2 * QemuLockCnt implementation
4 * Copyright Red Hat, Inc. 2017
7 * Paolo Bonzini <pbonzini@redhat.com>
9 #include "qemu/osdep.h"
10 #include "qemu/thread.h"
11 #include "qemu/atomic.h"
15 #include "qemu/futex.h"
17 /* On Linux, bits 0-1 are a futex-based lock, bits 2-31 are the counter.
18 * For the mutex algorithm see Ulrich Drepper's "Futexes Are Tricky" (ok,
19 * this is not the most relaxing citation I could make...). It is similar
20 * to mutex2 in the paper.
23 #define QEMU_LOCKCNT_STATE_MASK 3
24 #define QEMU_LOCKCNT_STATE_FREE 0 /* free, uncontended */
25 #define QEMU_LOCKCNT_STATE_LOCKED 1 /* locked, uncontended */
26 #define QEMU_LOCKCNT_STATE_WAITING 2 /* locked, contended */
28 #define QEMU_LOCKCNT_COUNT_STEP 4
29 #define QEMU_LOCKCNT_COUNT_SHIFT 2
31 void qemu_lockcnt_init(QemuLockCnt
*lockcnt
)
36 void qemu_lockcnt_destroy(QemuLockCnt
*lockcnt
)
40 /* *val is the current value of lockcnt->count.
42 * If the lock is free, try a cmpxchg from *val to new_if_free; return
43 * true and set *val to the old value found by the cmpxchg in
46 * If the lock is taken, wait for it to be released and return false
47 * *without trying again to take the lock*. Again, set *val to the
48 * new value of lockcnt->count.
50 * If *waited is true on return, new_if_free's bottom two bits must not
51 * be QEMU_LOCKCNT_STATE_LOCKED on subsequent calls, because the caller
52 * does not know if there are other waiters. Furthermore, after *waited
53 * is set the caller has effectively acquired the lock. If it returns
54 * with the lock not taken, it must wake another futex waiter.
56 static bool qemu_lockcnt_cmpxchg_or_wait(QemuLockCnt
*lockcnt
, int *val
,
57 int new_if_free
, bool *waited
)
59 /* Fast path for when the lock is free. */
60 if ((*val
& QEMU_LOCKCNT_STATE_MASK
) == QEMU_LOCKCNT_STATE_FREE
) {
63 trace_lockcnt_fast_path_attempt(lockcnt
, expected
, new_if_free
);
64 *val
= qatomic_cmpxchg(&lockcnt
->count
, expected
, new_if_free
);
65 if (*val
== expected
) {
66 trace_lockcnt_fast_path_success(lockcnt
, expected
, new_if_free
);
72 /* The slow path moves from locked to waiting if necessary, then
73 * does a futex wait. Both steps can be repeated ad nauseam,
74 * only getting out of the loop if we can have another shot at the
75 * fast path. Once we can, get out to compute the new destination
76 * value for the fast path.
78 while ((*val
& QEMU_LOCKCNT_STATE_MASK
) != QEMU_LOCKCNT_STATE_FREE
) {
79 if ((*val
& QEMU_LOCKCNT_STATE_MASK
) == QEMU_LOCKCNT_STATE_LOCKED
) {
81 int new = expected
- QEMU_LOCKCNT_STATE_LOCKED
+ QEMU_LOCKCNT_STATE_WAITING
;
83 trace_lockcnt_futex_wait_prepare(lockcnt
, expected
, new);
84 *val
= qatomic_cmpxchg(&lockcnt
->count
, expected
, new);
85 if (*val
== expected
) {
91 if ((*val
& QEMU_LOCKCNT_STATE_MASK
) == QEMU_LOCKCNT_STATE_WAITING
) {
93 trace_lockcnt_futex_wait(lockcnt
, *val
);
94 qemu_futex_wait(&lockcnt
->count
, *val
);
95 *val
= qatomic_read(&lockcnt
->count
);
96 trace_lockcnt_futex_wait_resume(lockcnt
, *val
);
105 static void lockcnt_wake(QemuLockCnt
*lockcnt
)
107 trace_lockcnt_futex_wake(lockcnt
);
108 qemu_futex_wake(&lockcnt
->count
, 1);
111 void qemu_lockcnt_inc(QemuLockCnt
*lockcnt
)
113 int val
= qatomic_read(&lockcnt
->count
);
117 if (val
>= QEMU_LOCKCNT_COUNT_STEP
) {
119 val
= qatomic_cmpxchg(&lockcnt
->count
, val
,
120 val
+ QEMU_LOCKCNT_COUNT_STEP
);
121 if (val
== expected
) {
125 /* The fast path is (0, unlocked)->(1, unlocked). */
126 if (qemu_lockcnt_cmpxchg_or_wait(lockcnt
, &val
, QEMU_LOCKCNT_COUNT_STEP
,
133 /* If we were woken by another thread, we should also wake one because
134 * we are effectively releasing the lock that was given to us. This is
135 * the case where qemu_lockcnt_lock would leave QEMU_LOCKCNT_STATE_WAITING
136 * in the low bits, and qemu_lockcnt_inc_and_unlock would find it and
140 lockcnt_wake(lockcnt
);
144 void qemu_lockcnt_dec(QemuLockCnt
*lockcnt
)
146 qatomic_sub(&lockcnt
->count
, QEMU_LOCKCNT_COUNT_STEP
);
149 /* Decrement a counter, and return locked if it is decremented to zero.
150 * If the function returns true, it is impossible for the counter to
151 * become nonzero until the next qemu_lockcnt_unlock.
153 bool qemu_lockcnt_dec_and_lock(QemuLockCnt
*lockcnt
)
155 int val
= qatomic_read(&lockcnt
->count
);
156 int locked_state
= QEMU_LOCKCNT_STATE_LOCKED
;
160 if (val
>= 2 * QEMU_LOCKCNT_COUNT_STEP
) {
162 val
= qatomic_cmpxchg(&lockcnt
->count
, val
,
163 val
- QEMU_LOCKCNT_COUNT_STEP
);
164 if (val
== expected
) {
168 /* If count is going 1->0, take the lock. The fast path is
169 * (1, unlocked)->(0, locked) or (1, unlocked)->(0, waiting).
171 if (qemu_lockcnt_cmpxchg_or_wait(lockcnt
, &val
, locked_state
, &waited
)) {
176 /* At this point we do not know if there are more waiters. Assume
179 locked_state
= QEMU_LOCKCNT_STATE_WAITING
;
184 /* If we were woken by another thread, but we're returning in unlocked
185 * state, we should also wake a thread because we are effectively
186 * releasing the lock that was given to us. This is the case where
187 * qemu_lockcnt_lock would leave QEMU_LOCKCNT_STATE_WAITING in the low
188 * bits, and qemu_lockcnt_unlock would find it and wake someone.
191 lockcnt_wake(lockcnt
);
196 /* If the counter is one, decrement it and return locked. Otherwise do
199 * If the function returns true, it is impossible for the counter to
200 * become nonzero until the next qemu_lockcnt_unlock.
202 bool qemu_lockcnt_dec_if_lock(QemuLockCnt
*lockcnt
)
204 int val
= qatomic_read(&lockcnt
->count
);
205 int locked_state
= QEMU_LOCKCNT_STATE_LOCKED
;
208 while (val
< 2 * QEMU_LOCKCNT_COUNT_STEP
) {
209 /* If count is going 1->0, take the lock. The fast path is
210 * (1, unlocked)->(0, locked) or (1, unlocked)->(0, waiting).
212 if (qemu_lockcnt_cmpxchg_or_wait(lockcnt
, &val
, locked_state
, &waited
)) {
217 /* At this point we do not know if there are more waiters. Assume
220 locked_state
= QEMU_LOCKCNT_STATE_WAITING
;
224 /* If we were woken by another thread, but we're returning in unlocked
225 * state, we should also wake a thread because we are effectively
226 * releasing the lock that was given to us. This is the case where
227 * qemu_lockcnt_lock would leave QEMU_LOCKCNT_STATE_WAITING in the low
228 * bits, and qemu_lockcnt_inc_and_unlock would find it and wake someone.
231 lockcnt_wake(lockcnt
);
236 void qemu_lockcnt_lock(QemuLockCnt
*lockcnt
)
238 int val
= qatomic_read(&lockcnt
->count
);
239 int step
= QEMU_LOCKCNT_STATE_LOCKED
;
242 /* The third argument is only used if the low bits of val are 0
243 * (QEMU_LOCKCNT_STATE_FREE), so just blindly mix in the desired
246 while (!qemu_lockcnt_cmpxchg_or_wait(lockcnt
, &val
, val
+ step
, &waited
)) {
248 /* At this point we do not know if there are more waiters. Assume
251 step
= QEMU_LOCKCNT_STATE_WAITING
;
256 void qemu_lockcnt_inc_and_unlock(QemuLockCnt
*lockcnt
)
258 int expected
, new, val
;
260 val
= qatomic_read(&lockcnt
->count
);
263 new = (val
+ QEMU_LOCKCNT_COUNT_STEP
) & ~QEMU_LOCKCNT_STATE_MASK
;
264 trace_lockcnt_unlock_attempt(lockcnt
, val
, new);
265 val
= qatomic_cmpxchg(&lockcnt
->count
, val
, new);
266 } while (val
!= expected
);
268 trace_lockcnt_unlock_success(lockcnt
, val
, new);
269 if (val
& QEMU_LOCKCNT_STATE_WAITING
) {
270 lockcnt_wake(lockcnt
);
274 void qemu_lockcnt_unlock(QemuLockCnt
*lockcnt
)
276 int expected
, new, val
;
278 val
= qatomic_read(&lockcnt
->count
);
281 new = val
& ~QEMU_LOCKCNT_STATE_MASK
;
282 trace_lockcnt_unlock_attempt(lockcnt
, val
, new);
283 val
= qatomic_cmpxchg(&lockcnt
->count
, val
, new);
284 } while (val
!= expected
);
286 trace_lockcnt_unlock_success(lockcnt
, val
, new);
287 if (val
& QEMU_LOCKCNT_STATE_WAITING
) {
288 lockcnt_wake(lockcnt
);
292 unsigned qemu_lockcnt_count(QemuLockCnt
*lockcnt
)
294 return qatomic_read(&lockcnt
->count
) >> QEMU_LOCKCNT_COUNT_SHIFT
;
297 void qemu_lockcnt_init(QemuLockCnt
*lockcnt
)
299 qemu_mutex_init(&lockcnt
->mutex
);
303 void qemu_lockcnt_destroy(QemuLockCnt
*lockcnt
)
305 qemu_mutex_destroy(&lockcnt
->mutex
);
308 void qemu_lockcnt_inc(QemuLockCnt
*lockcnt
)
312 old
= qatomic_read(&lockcnt
->count
);
314 qemu_lockcnt_lock(lockcnt
);
315 qemu_lockcnt_inc_and_unlock(lockcnt
);
318 if (qatomic_cmpxchg(&lockcnt
->count
, old
, old
+ 1) == old
) {
325 void qemu_lockcnt_dec(QemuLockCnt
*lockcnt
)
327 qatomic_dec(&lockcnt
->count
);
330 /* Decrement a counter, and return locked if it is decremented to zero.
331 * It is impossible for the counter to become nonzero while the mutex
334 bool qemu_lockcnt_dec_and_lock(QemuLockCnt
*lockcnt
)
336 int val
= qatomic_read(&lockcnt
->count
);
338 int old
= qatomic_cmpxchg(&lockcnt
->count
, val
, val
- 1);
347 qemu_lockcnt_lock(lockcnt
);
348 if (qatomic_fetch_dec(&lockcnt
->count
) == 1) {
352 qemu_lockcnt_unlock(lockcnt
);
356 /* Decrement a counter and return locked if it is decremented to zero.
357 * Otherwise do nothing.
359 * It is impossible for the counter to become nonzero while the mutex
362 bool qemu_lockcnt_dec_if_lock(QemuLockCnt
*lockcnt
)
364 /* No need for acquire semantics if we return false. */
365 int val
= qatomic_read(&lockcnt
->count
);
370 qemu_lockcnt_lock(lockcnt
);
371 if (qatomic_fetch_dec(&lockcnt
->count
) == 1) {
375 qemu_lockcnt_inc_and_unlock(lockcnt
);
379 void qemu_lockcnt_lock(QemuLockCnt
*lockcnt
)
381 qemu_mutex_lock(&lockcnt
->mutex
);
384 void qemu_lockcnt_inc_and_unlock(QemuLockCnt
*lockcnt
)
386 qatomic_inc(&lockcnt
->count
);
387 qemu_mutex_unlock(&lockcnt
->mutex
);
390 void qemu_lockcnt_unlock(QemuLockCnt
*lockcnt
)
392 qemu_mutex_unlock(&lockcnt
->mutex
);
395 unsigned qemu_lockcnt_count(QemuLockCnt
*lockcnt
)
397 return qatomic_read(&lockcnt
->count
);