannot be called under any spinlock */
	178	void sched_autogroup_create_attach(struct task_struct *p)
	179	{
	180	struct autogroup *ag = autogroup_create();
	181
	182	autogroup_move_group(p, ag);
	183	/* drop extra refrence added by autogroup_create() */
	184	autogroup_kref_put(ag);
	185	}
	186	EXPORT_SYMBOL(sched_autogroup_create_attach);
	187
	188	/* Cannot be called under siglock. Currently has no users */
	189	void sched_autogroup_detach(struct task_struct *p)
	190	{
	191	autogroup_move_group(p, &autogroup_default);
	192	}
	193	EXPORT_SYMBOL(sched_autogroup_detach);
	194
	195	void sched_autogroup_fork(struct signal_struct *sig)
	196	{
4f821987	197	sig->autogroup = autogroup_task_get(current);
5091faa4 MG	198	}
	199
	200	void sched_autogroup_exit(struct signal_struct *sig)
	201	{
	202	autogroup_kref_put(sig->autogroup);
	203	}
	204
	205	static int __init setup_autogroup(char *str)
	206	{
	207	sysctl_sched_autogroup_enabled = 0;
	208
	209	return 1;
	210	}
	211
	212	__setup("noautogroup", setup_autogroup);
	213
	214	#ifdef CONFIG_PROC_FS
	215
5091faa4 MG	216	int proc_sched_autogroup_set_nice(struct task_struct p, int nice)
	217	{
	218	static unsigned long next = INITIAL_JIFFIES;
	219	struct autogroup *ag;
	220	int err;
	221
	222	if (nice < -20 \|\| nice > 19)
	223	return -EINVAL;
	224
	225	err = security_task_setnice(current, *nice);
	226	if (err)
	227	return err;
	228
	229	if (nice < 0 && !can_nice(current, nice))
	230	return -EPERM;
	231
	232	/* this is a heavy operation taking global locks.. */
	233	if (!capable(CAP_SYS_ADMIN) && time_before(jiffies, next))
	234	return -EAGAIN;
	235
	236	next = HZ / 10 + jiffies;
4f821987	237	ag = autogroup_task_get(p);
5091faa4 MG	238
	239	down_write(&ag->lock);
	240	err = sched_group_set_shares(ag->tg, prio_to_weight[*nice + 20]);
	241	if (!err)
	242	ag->nice = *nice;
	243	up_write(&ag->lock);
	244
	245	autogroup_kref_put(ag);
	246
	247	return err;
	248	}
	249
	250	void proc_sched_autogroup_show_task(struct task_struct p, struct seq_file m)
	251	{
4f821987	252	struct autogroup *ag = autogroup_task_get(p);
5091faa4 MG	253
	254	down_read(&ag->lock);
	255	seq_printf(m, "/autogroup-%ld nice %d\n", ag->id, ag->nice);
	256	up_read(&ag->lock);
	257
	258	autogroup_kref_put(ag);
	259	}
	260	#endif /* CONFIG_PROC_FS */
	261
	262	#ifdef CONFIG_SCHED_DEBUG
	263	static inline int autogroup_path(struct task_group tg, char buf, int buflen)
	264	{
8ecedd7a BR	265	int enabled = ACCESS_ONCE(sysctl_sched_autogroup_enabled);
	266
	267	if (!enabled \|\| !tg->autogroup)
	268	return 0;
	269
5091faa4 MG	270	return snprintf(buf, buflen, "%s-%ld", "/autogroup", tg->autogroup->id);
	271	}
	272	#endif /* CONFIG_SCHED_DEBUG */
	273
	274	#endif /* CONFIG_SCHED_AUTOGROUP */