zephyr/kernel/mutex.c
Andy Ross 7832738ae9 kernel/timeout: Make timeout arguments an opaque type
Add a k_timeout_t type, and use it everywhere that kernel API
functions were accepting a millisecond timeout argument.  Instead of
forcing milliseconds everywhere (which are often not integrally
representable as system ticks), do the conversion to ticks at the
point where the timeout is created.  This avoids an extra unit
conversion in some application code, and allows us to express the
timeout in units other than milliseconds to achieve greater precision.

The existing K_MSEC() et. al. macros now return initializers for a
k_timeout_t.

The K_NO_WAIT and K_FOREVER constants have now become k_timeout_t
values, which means they cannot be operated on as integers.
Applications which have their own APIs that need to inspect these
vs. user-provided timeouts can now use a K_TIMEOUT_EQ() predicate to
test for equality.

Timer drivers, which receive an integer tick count in ther
z_clock_set_timeout() functions, now use the integer-valued
K_TICKS_FOREVER constant instead of K_FOREVER.

For the initial release, to preserve source compatibility, a
CONFIG_LEGACY_TIMEOUT_API kconfig is provided.  When true, the
k_timeout_t will remain a compatible 32 bit value that will work with
any legacy Zephyr application.

Some subsystems present timeout (or timeout-like) values to their own
users as APIs that would re-use the kernel's own constants and
conventions.  These will require some minor design work to adapt to
the new scheme (in most cases just using k_timeout_t directly in their
own API), and they have not been changed in this patch, instead
selecting CONFIG_LEGACY_TIMEOUT_API via kconfig.  These subsystems
include: CAN Bus, the Microbit display driver, I2S, LoRa modem
drivers, the UART Async API, Video hardware drivers, the console
subsystem, and the network buffer abstraction.

k_sleep() now takes a k_timeout_t argument, with a k_msleep() variant
provided that works identically to the original API.

Most of the changes here are just type/configuration management and
documentation, but there are logic changes in mempool, where a loop
that used a timeout numerically has been reworked using a new
z_timeout_end_calc() predicate.  Also in queue.c, a (when POLL was
enabled) a similar loop was needlessly used to try to retry the
k_poll() call after a spurious failure.  But k_poll() does not fail
spuriously, so the loop was removed.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2020-03-31 19:40:47 -04:00

289 lines
7.2 KiB
C

/*
* Copyright (c) 2016 Wind River Systems, Inc.
*
* SPDX-License-Identifier: Apache-2.0
*/
/**
* @file @brief mutex kernel services
*
* This module contains routines for handling mutex locking and unlocking.
*
* Mutexes implement a priority inheritance algorithm that boosts the priority
* level of the owning thread to match the priority level of the highest
* priority thread waiting on the mutex.
*
* Each mutex that contributes to priority inheritance must be released in the
* reverse order in which it was acquired. Furthermore each subsequent mutex
* that contributes to raising the owning thread's priority level must be
* acquired at a point after the most recent "bumping" of the priority level.
*
* For example, if thread A has two mutexes contributing to the raising of its
* priority level, the second mutex M2 must be acquired by thread A after
* thread A's priority level was bumped due to owning the first mutex M1.
* When releasing the mutex, thread A must release M2 before it releases M1.
* Failure to follow this nested model may result in threads running at
* unexpected priority levels (too high, or too low).
*/
#include <kernel.h>
#include <kernel_structs.h>
#include <toolchain.h>
#include <linker/sections.h>
#include <ksched.h>
#include <wait_q.h>
#include <sys/dlist.h>
#include <debug/object_tracing_common.h>
#include <errno.h>
#include <init.h>
#include <syscall_handler.h>
#include <tracing/tracing.h>
#include <sys/check.h>
/* We use a global spinlock here because some of the synchronization
* is protecting things like owner thread priorities which aren't
* "part of" a single k_mutex. Should move those bits of the API
* under the scheduler lock so we can break this up.
*/
static struct k_spinlock lock;
#ifdef CONFIG_OBJECT_TRACING
struct k_mutex *_trace_list_k_mutex;
/*
* Complete initialization of statically defined mutexes.
*/
static int init_mutex_module(struct device *dev)
{
ARG_UNUSED(dev);
Z_STRUCT_SECTION_FOREACH(k_mutex, mutex) {
SYS_TRACING_OBJ_INIT(k_mutex, mutex);
}
return 0;
}
SYS_INIT(init_mutex_module, PRE_KERNEL_1, CONFIG_KERNEL_INIT_PRIORITY_OBJECTS);
#endif /* CONFIG_OBJECT_TRACING */
int z_impl_k_mutex_init(struct k_mutex *mutex)
{
mutex->owner = NULL;
mutex->lock_count = 0U;
sys_trace_void(SYS_TRACE_ID_MUTEX_INIT);
z_waitq_init(&mutex->wait_q);
SYS_TRACING_OBJ_INIT(k_mutex, mutex);
z_object_init(mutex);
sys_trace_end_call(SYS_TRACE_ID_MUTEX_INIT);
return 0;
}
#ifdef CONFIG_USERSPACE
static inline int z_vrfy_k_mutex_init(struct k_mutex *mutex)
{
Z_OOPS(Z_SYSCALL_OBJ_INIT(mutex, K_OBJ_MUTEX));
return z_impl_k_mutex_init(mutex);
}
#include <syscalls/k_mutex_init_mrsh.c>
#endif
static s32_t new_prio_for_inheritance(s32_t target, s32_t limit)
{
int new_prio = z_is_prio_higher(target, limit) ? target : limit;
new_prio = z_get_new_prio_with_ceiling(new_prio);
return new_prio;
}
static bool adjust_owner_prio(struct k_mutex *mutex, s32_t new_prio)
{
if (mutex->owner->base.prio != new_prio) {
K_DEBUG("%p (ready (y/n): %c) prio changed to %d (was %d)\n",
mutex->owner, z_is_thread_ready(mutex->owner) ?
'y' : 'n',
new_prio, mutex->owner->base.prio);
return z_set_prio(mutex->owner, new_prio);
}
return false;
}
int z_impl_k_mutex_lock(struct k_mutex *mutex, k_timeout_t timeout)
{
int new_prio;
k_spinlock_key_t key;
bool resched = false;
sys_trace_void(SYS_TRACE_ID_MUTEX_LOCK);
key = k_spin_lock(&lock);
if (likely((mutex->lock_count == 0U) || (mutex->owner == _current))) {
mutex->owner_orig_prio = (mutex->lock_count == 0U) ?
_current->base.prio :
mutex->owner_orig_prio;
mutex->lock_count++;
mutex->owner = _current;
K_DEBUG("%p took mutex %p, count: %d, orig prio: %d\n",
_current, mutex, mutex->lock_count,
mutex->owner_orig_prio);
k_spin_unlock(&lock, key);
sys_trace_end_call(SYS_TRACE_ID_MUTEX_LOCK);
return 0;
}
if (unlikely(K_TIMEOUT_EQ(timeout, K_NO_WAIT))) {
k_spin_unlock(&lock, key);
sys_trace_end_call(SYS_TRACE_ID_MUTEX_LOCK);
return -EBUSY;
}
new_prio = new_prio_for_inheritance(_current->base.prio,
mutex->owner->base.prio);
K_DEBUG("adjusting prio up on mutex %p\n", mutex);
if (z_is_prio_higher(new_prio, mutex->owner->base.prio)) {
resched = adjust_owner_prio(mutex, new_prio);
}
int got_mutex = z_pend_curr(&lock, key, &mutex->wait_q, timeout);
K_DEBUG("on mutex %p got_mutex value: %d\n", mutex, got_mutex);
K_DEBUG("%p got mutex %p (y/n): %c\n", _current, mutex,
got_mutex ? 'y' : 'n');
if (got_mutex == 0) {
sys_trace_end_call(SYS_TRACE_ID_MUTEX_LOCK);
return 0;
}
/* timed out */
K_DEBUG("%p timeout on mutex %p\n", _current, mutex);
key = k_spin_lock(&lock);
struct k_thread *waiter = z_waitq_head(&mutex->wait_q);
new_prio = (waiter != NULL) ?
new_prio_for_inheritance(waiter->base.prio, mutex->owner_orig_prio) :
mutex->owner_orig_prio;
K_DEBUG("adjusting prio down on mutex %p\n", mutex);
resched = adjust_owner_prio(mutex, new_prio) || resched;
if (resched) {
z_reschedule(&lock, key);
} else {
k_spin_unlock(&lock, key);
}
sys_trace_end_call(SYS_TRACE_ID_MUTEX_LOCK);
return -EAGAIN;
}
#ifdef CONFIG_USERSPACE
static inline int z_vrfy_k_mutex_lock(struct k_mutex *mutex,
k_timeout_t timeout)
{
Z_OOPS(Z_SYSCALL_OBJ(mutex, K_OBJ_MUTEX));
return z_impl_k_mutex_lock(mutex, timeout);
}
#include <syscalls/k_mutex_lock_mrsh.c>
#endif
int z_impl_k_mutex_unlock(struct k_mutex *mutex)
{
struct k_thread *new_owner;
CHECKIF(mutex->owner == NULL) {
return -EINVAL;
}
/*
* The current thread does not own the mutex.
*/
CHECKIF(mutex->owner != _current) {
return -EPERM;
}
/*
* Attempt to unlock a mutex which is unlocked. mutex->lock_count
* cannot be zero if the current thread is equal to mutex->owner,
* therefore no underflow check is required. Use assert to catch
* undefined behavior.
*/
__ASSERT_NO_MSG(mutex->lock_count > 0U);
sys_trace_void(SYS_TRACE_ID_MUTEX_UNLOCK);
z_sched_lock();
K_DEBUG("mutex %p lock_count: %d\n", mutex, mutex->lock_count);
/*
* If we are the owner and count is greater than 1, then decrement
* the count and return and keep current thread as the owner.
*/
if (mutex->lock_count - 1U != 0U) {
mutex->lock_count--;
goto k_mutex_unlock_return;
}
k_spinlock_key_t key = k_spin_lock(&lock);
adjust_owner_prio(mutex, mutex->owner_orig_prio);
/* Get the new owner, if any */
new_owner = z_unpend_first_thread(&mutex->wait_q);
mutex->owner = new_owner;
K_DEBUG("new owner of mutex %p: %p (prio: %d)\n",
mutex, new_owner, new_owner ? new_owner->base.prio : -1000);
if (new_owner != NULL) {
/*
* new owner is already of higher or equal prio than first
* waiter since the wait queue is priority-based: no need to
* ajust its priority
*/
mutex->owner_orig_prio = new_owner->base.prio;
arch_thread_return_value_set(new_owner, 0);
z_ready_thread(new_owner);
z_reschedule(&lock, key);
} else {
mutex->lock_count = 0U;
k_spin_unlock(&lock, key);
}
k_mutex_unlock_return:
k_sched_unlock();
sys_trace_end_call(SYS_TRACE_ID_MUTEX_UNLOCK);
return 0;
}
#ifdef CONFIG_USERSPACE
static inline int z_vrfy_k_mutex_unlock(struct k_mutex *mutex)
{
Z_OOPS(Z_SYSCALL_OBJ(mutex, K_OBJ_MUTEX));
return z_impl_k_mutex_unlock(mutex);
}
#include <syscalls/k_mutex_unlock_mrsh.c>
#endif