blob: 852a99166144ff48ed4c9b510e1a91a59ea4ef33 [file] [log] [blame]
// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) 2020-21 Intel Corporation.
*/
#include "iosm_ipc_imem.h"
#include "iosm_ipc_task_queue.h"
/* Actual tasklet function, will be called whenever tasklet is scheduled.
* Calls event handler involves callback for each element in the message queue
*/
static void ipc_task_queue_handler(unsigned long data)
{
struct ipc_task_queue *ipc_task = (struct ipc_task_queue *)data;
unsigned int q_rpos = ipc_task->q_rpos;
/* Loop over the input queue contents. */
while (q_rpos != ipc_task->q_wpos) {
/* Get the current first queue element. */
struct ipc_task_queue_args *args = &ipc_task->args[q_rpos];
/* Process the input message. */
if (args->func)
args->response = args->func(args->ipc_imem, args->arg,
args->msg, args->size);
/* Signal completion for synchronous calls */
if (args->completion)
complete(args->completion);
/* Free message if copy was allocated. */
if (args->is_copy)
kfree(args->msg);
/* Set invalid queue element. Technically
* spin_lock_irqsave is not required here as
* the array element has been processed already
* so we can assume that immediately after processing
* ipc_task element, queue will not rotate again to
* ipc_task same element within such short time.
*/
args->completion = NULL;
args->func = NULL;
args->msg = NULL;
args->size = 0;
args->is_copy = false;
/* calculate the new read ptr and update the volatile read
* ptr
*/
q_rpos = (q_rpos + 1) % IPC_THREAD_QUEUE_SIZE;
ipc_task->q_rpos = q_rpos;
}
}
/* Free memory alloc and trigger completions left in the queue during dealloc */
static void ipc_task_queue_cleanup(struct ipc_task_queue *ipc_task)
{
unsigned int q_rpos = ipc_task->q_rpos;
while (q_rpos != ipc_task->q_wpos) {
struct ipc_task_queue_args *args = &ipc_task->args[q_rpos];
if (args->completion)
complete(args->completion);
if (args->is_copy)
kfree(args->msg);
q_rpos = (q_rpos + 1) % IPC_THREAD_QUEUE_SIZE;
ipc_task->q_rpos = q_rpos;
}
}
/* Add a message to the queue and trigger the ipc_task. */
static int
ipc_task_queue_add_task(struct iosm_imem *ipc_imem,
int arg, void *msg,
int (*func)(struct iosm_imem *ipc_imem, int arg,
void *msg, size_t size),
size_t size, bool is_copy, bool wait)
{
struct tasklet_struct *ipc_tasklet = ipc_imem->ipc_task->ipc_tasklet;
struct ipc_task_queue *ipc_task = &ipc_imem->ipc_task->ipc_queue;
struct completion completion;
unsigned int pos, nextpos;
unsigned long flags;
int result = -EIO;
init_completion(&completion);
/* tasklet send may be called from both interrupt or thread
* context, therefore protect queue operation by spinlock
*/
spin_lock_irqsave(&ipc_task->q_lock, flags);
pos = ipc_task->q_wpos;
nextpos = (pos + 1) % IPC_THREAD_QUEUE_SIZE;
/* Get next queue position. */
if (nextpos != ipc_task->q_rpos) {
/* Get the reference to the queue element and save the passed
* values.
*/
ipc_task->args[pos].arg = arg;
ipc_task->args[pos].msg = msg;
ipc_task->args[pos].func = func;
ipc_task->args[pos].ipc_imem = ipc_imem;
ipc_task->args[pos].size = size;
ipc_task->args[pos].is_copy = is_copy;
ipc_task->args[pos].completion = wait ? &completion : NULL;
ipc_task->args[pos].response = -1;
/* apply write barrier so that ipc_task->q_rpos elements
* are updated before ipc_task->q_wpos is being updated.
*/
smp_wmb();
/* Update the status of the free queue space. */
ipc_task->q_wpos = nextpos;
result = 0;
}
spin_unlock_irqrestore(&ipc_task->q_lock, flags);
if (result == 0) {
tasklet_schedule(ipc_tasklet);
if (wait) {
wait_for_completion(&completion);
result = ipc_task->args[pos].response;
}
} else {
dev_err(ipc_imem->ipc_task->dev, "queue is full");
}
return result;
}
int ipc_task_queue_send_task(struct iosm_imem *imem,
int (*func)(struct iosm_imem *ipc_imem, int arg,
void *msg, size_t size),
int arg, void *msg, size_t size, bool wait)
{
bool is_copy = false;
void *copy = msg;
int ret = -ENOMEM;
if (size > 0) {
copy = kmemdup(msg, size, GFP_ATOMIC);
if (!copy)
goto out;
is_copy = true;
}
ret = ipc_task_queue_add_task(imem, arg, copy, func,
size, is_copy, wait);
if (ret < 0) {
dev_err(imem->ipc_task->dev,
"add task failed for %ps %d, %p, %zu, %d", func, arg,
copy, size, is_copy);
if (is_copy)
kfree(copy);
goto out;
}
ret = 0;
out:
return ret;
}
int ipc_task_init(struct ipc_task *ipc_task)
{
struct ipc_task_queue *ipc_queue = &ipc_task->ipc_queue;
ipc_task->ipc_tasklet = kzalloc(sizeof(*ipc_task->ipc_tasklet),
GFP_KERNEL);
if (!ipc_task->ipc_tasklet)
return -ENOMEM;
/* Initialize the spinlock needed to protect the message queue of the
* ipc_task
*/
spin_lock_init(&ipc_queue->q_lock);
tasklet_init(ipc_task->ipc_tasklet, ipc_task_queue_handler,
(unsigned long)ipc_queue);
return 0;
}
void ipc_task_deinit(struct ipc_task *ipc_task)
{
tasklet_kill(ipc_task->ipc_tasklet);
kfree(ipc_task->ipc_tasklet);
/* This will free/complete any outstanding messages,
* without calling the actual handler
*/
ipc_task_queue_cleanup(&ipc_task->ipc_queue);
}