[net-next-2.6.git] / kernel / taskstats.c

/*
 * taskstats.c - Export per-task statistics to userland
 *
 * Copyright (C) Shailabh Nagar, IBM Corp. 2006
 *           (C) Balbir Singh,   IBM Corp. 2006
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 */

#include <linux/kernel.h>
#include <linux/taskstats_kern.h>
#include <linux/tsacct_kern.h>
#include <linux/delayacct.h>
#include <linux/tsacct_kern.h>
#include <linux/cpumask.h>
#include <linux/percpu.h>
#include <net/genetlink.h>
#include <asm/atomic.h>

/*
 * Maximum length of a cpumask that can be specified in
 * the TASKSTATS_CMD_ATTR_REGISTER/DEREGISTER_CPUMASK attribute
 */
#define TASKSTATS_CPUMASK_MAXLEN	(100+6*NR_CPUS)

static DEFINE_PER_CPU(__u32, taskstats_seqnum) = { 0 };
static int family_registered;
struct kmem_cache *taskstats_cache;

static struct genl_family family = {
	.id		= GENL_ID_GENERATE,
	.name		= TASKSTATS_GENL_NAME,
	.version	= TASKSTATS_GENL_VERSION,
	.maxattr	= TASKSTATS_CMD_ATTR_MAX,
};

static struct nla_policy taskstats_cmd_get_policy[TASKSTATS_CMD_ATTR_MAX+1]
__read_mostly = {
	[TASKSTATS_CMD_ATTR_PID]  = { .type = NLA_U32 },
	[TASKSTATS_CMD_ATTR_TGID] = { .type = NLA_U32 },
	[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK] = { .type = NLA_STRING },
	[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK] = { .type = NLA_STRING },};

struct listener {
	struct list_head list;
	pid_t pid;
	char valid;
};

struct listener_list {
	struct rw_semaphore sem;
	struct list_head list;
};
static DEFINE_PER_CPU(struct listener_list, listener_array);

enum actions {
	REGISTER,
	DEREGISTER,
	CPU_DONT_CARE
};

static int prepare_reply(struct genl_info *info, u8 cmd, struct sk_buff **skbp,
				size_t size)
{
	struct sk_buff *skb;
	void *reply;

	/*
	 * If new attributes are added, please revisit this allocation
	 */
	skb = genlmsg_new(size, GFP_KERNEL);
	if (!skb)
		return -ENOMEM;

	if (!info) {
		int seq = get_cpu_var(taskstats_seqnum)++;
		put_cpu_var(taskstats_seqnum);

		reply = genlmsg_put(skb, 0, seq, &family, 0, cmd);
	} else
		reply = genlmsg_put_reply(skb, info, &family, 0, cmd);
	if (reply == NULL) {
		nlmsg_free(skb);
		return -EINVAL;
	}

	*skbp = skb;
	return 0;
}

/*
 * Send taskstats data in @skb to listener with nl_pid @pid
 */
static int send_reply(struct sk_buff *skb, pid_t pid)
{
	struct genlmsghdr *genlhdr = nlmsg_data(nlmsg_hdr(skb));
	void *reply = genlmsg_data(genlhdr);
	int rc;

	rc = genlmsg_end(skb, reply);
	if (rc < 0) {
		nlmsg_free(skb);
		return rc;
	}

	return genlmsg_unicast(skb, pid);
}

/*
 * Send taskstats data in @skb to listeners registered for @cpu's exit data
 */
static void send_cpu_listeners(struct sk_buff *skb,
					struct listener_list *listeners)
{
	struct genlmsghdr *genlhdr = nlmsg_data(nlmsg_hdr(skb));
	struct listener *s, *tmp;
	struct sk_buff *skb_next, *skb_cur = skb;
	void *reply = genlmsg_data(genlhdr);
	int rc, delcount = 0;

	rc = genlmsg_end(skb, reply);
	if (rc < 0) {
		nlmsg_free(skb);
		return;
	}

	rc = 0;
	down_read(&listeners->sem);
	list_for_each_entry(s, &listeners->list, list) {
		skb_next = NULL;
		if (!list_is_last(&s->list, &listeners->list)) {
			skb_next = skb_clone(skb_cur, GFP_KERNEL);
			if (!skb_next)
				break;
		}
		rc = genlmsg_unicast(skb_cur, s->pid);
		if (rc == -ECONNREFUSED) {
			s->valid = 0;
			delcount++;
		}
		skb_cur = skb_next;
	}
	up_read(&listeners->sem);

	if (skb_cur)
		nlmsg_free(skb_cur);

	if (!delcount)
		return;

	/* Delete invalidated entries */
	down_write(&listeners->sem);
	list_for_each_entry_safe(s, tmp, &listeners->list, list) {
		if (!s->valid) {
			list_del(&s->list);
			kfree(s);
		}
	}
	up_write(&listeners->sem);
}

static int fill_pid(pid_t pid, struct task_struct *tsk,
		struct taskstats *stats)
{
	int rc = 0;

	if (!tsk) {
		rcu_read_lock();
		tsk = find_task_by_pid(pid);
		if (tsk)
			get_task_struct(tsk);
		rcu_read_unlock();
		if (!tsk)
			return -ESRCH;
	} else
		get_task_struct(tsk);

	memset(stats, 0, sizeof(*stats));
	/*
	 * Each accounting subsystem adds calls to its functions to
	 * fill in relevant parts of struct taskstsats as follows
	 *
	 *	per-task-foo(stats, tsk);
	 */

	delayacct_add_tsk(stats, tsk);

	/* fill in basic acct fields */
	stats->version = TASKSTATS_VERSION;
	bacct_add_tsk(stats, tsk);

	/* fill in extended acct fields */
	xacct_add_tsk(stats, tsk);

	/* Define err: label here if needed */
	put_task_struct(tsk);
	return rc;

}

static int fill_tgid(pid_t tgid, struct task_struct *first,
		struct taskstats *stats)
{
	struct task_struct *tsk;
	unsigned long flags;
	int rc = -ESRCH;

	/*
	 * Add additional stats from live tasks except zombie thread group
	 * leaders who are already counted with the dead tasks
	 */
	rcu_read_lock();
	if (!first)
		first = find_task_by_pid(tgid);

	if (!first || !lock_task_sighand(first, &flags))
		goto out;

	if (first->signal->stats)
		memcpy(stats, first->signal->stats, sizeof(*stats));
	else
		memset(stats, 0, sizeof(*stats));

	tsk = first;
	do {
		if (tsk->exit_state)
			continue;
		/*
		 * Accounting subsystem can call its functions here to
		 * fill in relevant parts of struct taskstsats as follows
		 *
		 *	per-task-foo(stats, tsk);
		 */
		delayacct_add_tsk(stats, tsk);

	} while_each_thread(first, tsk);

	unlock_task_sighand(first, &flags);
	rc = 0;
out:
	rcu_read_unlock();

	stats->version = TASKSTATS_VERSION;
	/*
	 * Accounting subsytems can also add calls here to modify
	 * fields of taskstats.
	 */
	return rc;
}


static void fill_tgid_exit(struct task_struct *tsk)
{
	unsigned long flags;

	spin_lock_irqsave(&tsk->sighand->siglock, flags);
	if (!tsk->signal->stats)
		goto ret;

	/*
	 * Each accounting subsystem calls its functions here to
	 * accumalate its per-task stats for tsk, into the per-tgid structure
	 *
	 *	per-task-foo(tsk->signal->stats, tsk);
	 */
	delayacct_add_tsk(tsk->signal->stats, tsk);
ret:
	spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
	return;
}

static int add_del_listener(pid_t pid, cpumask_t *maskp, int isadd)
{
	struct listener_list *listeners;
	struct listener *s, *tmp;
	unsigned int cpu;
	cpumask_t mask = *maskp;

	if (!cpus_subset(mask, cpu_possible_map))
		return -EINVAL;

	if (isadd == REGISTER) {
		for_each_cpu_mask(cpu, mask) {
			s = kmalloc_node(sizeof(struct listener), GFP_KERNEL,
					 cpu_to_node(cpu));
			if (!s)
				goto cleanup;
			s->pid = pid;
			INIT_LIST_HEAD(&s->list);
			s->valid = 1;

			listeners = &per_cpu(listener_array, cpu);
			down_write(&listeners->sem);
			list_add(&s->list, &listeners->list);
			up_write(&listeners->sem);
		}
		return 0;
	}

	/* Deregister or cleanup */
cleanup:
	for_each_cpu_mask(cpu, mask) {
		listeners = &per_cpu(listener_array, cpu);
		down_write(&listeners->sem);
		list_for_each_entry_safe(s, tmp, &listeners->list, list) {
			if (s->pid == pid) {
				list_del(&s->list);
				kfree(s);
				break;
			}
		}
		up_write(&listeners->sem);
	}
	return 0;
}

static int parse(struct nlattr *na, cpumask_t *mask)
{
	char *data;
	int len;
	int ret;

	if (na == NULL)
		return 1;
	len = nla_len(na);
	if (len > TASKSTATS_CPUMASK_MAXLEN)
		return -E2BIG;
	if (len < 1)
		return -EINVAL;
	data = kmalloc(len, GFP_KERNEL);
	if (!data)
		return -ENOMEM;
	nla_strlcpy(data, na, len);
	ret = cpulist_parse(data, *mask);
	kfree(data);
	return ret;
}

static struct taskstats *mk_reply(struct sk_buff *skb, int type, u32 pid)
{
	struct nlattr *na, *ret;
	int aggr;

	aggr = (type == TASKSTATS_TYPE_PID)
			? TASKSTATS_TYPE_AGGR_PID
			: TASKSTATS_TYPE_AGGR_TGID;

	na = nla_nest_start(skb, aggr);
	if (!na)
		goto err;
	if (nla_put(skb, type, sizeof(pid), &pid) < 0)
		goto err;
	ret = nla_reserve(skb, TASKSTATS_TYPE_STATS, sizeof(struct taskstats));
	if (!ret)
		goto err;
	nla_nest_end(skb, na);

	return nla_data(ret);
err:
	return NULL;
}

static int taskstats_user_cmd(struct sk_buff *skb, struct genl_info *info)
{
	int rc = 0;
	struct sk_buff *rep_skb;
	struct taskstats *stats;
	size_t size;
	cpumask_t mask;

	rc = parse(info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK], &mask);
	if (rc < 0)
		return rc;
	if (rc == 0)
		return add_del_listener(info->snd_pid, &mask, REGISTER);

	rc = parse(info->attrs[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK], &mask);
	if (rc < 0)
		return rc;
	if (rc == 0)
		return add_del_listener(info->snd_pid, &mask, DEREGISTER);

	/*
	 * Size includes space for nested attributes
	 */
	size = nla_total_size(sizeof(u32)) +
		nla_total_size(sizeof(struct taskstats)) + nla_total_size(0);

	rc = prepare_reply(info, TASKSTATS_CMD_NEW, &rep_skb, size);
	if (rc < 0)
		return rc;

	rc = -EINVAL;
	if (info->attrs[TASKSTATS_CMD_ATTR_PID]) {
		u32 pid = nla_get_u32(info->attrs[TASKSTATS_CMD_ATTR_PID]);
		stats = mk_reply(rep_skb, TASKSTATS_TYPE_PID, pid);
		if (!stats)
			goto err;

		rc = fill_pid(pid, NULL, stats);
		if (rc < 0)
			goto err;
	} else if (info->attrs[TASKSTATS_CMD_ATTR_TGID]) {
		u32 tgid = nla_get_u32(info->attrs[TASKSTATS_CMD_ATTR_TGID]);
		stats = mk_reply(rep_skb, TASKSTATS_TYPE_TGID, tgid);
		if (!stats)
			goto err;

		rc = fill_tgid(tgid, NULL, stats);
		if (rc < 0)
			goto err;
	} else
		goto err;

	return send_reply(rep_skb, info->snd_pid);
err:
	nlmsg_free(rep_skb);
	return rc;
}

static struct taskstats *taskstats_tgid_alloc(struct task_struct *tsk)
{
	struct signal_struct *sig = tsk->signal;
	struct taskstats *stats;

	if (sig->stats || thread_group_empty(tsk))
		goto ret;

	/* No problem if kmem_cache_zalloc() fails */
	stats = kmem_cache_zalloc(taskstats_cache, GFP_KERNEL);

	spin_lock_irq(&tsk->sighand->siglock);
	if (!sig->stats) {
		sig->stats = stats;
		stats = NULL;
	}
	spin_unlock_irq(&tsk->sighand->siglock);

	if (stats)
		kmem_cache_free(taskstats_cache, stats);
ret:
	return sig->stats;
}

/* Send pid data out on exit */
void taskstats_exit(struct task_struct *tsk, int group_dead)
{
	int rc;
	struct listener_list *listeners;
	struct taskstats *stats;
	struct sk_buff *rep_skb;
	size_t size;
	int is_thread_group;

	if (!family_registered)
		return;

	/*
	 * Size includes space for nested attributes
	 */
	size = nla_total_size(sizeof(u32)) +
		nla_total_size(sizeof(struct taskstats)) + nla_total_size(0);

	is_thread_group = !!taskstats_tgid_alloc(tsk);
	if (is_thread_group) {
		/* PID + STATS + TGID + STATS */
		size = 2 * size;
		/* fill the tsk->signal->stats structure */
		fill_tgid_exit(tsk);
	}

	listeners = &__raw_get_cpu_var(listener_array);
	if (list_empty(&listeners->list))
		return;

	rc = prepare_reply(NULL, TASKSTATS_CMD_NEW, &rep_skb, size);
	if (rc < 0)
		return;

	stats = mk_reply(rep_skb, TASKSTATS_TYPE_PID, tsk->pid);
	if (!stats)
		goto err;

	rc = fill_pid(tsk->pid, tsk, stats);
	if (rc < 0)
		goto err;

	/*
	 * Doesn't matter if tsk is the leader or the last group member leaving
	 */
	if (!is_thread_group || !group_dead)
		goto send;

	stats = mk_reply(rep_skb, TASKSTATS_TYPE_TGID, tsk->tgid);
	if (!stats)
		goto err;

	memcpy(stats, tsk->signal->stats, sizeof(*stats));

send:
	send_cpu_listeners(rep_skb, listeners);
	return;
err:
	nlmsg_free(rep_skb);
}

static struct genl_ops taskstats_ops = {
	.cmd		= TASKSTATS_CMD_GET,
	.doit		= taskstats_user_cmd,
	.policy		= taskstats_cmd_get_policy,
};

/* Needed early in initialization */
void __init taskstats_init_early(void)
{
	unsigned int i;

	taskstats_cache = KMEM_CACHE(taskstats, SLAB_PANIC);
	for_each_possible_cpu(i) {
		INIT_LIST_HEAD(&(per_cpu(listener_array, i).list));
		init_rwsem(&(per_cpu(listener_array, i).sem));
	}
}

static int __init taskstats_init(void)
{
	int rc;

	rc = genl_register_family(&family);
	if (rc)
		return rc;

	rc = genl_register_ops(&family, &taskstats_ops);
	if (rc < 0)
		goto err;

	family_registered = 1;
	return 0;
err:
	genl_unregister_family(&family);
	return rc;
}

/*
 * late initcall ensures initialization of statistics collection
 * mechanisms precedes initialization of the taskstats interface
 */
late_initcall(taskstats_init);
Commit	Line	Data
c757249a SN	1	/*
	2	* taskstats.c - Export per-task statistics to userland
	3	*
	4	* Copyright (C) Shailabh Nagar, IBM Corp. 2006
	5	* (C) Balbir Singh, IBM Corp. 2006
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License as published by
	9	* the Free Software Foundation; either version 2 of the License, or
	10	* (at your option) any later version.
	11	*
	12	* This program is distributed in the hope that it will be useful,
	13	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	14	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	15	* GNU General Public License for more details.
	16	*
	17	*/
	18
	19	#include <linux/kernel.h>
	20	#include <linux/taskstats_kern.h>
f3cef7a9	21	#include <linux/tsacct_kern.h>
6f44993f	22	#include <linux/delayacct.h>
9acc1853	23	#include <linux/tsacct_kern.h>
f9fd8914 SN	24	#include <linux/cpumask.h>
f9fd8914 SN	25	#include <linux/percpu.h>
c757249a SN	26	#include <net/genetlink.h>
	27	#include <asm/atomic.h>
	28
f9fd8914 SN	29	/*
	30	* Maximum length of a cpumask that can be specified in
	31	* the TASKSTATS_CMD_ATTR_REGISTER/DEREGISTER_CPUMASK attribute
	32	*/
	33	#define TASKSTATS_CPUMASK_MAXLEN (100+6*NR_CPUS)
	34
c757249a SN	35	static DEFINE_PER_CPU(__u32, taskstats_seqnum) = { 0 };
c757249a SN	36	static int family_registered;
e18b890b	37	struct kmem_cache *taskstats_cache;
c757249a SN	38
	39	static struct genl_family family = {
	40	.id = GENL_ID_GENERATE,
	41	.name = TASKSTATS_GENL_NAME,
	42	.version = TASKSTATS_GENL_VERSION,
	43	.maxattr = TASKSTATS_CMD_ATTR_MAX,
	44	};
	45
	46	static struct nla_policy taskstats_cmd_get_policy[TASKSTATS_CMD_ATTR_MAX+1]
	47	__read_mostly = {
	48	[TASKSTATS_CMD_ATTR_PID] = { .type = NLA_U32 },
	49	[TASKSTATS_CMD_ATTR_TGID] = { .type = NLA_U32 },
f9fd8914 SN	50	[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK] = { .type = NLA_STRING },
	51	[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK] = { .type = NLA_STRING },};
	52
	53	struct listener {
	54	struct list_head list;
	55	pid_t pid;
bb129994	56	char valid;
c757249a SN	57	};
c757249a SN	58
f9fd8914 SN	59	struct listener_list {
	60	struct rw_semaphore sem;
	61	struct list_head list;
	62	};
	63	static DEFINE_PER_CPU(struct listener_list, listener_array);
	64
	65	enum actions {
	66	REGISTER,
	67	DEREGISTER,
	68	CPU_DONT_CARE
	69	};
c757249a SN	70
c757249a SN	71	static int prepare_reply(struct genl_info info, u8 cmd, struct sk_buff *skbp,
37167485	72	size_t size)
c757249a SN	73	{
	74	struct sk_buff *skb;
	75	void *reply;
	76
	77	/*
	78	* If new attributes are added, please revisit this allocation
	79	*/
3dabc715	80	skb = genlmsg_new(size, GFP_KERNEL);
c757249a SN	81	if (!skb)
	82	return -ENOMEM;
	83
	84	if (!info) {
	85	int seq = get_cpu_var(taskstats_seqnum)++;
	86	put_cpu_var(taskstats_seqnum);
	87
17c157c8	88	reply = genlmsg_put(skb, 0, seq, &family, 0, cmd);
c757249a	89	} else
17c157c8	90	reply = genlmsg_put_reply(skb, info, &family, 0, cmd);
c757249a SN	91	if (reply == NULL) {
	92	nlmsg_free(skb);
	93	return -EINVAL;
	94	}
	95
	96	*skbp = skb;
c757249a SN	97	return 0;
	98	}
	99
f9fd8914 SN	100	/*
	101	* Send taskstats data in @skb to listener with nl_pid @pid
	102	*/
	103	static int send_reply(struct sk_buff *skb, pid_t pid)
c757249a	104	{
b529ccf2	105	struct genlmsghdr *genlhdr = nlmsg_data(nlmsg_hdr(skb));
f9fd8914	106	void *reply = genlmsg_data(genlhdr);
c757249a SN	107	int rc;
c757249a SN	108
c757249a SN	109	rc = genlmsg_end(skb, reply);
	110	if (rc < 0) {
	111	nlmsg_free(skb);
	112	return rc;
	113	}
	114
c757249a SN	115	return genlmsg_unicast(skb, pid);
	116	}
	117
f9fd8914 SN	118	/*
	119	* Send taskstats data in @skb to listeners registered for @cpu's exit data
	120	*/
115085ea ON	121	static void send_cpu_listeners(struct sk_buff *skb,
115085ea ON	122	struct listener_list *listeners)
f9fd8914	123	{
b529ccf2	124	struct genlmsghdr *genlhdr = nlmsg_data(nlmsg_hdr(skb));
f9fd8914 SN	125	struct listener s, tmp;
	126	struct sk_buff skb_next, skb_cur = skb;
	127	void *reply = genlmsg_data(genlhdr);
d94a0415	128	int rc, delcount = 0;
f9fd8914 SN	129
	130	rc = genlmsg_end(skb, reply);
	131	if (rc < 0) {
	132	nlmsg_free(skb);
d94a0415	133	return;
f9fd8914 SN	134	}
	135
	136	rc = 0;
bb129994	137	down_read(&listeners->sem);
d94a0415	138	list_for_each_entry(s, &listeners->list, list) {
f9fd8914 SN	139	skb_next = NULL;
	140	if (!list_is_last(&s->list, &listeners->list)) {
	141	skb_next = skb_clone(skb_cur, GFP_KERNEL);
d94a0415	142	if (!skb_next)
f9fd8914	143	break;
f9fd8914	144	}
d94a0415 SN	145	rc = genlmsg_unicast(skb_cur, s->pid);
d94a0415 SN	146	if (rc == -ECONNREFUSED) {
bb129994 SN	147	s->valid = 0;
bb129994 SN	148	delcount++;
f9fd8914 SN	149	}
	150	skb_cur = skb_next;
	151	}
bb129994	152	up_read(&listeners->sem);
f9fd8914	153
d94a0415 SN	154	if (skb_cur)
	155	nlmsg_free(skb_cur);
	156
bb129994	157	if (!delcount)
d94a0415	158	return;
bb129994 SN	159
	160	/* Delete invalidated entries */
	161	down_write(&listeners->sem);
	162	list_for_each_entry_safe(s, tmp, &listeners->list, list) {
	163	if (!s->valid) {
	164	list_del(&s->list);
	165	kfree(s);
	166	}
	167	}
	168	up_write(&listeners->sem);
f9fd8914 SN	169	}
f9fd8914 SN	170
a98b6094	171	static int fill_pid(pid_t pid, struct task_struct *tsk,
c757249a SN	172	struct taskstats *stats)
c757249a SN	173	{
7d94dddd	174	int rc = 0;
c757249a	175
a98b6094 ON	176	if (!tsk) {
a98b6094 ON	177	rcu_read_lock();
c757249a	178	tsk = find_task_by_pid(pid);
a98b6094 ON	179	if (tsk)
	180	get_task_struct(tsk);
	181	rcu_read_unlock();
	182	if (!tsk)
c757249a	183	return -ESRCH;
c757249a SN	184	} else
	185	get_task_struct(tsk);
	186
51de4d90	187	memset(stats, 0, sizeof(*stats));
c757249a SN	188	/*
	189	* Each accounting subsystem adds calls to its functions to
	190	* fill in relevant parts of struct taskstsats as follows
	191	*
7d94dddd	192	* per-task-foo(stats, tsk);
c757249a SN	193	*/
c757249a SN	194
7d94dddd	195	delayacct_add_tsk(stats, tsk);
f3cef7a9 JL	196
f3cef7a9 JL	197	/* fill in basic acct fields */
6f44993f	198	stats->version = TASKSTATS_VERSION;
f3cef7a9	199	bacct_add_tsk(stats, tsk);
6f44993f	200
9acc1853 JL	201	/* fill in extended acct fields */
	202	xacct_add_tsk(stats, tsk);
	203
6f44993f	204	/* Define err: label here if needed */
c757249a SN	205	put_task_struct(tsk);
	206	return rc;
	207
	208	}
	209
a98b6094	210	static int fill_tgid(pid_t tgid, struct task_struct *first,
c757249a SN	211	struct taskstats *stats)
c757249a SN	212	{
a98b6094	213	struct task_struct *tsk;
ad4ecbcb	214	unsigned long flags;
a98b6094	215	int rc = -ESRCH;
c757249a	216
ad4ecbcb SN	217	/*
	218	* Add additional stats from live tasks except zombie thread group
	219	* leaders who are already counted with the dead tasks
	220	*/
a98b6094 ON	221	rcu_read_lock();
a98b6094 ON	222	if (!first)
c757249a	223	first = find_task_by_pid(tgid);
ad4ecbcb	224
a98b6094 ON	225	if (!first \|\| !lock_task_sighand(first, &flags))
a98b6094 ON	226	goto out;
ad4ecbcb	227
a98b6094 ON	228	if (first->signal->stats)
a98b6094 ON	229	memcpy(stats, first->signal->stats, sizeof(*stats));
51de4d90 ON	230	else
51de4d90 ON	231	memset(stats, 0, sizeof(*stats));
fca178c0	232
a98b6094	233	tsk = first;
c757249a	234	do {
d7c3f5f2	235	if (tsk->exit_state)
ad4ecbcb	236	continue;
c757249a	237	/*
ad4ecbcb	238	* Accounting subsystem can call its functions here to
c757249a SN	239	* fill in relevant parts of struct taskstsats as follows
c757249a SN	240	*
ad4ecbcb	241	* per-task-foo(stats, tsk);
c757249a	242	*/
ad4ecbcb	243	delayacct_add_tsk(stats, tsk);
6f44993f	244
c757249a	245	} while_each_thread(first, tsk);
6f44993f	246
a98b6094 ON	247	unlock_task_sighand(first, &flags);
	248	rc = 0;
	249	out:
	250	rcu_read_unlock();
	251
	252	stats->version = TASKSTATS_VERSION;
c757249a	253	/*
ad4ecbcb SN	254	* Accounting subsytems can also add calls here to modify
ad4ecbcb SN	255	* fields of taskstats.
c757249a	256	*/
a98b6094	257	return rc;
ad4ecbcb SN	258	}
	259
	260
	261	static void fill_tgid_exit(struct task_struct *tsk)
	262	{
	263	unsigned long flags;
	264
b8534d7b	265	spin_lock_irqsave(&tsk->sighand->siglock, flags);
ad4ecbcb SN	266	if (!tsk->signal->stats)
	267	goto ret;
	268
	269	/*
	270	* Each accounting subsystem calls its functions here to
	271	* accumalate its per-task stats for tsk, into the per-tgid structure
	272	*
	273	* per-task-foo(tsk->signal->stats, tsk);
	274	*/
	275	delayacct_add_tsk(tsk->signal->stats, tsk);
	276	ret:
b8534d7b	277	spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
ad4ecbcb	278	return;
c757249a SN	279	}
c757249a SN	280
f9fd8914 SN	281	static int add_del_listener(pid_t pid, cpumask_t *maskp, int isadd)
	282	{
	283	struct listener_list *listeners;
	284	struct listener s, tmp;
	285	unsigned int cpu;
	286	cpumask_t mask = *maskp;
ad4ecbcb	287
f9fd8914 SN	288	if (!cpus_subset(mask, cpu_possible_map))
	289	return -EINVAL;
	290
	291	if (isadd == REGISTER) {
	292	for_each_cpu_mask(cpu, mask) {
	293	s = kmalloc_node(sizeof(struct listener), GFP_KERNEL,
	294	cpu_to_node(cpu));
	295	if (!s)
	296	goto cleanup;
	297	s->pid = pid;
	298	INIT_LIST_HEAD(&s->list);
bb129994	299	s->valid = 1;
f9fd8914 SN	300
	301	listeners = &per_cpu(listener_array, cpu);
	302	down_write(&listeners->sem);
	303	list_add(&s->list, &listeners->list);
	304	up_write(&listeners->sem);
	305	}
	306	return 0;
	307	}
	308
	309	/* Deregister or cleanup */
	310	cleanup:
	311	for_each_cpu_mask(cpu, mask) {
	312	listeners = &per_cpu(listener_array, cpu);
	313	down_write(&listeners->sem);
	314	list_for_each_entry_safe(s, tmp, &listeners->list, list) {
	315	if (s->pid == pid) {
	316	list_del(&s->list);
	317	kfree(s);
	318	break;
	319	}
	320	}
	321	up_write(&listeners->sem);
	322	}
	323	return 0;
	324	}
	325
	326	static int parse(struct nlattr na, cpumask_t mask)
	327	{
	328	char *data;
	329	int len;
	330	int ret;
	331
	332	if (na == NULL)
	333	return 1;
	334	len = nla_len(na);
	335	if (len > TASKSTATS_CPUMASK_MAXLEN)
	336	return -E2BIG;
	337	if (len < 1)
	338	return -EINVAL;
	339	data = kmalloc(len, GFP_KERNEL);
	340	if (!data)
	341	return -ENOMEM;
	342	nla_strlcpy(data, na, len);
	343	ret = cpulist_parse(data, *mask);
	344	kfree(data);
	345	return ret;
	346	}
	347
51de4d90	348	static struct taskstats mk_reply(struct sk_buff skb, int type, u32 pid)
68062b86	349	{
51de4d90	350	struct nlattr na, ret;
68062b86 ON	351	int aggr;
68062b86 ON	352
37167485 ON	353	aggr = (type == TASKSTATS_TYPE_PID)
	354	? TASKSTATS_TYPE_AGGR_PID
	355	: TASKSTATS_TYPE_AGGR_TGID;
68062b86 ON	356
68062b86 ON	357	na = nla_nest_start(skb, aggr);
37167485 ON	358	if (!na)
37167485 ON	359	goto err;
51de4d90 ON	360	if (nla_put(skb, type, sizeof(pid), &pid) < 0)
	361	goto err;
	362	ret = nla_reserve(skb, TASKSTATS_TYPE_STATS, sizeof(struct taskstats));
	363	if (!ret)
	364	goto err;
68062b86 ON	365	nla_nest_end(skb, na);
68062b86 ON	366
51de4d90 ON	367	return nla_data(ret);
	368	err:
	369	return NULL;
68062b86 ON	370	}
68062b86 ON	371
f9fd8914	372	static int taskstats_user_cmd(struct sk_buff skb, struct genl_info info)
c757249a SN	373	{
	374	int rc = 0;
	375	struct sk_buff *rep_skb;
51de4d90	376	struct taskstats *stats;
c757249a	377	size_t size;
f9fd8914 SN	378	cpumask_t mask;
	379
	380	rc = parse(info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK], &mask);
	381	if (rc < 0)
	382	return rc;
	383	if (rc == 0)
	384	return add_del_listener(info->snd_pid, &mask, REGISTER);
	385
	386	rc = parse(info->attrs[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK], &mask);
	387	if (rc < 0)
	388	return rc;
	389	if (rc == 0)
	390	return add_del_listener(info->snd_pid, &mask, DEREGISTER);
c757249a SN	391
	392	/*
	393	* Size includes space for nested attributes
	394	*/
	395	size = nla_total_size(sizeof(u32)) +
	396	nla_total_size(sizeof(struct taskstats)) + nla_total_size(0);
	397
37167485	398	rc = prepare_reply(info, TASKSTATS_CMD_NEW, &rep_skb, size);
c757249a SN	399	if (rc < 0)
	400	return rc;
	401
51de4d90	402	rc = -EINVAL;
c757249a SN	403	if (info->attrs[TASKSTATS_CMD_ATTR_PID]) {
c757249a SN	404	u32 pid = nla_get_u32(info->attrs[TASKSTATS_CMD_ATTR_PID]);
51de4d90 ON	405	stats = mk_reply(rep_skb, TASKSTATS_TYPE_PID, pid);
51de4d90 ON	406	if (!stats)
37167485	407	goto err;
c757249a	408
51de4d90 ON	409	rc = fill_pid(pid, NULL, stats);
51de4d90 ON	410	if (rc < 0)
37167485	411	goto err;
c757249a SN	412	} else if (info->attrs[TASKSTATS_CMD_ATTR_TGID]) {
c757249a SN	413	u32 tgid = nla_get_u32(info->attrs[TASKSTATS_CMD_ATTR_TGID]);
51de4d90 ON	414	stats = mk_reply(rep_skb, TASKSTATS_TYPE_TGID, tgid);
51de4d90 ON	415	if (!stats)
37167485	416	goto err;
c757249a	417
51de4d90 ON	418	rc = fill_tgid(tgid, NULL, stats);
51de4d90 ON	419	if (rc < 0)
37167485	420	goto err;
51de4d90	421	} else
c757249a	422	goto err;
c757249a	423
f9fd8914	424	return send_reply(rep_skb, info->snd_pid);
c757249a SN	425	err:
	426	nlmsg_free(rep_skb);
	427	return rc;
	428	}
	429
34ec1234 ON	430	static struct taskstats taskstats_tgid_alloc(struct task_struct tsk)
	431	{
	432	struct signal_struct *sig = tsk->signal;
	433	struct taskstats *stats;
	434
	435	if (sig->stats \|\| thread_group_empty(tsk))
	436	goto ret;
	437
	438	/* No problem if kmem_cache_zalloc() fails */
	439	stats = kmem_cache_zalloc(taskstats_cache, GFP_KERNEL);
	440
	441	spin_lock_irq(&tsk->sighand->siglock);
	442	if (!sig->stats) {
	443	sig->stats = stats;
	444	stats = NULL;
	445	}
	446	spin_unlock_irq(&tsk->sighand->siglock);
	447
	448	if (stats)
	449	kmem_cache_free(taskstats_cache, stats);
	450	ret:
	451	return sig->stats;
	452	}
	453
c757249a	454	/* Send pid data out on exit */
115085ea	455	void taskstats_exit(struct task_struct *tsk, int group_dead)
c757249a SN	456	{
c757249a SN	457	int rc;
115085ea	458	struct listener_list *listeners;
51de4d90	459	struct taskstats *stats;
c757249a	460	struct sk_buff *rep_skb;
c757249a SN	461	size_t size;
c757249a SN	462	int is_thread_group;
c757249a	463
4a279ff1	464	if (!family_registered)
c757249a SN	465	return;
c757249a SN	466
c757249a SN	467	/*
	468	* Size includes space for nested attributes
	469	*/
	470	size = nla_total_size(sizeof(u32)) +
	471	nla_total_size(sizeof(struct taskstats)) + nla_total_size(0);
	472
34ec1234	473	is_thread_group = !!taskstats_tgid_alloc(tsk);
4a279ff1 ON	474	if (is_thread_group) {
	475	/* PID + STATS + TGID + STATS */
	476	size = 2 * size;
	477	/* fill the tsk->signal->stats structure */
	478	fill_tgid_exit(tsk);
	479	}
	480
115085ea ON	481	listeners = &__raw_get_cpu_var(listener_array);
	482	if (list_empty(&listeners->list))
	483	return;
	484
37167485	485	rc = prepare_reply(NULL, TASKSTATS_CMD_NEW, &rep_skb, size);
c757249a	486	if (rc < 0)
51de4d90	487	return;
c757249a	488
51de4d90 ON	489	stats = mk_reply(rep_skb, TASKSTATS_TYPE_PID, tsk->pid);
51de4d90 ON	490	if (!stats)
37167485	491	goto err;
c757249a	492
51de4d90 ON	493	rc = fill_pid(tsk->pid, tsk, stats);
51de4d90 ON	494	if (rc < 0)
37167485	495	goto err;
c757249a	496
c757249a	497	/*
ad4ecbcb	498	* Doesn't matter if tsk is the leader or the last group member leaving
c757249a	499	*/
68062b86	500	if (!is_thread_group \|\| !group_dead)
ad4ecbcb	501	goto send;
c757249a	502
51de4d90 ON	503	stats = mk_reply(rep_skb, TASKSTATS_TYPE_TGID, tsk->tgid);
51de4d90 ON	504	if (!stats)
37167485	505	goto err;
51de4d90 ON	506
51de4d90 ON	507	memcpy(stats, tsk->signal->stats, sizeof(*stats));
c757249a	508
ad4ecbcb	509	send:
115085ea	510	send_cpu_listeners(rep_skb, listeners);
ad4ecbcb	511	return;
37167485	512	err:
c757249a	513	nlmsg_free(rep_skb);
c757249a SN	514	}
	515
	516	static struct genl_ops taskstats_ops = {
	517	.cmd = TASKSTATS_CMD_GET,
f9fd8914	518	.doit = taskstats_user_cmd,
c757249a SN	519	.policy = taskstats_cmd_get_policy,
	520	};
	521
	522	/* Needed early in initialization */
	523	void __init taskstats_init_early(void)
	524	{
f9fd8914 SN	525	unsigned int i;
f9fd8914 SN	526
0a31bd5f	527	taskstats_cache = KMEM_CACHE(taskstats, SLAB_PANIC);
f9fd8914 SN	528	for_each_possible_cpu(i) {
	529	INIT_LIST_HEAD(&(per_cpu(listener_array, i).list));
	530	init_rwsem(&(per_cpu(listener_array, i).sem));
	531	}
c757249a SN	532	}
	533
	534	static int __init taskstats_init(void)
	535	{
	536	int rc;
	537
	538	rc = genl_register_family(&family);
	539	if (rc)
	540	return rc;
	541
	542	rc = genl_register_ops(&family, &taskstats_ops);
	543	if (rc < 0)
	544	goto err;
	545
	546	family_registered = 1;
	547	return 0;
	548	err:
	549	genl_unregister_family(&family);
	550	return rc;
	551	}
	552
	553	/*
	554	* late initcall ensures initialization of statistics collection
	555	* mechanisms precedes initialization of the taskstats interface
	556	*/
	557	late_initcall(taskstats_init);