adafruit-mirror/zephyr
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 9

kernel: pipes: Add spin lock/unlock barrier pair

Browse source 
Adds a spin lock/unlock barrier pair after a pipe thread wakes.
After the list of waiting threads is generated, it is possible for
threads on that list to timeout and be removed from the wait queue.
However, since that list was generated before the timeout occurred,
the timed-out thread must wait until the copying is done (the
pipe's spin-lock has been released).

Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
This commit is contained in:
Peter Mitsis 2023-01-05 16:16:13 -05:00 committed by Stephanos Ioannidis
parent 31dfd84fd5
commit 59cb96e802
1 changed files with 20 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

20
kernel/pipes.c
View file

@ -478,6 +478,16 @@ int z_impl_k_pipe_put(struct k_pipe *pipe, void *data, size_t bytes_to_write,
z_sched_wait(&pipe->lock, key, &pipe->wait_q.writers, timeout, NULL); z_sched_wait(&pipe->lock, key, &pipe->wait_q.writers, timeout, NULL);
/*
* On SMP systems, threads in the processing list may timeout before
* the data has finished copying. The following spin lock/unlock pair
* prevents those threads from executing further until the data copying
* is complete.
*/
key = k_spin_lock(&pipe->lock);
k_spin_unlock(&pipe->lock, key);
*bytes_written = bytes_to_write - src_desc->bytes_to_xfer; *bytes_written = bytes_to_write - src_desc->bytes_to_xfer;
int ret = pipe_return_code(min_xfer, src_desc->bytes_to_xfer, int ret = pipe_return_code(min_xfer, src_desc->bytes_to_xfer,
@ -648,6 +658,16 @@ static int pipe_get_internal(k_spinlock_key_t key, struct k_pipe *pipe,
z_sched_wait(&pipe->lock, key, &pipe->wait_q.readers, timeout, NULL); z_sched_wait(&pipe->lock, key, &pipe->wait_q.readers, timeout, NULL);
/*
* On SMP systems, threads in the processing list may timeout before
* the data has finished copying. The following spin lock/unlock pair
* prevents those threads from executing further until the data copying
* is complete.
*/
key = k_spin_lock(&pipe->lock);
k_spin_unlock(&pipe->lock, key);
*bytes_read = bytes_to_read - dest_desc->bytes_to_xfer; *bytes_read = bytes_to_read - dest_desc->bytes_to_xfer;
int ret = pipe_return_code(min_xfer, dest_desc->bytes_to_xfer, int ret = pipe_return_code(min_xfer, dest_desc->bytes_to_xfer,