mbed-os/targets/TARGET_STM/stm_spi_api.c

800 lines
24 KiB
C

/* mbed Microcontroller Library
*******************************************************************************
* Copyright (c) 2015, STMicroelectronics
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* 2. 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.
* 3. Neither the name of STMicroelectronics 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 HOLDER 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.
*******************************************************************************
*/
#include "mbed_assert.h"
#include "mbed_error.h"
#include "mbed_debug.h"
#include "spi_api.h"
#if DEVICE_SPI
#include <stdbool.h>
#include <math.h>
#include <string.h>
#include "cmsis.h"
#include "pinmap.h"
#include "PeripheralPins.h"
#include "spi_device.h"
#if DEVICE_SPI_ASYNCH
#define SPI_INST(obj) ((SPI_TypeDef *)(obj->spi.spi))
#else
#define SPI_INST(obj) ((SPI_TypeDef *)(obj->spi))
#endif
#if DEVICE_SPI_ASYNCH
#define SPI_S(obj) (( struct spi_s *)(&(obj->spi)))
#else
#define SPI_S(obj) (( struct spi_s *)(obj))
#endif
#ifndef DEBUG_STDIO
# define DEBUG_STDIO 0
#endif
#if DEBUG_STDIO
# include <stdio.h>
# define DEBUG_PRINTF(...) do { printf(__VA_ARGS__); } while(0)
#else
# define DEBUG_PRINTF(...) {}
#endif
/* Consider 10ms as the default timeout for sending/receving 1 byte */
#define TIMEOUT_1_BYTE 10
#if defined(SPI_FLAG_FRLVL) // STM32F0 STM32F3 STM32F7 STM32L4
extern HAL_StatusTypeDef HAL_SPIEx_FlushRxFifo(SPI_HandleTypeDef *hspi);
#endif
void init_spi(spi_t *obj)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
__HAL_SPI_DISABLE(handle);
DEBUG_PRINTF("init_spi: instance=0x%8X\r\n", (int)handle->Instance);
if (HAL_SPI_Init(handle) != HAL_OK) {
error("Cannot initialize SPI");
}
/* In case of standard 4 wires SPI,PI can be kept enabled all time
* and SCK will only be generated during the write operations. But in case
* of 3 wires, it should be only enabled during rd/wr unitary operations,
* which is handled inside STM32 HAL layer.
*/
if (handle->Init.Direction == SPI_DIRECTION_2LINES) {
__HAL_SPI_ENABLE(handle);
}
}
SPIName spi_get_peripheral_name(PinName mosi, PinName miso, PinName sclk)
{
SPIName spi_mosi = (SPIName)pinmap_peripheral(mosi, PinMap_SPI_MOSI);
SPIName spi_miso = (SPIName)pinmap_peripheral(miso, PinMap_SPI_MISO);
SPIName spi_sclk = (SPIName)pinmap_peripheral(sclk, PinMap_SPI_SCLK);
SPIName spi_per;
// If 3 wire SPI is used, the miso is not connected.
if (miso == NC) {
spi_per = (SPIName)pinmap_merge(spi_mosi, spi_sclk);
} else {
SPIName spi_data = (SPIName)pinmap_merge(spi_mosi, spi_miso);
spi_per = (SPIName)pinmap_merge(spi_data, spi_sclk);
}
return spi_per;
}
void spi_init(spi_t *obj, PinName mosi, PinName miso, PinName sclk, PinName ssel)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
// Determine the SPI to use
SPIName spi_mosi = (SPIName)pinmap_peripheral(mosi, PinMap_SPI_MOSI);
SPIName spi_miso = (SPIName)pinmap_peripheral(miso, PinMap_SPI_MISO);
SPIName spi_sclk = (SPIName)pinmap_peripheral(sclk, PinMap_SPI_SCLK);
SPIName spi_ssel = (SPIName)pinmap_peripheral(ssel, PinMap_SPI_SSEL);
SPIName spi_data = (SPIName)pinmap_merge(spi_mosi, spi_miso);
SPIName spi_cntl = (SPIName)pinmap_merge(spi_sclk, spi_ssel);
spiobj->spi = (SPIName)pinmap_merge(spi_data, spi_cntl);
MBED_ASSERT(spiobj->spi != (SPIName)NC);
#if defined SPI1_BASE
// Enable SPI clock
if (spiobj->spi == SPI_1) {
__HAL_RCC_SPI1_CLK_ENABLE();
spiobj->spiIRQ = SPI1_IRQn;
}
#endif
#if defined SPI2_BASE
if (spiobj->spi == SPI_2) {
__HAL_RCC_SPI2_CLK_ENABLE();
spiobj->spiIRQ = SPI2_IRQn;
}
#endif
#if defined SPI3_BASE
if (spiobj->spi == SPI_3) {
__HAL_RCC_SPI3_CLK_ENABLE();
spiobj->spiIRQ = SPI3_IRQn;
}
#endif
#if defined SPI4_BASE
if (spiobj->spi == SPI_4) {
__HAL_RCC_SPI4_CLK_ENABLE();
spiobj->spiIRQ = SPI4_IRQn;
}
#endif
#if defined SPI5_BASE
if (spiobj->spi == SPI_5) {
__HAL_RCC_SPI5_CLK_ENABLE();
spiobj->spiIRQ = SPI5_IRQn;
}
#endif
#if defined SPI6_BASE
if (spiobj->spi == SPI_6) {
__HAL_RCC_SPI6_CLK_ENABLE();
spiobj->spiIRQ = SPI6_IRQn;
}
#endif
// Configure the SPI pins
pinmap_pinout(mosi, PinMap_SPI_MOSI);
pinmap_pinout(miso, PinMap_SPI_MISO);
pinmap_pinout(sclk, PinMap_SPI_SCLK);
spiobj->pin_miso = miso;
spiobj->pin_mosi = mosi;
spiobj->pin_sclk = sclk;
spiobj->pin_ssel = ssel;
if (ssel != NC) {
pinmap_pinout(ssel, PinMap_SPI_SSEL);
handle->Init.NSS = SPI_NSS_HARD_OUTPUT;
#if defined(SPI_NSS_PULSE_ENABLE)
handle->Init.NSSPMode = SPI_NSS_PULSE_ENABLE;
#endif
} else {
handle->Init.NSS = SPI_NSS_SOFT;
#if defined(SPI_NSS_PULSE_DISABLE)
handle->Init.NSSPMode = SPI_NSS_PULSE_DISABLE;
#endif
}
/* Fill default value */
handle->Instance = SPI_INST(obj);
handle->Init.Mode = SPI_MODE_MASTER;
handle->Init.BaudRatePrescaler = SPI_BAUDRATEPRESCALER_256;
if (miso != NC) {
handle->Init.Direction = SPI_DIRECTION_2LINES;
} else {
handle->Init.Direction = SPI_DIRECTION_1LINE;
}
handle->Init.CLKPhase = SPI_PHASE_1EDGE;
handle->Init.CLKPolarity = SPI_POLARITY_LOW;
handle->Init.CRCCalculation = SPI_CRCCALCULATION_DISABLE;
handle->Init.CRCPolynomial = 7;
handle->Init.DataSize = SPI_DATASIZE_8BIT;
handle->Init.FirstBit = SPI_FIRSTBIT_MSB;
handle->Init.TIMode = SPI_TIMODE_DISABLE;
#if TARGET_STM32H7
handle->Init.MasterKeepIOState = SPI_MASTER_KEEP_IO_STATE_ENABLE;
handle->Init.FifoThreshold = SPI_FIFO_THRESHOLD_01DATA;
#endif
/*
* According the STM32 Datasheet for SPI peripheral we need to PULLDOWN
* or PULLUP the SCK pin according the polarity used.
*/
pin_mode(spiobj->pin_sclk, (handle->Init.CLKPolarity == SPI_POLARITY_LOW) ? PullDown: PullUp);
init_spi(obj);
}
void spi_free(spi_t *obj)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
DEBUG_PRINTF("spi_free\r\n");
__HAL_SPI_DISABLE(handle);
HAL_SPI_DeInit(handle);
#if defined SPI1_BASE
// Reset SPI and disable clock
if (spiobj->spi == SPI_1) {
__HAL_RCC_SPI1_FORCE_RESET();
__HAL_RCC_SPI1_RELEASE_RESET();
__HAL_RCC_SPI1_CLK_DISABLE();
}
#endif
#if defined SPI2_BASE
if (spiobj->spi == SPI_2) {
__HAL_RCC_SPI2_FORCE_RESET();
__HAL_RCC_SPI2_RELEASE_RESET();
__HAL_RCC_SPI2_CLK_DISABLE();
}
#endif
#if defined SPI3_BASE
if (spiobj->spi == SPI_3) {
__HAL_RCC_SPI3_FORCE_RESET();
__HAL_RCC_SPI3_RELEASE_RESET();
__HAL_RCC_SPI3_CLK_DISABLE();
}
#endif
#if defined SPI4_BASE
if (spiobj->spi == SPI_4) {
__HAL_RCC_SPI4_FORCE_RESET();
__HAL_RCC_SPI4_RELEASE_RESET();
__HAL_RCC_SPI4_CLK_DISABLE();
}
#endif
#if defined SPI5_BASE
if (spiobj->spi == SPI_5) {
__HAL_RCC_SPI5_FORCE_RESET();
__HAL_RCC_SPI5_RELEASE_RESET();
__HAL_RCC_SPI5_CLK_DISABLE();
}
#endif
#if defined SPI6_BASE
if (spiobj->spi == SPI_6) {
__HAL_RCC_SPI6_FORCE_RESET();
__HAL_RCC_SPI6_RELEASE_RESET();
__HAL_RCC_SPI6_CLK_DISABLE();
}
#endif
// Configure GPIOs
pin_function(spiobj->pin_miso, STM_PIN_DATA(STM_MODE_INPUT, GPIO_NOPULL, 0));
pin_function(spiobj->pin_mosi, STM_PIN_DATA(STM_MODE_INPUT, GPIO_NOPULL, 0));
pin_function(spiobj->pin_sclk, STM_PIN_DATA(STM_MODE_INPUT, GPIO_NOPULL, 0));
if (handle->Init.NSS != SPI_NSS_SOFT) {
pin_function(spiobj->pin_ssel, STM_PIN_DATA(STM_MODE_INPUT, GPIO_NOPULL, 0));
}
}
void spi_format(spi_t *obj, int bits, int mode, int slave)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
PinMode pull = 0;
DEBUG_PRINTF("spi_format, bits:%d, mode:%d, slave?:%d\r\n", bits, mode, slave);
// Save new values
handle->Init.DataSize = (bits == 16) ? SPI_DATASIZE_16BIT : SPI_DATASIZE_8BIT;
switch (mode) {
case 0:
handle->Init.CLKPolarity = SPI_POLARITY_LOW;
handle->Init.CLKPhase = SPI_PHASE_1EDGE;
break;
case 1:
handle->Init.CLKPolarity = SPI_POLARITY_LOW;
handle->Init.CLKPhase = SPI_PHASE_2EDGE;
break;
case 2:
handle->Init.CLKPolarity = SPI_POLARITY_HIGH;
handle->Init.CLKPhase = SPI_PHASE_1EDGE;
break;
default:
handle->Init.CLKPolarity = SPI_POLARITY_HIGH;
handle->Init.CLKPhase = SPI_PHASE_2EDGE;
break;
}
if (handle->Init.NSS != SPI_NSS_SOFT) {
handle->Init.NSS = (slave) ? SPI_NSS_HARD_INPUT : SPI_NSS_HARD_OUTPUT;
}
handle->Init.Mode = (slave) ? SPI_MODE_SLAVE : SPI_MODE_MASTER;
if (slave && (handle->Init.Direction == SPI_DIRECTION_1LINE)) {
/* SPI slave implemtation in MBED does not support the 3 wires SPI.
* (e.g. when MISO is not connected). So we're forcing slave in
* 2LINES mode. As MISO is not connected, slave will only read
* from master, and cannot write to it. Inform user.
*/
debug("3 wires SPI slave not supported - slave will only read\r\n");
handle->Init.Direction = SPI_DIRECTION_2LINES;
}
/*
* According the STM32 Datasheet for SPI peripheral we need to PULLDOWN
* or PULLUP the SCK pin according the polarity used.
*/
pull = (handle->Init.CLKPolarity == SPI_POLARITY_LOW) ? PullDown: PullUp;
pin_mode(spiobj->pin_sclk, pull);
init_spi(obj);
}
/*
* Only the IP clock input is family dependant so it computed
* separately in spi_get_clock_freq
*/
extern int spi_get_clock_freq(spi_t *obj);
static const uint32_t baudrate_prescaler_table[] = {SPI_BAUDRATEPRESCALER_2,
SPI_BAUDRATEPRESCALER_4,
SPI_BAUDRATEPRESCALER_8,
SPI_BAUDRATEPRESCALER_16,
SPI_BAUDRATEPRESCALER_32,
SPI_BAUDRATEPRESCALER_64,
SPI_BAUDRATEPRESCALER_128,
SPI_BAUDRATEPRESCALER_256
};
void spi_frequency(spi_t *obj, int hz)
{
struct spi_s *spiobj = SPI_S(obj);
int spi_hz = 0;
uint8_t prescaler_rank = 0;
uint8_t last_index = (sizeof(baudrate_prescaler_table) / sizeof(baudrate_prescaler_table[0])) - 1;
SPI_HandleTypeDef *handle = &(spiobj->handle);
/* Calculate the spi clock for prescaler_rank 0: SPI_BAUDRATEPRESCALER_2 */
spi_hz = spi_get_clock_freq(obj) / 2;
/* Define pre-scaler in order to get highest available frequency below requested frequency */
while ((spi_hz > hz) && (prescaler_rank < last_index)) {
spi_hz = spi_hz / 2;
prescaler_rank++;
}
/* Use the best fit pre-scaler */
handle->Init.BaudRatePrescaler = baudrate_prescaler_table[prescaler_rank];
/* In case maximum pre-scaler still gives too high freq, raise an error */
if (spi_hz > hz) {
DEBUG_PRINTF("WARNING: lowest SPI freq (%d) higher than requested (%d)\r\n", spi_hz, hz);
}
DEBUG_PRINTF("spi_frequency, request:%d, select:%d\r\n", hz, spi_hz);
init_spi(obj);
}
static inline int ssp_readable(spi_t *obj)
{
int status;
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
// Check if data is received
status = ((__HAL_SPI_GET_FLAG(handle, SPI_FLAG_RXNE) != RESET) ? 1 : 0);
return status;
}
static inline int ssp_writeable(spi_t *obj)
{
int status;
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
// Check if data is transmitted
status = ((__HAL_SPI_GET_FLAG(handle, SPI_FLAG_TXE) != RESET) ? 1 : 0);
return status;
}
static inline int ssp_busy(spi_t *obj)
{
int status;
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
#if TARGET_STM32H7
status = ((__HAL_SPI_GET_FLAG(handle, SPI_FLAG_RXWNE) != RESET) ? 1 : 0);
#else /* TARGET_STM32H7 */
status = ((__HAL_SPI_GET_FLAG(handle, SPI_FLAG_BSY) != RESET) ? 1 : 0);
#endif /* TARGET_STM32H7 */
return status;
}
int spi_master_write(spi_t *obj, int value)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
if (handle->Init.Direction == SPI_DIRECTION_1LINE) {
return HAL_SPI_Transmit(handle, (uint8_t *)&value, 1, TIMEOUT_1_BYTE);
}
#if defined(LL_SPI_RX_FIFO_TH_HALF)
/* Configure the default data size */
if (handle->Init.DataSize == SPI_DATASIZE_16BIT) {
LL_SPI_SetRxFIFOThreshold(SPI_INST(obj), LL_SPI_RX_FIFO_TH_HALF);
} else {
LL_SPI_SetRxFIFOThreshold(SPI_INST(obj), LL_SPI_RX_FIFO_TH_QUARTER);
}
#endif
/* Here we're using LL which means direct registers access
* There is no error management, so we may end up looping
* infinitely here in case of faulty device for instance,
* but this will increase performances significantly
*/
#if TARGET_STM32H7
/* Master transfer start */
LL_SPI_StartMasterTransfer(SPI_INST(obj));
/* Wait TXP flag to transmit data */
while (!LL_SPI_IsActiveFlag_TXP(SPI_INST(obj)));
#else
/* Wait TXE flag to transmit data */
while (!LL_SPI_IsActiveFlag_TXE(SPI_INST(obj)));
#endif /* TARGET_STM32H7 */
/* Transmit data */
if (handle->Init.DataSize == SPI_DATASIZE_16BIT) {
LL_SPI_TransmitData16(SPI_INST(obj), (uint16_t)value);
} else {
LL_SPI_TransmitData8(SPI_INST(obj), (uint8_t)value);
}
#if TARGET_STM32H7
/* Wait for RXP or end of Transfer */
while (!LL_SPI_IsActiveFlag_RXP(SPI_INST(obj)));
#else /* TARGET_STM32H7 */
/* Wait for RXNE flag before reading */
while (!LL_SPI_IsActiveFlag_RXNE(SPI_INST(obj)));
#endif /* TARGET_STM32H7 */
/* Read received data */
if (handle->Init.DataSize == SPI_DATASIZE_16BIT) {
return LL_SPI_ReceiveData16(SPI_INST(obj));
} else {
return LL_SPI_ReceiveData8(SPI_INST(obj));
}
}
int spi_master_block_write(spi_t *obj, const char *tx_buffer, int tx_length,
char *rx_buffer, int rx_length, char write_fill)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
int total = (tx_length > rx_length) ? tx_length : rx_length;
if (handle->Init.Direction == SPI_DIRECTION_2LINES) {
for (int i = 0; i < total; i++) {
char out = (i < tx_length) ? tx_buffer[i] : write_fill;
char in = spi_master_write(obj, out);
if (i < rx_length) {
rx_buffer[i] = in;
}
}
} else {
/* In case of 1 WIRE only, first handle TX, then Rx */
if (tx_length != 0) {
if (HAL_OK != HAL_SPI_Transmit(handle, (uint8_t *)tx_buffer, tx_length, tx_length * TIMEOUT_1_BYTE)) {
/* report an error */
total = 0;
}
}
if (rx_length != 0) {
if (HAL_OK != HAL_SPI_Receive(handle, (uint8_t *)rx_buffer, rx_length, rx_length * TIMEOUT_1_BYTE)) {
/* report an error */
total = 0;
}
}
}
return total;
}
int spi_slave_receive(spi_t *obj)
{
return ((ssp_readable(obj) && !ssp_busy(obj)) ? 1 : 0);
};
int spi_slave_read(spi_t *obj)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
while (!ssp_readable(obj));
if (handle->Init.DataSize == SPI_DATASIZE_16BIT) {
return LL_SPI_ReceiveData16(SPI_INST(obj));
} else {
return LL_SPI_ReceiveData8(SPI_INST(obj));
}
}
void spi_slave_write(spi_t *obj, int value)
{
SPI_TypeDef *spi = SPI_INST(obj);
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
while (!ssp_writeable(obj));
if (handle->Init.DataSize == SPI_DATASIZE_8BIT) {
// Force 8-bit access to the data register
uint8_t *p_spi_dr = 0;
#if TARGET_STM32H7
p_spi_dr = (uint8_t *) & (spi->TXDR);
#else /* TARGET_STM32H7 */
p_spi_dr = (uint8_t *) & (spi->DR);
#endif /* TARGET_STM32H7 */
*p_spi_dr = (uint8_t)value;
} else { // SPI_DATASIZE_16BIT
#if TARGET_STM32H7
spi->TXDR = (uint16_t)value;
#else /* TARGET_STM32H7 */
spi->DR = (uint16_t)value;
#endif /* TARGET_STM32H7 */
}
}
int spi_busy(spi_t *obj)
{
return ssp_busy(obj);
}
const PinMap *spi_master_mosi_pinmap()
{
return PinMap_SPI_MOSI;
}
const PinMap *spi_master_miso_pinmap()
{
return PinMap_SPI_MISO;
}
const PinMap *spi_master_clk_pinmap()
{
return PinMap_SPI_SCLK;
}
const PinMap *spi_master_cs_pinmap()
{
return PinMap_SPI_SSEL;
}
const PinMap *spi_slave_mosi_pinmap()
{
return PinMap_SPI_MOSI;
}
const PinMap *spi_slave_miso_pinmap()
{
return PinMap_SPI_MISO;
}
const PinMap *spi_slave_clk_pinmap()
{
return PinMap_SPI_SCLK;
}
const PinMap *spi_slave_cs_pinmap()
{
return PinMap_SPI_SSEL;
}
#if DEVICE_SPI_ASYNCH
typedef enum {
SPI_TRANSFER_TYPE_NONE = 0,
SPI_TRANSFER_TYPE_TX = 1,
SPI_TRANSFER_TYPE_RX = 2,
SPI_TRANSFER_TYPE_TXRX = 3,
} transfer_type_t;
/// @returns the number of bytes transferred, or `0` if nothing transferred
static int spi_master_start_asynch_transfer(spi_t *obj, transfer_type_t transfer_type, const void *tx, void *rx, size_t length)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
bool is16bit = (handle->Init.DataSize == SPI_DATASIZE_16BIT);
// the HAL expects number of transfers instead of number of bytes
// so for 16 bit transfer width the count needs to be halved
size_t words;
DEBUG_PRINTF("SPI inst=0x%8X Start: %u, %u\r\n", (int)handle->Instance, transfer_type, length);
obj->spi.transfer_type = transfer_type;
if (is16bit) {
words = length / 2;
} else {
words = length;
}
// enable the interrupt
IRQn_Type irq_n = spiobj->spiIRQ;
NVIC_DisableIRQ(irq_n);
NVIC_ClearPendingIRQ(irq_n);
NVIC_SetPriority(irq_n, 1);
NVIC_EnableIRQ(irq_n);
// flush FIFO
#if defined(SPI_FLAG_FRLVL) // STM32F0 STM32F3 STM32F7 STM32L4
HAL_SPIEx_FlushRxFifo(handle);
#endif
// enable the right hal transfer
int rc = 0;
switch (transfer_type) {
case SPI_TRANSFER_TYPE_TXRX:
rc = HAL_SPI_TransmitReceive_IT(handle, (uint8_t *)tx, (uint8_t *)rx, words);
break;
case SPI_TRANSFER_TYPE_TX:
rc = HAL_SPI_Transmit_IT(handle, (uint8_t *)tx, words);
break;
case SPI_TRANSFER_TYPE_RX:
// the receive function also "transmits" the receive buffer so in order
// to guarantee that 0xff is on the line, we explicitly memset it here
memset(rx, SPI_FILL_CHAR, length);
rc = HAL_SPI_Receive_IT(handle, (uint8_t *)rx, words);
break;
default:
length = 0;
}
if (rc) {
DEBUG_PRINTF("SPI: RC=%u\n", rc);
length = 0;
}
return length;
}
// asynchronous API
void spi_master_transfer(spi_t *obj, const void *tx, size_t tx_length, void *rx, size_t rx_length, uint8_t bit_width, uint32_t handler, uint32_t event, DMAUsage hint)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
// TODO: DMA usage is currently ignored
(void) hint;
// check which use-case we have
bool use_tx = (tx != NULL && tx_length > 0);
bool use_rx = (rx != NULL && rx_length > 0);
bool is16bit = (handle->Init.DataSize == SPI_DATASIZE_16BIT);
// don't do anything, if the buffers aren't valid
if (!use_tx && !use_rx) {
return;
}
// copy the buffers to the SPI object
obj->tx_buff.buffer = (void *) tx;
obj->tx_buff.length = tx_length;
obj->tx_buff.pos = 0;
obj->tx_buff.width = is16bit ? 16 : 8;
obj->rx_buff.buffer = rx;
obj->rx_buff.length = rx_length;
obj->rx_buff.pos = 0;
obj->rx_buff.width = obj->tx_buff.width;
obj->spi.event = event;
DEBUG_PRINTF("SPI: Transfer: %u, %u\n", tx_length, rx_length);
// register the thunking handler
IRQn_Type irq_n = spiobj->spiIRQ;
NVIC_SetVector(irq_n, (uint32_t)handler);
// enable the right hal transfer
if (use_tx && use_rx) {
// we cannot manage different rx / tx sizes, let's use smaller one
size_t size = (tx_length < rx_length) ? tx_length : rx_length;
if (tx_length != rx_length) {
DEBUG_PRINTF("SPI: Full duplex transfer only 1 size: %d\n", size);
obj->tx_buff.length = size;
obj->rx_buff.length = size;
}
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_TXRX, tx, rx, size);
} else if (use_tx) {
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_TX, tx, NULL, tx_length);
} else if (use_rx) {
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_RX, NULL, rx, rx_length);
}
}
inline uint32_t spi_irq_handler_asynch(spi_t *obj)
{
int event = 0;
// call the CubeF4 handler, this will update the handle
HAL_SPI_IRQHandler(&obj->spi.handle);
if (obj->spi.handle.State == HAL_SPI_STATE_READY) {
// When HAL SPI is back to READY state, check if there was an error
int error = obj->spi.handle.ErrorCode;
if (error != HAL_SPI_ERROR_NONE) {
// something went wrong and the transfer has definitely completed
event = SPI_EVENT_ERROR | SPI_EVENT_INTERNAL_TRANSFER_COMPLETE;
if (error & HAL_SPI_ERROR_OVR) {
// buffer overrun
event |= SPI_EVENT_RX_OVERFLOW;
}
} else {
// else we're done
event = SPI_EVENT_COMPLETE | SPI_EVENT_INTERNAL_TRANSFER_COMPLETE;
}
// enable the interrupt
NVIC_DisableIRQ(obj->spi.spiIRQ);
NVIC_ClearPendingIRQ(obj->spi.spiIRQ);
}
return (event & (obj->spi.event | SPI_EVENT_INTERNAL_TRANSFER_COMPLETE));
}
uint8_t spi_active(spi_t *obj)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
HAL_SPI_StateTypeDef state = HAL_SPI_GetState(handle);
switch (state) {
case HAL_SPI_STATE_RESET:
case HAL_SPI_STATE_READY:
case HAL_SPI_STATE_ERROR:
return 0;
default:
return 1;
}
}
void spi_abort_asynch(spi_t *obj)
{
struct spi_s *spiobj = SPI_S(obj);
SPI_HandleTypeDef *handle = &(spiobj->handle);
// disable interrupt
IRQn_Type irq_n = spiobj->spiIRQ;
NVIC_ClearPendingIRQ(irq_n);
NVIC_DisableIRQ(irq_n);
// clean-up
__HAL_SPI_DISABLE(handle);
HAL_SPI_DeInit(handle);
HAL_SPI_Init(handle);
__HAL_SPI_ENABLE(handle);
}
#endif //DEVICE_SPI_ASYNCH
#endif