^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /* kernel/rwsem.c: R/W semaphores, public implementation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * Written by David Howells (dhowells@redhat.com).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Derived from asm-i386/semaphore.h
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Writer lock-stealing by Alex Shi <alex.shi@intel.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * and Michel Lespinasse <walken@google.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * Optimistic spinning by Tim Chen <tim.c.chen@intel.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) * and Davidlohr Bueso <davidlohr@hp.com>. Based on mutexes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) * Rwsem count bit fields re-definition and rwsem rearchitecture by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) * Waiman Long <longman@redhat.com> and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) * Peter Zijlstra <peterz@infradead.org>.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/types.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/kernel.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <linux/sched.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <linux/sched/rt.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include <linux/sched/task.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include <linux/sched/debug.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include <linux/sched/wake_q.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include <linux/sched/signal.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include <linux/sched/clock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include <linux/export.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include <linux/rwsem.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include <linux/atomic.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include "lock_events.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) #include <trace/hooks/rwsem.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #include <trace/hooks/dtask.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * The least significant 3 bits of the owner value has the following
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * meanings when set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) * - Bit 0: RWSEM_READER_OWNED - The rwsem is owned by readers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) * - Bit 1: RWSEM_RD_NONSPINNABLE - Readers cannot spin on this lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) * - Bit 2: RWSEM_WR_NONSPINNABLE - Writers cannot spin on this lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) * When the rwsem is either owned by an anonymous writer, or it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) * reader-owned, but a spinning writer has timed out, both nonspinnable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) * bits will be set to disable optimistic spinning by readers and writers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) * In the later case, the last unlocking reader should then check the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) * writer nonspinnable bit and clear it only to give writers preference
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) * to acquire the lock via optimistic spinning, but not readers. Similar
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) * action is also done in the reader slowpath.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) * When a writer acquires a rwsem, it puts its task_struct pointer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) * into the owner field. It is cleared after an unlock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) * When a reader acquires a rwsem, it will also puts its task_struct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) * pointer into the owner field with the RWSEM_READER_OWNED bit set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) * On unlock, the owner field will largely be left untouched. So
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) * for a free or reader-owned rwsem, the owner value may contain
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) * information about the last reader that acquires the rwsem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * That information may be helpful in debugging cases where the system
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) * seems to hang on a reader owned rwsem especially if only one reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) * is involved. Ideally we would like to track all the readers that own
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) * a rwsem, but the overhead is simply too big.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) * Reader optimistic spinning is helpful when the reader critical section
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) * is short and there aren't that many readers around. It makes readers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) * relatively more preferred than writers. When a writer times out spinning
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) * on a reader-owned lock and set the nospinnable bits, there are two main
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) * reasons for that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * 1) The reader critical section is long, perhaps the task sleeps after
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) * acquiring the read lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * 2) There are just too many readers contending the lock causing it to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * take a while to service all of them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) * In the former case, long reader critical section will impede the progress
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) * of writers which is usually more important for system performance. In
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) * the later case, reader optimistic spinning tends to make the reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) * groups that contain readers that acquire the lock together smaller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) * leading to more of them. That may hurt performance in some cases. In
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) * other words, the setting of nonspinnable bits indicates that reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) * optimistic spinning may not be helpful for those workloads that cause
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) * it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * Therefore, any writers that had observed the setting of the writer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) * nonspinnable bit for a given rwsem after they fail to acquire the lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * via optimistic spinning will set the reader nonspinnable bit once they
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) * acquire the write lock. Similarly, readers that observe the setting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) * of reader nonspinnable bit at slowpath entry will set the reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) * nonspinnable bits when they acquire the read lock via the wakeup path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) * Once the reader nonspinnable bit is on, it will only be reset when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) * a writer is able to acquire the rwsem in the fast path or somehow a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) * reader or writer in the slowpath doesn't observe the nonspinable bit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) * This is to discourage reader optmistic spinning on that particular
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * rwsem and make writers more preferred. This adaptive disabling of reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) * optimistic spinning will alleviate the negative side effect of this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * feature.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) #define RWSEM_READER_OWNED (1UL << 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) #define RWSEM_RD_NONSPINNABLE (1UL << 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) #define RWSEM_WR_NONSPINNABLE (1UL << 2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) #define RWSEM_NONSPINNABLE (RWSEM_RD_NONSPINNABLE | RWSEM_WR_NONSPINNABLE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) #define RWSEM_OWNER_FLAGS_MASK (RWSEM_READER_OWNED | RWSEM_NONSPINNABLE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) #ifdef CONFIG_DEBUG_RWSEMS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) # define DEBUG_RWSEMS_WARN_ON(c, sem) do { \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) if (!debug_locks_silent && \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) WARN_ONCE(c, "DEBUG_RWSEMS_WARN_ON(%s): count = 0x%lx, magic = 0x%lx, owner = 0x%lx, curr 0x%lx, list %sempty\n",\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) #c, atomic_long_read(&(sem)->count), \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) (unsigned long) sem->magic, \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) atomic_long_read(&(sem)->owner), (long)current, \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) list_empty(&(sem)->wait_list) ? "" : "not ")) \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) debug_locks_off(); \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) } while (0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) # define DEBUG_RWSEMS_WARN_ON(c, sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) * On 64-bit architectures, the bit definitions of the count are:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) * Bit 0 - writer locked bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) * Bit 1 - waiters present bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) * Bit 2 - lock handoff bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) * Bits 3-7 - reserved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) * Bits 8-62 - 55-bit reader count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) * Bit 63 - read fail bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) * On 32-bit architectures, the bit definitions of the count are:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) * Bit 0 - writer locked bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) * Bit 1 - waiters present bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) * Bit 2 - lock handoff bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) * Bits 3-7 - reserved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) * Bits 8-30 - 23-bit reader count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) * Bit 31 - read fail bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) * It is not likely that the most significant bit (read fail bit) will ever
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) * be set. This guard bit is still checked anyway in the down_read() fastpath
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) * just in case we need to use up more of the reader bits for other purpose
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) * in the future.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) * atomic_long_fetch_add() is used to obtain reader lock, whereas
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) * atomic_long_cmpxchg() will be used to obtain writer lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) * There are three places where the lock handoff bit may be set or cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) * 1) rwsem_mark_wake() for readers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) * 2) rwsem_try_write_lock() for writers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) * 3) Error path of rwsem_down_write_slowpath().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) * For all the above cases, wait_lock will be held. A writer must also
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) * be the first one in the wait_list to be eligible for setting the handoff
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) * bit. So concurrent setting/clearing of handoff bit is not possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) #define RWSEM_WRITER_LOCKED (1UL << 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) #define RWSEM_FLAG_WAITERS (1UL << 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) #define RWSEM_FLAG_HANDOFF (1UL << 2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) #define RWSEM_FLAG_READFAIL (1UL << (BITS_PER_LONG - 1))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) #define RWSEM_READER_SHIFT 8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) #define RWSEM_READER_BIAS (1UL << RWSEM_READER_SHIFT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) #define RWSEM_READER_MASK (~(RWSEM_READER_BIAS - 1))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) #define RWSEM_WRITER_MASK RWSEM_WRITER_LOCKED
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) #define RWSEM_LOCK_MASK (RWSEM_WRITER_MASK|RWSEM_READER_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) #define RWSEM_READ_FAILED_MASK (RWSEM_WRITER_MASK|RWSEM_FLAG_WAITERS|\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) RWSEM_FLAG_HANDOFF|RWSEM_FLAG_READFAIL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) * All writes to owner are protected by WRITE_ONCE() to make sure that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) * store tearing can't happen as optimistic spinners may read and use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) * the owner value concurrently without lock. Read from owner, however,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) * may not need READ_ONCE() as long as the pointer value is only used
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) * for comparison and isn't being dereferenced.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) static inline void rwsem_set_owner(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) atomic_long_set(&sem->owner, (long)current);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) trace_android_vh_rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) static inline void rwsem_clear_owner(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) atomic_long_set(&sem->owner, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) * Test the flags in the owner field.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) static inline bool rwsem_test_oflags(struct rw_semaphore *sem, long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) return atomic_long_read(&sem->owner) & flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) * The task_struct pointer of the last owning reader will be left in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) * the owner field.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) * Note that the owner value just indicates the task has owned the rwsem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) * previously, it may not be the real owner or one of the real owners
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) * anymore when that field is examined, so take it with a grain of salt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) * The reader non-spinnable bit is preserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) static inline void __rwsem_set_reader_owned(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) struct task_struct *owner)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) unsigned long val = (unsigned long)owner | RWSEM_READER_OWNED |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) (atomic_long_read(&sem->owner) & RWSEM_RD_NONSPINNABLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) atomic_long_set(&sem->owner, val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) static inline void rwsem_set_reader_owned(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) __rwsem_set_reader_owned(sem, current);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) trace_android_vh_rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) * Return true if the rwsem is owned by a reader.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) static inline bool is_rwsem_reader_owned(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) #ifdef CONFIG_DEBUG_RWSEMS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) * Check the count to see if it is write-locked.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) long count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) if (count & RWSEM_WRITER_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) return rwsem_test_oflags(sem, RWSEM_READER_OWNED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) #ifdef CONFIG_DEBUG_RWSEMS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) * With CONFIG_DEBUG_RWSEMS configured, it will make sure that if there
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) * is a task pointer in owner of a reader-owned rwsem, it will be the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) * real owner or one of the real owners. The only exception is when the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) * unlock is done by up_read_non_owner().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) static inline void rwsem_clear_reader_owned(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) unsigned long val = atomic_long_read(&sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) while ((val & ~RWSEM_OWNER_FLAGS_MASK) == (unsigned long)current) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) if (atomic_long_try_cmpxchg(&sem->owner, &val,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) val & RWSEM_OWNER_FLAGS_MASK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) static inline void rwsem_clear_reader_owned(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) * Set the RWSEM_NONSPINNABLE bits if the RWSEM_READER_OWNED flag
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) * remains set. Otherwise, the operation will be aborted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) static inline void rwsem_set_nonspinnable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) unsigned long owner = atomic_long_read(&sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) if (!(owner & RWSEM_READER_OWNED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) if (owner & RWSEM_NONSPINNABLE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) } while (!atomic_long_try_cmpxchg(&sem->owner, &owner,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) owner | RWSEM_NONSPINNABLE));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) static inline bool rwsem_read_trylock(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) long cnt = atomic_long_add_return_acquire(RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) if (WARN_ON_ONCE(cnt < 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) rwsem_set_nonspinnable(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) return !(cnt & RWSEM_READ_FAILED_MASK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) * Return just the real task structure pointer of the owner
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) static inline struct task_struct *rwsem_owner(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) return (struct task_struct *)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) (atomic_long_read(&sem->owner) & ~RWSEM_OWNER_FLAGS_MASK);
^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) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) * Return the real task structure pointer of the owner and the embedded
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) * flags in the owner. pflags must be non-NULL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) static inline struct task_struct *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) rwsem_owner_flags(struct rw_semaphore *sem, unsigned long *pflags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) unsigned long owner = atomic_long_read(&sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) *pflags = owner & RWSEM_OWNER_FLAGS_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) return (struct task_struct *)(owner & ~RWSEM_OWNER_FLAGS_MASK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) * Guide to the rw_semaphore's count field.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) * When the RWSEM_WRITER_LOCKED bit in count is set, the lock is owned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) * by a writer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) * The lock is owned by readers when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) * (1) the RWSEM_WRITER_LOCKED isn't set in count,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) * (2) some of the reader bits are set in count, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) * (3) the owner field has RWSEM_READ_OWNED bit set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) * Having some reader bits set is not enough to guarantee a readers owned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) * lock as the readers may be in the process of backing out from the count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) * and a writer has just released the lock. So another writer may steal
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) * the lock immediately after that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) * Initialize an rwsem:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) void __init_rwsem(struct rw_semaphore *sem, const char *name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) struct lock_class_key *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) #ifdef CONFIG_DEBUG_LOCK_ALLOC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) * Make sure we are not reinitializing a held semaphore:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) debug_check_no_locks_freed((void *)sem, sizeof(*sem));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) lockdep_init_map_wait(&sem->dep_map, name, key, 0, LD_WAIT_SLEEP);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) #ifdef CONFIG_DEBUG_RWSEMS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) sem->magic = sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) atomic_long_set(&sem->count, RWSEM_UNLOCKED_VALUE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) raw_spin_lock_init(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) INIT_LIST_HEAD(&sem->wait_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) atomic_long_set(&sem->owner, 0L);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) #ifdef CONFIG_RWSEM_SPIN_ON_OWNER
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) osq_lock_init(&sem->osq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) trace_android_vh_rwsem_init(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) EXPORT_SYMBOL(__init_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) #define rwsem_first_waiter(sem) \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) list_first_entry(&sem->wait_list, struct rwsem_waiter, list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) enum rwsem_wake_type {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) RWSEM_WAKE_ANY, /* Wake whatever's at head of wait list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) RWSEM_WAKE_READERS, /* Wake readers only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) RWSEM_WAKE_READ_OWNED /* Waker thread holds the read lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) enum writer_wait_state {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) WRITER_NOT_FIRST, /* Writer is not first in wait list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) WRITER_FIRST, /* Writer is first in wait list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) WRITER_HANDOFF /* Writer is first & handoff needed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) * The typical HZ value is either 250 or 1000. So set the minimum waiting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) * time to at least 4ms or 1 jiffy (if it is higher than 4ms) in the wait
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) * queue before initiating the handoff protocol.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) #define RWSEM_WAIT_TIMEOUT DIV_ROUND_UP(HZ, 250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) * Magic number to batch-wakeup waiting readers, even when writers are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) * also present in the queue. This both limits the amount of work the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) * waking thread must do and also prevents any potential counter overflow,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) * however unlikely.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) #define MAX_READERS_WAKEUP 0x100
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) * handle the lock release when processes blocked on it that can now run
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) * - if we come here from up_xxxx(), then the RWSEM_FLAG_WAITERS bit must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) * have been set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) * - there must be someone on the queue
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) * - the wait_lock must be held by the caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) * - tasks are marked for wakeup, the caller must later invoke wake_up_q()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) * to actually wakeup the blocked task(s) and drop the reference count,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) * preferably when the wait_lock is released
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) * - woken process blocks are discarded from the list after having task zeroed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) * - writers are only marked woken if downgrading is false
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) static void rwsem_mark_wake(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) enum rwsem_wake_type wake_type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) struct wake_q_head *wake_q)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) struct rwsem_waiter *waiter, *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) long oldcount, woken = 0, adjustment = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) struct list_head wlist;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) lockdep_assert_held(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) * Take a peek at the queue head waiter such that we can determine
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) * the wakeup(s) to perform.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) waiter = rwsem_first_waiter(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) if (waiter->type == RWSEM_WAITING_FOR_WRITE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) if (wake_type == RWSEM_WAKE_ANY) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) * Mark writer at the front of the queue for wakeup.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) * Until the task is actually later awoken later by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) * the caller, other writers are able to steal it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) * Readers, on the other hand, will block as they
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) * will notice the queued writer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) wake_q_add(wake_q, waiter->task);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) lockevent_inc(rwsem_wake_writer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) * No reader wakeup if there are too many of them already.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) if (unlikely(atomic_long_read(&sem->count) < 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) * Writers might steal the lock before we grant it to the next reader.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) * We prefer to do the first reader grant before counting readers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) * so we can bail out early if a writer stole the lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) if (wake_type != RWSEM_WAKE_READ_OWNED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) struct task_struct *owner;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) adjustment = RWSEM_READER_BIAS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) oldcount = atomic_long_fetch_add(adjustment, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) if (unlikely(oldcount & RWSEM_WRITER_MASK)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) * When we've been waiting "too" long (for writers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) * to give up the lock), request a HANDOFF to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) * force the issue.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) if (!(oldcount & RWSEM_FLAG_HANDOFF) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) time_after(jiffies, waiter->timeout)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) adjustment -= RWSEM_FLAG_HANDOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) lockevent_inc(rwsem_rlock_handoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) atomic_long_add(-adjustment, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) * Set it to reader-owned to give spinners an early
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) * indication that readers now have the lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) * The reader nonspinnable bit seen at slowpath entry of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) * the reader is copied over.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) owner = waiter->task;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) if (waiter->last_rowner & RWSEM_RD_NONSPINNABLE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) owner = (void *)((unsigned long)owner | RWSEM_RD_NONSPINNABLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) lockevent_inc(rwsem_opt_norspin);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) __rwsem_set_reader_owned(sem, owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) * Grant up to MAX_READERS_WAKEUP read locks to all the readers in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) * queue. We know that the woken will be at least 1 as we accounted
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) * for above. Note we increment the 'active part' of the count by the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) * number of readers before waking any processes up.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) * This is an adaptation of the phase-fair R/W locks where at the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) * reader phase (first waiter is a reader), all readers are eligible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) * to acquire the lock at the same time irrespective of their order
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) * in the queue. The writers acquire the lock according to their
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) * order in the queue.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) * We have to do wakeup in 2 passes to prevent the possibility that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) * the reader count may be decremented before it is incremented. It
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) * is because the to-be-woken waiter may not have slept yet. So it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) * may see waiter->task got cleared, finish its critical section and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) * do an unlock before the reader count increment.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) * 1) Collect the read-waiters in a separate list, count them and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) * fully increment the reader count in rwsem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) * 2) For each waiters in the new list, clear waiter->task and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) * put them into wake_q to be woken up later.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) INIT_LIST_HEAD(&wlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) list_for_each_entry_safe(waiter, tmp, &sem->wait_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) if (waiter->type == RWSEM_WAITING_FOR_WRITE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) woken++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) list_move_tail(&waiter->list, &wlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) trace_android_vh_rwsem_mark_wake_readers(sem, waiter);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) * Limit # of readers that can be woken up per wakeup call.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) if (woken >= MAX_READERS_WAKEUP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) adjustment = woken * RWSEM_READER_BIAS - adjustment;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) lockevent_cond_inc(rwsem_wake_reader, woken);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) if (list_empty(&sem->wait_list)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) /* hit end of list above */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) adjustment -= RWSEM_FLAG_WAITERS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) * When we've woken a reader, we no longer need to force writers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) * to give up the lock and we can clear HANDOFF.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) if (woken && (atomic_long_read(&sem->count) & RWSEM_FLAG_HANDOFF))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) adjustment -= RWSEM_FLAG_HANDOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) if (adjustment)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) atomic_long_add(adjustment, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) /* 2nd pass */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) list_for_each_entry_safe(waiter, tmp, &wlist, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) struct task_struct *tsk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) tsk = waiter->task;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) get_task_struct(tsk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) * Ensure calling get_task_struct() before setting the reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) * waiter to nil such that rwsem_down_read_slowpath() cannot
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) * race with do_exit() by always holding a reference count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) * to the task to wakeup.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) smp_store_release(&waiter->task, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) * Ensure issuing the wakeup (either by us or someone else)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) * after setting the reader waiter to nil.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) wake_q_add_safe(wake_q, tsk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) * This function must be called with the sem->wait_lock held to prevent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) * race conditions between checking the rwsem wait list and setting the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) * sem->count accordingly.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) * If wstate is WRITER_HANDOFF, it will make sure that either the handoff
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) * bit is set or the lock is acquired with handoff bit cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) static inline bool rwsem_try_write_lock(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) enum writer_wait_state wstate)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) long count, new;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) lockdep_assert_held(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) bool has_handoff = !!(count & RWSEM_FLAG_HANDOFF);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) if (has_handoff && wstate == WRITER_NOT_FIRST)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) new = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) if (count & RWSEM_LOCK_MASK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) if (has_handoff || (wstate != WRITER_HANDOFF))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) new |= RWSEM_FLAG_HANDOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) new |= RWSEM_WRITER_LOCKED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) new &= ~RWSEM_FLAG_HANDOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (list_is_singular(&sem->wait_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) new &= ~RWSEM_FLAG_WAITERS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) } while (!atomic_long_try_cmpxchg_acquire(&sem->count, &count, new));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) * We have either acquired the lock with handoff bit cleared or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) * set the handoff bit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) if (new & RWSEM_FLAG_HANDOFF)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) #ifdef CONFIG_RWSEM_SPIN_ON_OWNER
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) * Try to acquire read lock before the reader is put on wait queue.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) * Lock acquisition isn't allowed if the rwsem is locked or a writer handoff
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) * is ongoing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) static inline bool rwsem_try_read_lock_unqueued(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) long count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) if (count & (RWSEM_WRITER_MASK | RWSEM_FLAG_HANDOFF))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) count = atomic_long_fetch_add_acquire(RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) if (!(count & (RWSEM_WRITER_MASK | RWSEM_FLAG_HANDOFF))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) lockevent_inc(rwsem_opt_rlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) /* Back out the change */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) atomic_long_add(-RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) * Try to acquire write lock before the writer has been put on wait queue.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) static inline bool rwsem_try_write_lock_unqueued(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) long count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) while (!(count & (RWSEM_LOCK_MASK|RWSEM_FLAG_HANDOFF))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) if (atomic_long_try_cmpxchg_acquire(&sem->count, &count,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) count | RWSEM_WRITER_LOCKED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) lockevent_inc(rwsem_opt_wlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) static inline bool owner_on_cpu(struct task_struct *owner)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) * As lock holder preemption issue, we both skip spinning if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) * task is not on cpu or its cpu is preempted
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) return owner->on_cpu && !vcpu_is_preempted(task_cpu(owner));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) static inline bool rwsem_can_spin_on_owner(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) unsigned long nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) struct task_struct *owner;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) bool ret = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) if (need_resched()) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) lockevent_inc(rwsem_opt_fail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) preempt_disable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) owner = rwsem_owner_flags(sem, &flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) * Don't check the read-owner as the entry may be stale.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) if ((flags & nonspinnable) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) (owner && !(flags & RWSEM_READER_OWNED) && !owner_on_cpu(owner)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) ret = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) preempt_enable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) lockevent_cond_inc(rwsem_opt_fail, !ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) * The rwsem_spin_on_owner() function returns the folowing 4 values
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) * depending on the lock owner state.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) * OWNER_NULL : owner is currently NULL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) * OWNER_WRITER: when owner changes and is a writer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) * OWNER_READER: when owner changes and the new owner may be a reader.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) * OWNER_NONSPINNABLE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) * when optimistic spinning has to stop because either the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) * owner stops running, is unknown, or its timeslice has
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) * been used up.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) enum owner_state {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) OWNER_NULL = 1 << 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) OWNER_WRITER = 1 << 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) OWNER_READER = 1 << 2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) OWNER_NONSPINNABLE = 1 << 3,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) #define OWNER_SPINNABLE (OWNER_NULL | OWNER_WRITER | OWNER_READER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) static inline enum owner_state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) rwsem_owner_state(struct task_struct *owner, unsigned long flags, unsigned long nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) if (flags & nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) return OWNER_NONSPINNABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) if (flags & RWSEM_READER_OWNED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) return OWNER_READER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) return owner ? OWNER_WRITER : OWNER_NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) static noinline enum owner_state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) rwsem_spin_on_owner(struct rw_semaphore *sem, unsigned long nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) struct task_struct *new, *owner;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) unsigned long flags, new_flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) enum owner_state state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) owner = rwsem_owner_flags(sem, &flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) state = rwsem_owner_state(owner, flags, nonspinnable);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) if (state != OWNER_WRITER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) return state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) for (;;) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) * When a waiting writer set the handoff flag, it may spin
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) * on the owner as well. Once that writer acquires the lock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) * we can spin on it. So we don't need to quit even when the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) * handoff bit is set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) new = rwsem_owner_flags(sem, &new_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) if ((new != owner) || (new_flags != flags)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) state = rwsem_owner_state(new, new_flags, nonspinnable);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) * Ensure we emit the owner->on_cpu, dereference _after_
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) * checking sem->owner still matches owner, if that fails,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) * owner might point to free()d memory, if it still matches,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) * the rcu_read_lock() ensures the memory stays valid.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) barrier();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) if (need_resched() || !owner_on_cpu(owner)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) state = OWNER_NONSPINNABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) cpu_relax();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) return state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) * Calculate reader-owned rwsem spinning threshold for writer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) * The more readers own the rwsem, the longer it will take for them to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) * wind down and free the rwsem. So the empirical formula used to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) * determine the actual spinning time limit here is:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) * Spinning threshold = (10 + nr_readers/2)us
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) * The limit is capped to a maximum of 25us (30 readers). This is just
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) * a heuristic and is subjected to change in the future.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) static inline u64 rwsem_rspin_threshold(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) long count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) int readers = count >> RWSEM_READER_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) u64 delta;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) if (readers > 30)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) readers = 30;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) delta = (20 + readers) * NSEC_PER_USEC / 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) return sched_clock() + delta;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) static bool rwsem_optimistic_spin(struct rw_semaphore *sem, bool wlock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) bool taken = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) int prev_owner_state = OWNER_NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) int loop = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) u64 rspin_threshold = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) unsigned long nonspinnable = wlock ? RWSEM_WR_NONSPINNABLE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) : RWSEM_RD_NONSPINNABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) preempt_disable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) /* sem->wait_lock should not be held when doing optimistic spinning */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) if (!osq_lock(&sem->osq))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) * Optimistically spin on the owner field and attempt to acquire the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) * lock whenever the owner changes. Spinning will be stopped when:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) * 1) the owning writer isn't running; or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) * 2) readers own the lock and spinning time has exceeded limit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) for (;;) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) enum owner_state owner_state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) owner_state = rwsem_spin_on_owner(sem, nonspinnable);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) if (!(owner_state & OWNER_SPINNABLE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) * Try to acquire the lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) taken = wlock ? rwsem_try_write_lock_unqueued(sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) : rwsem_try_read_lock_unqueued(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) if (taken)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) * Time-based reader-owned rwsem optimistic spinning
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) if (wlock && (owner_state == OWNER_READER)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) * Re-initialize rspin_threshold every time when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) * the owner state changes from non-reader to reader.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) * This allows a writer to steal the lock in between
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) * 2 reader phases and have the threshold reset at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) * the beginning of the 2nd reader phase.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) if (prev_owner_state != OWNER_READER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) if (rwsem_test_oflags(sem, nonspinnable))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) rspin_threshold = rwsem_rspin_threshold(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) loop = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) * Check time threshold once every 16 iterations to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) * avoid calling sched_clock() too frequently so
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) * as to reduce the average latency between the times
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) * when the lock becomes free and when the spinner
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) * is ready to do a trylock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) else if (!(++loop & 0xf) && (sched_clock() > rspin_threshold)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) rwsem_set_nonspinnable(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) lockevent_inc(rwsem_opt_nospin);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) * An RT task cannot do optimistic spinning if it cannot
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) * be sure the lock holder is running or live-lock may
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) * happen if the current task and the lock holder happen
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) * to run in the same CPU. However, aborting optimistic
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) * spinning while a NULL owner is detected may miss some
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) * opportunity where spinning can continue without causing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) * problem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) * There are 2 possible cases where an RT task may be able
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) * to continue spinning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) * 1) The lock owner is in the process of releasing the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) * lock, sem->owner is cleared but the lock has not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) * been released yet.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) * 2) The lock was free and owner cleared, but another
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) * task just comes in and acquire the lock before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) * we try to get it. The new owner may be a spinnable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) * writer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) * To take advantage of two scenarios listed agove, the RT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) * task is made to retry one more time to see if it can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) * acquire the lock or continue spinning on the new owning
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) * writer. Of course, if the time lag is long enough or the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) * new owner is not a writer or spinnable, the RT task will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) * quit spinning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) * If the owner is a writer, the need_resched() check is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) * done inside rwsem_spin_on_owner(). If the owner is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) * a writer, need_resched() check needs to be done here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) if (owner_state != OWNER_WRITER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) if (need_resched())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) if (rt_task(current) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) (prev_owner_state != OWNER_WRITER))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) prev_owner_state = owner_state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) * The cpu_relax() call is a compiler barrier which forces
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890) * everything in this loop to be re-loaded. We don't need
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) * memory barriers as we'll eventually observe the right
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) * values at the cost of a few extra spins.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) cpu_relax();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) osq_unlock(&sem->osq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) done:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) preempt_enable();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) lockevent_cond_inc(rwsem_opt_fail, !taken);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) return taken;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) * Clear the owner's RWSEM_WR_NONSPINNABLE bit if it is set. This should
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) * only be called when the reader count reaches 0.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) * This give writers better chance to acquire the rwsem first before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) * readers when the rwsem was being held by readers for a relatively long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) * period of time. Race can happen that an optimistic spinner may have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) * just stolen the rwsem and set the owner, but just clearing the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) * RWSEM_WR_NONSPINNABLE bit will do no harm anyway.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) static inline void clear_wr_nonspinnable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) if (rwsem_test_oflags(sem, RWSEM_WR_NONSPINNABLE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) atomic_long_andnot(RWSEM_WR_NONSPINNABLE, &sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) * This function is called when the reader fails to acquire the lock via
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) * optimistic spinning. In this case we will still attempt to do a trylock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) * when comparing the rwsem state right now with the state when entering
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) * the slowpath indicates that the reader is still in a valid reader phase.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) * This happens when the following conditions are true:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) * 1) The lock is currently reader owned, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) * 2) The lock is previously not reader-owned or the last read owner changes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) * In the former case, we have transitioned from a writer phase to a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) * reader-phase while spinning. In the latter case, it means the reader
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) * phase hasn't ended when we entered the optimistic spinning loop. In
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) * both cases, the reader is eligible to acquire the lock. This is the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) * secondary path where a read lock is acquired optimistically.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) * The reader non-spinnable bit wasn't set at time of entry or it will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) * not be here at all.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) static inline bool rwsem_reader_phase_trylock(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) unsigned long last_rowner)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) unsigned long owner = atomic_long_read(&sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) if (!(owner & RWSEM_READER_OWNED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) if (((owner ^ last_rowner) & ~RWSEM_OWNER_FLAGS_MASK) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) rwsem_try_read_lock_unqueued(sem)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) lockevent_inc(rwsem_opt_rlock2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) lockevent_add(rwsem_opt_fail, -1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) static inline bool rwsem_can_spin_on_owner(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) unsigned long nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) static inline bool rwsem_optimistic_spin(struct rw_semaphore *sem, bool wlock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) static inline void clear_wr_nonspinnable(struct rw_semaphore *sem) { }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) static inline bool rwsem_reader_phase_trylock(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) unsigned long last_rowner)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) static inline int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) rwsem_spin_on_owner(struct rw_semaphore *sem, unsigned long nonspinnable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) #define OWNER_NULL 1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) * Wait for the read lock to be granted
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) static struct rw_semaphore __sched *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986) rwsem_down_read_slowpath(struct rw_semaphore *sem, int state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) long count, adjustment = -RWSEM_READER_BIAS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) struct rwsem_waiter waiter;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) DEFINE_WAKE_Q(wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) bool wake = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) bool already_on_list = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) * Save the current read-owner of rwsem, if available, and the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) * reader nonspinnable bit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) waiter.last_rowner = atomic_long_read(&sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) if (!(waiter.last_rowner & RWSEM_READER_OWNED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) waiter.last_rowner &= RWSEM_RD_NONSPINNABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) if (!rwsem_can_spin_on_owner(sem, RWSEM_RD_NONSPINNABLE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) goto queue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) * Undo read bias from down_read() and do optimistic spinning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) atomic_long_add(-RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) adjustment = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) if (rwsem_optimistic_spin(sem, false)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) /* rwsem_optimistic_spin() implies ACQUIRE on success */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) * Wake up other readers in the wait list if the front
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) * waiter is a reader.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) if ((atomic_long_read(&sem->count) & RWSEM_FLAG_WAITERS)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) if (!list_empty(&sem->wait_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) rwsem_mark_wake(sem, RWSEM_WAKE_READ_OWNED,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) } else if (rwsem_reader_phase_trylock(sem, waiter.last_rowner)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) /* rwsem_reader_phase_trylock() implies ACQUIRE on success */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) queue:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) waiter.task = current;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) waiter.type = RWSEM_WAITING_FOR_READ;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) waiter.timeout = jiffies + RWSEM_WAIT_TIMEOUT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) if (list_empty(&sem->wait_list)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038) * In case the wait queue is empty and the lock isn't owned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) * by a writer or has the handoff bit set, this reader can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) * exit the slowpath and return immediately as its
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) * RWSEM_READER_BIAS has already been set in the count.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) if (adjustment && !(atomic_long_read(&sem->count) &
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) (RWSEM_WRITER_MASK | RWSEM_FLAG_HANDOFF))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) /* Provide lock ACQUIRE */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) smp_acquire__after_ctrl_dep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049) lockevent_inc(rwsem_rlock_fast);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) adjustment += RWSEM_FLAG_WAITERS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054) trace_android_vh_alter_rwsem_list_add(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) &waiter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) sem, &already_on_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) if (!already_on_list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) list_add_tail(&waiter.list, &sem->wait_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) /* we're now waiting on the lock, but no longer actively locking */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061) if (adjustment)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) count = atomic_long_add_return(adjustment, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067) * If there are no active locks, wake the front queued process(es).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) * If there are no writers and we are first in the queue,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070) * wake our own waiter to join the existing active readers !
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) if (!(count & RWSEM_LOCK_MASK)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073) clear_wr_nonspinnable(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) wake = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076) if (wake || (!(count & RWSEM_WRITER_MASK) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) (adjustment & RWSEM_FLAG_WAITERS)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) rwsem_mark_wake(sem, RWSEM_WAKE_ANY, &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) trace_android_vh_rwsem_wake(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) /* wait to be given the lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) trace_android_vh_rwsem_read_wait_start(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) for (;;) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) set_current_state(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088) if (!smp_load_acquire(&waiter.task)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) /* Matches rwsem_mark_wake()'s smp_store_release(). */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) if (signal_pending_state(state, current)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) if (waiter.task)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) goto out_nolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097) /* Ordered by sem->wait_lock against rwsem_mark_wake(). */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) schedule();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) lockevent_inc(rwsem_sleep_reader);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) __set_current_state(TASK_RUNNING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) trace_android_vh_rwsem_read_wait_finish(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) lockevent_inc(rwsem_rlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) out_nolock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) list_del(&waiter.list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) if (list_empty(&sem->wait_list)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) atomic_long_andnot(RWSEM_FLAG_WAITERS|RWSEM_FLAG_HANDOFF,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) __set_current_state(TASK_RUNNING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) trace_android_vh_rwsem_read_wait_finish(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) lockevent_inc(rwsem_rlock_fail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) return ERR_PTR(-EINTR);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) * This function is called by the a write lock owner. So the owner value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) * won't get changed by others.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) static inline void rwsem_disable_reader_optspin(struct rw_semaphore *sem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) bool disable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129) if (unlikely(disable)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) atomic_long_or(RWSEM_RD_NONSPINNABLE, &sem->owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131) lockevent_inc(rwsem_opt_norspin);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) * Wait until we successfully acquire the write lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) static struct rw_semaphore *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) rwsem_down_write_slowpath(struct rw_semaphore *sem, int state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141) long count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) bool disable_rspin;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) enum writer_wait_state wstate;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) struct rwsem_waiter waiter;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145) struct rw_semaphore *ret = sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) DEFINE_WAKE_Q(wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147) bool already_on_list = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) /* do optimistic spinning and steal lock if possible */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) if (rwsem_can_spin_on_owner(sem, RWSEM_WR_NONSPINNABLE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) rwsem_optimistic_spin(sem, true)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) /* rwsem_optimistic_spin() implies ACQUIRE on success */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) * Disable reader optimistic spinning for this rwsem after
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) * acquiring the write lock when the setting of the nonspinnable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) * bits are observed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) disable_rspin = atomic_long_read(&sem->owner) & RWSEM_NONSPINNABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) * Optimistic spinning failed, proceed to the slowpath
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) * and block until we can acquire the sem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) waiter.task = current;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168) waiter.type = RWSEM_WAITING_FOR_WRITE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) waiter.timeout = jiffies + RWSEM_WAIT_TIMEOUT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173) /* account for this before adding a new element to the list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) wstate = list_empty(&sem->wait_list) ? WRITER_FIRST : WRITER_NOT_FIRST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) trace_android_vh_alter_rwsem_list_add(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) &waiter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178) sem, &already_on_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) if (!already_on_list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) list_add_tail(&waiter.list, &sem->wait_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) /* we're now waiting on the lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) if (wstate == WRITER_NOT_FIRST) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) * If there were already threads queued before us and:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) * 1) there are no active locks, wake the front
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) * queued process(es) as the handoff bit might be set.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) * 2) there are no active writers and some readers, the lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) * must be read owned; so we try to wake any read lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) * waiters that were queued ahead of us.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) if (count & RWSEM_WRITER_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) goto wait;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) rwsem_mark_wake(sem, (count & RWSEM_READER_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) ? RWSEM_WAKE_READERS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) : RWSEM_WAKE_ANY, &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) if (!wake_q_empty(&wake_q)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) * We want to minimize wait_lock hold time especially
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204) * when a large number of readers are to be woken up.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208) wake_q_init(&wake_q); /* Used again, reinit */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212) atomic_long_or(RWSEM_FLAG_WAITERS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215) wait:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) trace_android_vh_rwsem_wake(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217) /* wait until we successfully acquire the lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218) trace_android_vh_rwsem_write_wait_start(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219) set_current_state(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) for (;;) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) if (rwsem_try_write_lock(sem, wstate)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222) /* rwsem_try_write_lock() implies ACQUIRE on success */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1223) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1224) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229) * After setting the handoff bit and failing to acquire
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230) * the lock, attempt to spin on owner to accelerate lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231) * transfer. If the previous owner is a on-cpu writer and it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) * has just released the lock, OWNER_NULL will be returned.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) * In this case, we attempt to acquire the lock again
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234) * without sleeping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) if (wstate == WRITER_HANDOFF &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) rwsem_spin_on_owner(sem, RWSEM_NONSPINNABLE) == OWNER_NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238) goto trylock_again;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240) /* Block until there are no active lockers. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241) for (;;) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) if (signal_pending_state(state, current))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243) goto out_nolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) schedule();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246) lockevent_inc(rwsem_sleep_writer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247) set_current_state(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249) * If HANDOFF bit is set, unconditionally do
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250) * a trylock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252) if (wstate == WRITER_HANDOFF)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) if ((wstate == WRITER_NOT_FIRST) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) (rwsem_first_waiter(sem) == &waiter))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) wstate = WRITER_FIRST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259) count = atomic_long_read(&sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) if (!(count & RWSEM_LOCK_MASK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) * The setting of the handoff bit is deferred
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265) * until rwsem_try_write_lock() is called.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267) if ((wstate == WRITER_FIRST) && (rt_task(current) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) time_after(jiffies, waiter.timeout))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) wstate = WRITER_HANDOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270) lockevent_inc(rwsem_wlock_handoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) trylock_again:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277) __set_current_state(TASK_RUNNING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278) trace_android_vh_rwsem_write_wait_finish(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279) list_del(&waiter.list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280) rwsem_disable_reader_optspin(sem, disable_rspin);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) lockevent_inc(rwsem_wlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) out_nolock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) __set_current_state(TASK_RUNNING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288) trace_android_vh_rwsem_write_wait_finish(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) raw_spin_lock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) list_del(&waiter.list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) if (unlikely(wstate == WRITER_HANDOFF))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) atomic_long_andnot(RWSEM_FLAG_HANDOFF, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295) if (list_empty(&sem->wait_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296) atomic_long_andnot(RWSEM_FLAG_WAITERS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) rwsem_mark_wake(sem, RWSEM_WAKE_ANY, &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299) raw_spin_unlock_irq(&sem->wait_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301) lockevent_inc(rwsem_wlock_fail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) return ERR_PTR(-EINTR);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) * handle waking up a waiter on the semaphore
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308) * - up_read/up_write has decremented the active part of count if we come here
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310) static struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem, long count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313) DEFINE_WAKE_Q(wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315) raw_spin_lock_irqsave(&sem->wait_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) if (!list_empty(&sem->wait_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318) rwsem_mark_wake(sem, RWSEM_WAKE_ANY, &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) trace_android_vh_rwsem_wake_finish(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1321) raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1322) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328) * downgrade a write lock into a read lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) * - caller incremented waiting part of count and discovered it still negative
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330) * - just wake up any readers at the front of the queue
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332) static struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335) DEFINE_WAKE_Q(wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337) raw_spin_lock_irqsave(&sem->wait_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) if (!list_empty(&sem->wait_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) rwsem_mark_wake(sem, RWSEM_WAKE_READ_OWNED, &wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) raw_spin_unlock_irqrestore(&sem->wait_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343) wake_up_q(&wake_q);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) return sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) * lock for reading
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) static inline void __down_read(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1352) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1353) if (!rwsem_read_trylock(sem)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1354) rwsem_down_read_slowpath(sem, TASK_UNINTERRUPTIBLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355) DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361) static inline int __down_read_interruptible(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) if (!rwsem_read_trylock(sem)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) if (IS_ERR(rwsem_down_read_slowpath(sem, TASK_INTERRUPTIBLE)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366) DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373) static inline int __down_read_killable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) if (!rwsem_read_trylock(sem)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376) if (IS_ERR(rwsem_down_read_slowpath(sem, TASK_KILLABLE)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378) DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385) static inline int __down_read_trylock(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) long tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389) DEBUG_RWSEMS_WARN_ON(sem->magic != sem, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392) * Optimize for the case when the rwsem is not locked at all.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) tmp = RWSEM_UNLOCKED_VALUE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) if (atomic_long_try_cmpxchg_acquire(&sem->count, &tmp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397) tmp + RWSEM_READER_BIAS)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1400) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1401) } while (!(tmp & RWSEM_READ_FAILED_MASK));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1402) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1403) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1404)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1405) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1406) * lock for writing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1407) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1408) static inline void __down_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1409) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1410) long tmp = RWSEM_UNLOCKED_VALUE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1411)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1412) if (unlikely(!atomic_long_try_cmpxchg_acquire(&sem->count, &tmp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1413) RWSEM_WRITER_LOCKED)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1414) rwsem_down_write_slowpath(sem, TASK_UNINTERRUPTIBLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1415) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1416) rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1417) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1418)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1419) static inline int __down_write_killable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1420) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1421) long tmp = RWSEM_UNLOCKED_VALUE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1422)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1423) if (unlikely(!atomic_long_try_cmpxchg_acquire(&sem->count, &tmp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1424) RWSEM_WRITER_LOCKED))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1425) if (IS_ERR(rwsem_down_write_slowpath(sem, TASK_KILLABLE)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1426) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1427) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1428) rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1429) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1430) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1431) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1432)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1433) static inline int __down_write_trylock(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1434) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1435) long tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1436)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1437) DEBUG_RWSEMS_WARN_ON(sem->magic != sem, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1438)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1439) tmp = RWSEM_UNLOCKED_VALUE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1440) if (atomic_long_try_cmpxchg_acquire(&sem->count, &tmp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1441) RWSEM_WRITER_LOCKED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1442) rwsem_set_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1443) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1444) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1445) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1446) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1447)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1448) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1449) * unlock after reading
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1450) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1451) static inline void __up_read(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1452) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1453) long tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1454)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1455) DEBUG_RWSEMS_WARN_ON(sem->magic != sem, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1456) DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1457)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1458) rwsem_clear_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1459) tmp = atomic_long_add_return_release(-RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1460) DEBUG_RWSEMS_WARN_ON(tmp < 0, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1461) if (unlikely((tmp & (RWSEM_LOCK_MASK|RWSEM_FLAG_WAITERS)) ==
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1462) RWSEM_FLAG_WAITERS)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1463) clear_wr_nonspinnable(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1464) rwsem_wake(sem, tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1465) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1466) trace_android_vh_rwsem_up_read_end(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1467) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1468)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1469) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1470) * unlock after writing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1471) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1472) static inline void __up_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1473) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1474) long tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1475)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1476) DEBUG_RWSEMS_WARN_ON(sem->magic != sem, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1477) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1478) * sem->owner may differ from current if the ownership is transferred
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1479) * to an anonymous writer by setting the RWSEM_NONSPINNABLE bits.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1480) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1481) DEBUG_RWSEMS_WARN_ON((rwsem_owner(sem) != current) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1482) !rwsem_test_oflags(sem, RWSEM_NONSPINNABLE), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1483)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1484) rwsem_clear_owner(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1485) tmp = atomic_long_fetch_add_release(-RWSEM_WRITER_LOCKED, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1486) if (unlikely(tmp & RWSEM_FLAG_WAITERS))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1487) rwsem_wake(sem, tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1488) trace_android_vh_rwsem_up_write_end(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1489) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1490)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1491) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1492) * downgrade write lock to read lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1493) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1494) static inline void __downgrade_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1495) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1496) long tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1497)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1498) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1499) * When downgrading from exclusive to shared ownership,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1500) * anything inside the write-locked region cannot leak
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1501) * into the read side. In contrast, anything in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1502) * read-locked region is ok to be re-ordered into the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1503) * write side. As such, rely on RELEASE semantics.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1504) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1505) DEBUG_RWSEMS_WARN_ON(rwsem_owner(sem) != current, sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1506) tmp = atomic_long_fetch_add_release(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1507) -RWSEM_WRITER_LOCKED+RWSEM_READER_BIAS, &sem->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1508) rwsem_set_reader_owned(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1509) if (tmp & RWSEM_FLAG_WAITERS)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1510) rwsem_downgrade_wake(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1511) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1512)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1513) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1514) * lock for reading
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1515) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1516) void __sched down_read(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1517) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1518) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1519) rwsem_acquire_read(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1521) LOCK_CONTENDED(sem, __down_read_trylock, __down_read);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1522) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1523) EXPORT_SYMBOL(down_read);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1525) int __sched down_read_interruptible(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1526) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1527) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1528) rwsem_acquire_read(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1530) if (LOCK_CONTENDED_RETURN(sem, __down_read_trylock, __down_read_interruptible)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1531) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1532) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1533) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1534)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1535) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1536) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1537) EXPORT_SYMBOL(down_read_interruptible);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1538)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1539) int __sched down_read_killable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1540) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1541) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1542) rwsem_acquire_read(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1543)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1544) if (LOCK_CONTENDED_RETURN(sem, __down_read_trylock, __down_read_killable)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1545) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1546) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1547) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1548)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1549) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1550) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1551) EXPORT_SYMBOL(down_read_killable);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1553) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1554) * trylock for reading -- returns 1 if successful, 0 if contention
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1555) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1556) int down_read_trylock(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1557) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1558) int ret = __down_read_trylock(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1559)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1560) if (ret == 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1561) rwsem_acquire_read(&sem->dep_map, 0, 1, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1562) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1563) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1564) EXPORT_SYMBOL(down_read_trylock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1566) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1567) * lock for writing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1568) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1569) void __sched down_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1570) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1571) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1572) rwsem_acquire(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1573) LOCK_CONTENDED(sem, __down_write_trylock, __down_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1574) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1575) EXPORT_SYMBOL(down_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1576)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1577) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1578) * lock for writing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1579) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1580) int __sched down_write_killable(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1581) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1582) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1583) rwsem_acquire(&sem->dep_map, 0, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1584)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1585) if (LOCK_CONTENDED_RETURN(sem, __down_write_trylock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1586) __down_write_killable)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1587) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1588) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1589) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1590)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1591) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1592) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1593) EXPORT_SYMBOL(down_write_killable);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1594)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1595) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1596) * trylock for writing -- returns 1 if successful, 0 if contention
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1597) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1598) int down_write_trylock(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1599) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1600) int ret = __down_write_trylock(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1601)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1602) if (ret == 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1603) rwsem_acquire(&sem->dep_map, 0, 1, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1605) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1606) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1607) EXPORT_SYMBOL(down_write_trylock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1609) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1610) * release a read lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1611) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1612) void up_read(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1613) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1614) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1615) __up_read(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1616) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1617) EXPORT_SYMBOL(up_read);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1618)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1619) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1620) * release a write lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1621) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1622) void up_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1623) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1624) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1625) trace_android_vh_rwsem_write_finished(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1626) __up_write(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1627) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1628) EXPORT_SYMBOL(up_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1629)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1630) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1631) * downgrade write lock to read lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1632) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1633) void downgrade_write(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1634) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1635) lock_downgrade(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1636) trace_android_vh_rwsem_write_finished(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1637) __downgrade_write(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1638) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1639) EXPORT_SYMBOL(downgrade_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1640)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1641) #ifdef CONFIG_DEBUG_LOCK_ALLOC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1642)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1643) void down_read_nested(struct rw_semaphore *sem, int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1644) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1645) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1646) rwsem_acquire_read(&sem->dep_map, subclass, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1647) LOCK_CONTENDED(sem, __down_read_trylock, __down_read);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1648) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1649) EXPORT_SYMBOL(down_read_nested);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1650)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1651) int down_read_killable_nested(struct rw_semaphore *sem, int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1652) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1653) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1654) rwsem_acquire_read(&sem->dep_map, subclass, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1655)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1656) if (LOCK_CONTENDED_RETURN(sem, __down_read_trylock, __down_read_killable)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1657) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1658) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1659) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1660)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1661) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1662) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1663) EXPORT_SYMBOL(down_read_killable_nested);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1664)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1665) void _down_write_nest_lock(struct rw_semaphore *sem, struct lockdep_map *nest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1666) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1667) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1668) rwsem_acquire_nest(&sem->dep_map, 0, 0, nest, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1669) LOCK_CONTENDED(sem, __down_write_trylock, __down_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1670) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1671) EXPORT_SYMBOL(_down_write_nest_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1672)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1673) void down_read_non_owner(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1674) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1675) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1676) __down_read(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1677) __rwsem_set_reader_owned(sem, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1678) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1679) EXPORT_SYMBOL(down_read_non_owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1680)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1681) void down_write_nested(struct rw_semaphore *sem, int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1682) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1683) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1684) rwsem_acquire(&sem->dep_map, subclass, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1685) LOCK_CONTENDED(sem, __down_write_trylock, __down_write);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1686) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1687) EXPORT_SYMBOL(down_write_nested);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1688)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1689) int __sched down_write_killable_nested(struct rw_semaphore *sem, int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1690) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1691) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1692) rwsem_acquire(&sem->dep_map, subclass, 0, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1693)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1694) if (LOCK_CONTENDED_RETURN(sem, __down_write_trylock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1695) __down_write_killable)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1696) rwsem_release(&sem->dep_map, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1697) return -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1698) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1699)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1700) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1701) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1702) EXPORT_SYMBOL(down_write_killable_nested);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1703)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1704) void up_read_non_owner(struct rw_semaphore *sem)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1705) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1706) DEBUG_RWSEMS_WARN_ON(!is_rwsem_reader_owned(sem), sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1707) __up_read(sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1708) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1709) EXPORT_SYMBOL(up_read_non_owner);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1710)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1711) #endif
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags