^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) #include <linux/atomic.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) #include <linux/percpu.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) #include <linux/wait.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) #include <linux/lockdep.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include <linux/percpu-rwsem.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include <linux/rcupdate.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include <linux/sched.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include <linux/sched/task.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/errno.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) int __percpu_init_rwsem(struct percpu_rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) const char *name, struct lock_class_key *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) sem->read_count = alloc_percpu(int);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) if (unlikely(!sem->read_count))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) rcu_sync_init(&sem->rss);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) rcuwait_init(&sem->writer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) init_waitqueue_head(&sem->waiters);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) atomic_set(&sem->block, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #ifdef CONFIG_DEBUG_LOCK_ALLOC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) debug_check_no_locks_freed((void *)sem, sizeof(*sem));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) lockdep_init_map(&sem->dep_map, name, key, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) EXPORT_SYMBOL_GPL(__percpu_init_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) void percpu_free_rwsem(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * XXX: temporary kludge. The error path in alloc_super()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * assumes that percpu_free_rwsem() is safe after kzalloc().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) if (!sem->read_count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) rcu_sync_dtor(&sem->rss);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) free_percpu(sem->read_count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) sem->read_count = NULL; /* catch use after free bugs */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) EXPORT_SYMBOL_GPL(percpu_free_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) static bool __percpu_down_read_trylock(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) this_cpu_inc(*sem->read_count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) * Due to having preemption disabled the decrement happens on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) * the same CPU as the increment, avoiding the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) * increment-on-one-CPU-and-decrement-on-another problem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) * If the reader misses the writer's assignment of sem->block, then the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) * writer is guaranteed to see the reader's increment.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * Conversely, any readers that increment their sem->read_count after
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) * the writer looks are guaranteed to see the sem->block value, which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) * in turn means that they are guaranteed to immediately decrement
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) * their sem->read_count, so that it doesn't matter that the writer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) * missed them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) smp_mb(); /* A matches D */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) * If !sem->block the critical section starts here, matched by the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * release in percpu_up_write().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) if (likely(!atomic_read_acquire(&sem->block)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) this_cpu_dec(*sem->read_count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) /* Prod writer to re-evaluate readers_active_check() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) rcuwait_wake_up(&sem->writer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) static inline bool __percpu_down_write_trylock(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) if (atomic_read(&sem->block))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) return atomic_xchg(&sem->block, 1) == 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) static bool __percpu_rwsem_trylock(struct percpu_rw_semaphore *sem, bool reader)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) if (reader) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) bool ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) preempt_disable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) ret = __percpu_down_read_trylock(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) preempt_enable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) return __percpu_down_write_trylock(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) * The return value of wait_queue_entry::func means:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) * <0 - error, wakeup is terminated and the error is returned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) * 0 - no wakeup, a next waiter is tried
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) * >0 - woken, if EXCLUSIVE, counted towards @nr_exclusive.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) * We use EXCLUSIVE for both readers and writers to preserve FIFO order,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) * and play games with the return value to allow waking multiple readers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) * Specifically, we wake readers until we've woken a single writer, or until a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) * trylock fails.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) static int percpu_rwsem_wake_function(struct wait_queue_entry *wq_entry,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) unsigned int mode, int wake_flags,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) bool reader = wq_entry->flags & WQ_FLAG_CUSTOM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) struct percpu_rw_semaphore *sem = key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) struct task_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) /* concurrent against percpu_down_write(), can get stolen */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) if (!__percpu_rwsem_trylock(sem, reader))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) p = get_task_struct(wq_entry->private);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) list_del_init(&wq_entry->entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) smp_store_release(&wq_entry->private, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) wake_up_process(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) put_task_struct(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) return !reader; /* wake (readers until) 1 writer */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) static void percpu_rwsem_wait(struct percpu_rw_semaphore *sem, bool reader)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) DEFINE_WAIT_FUNC(wq_entry, percpu_rwsem_wake_function);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) bool wait;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) spin_lock_irq(&sem->waiters.lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) * Serialize against the wakeup in percpu_up_write(), if we fail
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) * the trylock, the wakeup must see us on the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) wait = !__percpu_rwsem_trylock(sem, reader);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) if (wait) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) wq_entry.flags |= WQ_FLAG_EXCLUSIVE | reader * WQ_FLAG_CUSTOM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) __add_wait_queue_entry_tail(&sem->waiters, &wq_entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) spin_unlock_irq(&sem->waiters.lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) while (wait) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) set_current_state(TASK_UNINTERRUPTIBLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) if (!smp_load_acquire(&wq_entry.private))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) schedule();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) __set_current_state(TASK_RUNNING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) bool __percpu_down_read(struct percpu_rw_semaphore *sem, bool try)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) if (__percpu_down_read_trylock(sem))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) if (try)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) preempt_enable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) percpu_rwsem_wait(sem, /* .reader = */ true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) preempt_disable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) EXPORT_SYMBOL_GPL(__percpu_down_read);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) #define per_cpu_sum(var) \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) ({ \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) typeof(var) __sum = 0; \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) int cpu; \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) compiletime_assert_atomic_type(__sum); \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) for_each_possible_cpu(cpu) \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) __sum += per_cpu(var, cpu); \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) __sum; \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) })
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) * Return true if the modular sum of the sem->read_count per-CPU variable is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) * zero. If this sum is zero, then it is stable due to the fact that if any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) * newly arriving readers increment a given counter, they will immediately
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) * decrement that same counter.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) * Assumes sem->block is set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) static bool readers_active_check(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) if (per_cpu_sum(*sem->read_count) != 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) * If we observed the decrement; ensure we see the entire critical
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * section.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) smp_mb(); /* C matches B */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) void percpu_down_write(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) rwsem_acquire(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) /* Notify readers to take the slow path. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) rcu_sync_enter(&sem->rss);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) * Try set sem->block; this provides writer-writer exclusion.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) * Having sem->block set makes new readers block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) if (!__percpu_down_write_trylock(sem))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) percpu_rwsem_wait(sem, /* .reader = */ false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) /* smp_mb() implied by __percpu_down_write_trylock() on success -- D matches A */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) * If they don't see our store of sem->block, then we are guaranteed to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) * see their sem->read_count increment, and therefore will wait for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) * them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) /* Wait for all active readers to complete. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) rcuwait_wait_event(&sem->writer, readers_active_check(sem), TASK_UNINTERRUPTIBLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) EXPORT_SYMBOL_GPL(percpu_down_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) void percpu_up_write(struct percpu_rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) * Signal the writer is done, no fast path yet.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) * One reason that we cannot just immediately flip to readers_fast is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) * that new readers might fail to see the results of this writer's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) * critical section.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) * Therefore we force it through the slow path which guarantees an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) * acquire and thereby guarantees the critical section's consistency.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) atomic_set_release(&sem->block, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) * Prod any pending reader/writer to make progress.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) __wake_up(&sem->waiters, TASK_NORMAL, 1, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) * Once this completes (at least one RCU-sched grace period hence) the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) * reader fast path will be available again. Safe to use outside the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) * exclusive write lock because its counting.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) rcu_sync_exit(&sem->rss);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) EXPORT_SYMBOL_GPL(percpu_up_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) static LIST_HEAD(destroy_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) static DEFINE_SPINLOCK(destroy_list_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) static void destroy_list_workfn(struct work_struct *work)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) struct percpu_rw_semaphore_atomic *sem, *sem2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) LIST_HEAD(to_destroy);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) spin_lock(&destroy_list_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) list_splice_init(&destroy_list, &to_destroy);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) spin_unlock(&destroy_list_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) if (list_empty(&to_destroy))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) list_for_each_entry_safe(sem, sem2, &to_destroy, destroy_list_entry) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) percpu_free_rwsem(&sem->rw_sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) kfree(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) static DECLARE_WORK(destroy_list_work, destroy_list_workfn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) void percpu_rwsem_async_destroy(struct percpu_rw_semaphore_atomic *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) spin_lock(&destroy_list_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) list_add_tail(&sem->destroy_list_entry, &destroy_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) spin_unlock(&destroy_list_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) schedule_work(&destroy_list_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) }