acrn-hypervisor/hypervisor/debug/sbuf.c

195 lines
5.2 KiB
C

/*
* SHARED BUFFER
*
* Copyright (C) 2017 Intel Corporation. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* * Neither the name of Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Li Fei <fei1.li@intel.com>
*
*/
#include <hv_lib.h>
#include <acrn_common.h>
#include <hv_arch.h>
#include <hv_debug.h>
DEFINE_CPU_DATA(uint64_t * [ACRN_SBUF_ID_MAX], sbuf);
static inline bool sbuf_is_empty(struct shared_buf *sbuf)
{
return (sbuf->head == sbuf->tail);
}
static inline uint32_t sbuf_next_ptr(uint32_t pos,
uint32_t span, uint32_t scope)
{
pos += span;
pos = (pos >= scope) ? (pos - scope) : pos;
return pos;
}
static inline uint32_t sbuf_calculate_allocate_size(uint32_t ele_num,
uint32_t ele_size)
{
uint64_t sbuf_allocate_size;
sbuf_allocate_size = ele_num * ele_size;
sbuf_allocate_size += SBUF_HEAD_SIZE;
if (sbuf_allocate_size > SBUF_MAX_SIZE) {
pr_err("%s, num=0x%x, size=0x%x exceed 0x%x",
__func__, ele_num, ele_size, SBUF_MAX_SIZE);
return 0;
}
return sbuf_allocate_size;
}
struct shared_buf *sbuf_allocate(uint32_t ele_num, uint32_t ele_size)
{
struct shared_buf *sbuf;
uint32_t sbuf_allocate_size;
if (!ele_num || !ele_size) {
pr_err("%s invalid parameter!", __func__);
return NULL;
}
sbuf_allocate_size = sbuf_calculate_allocate_size(ele_num, ele_size);
if (!sbuf_allocate_size)
return NULL;
sbuf = malloc(sbuf_allocate_size);
if (sbuf == NULL) {
pr_err("%s no memory!", __func__);
return NULL;
}
memset(sbuf, 0, SBUF_HEAD_SIZE);
sbuf->ele_num = ele_num;
sbuf->ele_size = ele_size;
sbuf->size = ele_num * ele_size;
sbuf->magic = SBUF_MAGIC;
pr_info("%s ele_num=0x%x, ele_size=0x%x allocated",
__func__, ele_num, ele_size);
return sbuf;
}
void sbuf_free(struct shared_buf *sbuf)
{
if ((sbuf == NULL) || sbuf->magic != SBUF_MAGIC) {
pr_err("%s invalid parameter!", __func__);
return;
}
sbuf->magic = 0;
free(sbuf);
}
int sbuf_get(struct shared_buf *sbuf, uint8_t *data)
{
const void *from;
if ((sbuf == NULL) || (data == NULL))
return -EINVAL;
if (sbuf_is_empty(sbuf)) {
/* no data available */
return 0;
}
from = (void *)sbuf + SBUF_HEAD_SIZE + sbuf->head;
memcpy_s((void *)data, sbuf->ele_size, from, sbuf->ele_size);
sbuf->head = sbuf_next_ptr(sbuf->head, sbuf->ele_size, sbuf->size);
return sbuf->ele_size;
}
/**
* The high caller should guarantee each time there must have
* sbuf->ele_size data can be write form data and this function
* should guarantee execution atomically.
*
* flag:
* If OVERWRITE_EN set, buf can store (ele_num - 1) elements at most.
* Should use lock to guarantee that only one read or write at
* the same time.
* if OVERWRITE_EN not set, buf can store (ele_num - 1) elements
* at most. Shouldn't modify the sbuf->head.
*
* return:
* ele_size: write succeeded.
* 0: no write, buf is full
* negative: failed.
*/
int sbuf_put(struct shared_buf *sbuf, uint8_t *data)
{
void *to;
uint32_t next_tail;
bool trigger_overwrite = false;
if ((sbuf == NULL) || (data == NULL))
return -EINVAL;
next_tail = sbuf_next_ptr(sbuf->tail, sbuf->ele_size, sbuf->size);
/* if this write would trigger overrun */
if (next_tail == sbuf->head) {
/* accumulate overrun count if necessary */
sbuf->overrun_cnt += sbuf->flags & OVERRUN_CNT_EN;
if (!(sbuf->flags & OVERWRITE_EN)) {
/* if not enable over write, return here. */
return 0;
}
trigger_overwrite = true;
}
to = (void *)sbuf + SBUF_HEAD_SIZE + sbuf->tail;
memcpy_s(to, sbuf->ele_size, data, sbuf->ele_size);
if (trigger_overwrite) {
sbuf->head = sbuf_next_ptr(sbuf->head,
sbuf->ele_size, sbuf->size);
}
sbuf->tail = next_tail;
return sbuf->ele_size;
}
int sbuf_share_setup(uint32_t pcpu_id, uint32_t sbuf_id, uint64_t *hva)
{
if (pcpu_id >= (uint32_t) phy_cpu_num ||
sbuf_id >= ACRN_SBUF_ID_MAX)
return -EINVAL;
per_cpu(sbuf, pcpu_id)[sbuf_id] = hva;
return 0;
}