summaryrefslogtreecommitdiff
path: root/kernel/locking/qspinlock_paravirt.h
blob: 04ab18151cc8fa174a5859124ee07c144f33d505 (plain)
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
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
#ifndef _GEN_PV_LOCK_SLOWPATH
#error "do not include this file"
#endif

#include <linux/hash.h>
#include <linux/bootmem.h>

/*
 * Implement paravirt qspinlocks; the general idea is to halt the vcpus instead
 * of spinning them.
 *
 * This relies on the architecture to provide two paravirt hypercalls:
 *
 *   pv_wait(u8 *ptr, u8 val) -- suspends the vcpu if *ptr == val
 *   pv_kick(cpu)             -- wakes a suspended vcpu
 *
 * Using these we implement __pv_queued_spin_lock_slowpath() and
 * __pv_queued_spin_unlock() to replace native_queued_spin_lock_slowpath() and
 * native_queued_spin_unlock().
 */

#define _Q_SLOW_VAL	(3U << _Q_LOCKED_OFFSET)

enum vcpu_state {
	vcpu_running = 0,
	vcpu_halted,
};

struct pv_node {
	struct mcs_spinlock	mcs;
	struct mcs_spinlock	__res[3];

	int			cpu;
	u8			state;
};

/*
 * Lock and MCS node addresses hash table for fast lookup
 *
 * Hashing is done on a per-cacheline basis to minimize the need to access
 * more than one cacheline.
 *
 * Dynamically allocate a hash table big enough to hold at least 4X the
 * number of possible cpus in the system. Allocation is done on page
 * granularity. So the minimum number of hash buckets should be at least
 * 256 (64-bit) or 512 (32-bit) to fully utilize a 4k page.
 *
 * Since we should not be holding locks from NMI context (very rare indeed) the
 * max load factor is 0.75, which is around the point where open addressing
 * breaks down.
 *
 */
struct pv_hash_entry {
	struct qspinlock *lock;
	struct pv_node   *node;
};

#define PV_HE_PER_LINE	(SMP_CACHE_BYTES / sizeof(struct pv_hash_entry))
#define PV_HE_MIN	(PAGE_SIZE / sizeof(struct pv_hash_entry))

static struct pv_hash_entry *pv_lock_hash;
static unsigned int pv_lock_hash_bits __read_mostly;

/*
 * Allocate memory for the PV qspinlock hash buckets
 *
 * This function should be called from the paravirt spinlock initialization
 * routine.
 */
void __init __pv_init_lock_hash(void)
{
	int pv_hash_size = ALIGN(4 * num_possible_cpus(), PV_HE_PER_LINE);

	if (pv_hash_size < PV_HE_MIN)
		pv_hash_size = PV_HE_MIN;

	/*
	 * Allocate space from bootmem which should be page-size aligned
	 * and hence cacheline aligned.
	 */
	pv_lock_hash = alloc_large_system_hash("PV qspinlock",
					       sizeof(struct pv_hash_entry),
					       pv_hash_size, 0, HASH_EARLY,
					       &pv_lock_hash_bits, NULL,
					       pv_hash_size, pv_hash_size);
}

#define for_each_hash_entry(he, offset, hash)						\
	for (hash &= ~(PV_HE_PER_LINE - 1), he = &pv_lock_hash[hash], offset = 0;	\
	     offset < (1 << pv_lock_hash_bits);						\
	     offset++, he = &pv_lock_hash[(hash + offset) & ((1 << pv_lock_hash_bits) - 1)])

static struct qspinlock **pv_hash(struct qspinlock *lock, struct pv_node *node)
{
	unsigned long offset, hash = hash_ptr(lock, pv_lock_hash_bits);
	struct pv_hash_entry *he;

	for_each_hash_entry(he, offset, hash) {
		if (!cmpxchg(&he->lock, NULL, lock)) {
			WRITE_ONCE(he->node, node);
			return &he->lock;
		}
	}
	/*
	 * Hard assume there is a free entry for us.
	 *
	 * This is guaranteed by ensuring every blocked lock only ever consumes
	 * a single entry, and since we only have 4 nesting levels per CPU
	 * and allocated 4*nr_possible_cpus(), this must be so.
	 *
	 * The single entry is guaranteed by having the lock owner unhash
	 * before it releases.
	 */
	BUG();
}

static struct pv_node *pv_unhash(struct qspinlock *lock)
{
	unsigned long offset, hash = hash_ptr(lock, pv_lock_hash_bits);
	struct pv_hash_entry *he;
	struct pv_node *node;

	for_each_hash_entry(he, offset, hash) {
		if (READ_ONCE(he->lock) == lock) {
			node = READ_ONCE(he->node);
			WRITE_ONCE(he->lock, NULL);
			return node;
		}
	}
	/*
	 * Hard assume we'll find an entry.
	 *
	 * This guarantees a limited lookup time and is itself guaranteed by
	 * having the lock owner do the unhash -- IFF the unlock sees the
	 * SLOW flag, there MUST be a hash entry.
	 */
	BUG();
}

/*
 * Initialize the PV part of the mcs_spinlock node.
 */
static void pv_init_node(struct mcs_spinlock *node)
{
	struct pv_node *pn = (struct pv_node *)node;

	BUILD_BUG_ON(sizeof(struct pv_node) > 5*sizeof(struct mcs_spinlock));

	pn->cpu = smp_processor_id();
	pn->state = vcpu_running;
}

/*
 * Wait for node->locked to become true, halt the vcpu after a short spin.
 * pv_kick_node() is used to wake the vcpu again.
 */
static void pv_wait_node(struct mcs_spinlock *node)
{
	struct pv_node *pn = (struct pv_node *)node;
	int loop;

	for (;;) {
		for (loop = SPIN_THRESHOLD; loop; loop--) {
			if (READ_ONCE(node->locked))
				return;
			cpu_relax();
		}

		/*
		 * Order pn->state vs pn->locked thusly:
		 *
		 * [S] pn->state = vcpu_halted	  [S] next->locked = 1
		 *     MB			      MB
		 * [L] pn->locked		[RmW] pn->state = vcpu_running
		 *
		 * Matches the xchg() from pv_kick_node().
		 */
		smp_store_mb(pn->state, vcpu_halted);

		if (!READ_ONCE(node->locked))
			pv_wait(&pn->state, vcpu_halted);

		/*
		 * Reset the vCPU state to avoid unncessary CPU kicking
		 */
		WRITE_ONCE(pn->state, vcpu_running);

		/*
		 * If the locked flag is still not set after wakeup, it is a
		 * spurious wakeup and the vCPU should wait again. However,
		 * there is a pretty high overhead for CPU halting and kicking.
		 * So it is better to spin for a while in the hope that the
		 * MCS lock will be released soon.
		 */
	}
	/*
	 * By now our node->locked should be 1 and our caller will not actually
	 * spin-wait for it. We do however rely on our caller to do a
	 * load-acquire for us.
	 */
}

/*
 * Called after setting next->locked = 1, used to wake those stuck in
 * pv_wait_node().
 */
static void pv_kick_node(struct mcs_spinlock *node)
{
	struct pv_node *pn = (struct pv_node *)node;

	/*
	 * Note that because node->locked is already set, this actual
	 * mcs_spinlock entry could be re-used already.
	 *
	 * This should be fine however, kicking people for no reason is
	 * harmless.
	 *
	 * See the comment in pv_wait_node().
	 */
	if (xchg(&pn->state, vcpu_running) == vcpu_halted)
		pv_kick(pn->cpu);
}

/*
 * Wait for l->locked to become clear; halt the vcpu after a short spin.
 * __pv_queued_spin_unlock() will wake us.
 */
static void pv_wait_head(struct qspinlock *lock, struct mcs_spinlock *node)
{
	struct pv_node *pn = (struct pv_node *)node;
	struct __qspinlock *l = (void *)lock;
	struct qspinlock **lp = NULL;
	int loop;

	for (;;) {
		for (loop = SPIN_THRESHOLD; loop; loop--) {
			if (!READ_ONCE(l->locked))
				return;
			cpu_relax();
		}

		WRITE_ONCE(pn->state, vcpu_halted);
		if (!lp) { /* ONCE */
			lp = pv_hash(lock, pn);
			/*
			 * lp must be set before setting _Q_SLOW_VAL
			 *
			 * [S] lp = lock                [RmW] l = l->locked = 0
			 *     MB                             MB
			 * [S] l->locked = _Q_SLOW_VAL  [L]   lp
			 *
			 * Matches the cmpxchg() in __pv_queued_spin_unlock().
			 */
			if (!cmpxchg(&l->locked, _Q_LOCKED_VAL, _Q_SLOW_VAL)) {
				/*
				 * The lock is free and _Q_SLOW_VAL has never
				 * been set. Therefore we need to unhash before
				 * getting the lock.
				 */
				WRITE_ONCE(*lp, NULL);
				return;
			}
		}
		pv_wait(&l->locked, _Q_SLOW_VAL);

		/*
		 * The unlocker should have freed the lock before kicking the
		 * CPU. So if the lock is still not free, it is a spurious
		 * wakeup and so the vCPU should wait again after spinning for
		 * a while.
		 */
	}

	/*
	 * Lock is unlocked now; the caller will acquire it without waiting.
	 * As with pv_wait_node() we rely on the caller to do a load-acquire
	 * for us.
	 */
}

/*
 * PV version of the unlock function to be used in stead of
 * queued_spin_unlock().
 */
__visible void __pv_queued_spin_unlock(struct qspinlock *lock)
{
	struct __qspinlock *l = (void *)lock;
	struct pv_node *node;

	/*
	 * We must not unlock if SLOW, because in that case we must first
	 * unhash. Otherwise it would be possible to have multiple @lock
	 * entries, which would be BAD.
	 */
	if (likely(cmpxchg(&l->locked, _Q_LOCKED_VAL, 0) == _Q_LOCKED_VAL))
		return;

	/*
	 * Since the above failed to release, this must be the SLOW path.
	 * Therefore start by looking up the blocked node and unhashing it.
	 */
	node = pv_unhash(lock);

	/*
	 * Now that we have a reference to the (likely) blocked pv_node,
	 * release the lock.
	 */
	smp_store_release(&l->locked, 0);

	/*
	 * At this point the memory pointed at by lock can be freed/reused,
	 * however we can still use the pv_node to kick the CPU.
	 */
	if (READ_ONCE(node->state) == vcpu_halted)
		pv_kick(node->cpu);
}
/*
 * Include the architecture specific callee-save thunk of the
 * __pv_queued_spin_unlock(). This thunk is put together with
 * __pv_queued_spin_unlock() near the top of the file to make sure
 * that the callee-save thunk and the real unlock function are close
 * to each other sharing consecutive instruction cachelines.
 */
#include <asm/qspinlock_paravirt.h>