michaelh/zephyr
1
0
Fork 0
Code Releases Activity

kernel/pipe: Spinlockify

Browse source 
One spinlock per pipe object.  Also removed some vestigial locking
around _ready_thread().  That call is internally synchronized now.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This commit is contained in:
Andy Ross 2019-02-05 16:10:18 -08:00 • committed by Anas Nashif
commit f582b55dd6
2 changed files with 16 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

1
include/kernel.h
View file

@ -3631,6 +3631,7 @@ struct k_pipe {
size_t bytes_used; /**< # bytes used in buffer */ size_t bytes_used; /**< # bytes used in buffer */
size_t read_index; /**< Where in buffer to read from */ size_t read_index; /**< Where in buffer to read from */
size_t write_index; /**< Where in buffer to write */ size_t write_index; /**< Where in buffer to write */
struct k_spinlock lock; /**< Synchronization lock */
struct { struct {
_wait_q_t readers; /**< Reader wait queue */ _wait_q_t readers; /**< Reader wait queue */

35
kernel/pipes.c
View file

@ -413,8 +413,6 @@ static int pipe_return_code(size_t min_xfer, size_t bytes_remaining,
*/ */
static void pipe_thread_ready(struct k_thread *thread) static void pipe_thread_ready(struct k_thread *thread)
{ {
unsigned int key;
#if (CONFIG_NUM_PIPE_ASYNC_MSGS > 0) #if (CONFIG_NUM_PIPE_ASYNC_MSGS > 0)
if ((thread->base.thread_state & _THREAD_DUMMY) != 0) { if ((thread->base.thread_state & _THREAD_DUMMY) != 0) {
pipe_async_finish((struct k_pipe_async *)thread); pipe_async_finish((struct k_pipe_async *)thread);
@ -422,9 +420,7 @@ static void pipe_thread_ready(struct k_thread *thread)
} }
#endif #endif
key = irq_lock();
_ready_thread(thread); _ready_thread(thread);
irq_unlock(key);
} }
/** /**
@ -438,7 +434,6 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
struct k_thread *reader; struct k_thread *reader;
struct k_pipe_desc *desc; struct k_pipe_desc *desc;
sys_dlist_t xfer_list; sys_dlist_t xfer_list;
unsigned int key;
size_t num_bytes_written = 0; size_t num_bytes_written = 0;
size_t bytes_copied; size_t bytes_copied;
@ -446,7 +441,7 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
ARG_UNUSED(async_desc); ARG_UNUSED(async_desc);
#endif #endif
key = irq_lock(); k_spinlock_key_t key = k_spin_lock(&pipe->lock);
/* /*
* Create a list of "working readers" into which the data will be * Create a list of "working readers" into which the data will be
@ -456,13 +451,13 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
if (!pipe_xfer_prepare(&xfer_list, &reader, &pipe->wait_q.readers, if (!pipe_xfer_prepare(&xfer_list, &reader, &pipe->wait_q.readers,
pipe->size - pipe->bytes_used, bytes_to_write, pipe->size - pipe->bytes_used, bytes_to_write,
min_xfer, timeout)) { min_xfer, timeout)) {
irq_unlock(key); k_spin_unlock(&pipe->lock, key);
*bytes_written = 0; *bytes_written = 0;
return -EIO; return -EIO;
} }
_sched_lock(); _sched_lock();
irq_unlock(key); k_spin_unlock(&pipe->lock, key);
/* /*
* 1. 'xfer_list' currently contains a list of reader threads that can * 1. 'xfer_list' currently contains a list of reader threads that can
@ -489,9 +484,7 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
desc->bytes_to_xfer -= bytes_copied; desc->bytes_to_xfer -= bytes_copied;
/* The thread's read request has been satisfied. Ready it. */ /* The thread's read request has been satisfied. Ready it. */
key = irq_lock();
_ready_thread(thread); _ready_thread(thread);
irq_unlock(key);
thread = (struct k_thread *)sys_dlist_get(&xfer_list); thread = (struct k_thread *)sys_dlist_get(&xfer_list);
} }
@ -539,7 +532,7 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
* Lock interrupts and unlock the scheduler before * Lock interrupts and unlock the scheduler before
* manipulating the writers wait_q. * manipulating the writers wait_q.
*/ */
key = irq_lock(); k_spinlock_key_t key = k_spin_lock(&pipe->lock);
_sched_unlock_no_reschedule(); _sched_unlock_no_reschedule();
async_desc->desc.buffer = data + num_bytes_written; async_desc->desc.buffer = data + num_bytes_written;
@ -548,7 +541,7 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
_pend_thread((struct k_thread *) &async_desc->thread, _pend_thread((struct k_thread *) &async_desc->thread,
&pipe->wait_q.writers, K_FOREVER); &pipe->wait_q.writers, K_FOREVER);
_reschedule_irqlock(key); _reschedule(&pipe->lock, key);
return 0; return 0;
} }
#endif #endif
@ -564,9 +557,10 @@ int _k_pipe_put_internal(struct k_pipe *pipe, struct k_pipe_async *async_desc,
* Lock interrupts and unlock the scheduler before * Lock interrupts and unlock the scheduler before
* manipulating the writers wait_q. * manipulating the writers wait_q.
*/ */
key = irq_lock(); k_spinlock_key_t key = k_spin_lock(&pipe->lock);
_sched_unlock_no_reschedule(); _sched_unlock_no_reschedule();
(void)_pend_curr_irqlock(key, &pipe->wait_q.writers, timeout); (void)_pend_curr(&pipe->lock, key,
&pipe->wait_q.writers, timeout);
} else { } else {
k_sched_unlock(); k_sched_unlock();
} }
@ -583,14 +577,13 @@ int _impl_k_pipe_get(struct k_pipe *pipe, void *data, size_t bytes_to_read,
struct k_thread *writer; struct k_thread *writer;
struct k_pipe_desc *desc; struct k_pipe_desc *desc;
sys_dlist_t xfer_list; sys_dlist_t xfer_list;
unsigned int key;
size_t num_bytes_read = 0; size_t num_bytes_read = 0;
size_t bytes_copied; size_t bytes_copied;
__ASSERT(min_xfer <= bytes_to_read, ""); __ASSERT(min_xfer <= bytes_to_read, "");
__ASSERT(bytes_read != NULL, ""); __ASSERT(bytes_read != NULL, "");
key = irq_lock(); k_spinlock_key_t key = k_spin_lock(&pipe->lock);
/* /*
* Create a list of "working readers" into which the data will be * Create a list of "working readers" into which the data will be
@ -600,13 +593,13 @@ int _impl_k_pipe_get(struct k_pipe *pipe, void *data, size_t bytes_to_read,
if (!pipe_xfer_prepare(&xfer_list, &writer, &pipe->wait_q.writers, if (!pipe_xfer_prepare(&xfer_list, &writer, &pipe->wait_q.writers,
pipe->bytes_used, bytes_to_read, pipe->bytes_used, bytes_to_read,
min_xfer, timeout)) { min_xfer, timeout)) {
irq_unlock(key); k_spin_unlock(&pipe->lock, key);
*bytes_read = 0; *bytes_read = 0;
return -EIO; return -EIO;
} }
_sched_lock(); _sched_lock();
irq_unlock(key); k_spin_unlock(&pipe->lock, key);
num_bytes_read = pipe_buffer_get(pipe, data, bytes_to_read); num_bytes_read = pipe_buffer_get(pipe, data, bytes_to_read);
@ -706,9 +699,11 @@ int _impl_k_pipe_get(struct k_pipe *pipe, void *data, size_t bytes_to_read,
if (timeout != K_NO_WAIT) { if (timeout != K_NO_WAIT) {
_current->base.swap_data = &pipe_desc; _current->base.swap_data = &pipe_desc;
key = irq_lock(); k_spinlock_key_t key = k_spin_lock(&pipe->lock);
_sched_unlock_no_reschedule(); _sched_unlock_no_reschedule();
(void)_pend_curr_irqlock(key, &pipe->wait_q.readers, timeout); (void)_pend_curr(&pipe->lock, key,
&pipe->wait_q.readers, timeout);
} else { } else {
k_sched_unlock(); k_sched_unlock();
} }