Blame view
lib/percpu-rwsem.c
4.99 KB
9390ef0c8 percpu_rw_semapho... |
1 |
#include <linux/atomic.h> |
a1fd3e24d percpu_rw_semapho... |
2 3 4 |
#include <linux/rwsem.h> #include <linux/percpu.h> #include <linux/wait.h> |
8ebe34731 percpu_rw_semapho... |
5 |
#include <linux/lockdep.h> |
a1fd3e24d percpu_rw_semapho... |
6 7 8 9 |
#include <linux/percpu-rwsem.h> #include <linux/rcupdate.h> #include <linux/sched.h> #include <linux/errno.h> |
8ebe34731 percpu_rw_semapho... |
10 11 |
int __percpu_init_rwsem(struct percpu_rw_semaphore *brw, const char *name, struct lock_class_key *rwsem_key) |
a1fd3e24d percpu_rw_semapho... |
12 13 14 15 |
{ brw->fast_read_ctr = alloc_percpu(int); if (unlikely(!brw->fast_read_ctr)) return -ENOMEM; |
8ebe34731 percpu_rw_semapho... |
16 17 |
/* ->rw_sem represents the whole percpu_rw_semaphore for lockdep */ __init_rwsem(&brw->rw_sem, name, rwsem_key); |
9390ef0c8 percpu_rw_semapho... |
18 |
atomic_set(&brw->write_ctr, 0); |
a1fd3e24d percpu_rw_semapho... |
19 20 21 22 23 24 25 26 27 28 29 30 31 |
atomic_set(&brw->slow_read_ctr, 0); init_waitqueue_head(&brw->write_waitq); return 0; } void percpu_free_rwsem(struct percpu_rw_semaphore *brw) { free_percpu(brw->fast_read_ctr); brw->fast_read_ctr = NULL; /* catch use after free bugs */ } /* * This is the fast-path for down_read/up_read, it only needs to ensure |
9390ef0c8 percpu_rw_semapho... |
32 |
* there is no pending writer (atomic_read(write_ctr) == 0) and inc/dec the |
a1fd3e24d percpu_rw_semapho... |
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
* fast per-cpu counter. The writer uses synchronize_sched_expedited() to * serialize with the preempt-disabled section below. * * The nontrivial part is that we should guarantee acquire/release semantics * in case when * * R_W: down_write() comes after up_read(), the writer should see all * changes done by the reader * or * W_R: down_read() comes after up_write(), the reader should see all * changes done by the writer * * If this helper fails the callers rely on the normal rw_semaphore and * atomic_dec_and_test(), so in this case we have the necessary barriers. * |
9390ef0c8 percpu_rw_semapho... |
48 |
* But if it succeeds we do not have any barriers, atomic_read(write_ctr) or |
a1fd3e24d percpu_rw_semapho... |
49 50 51 52 53 54 55 56 57 |
* __this_cpu_add() below can be reordered with any LOAD/STORE done by the * reader inside the critical section. See the comments in down_write and * up_write below. */ static bool update_fast_ctr(struct percpu_rw_semaphore *brw, unsigned int val) { bool success = false; preempt_disable(); |
9390ef0c8 percpu_rw_semapho... |
58 |
if (likely(!atomic_read(&brw->write_ctr))) { |
a1fd3e24d percpu_rw_semapho... |
59 60 61 62 63 64 65 66 67 68 69 |
__this_cpu_add(*brw->fast_read_ctr, val); success = true; } preempt_enable(); return success; } /* * Like the normal down_read() this is not recursive, the writer can * come after the first percpu_down_read() and create the deadlock. |
8ebe34731 percpu_rw_semapho... |
70 71 72 73 |
* * Note: returns with lock_is_held(brw->rw_sem) == T for lockdep, * percpu_up_read() does rwsem_release(). This pairs with the usage * of ->rw_sem in percpu_down/up_write(). |
a1fd3e24d percpu_rw_semapho... |
74 75 76 |
*/ void percpu_down_read(struct percpu_rw_semaphore *brw) { |
8ebe34731 percpu_rw_semapho... |
77 78 79 |
might_sleep(); if (likely(update_fast_ctr(brw, +1))) { rwsem_acquire_read(&brw->rw_sem.dep_map, 0, 0, _RET_IP_); |
a1fd3e24d percpu_rw_semapho... |
80 |
return; |
8ebe34731 percpu_rw_semapho... |
81 |
} |
a1fd3e24d percpu_rw_semapho... |
82 83 84 |
down_read(&brw->rw_sem); atomic_inc(&brw->slow_read_ctr); |
8ebe34731 percpu_rw_semapho... |
85 86 |
/* avoid up_read()->rwsem_release() */ __up_read(&brw->rw_sem); |
a1fd3e24d percpu_rw_semapho... |
87 88 89 90 |
} void percpu_up_read(struct percpu_rw_semaphore *brw) { |
8ebe34731 percpu_rw_semapho... |
91 |
rwsem_release(&brw->rw_sem.dep_map, 1, _RET_IP_); |
a1fd3e24d percpu_rw_semapho... |
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 |
if (likely(update_fast_ctr(brw, -1))) return; /* false-positive is possible but harmless */ if (atomic_dec_and_test(&brw->slow_read_ctr)) wake_up_all(&brw->write_waitq); } static int clear_fast_ctr(struct percpu_rw_semaphore *brw) { unsigned int sum = 0; int cpu; for_each_possible_cpu(cpu) { sum += per_cpu(*brw->fast_read_ctr, cpu); per_cpu(*brw->fast_read_ctr, cpu) = 0; } return sum; } /* |
9390ef0c8 percpu_rw_semapho... |
114 115 |
* A writer increments ->write_ctr to force the readers to switch to the * slow mode, note the atomic_read() check in update_fast_ctr(). |
a1fd3e24d percpu_rw_semapho... |
116 117 118 119 120 121 122 123 124 125 |
* * After that the readers can only inc/dec the slow ->slow_read_ctr counter, * ->fast_read_ctr is stable. Once the writer moves its sum into the slow * counter it represents the number of active readers. * * Finally the writer takes ->rw_sem for writing and blocks the new readers, * then waits until the slow counter becomes zero. */ void percpu_down_write(struct percpu_rw_semaphore *brw) { |
9390ef0c8 percpu_rw_semapho... |
126 127 |
/* tell update_fast_ctr() there is a pending writer */ atomic_inc(&brw->write_ctr); |
a1fd3e24d percpu_rw_semapho... |
128 |
/* |
9390ef0c8 percpu_rw_semapho... |
129 |
* 1. Ensures that write_ctr != 0 is visible to any down_read/up_read |
a1fd3e24d percpu_rw_semapho... |
130 131 132 133 134 135 136 137 138 139 |
* so that update_fast_ctr() can't succeed. * * 2. Ensures we see the result of every previous this_cpu_add() in * update_fast_ctr(). * * 3. Ensures that if any reader has exited its critical section via * fast-path, it executes a full memory barrier before we return. * See R_W case in the comment above update_fast_ctr(). */ synchronize_sched_expedited(); |
9390ef0c8 percpu_rw_semapho... |
140 141 |
/* exclude other writers, and block the new readers completely */ down_write(&brw->rw_sem); |
a1fd3e24d percpu_rw_semapho... |
142 143 |
/* nobody can use fast_read_ctr, move its sum into slow_read_ctr */ atomic_add(clear_fast_ctr(brw), &brw->slow_read_ctr); |
a1fd3e24d percpu_rw_semapho... |
144 145 146 147 148 149 |
/* wait for all readers to complete their percpu_up_read() */ wait_event(brw->write_waitq, !atomic_read(&brw->slow_read_ctr)); } void percpu_up_write(struct percpu_rw_semaphore *brw) { |
9390ef0c8 percpu_rw_semapho... |
150 |
/* release the lock, but the readers can't use the fast-path */ |
a1fd3e24d percpu_rw_semapho... |
151 |
up_write(&brw->rw_sem); |
a1fd3e24d percpu_rw_semapho... |
152 153 154 155 156 |
/* * Insert the barrier before the next fast-path in down_read, * see W_R case in the comment above update_fast_ctr(). */ synchronize_sched_expedited(); |
9390ef0c8 percpu_rw_semapho... |
157 158 |
/* the last writer unblocks update_fast_ctr() */ atomic_dec(&brw->write_ctr); |
a1fd3e24d percpu_rw_semapho... |
159 |
} |