percpu-refcount.h 6.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
/*
 * Percpu refcounts:
 * (C) 2012 Google, Inc.
 * Author: Kent Overstreet <koverstreet@google.com>
 *
 * This implements a refcount with similar semantics to atomic_t - atomic_inc(),
 * atomic_dec_and_test() - but percpu.
 *
 * There's one important difference between percpu refs and normal atomic_t
 * refcounts; you have to keep track of your initial refcount, and then when you
 * start shutting down you call percpu_ref_kill() _before_ dropping the initial
 * refcount.
 *
 * The refcount will have a range of 0 to ((1U << 31) - 1), i.e. one bit less
 * than an atomic_t - this is because of the way shutdown works, see
 * percpu_ref_kill()/PCPU_COUNT_BIAS.
 *
 * Before you call percpu_ref_kill(), percpu_ref_put() does not check for the
 * refcount hitting 0 - it can't, if it was in percpu mode. percpu_ref_kill()
 * puts the ref back in single atomic_t mode, collecting the per cpu refs and
 * issuing the appropriate barriers, and then marks the ref as shutting down so
 * that percpu_ref_put() will check for the ref hitting 0.  After it returns,
 * it's safe to drop the initial ref.
 *
 * USAGE:
 *
 * See fs/aio.c for some example usage; it's used there for struct kioctx, which
 * is created when userspaces calls io_setup(), and destroyed when userspace
 * calls io_destroy() or the process exits.
 *
 * In the aio code, kill_ioctx() is called when we wish to destroy a kioctx; it
 * calls percpu_ref_kill(), then hlist_del_rcu() and sychronize_rcu() to remove
 * the kioctx from the proccess's list of kioctxs - after that, there can't be
 * any new users of the kioctx (from lookup_ioctx()) and it's then safe to drop
 * the initial ref with percpu_ref_put().
 *
 * Code that does a two stage shutdown like this often needs some kind of
 * explicit synchronization to ensure the initial refcount can only be dropped
 * once - percpu_ref_kill() does this for you, it returns true once and false if
 * someone else already called it. The aio code uses it this way, but it's not
 * necessary if the code has some other mechanism to synchronize teardown.
 * around.
 */

#ifndef _LINUX_PERCPU_REFCOUNT_H
#define _LINUX_PERCPU_REFCOUNT_H

#include <linux/atomic.h>
#include <linux/kernel.h>
#include <linux/percpu.h>
#include <linux/rcupdate.h>
52
#include <linux/gfp.h>
53 54

struct percpu_ref;
T
Tejun Heo 已提交
55
typedef void (percpu_ref_func_t)(struct percpu_ref *);
56 57

struct percpu_ref {
58
	atomic_long_t		count;
59 60
	/*
	 * The low bit of the pointer indicates whether the ref is in percpu
61
	 * mode; if set, then get/put will manipulate the atomic_t.
62
	 */
63
	unsigned long		pcpu_count_ptr;
T
Tejun Heo 已提交
64
	percpu_ref_func_t	*release;
65
	percpu_ref_func_t	*confirm_kill;
66 67 68
	struct rcu_head		rcu;
};

69
int __must_check percpu_ref_init(struct percpu_ref *ref,
70
				 percpu_ref_func_t *release, gfp_t gfp);
71
void percpu_ref_exit(struct percpu_ref *ref);
72 73
void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
				 percpu_ref_func_t *confirm_kill);
74
void percpu_ref_reinit(struct percpu_ref *ref);
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89

/**
 * percpu_ref_kill - drop the initial ref
 * @ref: percpu_ref to kill
 *
 * Must be used to drop the initial ref on a percpu refcount; must be called
 * precisely once before shutdown.
 *
 * Puts @ref in non percpu mode, then does a call_rcu() before gathering up the
 * percpu counters and dropping the initial ref.
 */
static inline void percpu_ref_kill(struct percpu_ref *ref)
{
	return percpu_ref_kill_and_confirm(ref, NULL);
}
90 91 92

#define PCPU_REF_DEAD		1

93 94 95 96 97 98 99
/*
 * Internal helper.  Don't use outside percpu-refcount proper.  The
 * function doesn't return the pointer and let the caller test it for NULL
 * because doing so forces the compiler to generate two conditional
 * branches as it can't assume that @ref->pcpu_count is not NULL.
 */
static inline bool __pcpu_ref_alive(struct percpu_ref *ref,
100
				    unsigned long __percpu **pcpu_countp)
101
{
102
	unsigned long pcpu_ptr = ACCESS_ONCE(ref->pcpu_count_ptr);
103

104 105 106
	/* paired with smp_store_release() in percpu_ref_reinit() */
	smp_read_barrier_depends();

107 108 109
	if (unlikely(pcpu_ptr & PCPU_REF_DEAD))
		return false;

110
	*pcpu_countp = (unsigned long __percpu *)pcpu_ptr;
111 112
	return true;
}
113 114 115

/**
 * percpu_ref_get - increment a percpu refcount
T
Tejun Heo 已提交
116
 * @ref: percpu_ref to get
117
 *
118 119 120 121
 * Analagous to atomic_long_inc().
 *
 * This function is safe to call as long as @ref is between init and exit.
 */
122 123
static inline void percpu_ref_get(struct percpu_ref *ref)
{
124
	unsigned long __percpu *pcpu_count;
125

126
	rcu_read_lock_sched();
127

128
	if (__pcpu_ref_alive(ref, &pcpu_count))
129
		this_cpu_inc(*pcpu_count);
130
	else
131
		atomic_long_inc(&ref->count);
132

133
	rcu_read_unlock_sched();
134 135
}

136 137 138 139 140 141 142
/**
 * percpu_ref_tryget - try to increment a percpu refcount
 * @ref: percpu_ref to try-get
 *
 * Increment a percpu refcount unless its count already reached zero.
 * Returns %true on success; %false on failure.
 *
143
 * This function is safe to call as long as @ref is between init and exit.
144 145 146
 */
static inline bool percpu_ref_tryget(struct percpu_ref *ref)
{
147
	unsigned long __percpu *pcpu_count;
148
	int ret;
149 150 151

	rcu_read_lock_sched();

152
	if (__pcpu_ref_alive(ref, &pcpu_count)) {
T
Tejun Heo 已提交
153
		this_cpu_inc(*pcpu_count);
154 155
		ret = true;
	} else {
156
		ret = atomic_long_inc_not_zero(&ref->count);
157 158 159 160 161 162 163
	}

	rcu_read_unlock_sched();

	return ret;
}

164
/**
165
 * percpu_ref_tryget_live - try to increment a live percpu refcount
166 167 168 169 170
 * @ref: percpu_ref to try-get
 *
 * Increment a percpu refcount unless it has already been killed.  Returns
 * %true on success; %false on failure.
 *
171 172 173 174 175
 * Completion of percpu_ref_kill() in itself doesn't guarantee that this
 * function will fail.  For such guarantee, percpu_ref_kill_and_confirm()
 * should be used.  After the confirm_kill callback is invoked, it's
 * guaranteed that no new reference will be given out by
 * percpu_ref_tryget_live().
176
 *
177
 * This function is safe to call as long as @ref is between init and exit.
178
 */
179
static inline bool percpu_ref_tryget_live(struct percpu_ref *ref)
180
{
181
	unsigned long __percpu *pcpu_count;
182 183
	int ret = false;

184
	rcu_read_lock_sched();
185

186
	if (__pcpu_ref_alive(ref, &pcpu_count)) {
187
		this_cpu_inc(*pcpu_count);
188 189 190
		ret = true;
	}

191
	rcu_read_unlock_sched();
192 193 194 195

	return ret;
}

196 197
/**
 * percpu_ref_put - decrement a percpu refcount
T
Tejun Heo 已提交
198
 * @ref: percpu_ref to put
199 200 201
 *
 * Decrement the refcount, and if 0, call the release function (which was passed
 * to percpu_ref_init())
202 203
 *
 * This function is safe to call as long as @ref is between init and exit.
204 205 206
 */
static inline void percpu_ref_put(struct percpu_ref *ref)
{
207
	unsigned long __percpu *pcpu_count;
208

209
	rcu_read_lock_sched();
210

211
	if (__pcpu_ref_alive(ref, &pcpu_count))
212
		this_cpu_dec(*pcpu_count);
213
	else if (unlikely(atomic_long_dec_and_test(&ref->count)))
214 215
		ref->release(ref);

216
	rcu_read_unlock_sched();
217 218
}

219 220 221 222 223
/**
 * percpu_ref_is_zero - test whether a percpu refcount reached zero
 * @ref: percpu_ref to test
 *
 * Returns %true if @ref reached zero.
224 225
 *
 * This function is safe to call as long as @ref is between init and exit.
226 227 228
 */
static inline bool percpu_ref_is_zero(struct percpu_ref *ref)
{
229
	unsigned long __percpu *pcpu_count;
230 231 232

	if (__pcpu_ref_alive(ref, &pcpu_count))
		return false;
233
	return !atomic_long_read(&ref->count);
234 235
}

236
#endif