zephyr/kernel/queue.c

363 lines
8.1 KiB
C
Raw Normal View History

/*
* Copyright (c) 2010-2016 Wind River Systems, Inc.
*
* SPDX-License-Identifier: Apache-2.0
*/
/**
* @file
*
* @brief dynamic-size QUEUE object.
*/
#include <kernel.h>
#include <kernel_structs.h>
#include <debug/object_tracing_common.h>
#include <toolchain.h>
#include <linker/sections.h>
#include <wait_q.h>
#include <ksched.h>
#include <misc/sflist.h>
#include <init.h>
#include <syscall_handler.h>
extern struct k_queue _k_queue_list_start[];
extern struct k_queue _k_queue_list_end[];
struct alloc_node {
sys_sfnode_t node;
void *data;
};
void *z_queue_node_peek(sys_sfnode_t *node, bool needs_free)
{
void *ret;
if (node && sys_sfnode_flags_get(node)) {
/* If the flag is set, then the enqueue operation for this item
* did a behind-the scenes memory allocation of an alloc_node
* struct, which is what got put in the queue. Free it and pass
* back the data pointer.
*/
struct alloc_node *anode;
anode = CONTAINER_OF(node, struct alloc_node, node);
ret = anode->data;
if (needs_free) {
k_free(anode);
}
} else {
/* Data was directly placed in the queue, the first 4 bytes
* reserved for the linked list. User mode isn't allowed to
* do this, although it can get data sent this way.
*/
ret = (void *)node;
}
return ret;
}
#ifdef CONFIG_OBJECT_TRACING
struct k_queue *_trace_list_k_queue;
/*
* Complete initialization of statically defined queues.
*/
static int init_queue_module(struct device *dev)
{
ARG_UNUSED(dev);
struct k_queue *queue;
for (queue = _k_queue_list_start; queue < _k_queue_list_end; queue++) {
SYS_TRACING_OBJ_INIT(k_queue, queue);
}
return 0;
}
SYS_INIT(init_queue_module, PRE_KERNEL_1, CONFIG_KERNEL_INIT_PRIORITY_OBJECTS);
#endif /* CONFIG_OBJECT_TRACING */
void _impl_k_queue_init(struct k_queue *queue)
{
sys_sflist_init(&queue->data_q);
_waitq_init(&queue->wait_q);
#if defined(CONFIG_POLL)
sys_dlist_init(&queue->poll_events);
#endif
SYS_TRACING_OBJ_INIT(k_queue, queue);
_k_object_init(queue);
}
#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_queue_init, queue_ptr)
{
struct k_queue *queue = (struct k_queue *)queue_ptr;
Z_OOPS(Z_SYSCALL_OBJ_NEVER_INIT(queue, K_OBJ_QUEUE));
_impl_k_queue_init(queue);
return 0;
}
#endif
#if !defined(CONFIG_POLL)
static void prepare_thread_to_run(struct k_thread *thread, void *data)
{
_ready_thread(thread);
_set_thread_return_value_with_data(thread, 0, data);
}
#endif /* CONFIG_POLL */
#ifdef CONFIG_POLL
static inline void handle_poll_events(struct k_queue *queue, u32_t state)
{
_handle_obj_poll_events(&queue->poll_events, state);
}
#endif
void _impl_k_queue_cancel_wait(struct k_queue *queue)
{
unsigned int key = irq_lock();
#if !defined(CONFIG_POLL)
struct k_thread *first_pending_thread;
first_pending_thread = _unpend_first_thread(&queue->wait_q);
if (first_pending_thread) {
prepare_thread_to_run(first_pending_thread, NULL);
}
#else
handle_poll_events(queue, K_POLL_STATE_NOT_READY);
#endif /* !CONFIG_POLL */
_reschedule(key);
}
#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER1_SIMPLE_VOID(k_queue_cancel_wait, K_OBJ_QUEUE,
struct k_queue *);
#endif
static int queue_insert(struct k_queue *queue, void *prev, void *data,
bool alloc)
{
unsigned int key = irq_lock();
#if !defined(CONFIG_POLL)
struct k_thread *first_pending_thread;
first_pending_thread = _unpend_first_thread(&queue->wait_q);
if (first_pending_thread) {
prepare_thread_to_run(first_pending_thread, data);
_reschedule(key);
return 0;
}
#endif /* !CONFIG_POLL */
/* Only need to actually allocate if no threads are pending */
if (alloc) {
struct alloc_node *anode;
anode = z_thread_malloc(sizeof(*anode));
if (!anode) {
return -ENOMEM;
}
anode->data = data;
sys_sfnode_init(&anode->node, 0x1);
data = anode;
} else {
sys_sfnode_init(data, 0x0);
}
sys_sflist_insert(&queue->data_q, prev, data);
#if defined(CONFIG_POLL)
handle_poll_events(queue, K_POLL_STATE_DATA_AVAILABLE);
#endif /* CONFIG_POLL */
_reschedule(key);
return 0;
}
void k_queue_insert(struct k_queue *queue, void *prev, void *data)
{
(void)queue_insert(queue, prev, data, false);
}
void k_queue_append(struct k_queue *queue, void *data)
{
(void)queue_insert(queue, sys_sflist_peek_tail(&queue->data_q),
data, false);
}
void k_queue_prepend(struct k_queue *queue, void *data)
{
(void)queue_insert(queue, NULL, data, false);
}
int _impl_k_queue_alloc_append(struct k_queue *queue, void *data)
{
return queue_insert(queue, sys_sflist_peek_tail(&queue->data_q), data,
true);
}
#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_queue_alloc_append, queue, data)
{
Z_OOPS(Z_SYSCALL_OBJ(queue, K_OBJ_QUEUE));
return _impl_k_queue_alloc_append((struct k_queue *)queue,
(void *)data);
}
#endif
int _impl_k_queue_alloc_prepend(struct k_queue *queue, void *data)
{
return queue_insert(queue, NULL, data, true);
}
#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_queue_alloc_prepend, queue, data)
{
Z_OOPS(Z_SYSCALL_OBJ(queue, K_OBJ_QUEUE));
return _impl_k_queue_alloc_prepend((struct k_queue *)queue,
(void *)data);
}
#endif
void k_queue_append_list(struct k_queue *queue, void *head, void *tail)
{
__ASSERT(head && tail, "invalid head or tail");
unsigned int key = irq_lock();
#if !defined(CONFIG_POLL)
struct k_thread *thread;
while (head && ((thread = _unpend_first_thread(&queue->wait_q)))) {
prepare_thread_to_run(thread, head);
head = *(void **)head;
}
if (head) {
sys_sflist_append_list(&queue->data_q, head, tail);
}
#else
sys_sflist_append_list(&queue->data_q, head, tail);
handle_poll_events(queue, K_POLL_STATE_DATA_AVAILABLE);
#endif /* !CONFIG_POLL */
_reschedule(key);
}
void k_queue_merge_slist(struct k_queue *queue, sys_slist_t *list)
{
__ASSERT(!sys_slist_is_empty(list), "list must not be empty");
/*
* note: this works as long as:
* - the slist implementation keeps the next pointer as the first
* field of the node object type
* - list->tail->next = NULL.
* - sflist implementation only differs from slist by stuffing
* flag bytes in the lower order bits of the data pointer
* - source list is really an slist and not an sflist with flags set
*/
k_queue_append_list(queue, list->head, list->tail);
sys_slist_init(list);
}
#if defined(CONFIG_POLL)
static void *k_queue_poll(struct k_queue *queue, s32_t timeout)
{
struct k_poll_event event;
kernel/queue: Fix spurious NULL exit condition when using timeouts The queue loop when CONFIG_POLL is in used has an inherent race between the return of k_poll() and the inspection of the list where no lock can be held. Other contending readers of the same queue can sneak in and steal the item out of the list before the current thread gets to the sys_sflist_get() call, and the current loop will (if it has a timeout) spuriously return NULL before the timeout expires. It's not even a hard race to exercise. Consider three threads at different priorities: High (which can be an ISR too), Mid, and Low: 1. Mid and Low both enter k_queue_get() and sleep inside k_poll() on an empty queue. 2. High comes along and calls k_queue_insert(). The queue code then wakes up Mid, and reschedules, but because High is still running Mid doesn't get to run yet. 3. High inserts a SECOND item. The queue then unpends the next thread in the list (Low), and readies it to run. But as before, it won't be scheduled yet. 4. Now High sleeps (or if it's an interrupt, exits), and Mid gets to run. It dequeues and returns the item it was delivered normally. 5. But Mid is still running! So it re-enters the loop it's sitting in and calls k_queue_get() again, which sees and returns the second item in the queue synchronously. Then it calls it a third time and goes to sleep because the queue is empty. 6. Finally, Low wakes up to find an empty queue, and returns NULL despite the fact that the timeout hadn't expired. The fix is simple enough: check the timeout expiration inside the loop so we don't return early. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-06-04 09:25:14 -07:00
int err, elapsed = 0, done = 0;
unsigned int key;
void *val;
kernel/queue: Fix spurious NULL exit condition when using timeouts The queue loop when CONFIG_POLL is in used has an inherent race between the return of k_poll() and the inspection of the list where no lock can be held. Other contending readers of the same queue can sneak in and steal the item out of the list before the current thread gets to the sys_sflist_get() call, and the current loop will (if it has a timeout) spuriously return NULL before the timeout expires. It's not even a hard race to exercise. Consider three threads at different priorities: High (which can be an ISR too), Mid, and Low: 1. Mid and Low both enter k_queue_get() and sleep inside k_poll() on an empty queue. 2. High comes along and calls k_queue_insert(). The queue code then wakes up Mid, and reschedules, but because High is still running Mid doesn't get to run yet. 3. High inserts a SECOND item. The queue then unpends the next thread in the list (Low), and readies it to run. But as before, it won't be scheduled yet. 4. Now High sleeps (or if it's an interrupt, exits), and Mid gets to run. It dequeues and returns the item it was delivered normally. 5. But Mid is still running! So it re-enters the loop it's sitting in and calls k_queue_get() again, which sees and returns the second item in the queue synchronously. Then it calls it a third time and goes to sleep because the queue is empty. 6. Finally, Low wakes up to find an empty queue, and returns NULL despite the fact that the timeout hadn't expired. The fix is simple enough: check the timeout expiration inside the loop so we don't return early. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-06-04 09:25:14 -07:00
u32_t start;
k_poll_event_init(&event, K_POLL_TYPE_FIFO_DATA_AVAILABLE,
K_POLL_MODE_NOTIFY_ONLY, queue);
kernel/queue: Fix spurious NULL exit condition when using timeouts The queue loop when CONFIG_POLL is in used has an inherent race between the return of k_poll() and the inspection of the list where no lock can be held. Other contending readers of the same queue can sneak in and steal the item out of the list before the current thread gets to the sys_sflist_get() call, and the current loop will (if it has a timeout) spuriously return NULL before the timeout expires. It's not even a hard race to exercise. Consider three threads at different priorities: High (which can be an ISR too), Mid, and Low: 1. Mid and Low both enter k_queue_get() and sleep inside k_poll() on an empty queue. 2. High comes along and calls k_queue_insert(). The queue code then wakes up Mid, and reschedules, but because High is still running Mid doesn't get to run yet. 3. High inserts a SECOND item. The queue then unpends the next thread in the list (Low), and readies it to run. But as before, it won't be scheduled yet. 4. Now High sleeps (or if it's an interrupt, exits), and Mid gets to run. It dequeues and returns the item it was delivered normally. 5. But Mid is still running! So it re-enters the loop it's sitting in and calls k_queue_get() again, which sees and returns the second item in the queue synchronously. Then it calls it a third time and goes to sleep because the queue is empty. 6. Finally, Low wakes up to find an empty queue, and returns NULL despite the fact that the timeout hadn't expired. The fix is simple enough: check the timeout expiration inside the loop so we don't return early. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-06-04 09:25:14 -07:00
if (timeout != K_FOREVER) {
start = k_uptime_get_32();
}
do {
event.state = K_POLL_STATE_NOT_READY;
kernel/queue: Fix spurious NULL exit condition when using timeouts The queue loop when CONFIG_POLL is in used has an inherent race between the return of k_poll() and the inspection of the list where no lock can be held. Other contending readers of the same queue can sneak in and steal the item out of the list before the current thread gets to the sys_sflist_get() call, and the current loop will (if it has a timeout) spuriously return NULL before the timeout expires. It's not even a hard race to exercise. Consider three threads at different priorities: High (which can be an ISR too), Mid, and Low: 1. Mid and Low both enter k_queue_get() and sleep inside k_poll() on an empty queue. 2. High comes along and calls k_queue_insert(). The queue code then wakes up Mid, and reschedules, but because High is still running Mid doesn't get to run yet. 3. High inserts a SECOND item. The queue then unpends the next thread in the list (Low), and readies it to run. But as before, it won't be scheduled yet. 4. Now High sleeps (or if it's an interrupt, exits), and Mid gets to run. It dequeues and returns the item it was delivered normally. 5. But Mid is still running! So it re-enters the loop it's sitting in and calls k_queue_get() again, which sees and returns the second item in the queue synchronously. Then it calls it a third time and goes to sleep because the queue is empty. 6. Finally, Low wakes up to find an empty queue, and returns NULL despite the fact that the timeout hadn't expired. The fix is simple enough: check the timeout expiration inside the loop so we don't return early. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-06-04 09:25:14 -07:00
err = k_poll(&event, 1, timeout - elapsed);
if (err && err != -EAGAIN) {
return NULL;
}
/* sys_sflist_* aren't threadsafe, so must be always protected
* by irq_lock.
*/
key = irq_lock();
val = z_queue_node_peek(sys_sflist_get(&queue->data_q), true);
irq_unlock(key);
kernel/queue: Fix spurious NULL exit condition when using timeouts The queue loop when CONFIG_POLL is in used has an inherent race between the return of k_poll() and the inspection of the list where no lock can be held. Other contending readers of the same queue can sneak in and steal the item out of the list before the current thread gets to the sys_sflist_get() call, and the current loop will (if it has a timeout) spuriously return NULL before the timeout expires. It's not even a hard race to exercise. Consider three threads at different priorities: High (which can be an ISR too), Mid, and Low: 1. Mid and Low both enter k_queue_get() and sleep inside k_poll() on an empty queue. 2. High comes along and calls k_queue_insert(). The queue code then wakes up Mid, and reschedules, but because High is still running Mid doesn't get to run yet. 3. High inserts a SECOND item. The queue then unpends the next thread in the list (Low), and readies it to run. But as before, it won't be scheduled yet. 4. Now High sleeps (or if it's an interrupt, exits), and Mid gets to run. It dequeues and returns the item it was delivered normally. 5. But Mid is still running! So it re-enters the loop it's sitting in and calls k_queue_get() again, which sees and returns the second item in the queue synchronously. Then it calls it a third time and goes to sleep because the queue is empty. 6. Finally, Low wakes up to find an empty queue, and returns NULL despite the fact that the timeout hadn't expired. The fix is simple enough: check the timeout expiration inside the loop so we don't return early. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-06-04 09:25:14 -07:00
if (!val && timeout != K_FOREVER) {
elapsed = k_uptime_get_32() - start;
done = elapsed > timeout;
}
} while (!val && !done);
return val;
}
#endif /* CONFIG_POLL */
void *_impl_k_queue_get(struct k_queue *queue, s32_t timeout)
{
unsigned int key;
void *data;
key = irq_lock();
if (likely(!sys_sflist_is_empty(&queue->data_q))) {
sys_sfnode_t *node;
node = sys_sflist_get_not_empty(&queue->data_q);
data = z_queue_node_peek(node, true);
irq_unlock(key);
return data;
}
if (timeout == K_NO_WAIT) {
irq_unlock(key);
return NULL;
}
#if defined(CONFIG_POLL)
irq_unlock(key);
return k_queue_poll(queue, timeout);
#else
int ret = _pend_current_thread(key, &queue->wait_q, timeout);
return ret ? NULL : _current->base.swap_data;
#endif /* CONFIG_POLL */
}
#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_queue_get, queue, timeout_p)
{
s32_t timeout = timeout_p;
Z_OOPS(Z_SYSCALL_OBJ(queue, K_OBJ_QUEUE));
return (u32_t)_impl_k_queue_get((struct k_queue *)queue, timeout);
}
Z_SYSCALL_HANDLER1_SIMPLE(k_queue_is_empty, K_OBJ_QUEUE, struct k_queue *);
Z_SYSCALL_HANDLER1_SIMPLE(k_queue_peek_head, K_OBJ_QUEUE, struct k_queue *);
Z_SYSCALL_HANDLER1_SIMPLE(k_queue_peek_tail, K_OBJ_QUEUE, struct k_queue *);
#endif /* CONFIG_USERSPACE */