blob: 2df61244ece26c6162a6da96b374e668745c481b [file] [log] [blame]
/*
* Copyright 2022-2024 NXP
*
* SPDX-License-Identifier: Apache-2.0
*/
#define DT_DRV_COMPAT nxp_s32_spi
#include <zephyr/drivers/clock_control.h>
#include <zephyr/drivers/pinctrl.h>
#include "spi_nxp_s32.h"
static bool spi_nxp_s32_last_packet(struct spi_nxp_s32_data *data)
{
struct spi_context *ctx = &data->ctx;
if (ctx->tx_count <= 1U && ctx->rx_count <= 1U) {
if (!spi_context_tx_on(ctx) && (data->transfer_len == ctx->rx_len)) {
return true;
}
if (!spi_context_rx_on(ctx) && (data->transfer_len == ctx->tx_len)) {
return true;
}
if ((ctx->rx_len == ctx->tx_len) && (data->transfer_len == ctx->tx_len)) {
return true;
}
}
return false;
}
static inline bool spi_nxp_s32_transfer_done(struct spi_context *ctx)
{
return !spi_context_tx_on(ctx) && !spi_context_rx_on(ctx);
}
static int spi_nxp_s32_transfer_next_packet(const struct device *dev)
{
const struct spi_nxp_s32_config *config = dev->config;
struct spi_nxp_s32_data *data = dev->data;
Spi_Ip_StatusType status;
Spi_Ip_CallbackType data_cb;
Spi_Ip_TransferAdjustmentType param;
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
data_cb = config->cb;
#else
data_cb = NULL;
#endif /* CONFIG_NXP_S32_SPI_INTERRUPT */
data->transfer_len = spi_context_max_continuous_chunk(&data->ctx);
data->transfer_len = MIN(data->transfer_len,
SPI_NXP_S32_MAX_BYTES_PER_PACKAGE(data->bytes_per_frame));
param.KeepCs = !spi_nxp_s32_last_packet(data);
param.DeviceParams = NULL;
Spi_Ip_UpdateTransferParam(&data->transfer_cfg, &param);
status = Spi_Ip_AsyncTransmit(&data->transfer_cfg, (uint8_t *)data->ctx.tx_buf,
data->ctx.rx_buf, data->transfer_len, data_cb);
if (status) {
LOG_ERR("Transfer could not start");
return -EIO;
}
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
return 0;
#else
while (Spi_Ip_GetStatus(config->spi_hw_cfg->Instance) == SPI_IP_BUSY) {
Spi_Ip_ManageBuffers(config->spi_hw_cfg->Instance);
}
if (Spi_Ip_GetStatus(config->spi_hw_cfg->Instance) == SPI_IP_FAULT) {
return -EIO;
}
return 0;
#endif /* CONFIG_NXP_S32_SPI_INTERRUPT */
}
/*
* The function to get Scaler and Prescaler for corresponding registers
* to configure the baudrate for the transmission. The real frequency is
* computated to ensure it will always equal or the nearest approximation
* lower to the expected one.
*/
static void spi_nxp_s32_getbestfreq(uint32_t clock_frequency,
uint32_t requested_baud,
struct spi_nxp_s32_baudrate_param *best_baud)
{
uint8_t scaler;
uint8_t prescaler;
uint32_t low, high;
uint32_t curr_freq;
uint32_t best_freq = 0U;
static const uint8_t prescaler_arr[SPI_NXP_S32_NUM_PRESCALER] = {2U, 3U, 5U, 7U};
static const uint16_t scaller_arr[SPI_NXP_S32_NUM_SCALER] = {
2U, 4U, 6U, 8U, 16U, 32U, 64U, 128U, 256U, 512U, 1024U, 2048U,
4096U, 8192U, 16384U, 32768U
};
for (prescaler = 0U; prescaler < SPI_NXP_S32_NUM_PRESCALER; prescaler++) {
low = 0U;
high = SPI_NXP_S32_NUM_SCALER - 1U;
/* Implement golden section search algorithm */
do {
scaler = (low + high) / 2U;
curr_freq = clock_frequency * 1U /
(prescaler_arr[prescaler] * scaller_arr[scaler]);
/*
* If the scaler make current frequency higher than the
* expected one, skip the next step
*/
if (curr_freq > requested_baud) {
low = scaler;
continue;
} else {
high = scaler;
}
if ((requested_baud - best_freq) > (requested_baud - curr_freq)) {
best_freq = curr_freq;
best_baud->prescaler = prescaler;
best_baud->scaler = scaler;
}
if (best_freq == requested_baud) {
break;
}
} while ((high - low) > 1U);
if ((high - low) <= 1U) {
if (high == scaler) {
/* use low value */
scaler = low;
} else {
scaler = high;
}
curr_freq = clock_frequency * 1U /
(prescaler_arr[prescaler] * scaller_arr[scaler]);
if (curr_freq <= requested_baud) {
if ((requested_baud - best_freq) > (requested_baud - curr_freq)) {
best_freq = curr_freq;
best_baud->prescaler = prescaler;
best_baud->scaler = scaler;
}
}
}
if (best_freq == requested_baud) {
break;
}
}
best_baud->frequency = best_freq;
}
/*
* The function to get Scaler and Prescaler for corresponding registers
* to configure the delay for the transmission. The real delay is computated
* to ensure it will always equal or the nearest approximation higher to
* the expected one. In the worst case, use the delay as much as possible.
*/
static void spi_nxp_s32_getbestdelay(uint32_t clock_frequency, uint32_t requested_delay,
uint8_t *best_scaler, uint8_t *best_prescaler)
{
uint32_t current_delay;
uint8_t scaler, prescaler;
uint32_t low, high;
uint32_t best_delay = 0xFFFFFFFFU;
/* The scaler array is a power of two, so do not need to be defined */
static const uint8_t prescaler_arr[SPI_NXP_S32_NUM_PRESCALER] = {1U, 3U, 5U, 7U};
clock_frequency = clock_frequency / MHZ(1);
for (prescaler = 0; prescaler < SPI_NXP_S32_NUM_PRESCALER; prescaler++) {
low = 0U;
high = SPI_NXP_S32_NUM_SCALER - 1U;
do {
scaler = (low + high) / 2U;
current_delay = NSEC_PER_USEC * prescaler_arr[prescaler]
* (1U << (scaler + 1)) / clock_frequency;
/*
* If the scaler make current delay smaller than
* the expected one, skip the next step
*/
if (current_delay < requested_delay) {
low = scaler;
continue;
} else {
high = scaler;
}
if ((best_delay - requested_delay) > (current_delay - requested_delay)) {
best_delay = current_delay;
*best_prescaler = prescaler;
*best_scaler = scaler;
}
if (best_delay == requested_delay) {
break;
}
} while ((high - low) > 1U);
if ((high - low) <= 1U) {
if (high == scaler) {
/* use low value */
scaler = low;
} else {
scaler = high;
}
current_delay = NSEC_PER_USEC * prescaler_arr[prescaler]
* (1U << (scaler + 1)) / clock_frequency;
if (current_delay >= requested_delay) {
if ((best_delay - requested_delay) >
(current_delay - requested_delay)) {
best_delay = current_delay;
*best_prescaler = prescaler;
*best_scaler = scaler;
}
}
}
if (best_delay == requested_delay) {
break;
}
}
if (best_delay == 0xFFFFFFFFU) {
/* Use the delay as much as possible */
*best_prescaler = SPI_NXP_S32_NUM_PRESCALER - 1U;
*best_scaler = SPI_NXP_S32_NUM_SCALER - 1U;
}
}
static int spi_nxp_s32_configure(const struct device *dev,
const struct spi_config *spi_cfg)
{
const struct spi_nxp_s32_config *config = dev->config;
struct spi_nxp_s32_data *data = dev->data;
bool clk_phase, clk_polarity;
bool lsb, hold_cs;
bool slave_mode, cs_active_high;
uint8_t frame_size;
struct spi_nxp_s32_baudrate_param best_baud = {0};
uint32_t clock_rate;
int err;
if (spi_context_configured(&data->ctx, spi_cfg)) {
/* This configuration is already in use */
return 0;
}
err = clock_control_get_rate(config->clock_dev, config->clock_subsys, &clock_rate);
if (err) {
LOG_ERR("Failed to get clock frequency");
return err;
}
clk_phase = !!(SPI_MODE_GET(spi_cfg->operation) & SPI_MODE_CPHA);
clk_polarity = !!(SPI_MODE_GET(spi_cfg->operation) & SPI_MODE_CPOL);
hold_cs = !!(spi_cfg->operation & SPI_HOLD_ON_CS);
lsb = !!(spi_cfg->operation & SPI_TRANSFER_LSB);
slave_mode = !!(SPI_OP_MODE_GET(spi_cfg->operation));
frame_size = SPI_WORD_SIZE_GET(spi_cfg->operation);
cs_active_high = !!(spi_cfg->operation & SPI_CS_ACTIVE_HIGH);
if (slave_mode == (!!(config->spi_hw_cfg->Mcr & SPI_MCR_MSTR_MASK))) {
LOG_ERR("SPI mode (master/slave) must be same as configured in DT");
return -ENOTSUP;
}
if (slave_mode && !IS_ENABLED(CONFIG_SPI_SLAVE)) {
LOG_ERR("Kconfig for enable SPI in slave mode is not enabled");
return -ENOTSUP;
}
if (slave_mode && lsb) {
LOG_ERR("SPI does not support to shifting out with LSB in slave mode");
return -ENOTSUP;
}
if (spi_cfg->slave >= config->num_cs) {
LOG_ERR("Slave %d excess the allowed maximum value (%d)",
spi_cfg->slave, config->num_cs - 1);
return -ENOTSUP;
}
if (frame_size > 32U) {
LOG_ERR("Unsupported frame size %d bits", frame_size);
return -ENOTSUP;
}
if ((spi_cfg->operation & SPI_LINES_MASK) != SPI_LINES_SINGLE) {
LOG_ERR("Only single line mode is supported");
return -ENOTSUP;
}
if (spi_cfg->operation & SPI_MODE_LOOP) {
LOG_ERR("Loopback mode is not supported");
return -ENOTSUP;
}
if (cs_active_high && !spi_cs_is_gpio(spi_cfg)) {
LOG_ERR("For CS has active state is high, a GPIO pin must be used to"
" control CS line instead");
return -ENOTSUP;
}
if (!slave_mode) {
if ((spi_cfg->frequency < SPI_NXP_S32_MIN_FREQ) ||
(spi_cfg->frequency > SPI_NXP_S32_MAX_FREQ)) {
LOG_ERR("The frequency is out of range");
return -ENOTSUP;
}
spi_nxp_s32_getbestfreq(clock_rate, spi_cfg->frequency, &best_baud);
data->transfer_cfg.Ctar &= ~(SPI_CTAR_BR_MASK | SPI_CTAR_PBR_MASK);
data->transfer_cfg.Ctar |= SPI_CTAR_BR(best_baud.scaler) |
SPI_CTAR_PBR(best_baud.prescaler);
data->transfer_cfg.PushrCmd &= ~((SPI_PUSHR_CONT_MASK | SPI_PUSHR_PCS_MASK) >> 16U);
if (!spi_cs_is_gpio(spi_cfg)) {
/* Use inner CS signal from SPI module */
data->transfer_cfg.PushrCmd |= hold_cs << 15U;
data->transfer_cfg.PushrCmd |= (1U << spi_cfg->slave);
}
}
data->transfer_cfg.Ctar &= ~(SPI_CTAR_CPHA_MASK | SPI_CTAR_CPOL_MASK);
data->transfer_cfg.Ctar |= SPI_CTAR_CPHA(clk_phase) | SPI_CTAR_CPOL(clk_polarity);
Spi_Ip_UpdateFrameSize(&data->transfer_cfg, frame_size);
Spi_Ip_UpdateLsb(&data->transfer_cfg, lsb);
data->ctx.config = spi_cfg;
data->bytes_per_frame = SPI_NXP_S32_BYTE_PER_FRAME(frame_size);
if (slave_mode) {
LOG_DBG("SPI configuration: cpol = %u, cpha = %u,"
" lsb = %u, frame_size = %u, mode: slave",
clk_polarity, clk_phase, lsb, frame_size);
} else {
LOG_DBG("SPI configuration: frequency = %uHz, cpol = %u,"
" cpha = %u, lsb = %u, hold_cs = %u, frame_size = %u,"
" mode: master, CS = %u\n",
best_baud.frequency, clk_polarity, clk_phase,
lsb, hold_cs, frame_size, spi_cfg->slave);
}
return 0;
}
static int transceive(const struct device *dev,
const struct spi_config *spi_cfg,
const struct spi_buf_set *tx_bufs,
const struct spi_buf_set *rx_bufs,
bool asynchronous,
spi_callback_t cb,
void *userdata)
{
struct spi_nxp_s32_data *data = dev->data;
struct spi_context *context = &data->ctx;
int ret;
if (!tx_bufs && !rx_bufs) {
return 0;
}
#ifndef CONFIG_NXP_S32_SPI_INTERRUPT
if (asynchronous) {
return -ENOTSUP;
}
#endif /* CONFIG_NXP_S32_SPI_INTERRUPT */
spi_context_lock(context, asynchronous, cb, userdata, spi_cfg);
ret = spi_nxp_s32_configure(dev, spi_cfg);
if (ret) {
LOG_ERR("An error occurred in the SPI configuration");
spi_context_release(context, ret);
return ret;
}
spi_context_buffers_setup(context, tx_bufs, rx_bufs, 1U);
if (spi_nxp_s32_transfer_done(context)) {
spi_context_release(context, 0);
return 0;
}
spi_context_cs_control(context, true);
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
ret = spi_nxp_s32_transfer_next_packet(dev);
if (!ret) {
ret = spi_context_wait_for_completion(context);
} else {
spi_context_cs_control(context, false);
}
#else
do {
ret = spi_nxp_s32_transfer_next_packet(dev);
if (!ret) {
spi_context_update_tx(context, 1U, data->transfer_len);
spi_context_update_rx(context, 1U, data->transfer_len);
}
} while (!ret && !spi_nxp_s32_transfer_done(context));
spi_context_cs_control(context, false);
#ifdef CONFIG_SPI_SLAVE
if (spi_context_is_slave(context) && !ret) {
ret = data->ctx.recv_frames;
}
#endif /* CONFIG_SPI_SLAVE */
#endif /* CONFIG_NXP_S32_SPI_INTERRUPT */
spi_context_release(context, ret);
return ret;
}
static int spi_nxp_s32_transceive(const struct device *dev,
const struct spi_config *spi_cfg,
const struct spi_buf_set *tx_bufs,
const struct spi_buf_set *rx_bufs)
{
return transceive(dev, spi_cfg, tx_bufs, rx_bufs, false, NULL, NULL);
}
#ifdef CONFIG_SPI_ASYNC
static int spi_nxp_s32_transceive_async(const struct device *dev,
const struct spi_config *spi_cfg,
const struct spi_buf_set *tx_bufs,
const struct spi_buf_set *rx_bufs,
spi_callback_t callback,
void *userdata)
{
return transceive(dev, spi_cfg, tx_bufs, rx_bufs, true, callback, userdata);
}
#endif /* CONFIG_SPI_ASYNC */
static int spi_nxp_s32_release(const struct device *dev,
const struct spi_config *spi_cfg)
{
struct spi_nxp_s32_data *data = dev->data;
(void)spi_cfg;
spi_context_unlock_unconditionally(&data->ctx);
return 0;
}
static int spi_nxp_s32_init(const struct device *dev)
{
const struct spi_nxp_s32_config *config = dev->config;
struct spi_nxp_s32_data *data = dev->data;
uint32_t clock_rate;
uint8_t scaler, prescaler;
uint32_t ctar = 0;
int ret = 0;
if (!device_is_ready(config->clock_dev)) {
LOG_ERR("Clock control device not ready");
return -ENODEV;
}
ret = clock_control_on(config->clock_dev, config->clock_subsys);
if (ret) {
LOG_ERR("Failed to enable clock");
return ret;
}
ret = clock_control_get_rate(config->clock_dev, config->clock_subsys, &clock_rate);
if (ret) {
LOG_ERR("Failed to get clock frequency");
return ret;
}
ret = pinctrl_apply_state(config->pincfg, PINCTRL_STATE_DEFAULT);
if (ret < 0) {
return ret;
}
if (Spi_Ip_Init(config->spi_hw_cfg)) {
return -EBUSY;
}
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
if (Spi_Ip_UpdateTransferMode(config->spi_hw_cfg->Instance, SPI_IP_INTERRUPT)) {
return -EBUSY;
}
config->irq_config_func(dev);
#endif /* CONFIG_NXP_S32_SPI_INTERRUPT */
/*
* Update the delay timings configuration that are
* applied for all inner CS signals of SPI module.
*/
spi_nxp_s32_getbestdelay(clock_rate, config->sck_cs_delay, &scaler, &prescaler);
ctar |= SPI_CTAR_ASC(scaler) | SPI_CTAR_PASC(prescaler);
spi_nxp_s32_getbestdelay(clock_rate, config->cs_sck_delay, &scaler, &prescaler);
ctar |= SPI_CTAR_CSSCK(scaler) | SPI_CTAR_PCSSCK(prescaler);
spi_nxp_s32_getbestdelay(clock_rate, config->cs_cs_delay, &scaler, &prescaler);
ctar |= SPI_CTAR_DT(scaler) | SPI_CTAR_PDT(prescaler);
data->transfer_cfg.Ctar |= ctar;
data->transfer_cfg.DeviceParams = &data->transfer_params;
ret = spi_context_cs_configure_all(&data->ctx);
if (ret < 0) {
return ret;
}
spi_context_unlock_unconditionally(&data->ctx);
return 0;
}
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
void spi_nxp_s32_isr(const struct device *dev)
{
const struct spi_nxp_s32_config *config = dev->config;
Spi_Ip_IrqHandler(config->spi_hw_cfg->Instance);
}
static void spi_nxp_s32_transfer_callback(const struct device *dev, Spi_Ip_EventType event)
{
struct spi_nxp_s32_data *data = dev->data;
int ret = 0;
if (event == SPI_IP_EVENT_END_TRANSFER) {
spi_context_update_tx(&data->ctx, 1U, data->transfer_len);
spi_context_update_rx(&data->ctx, 1U, data->transfer_len);
if (spi_nxp_s32_transfer_done(&data->ctx)) {
spi_context_complete(&data->ctx, dev, 0);
spi_context_cs_control(&data->ctx, false);
} else {
ret = spi_nxp_s32_transfer_next_packet(dev);
}
} else {
LOG_ERR("Failing in transfer_callback");
ret = -EIO;
}
if (ret) {
spi_context_complete(&data->ctx, dev, ret);
spi_context_cs_control(&data->ctx, false);
}
}
#endif /*CONFIG_NXP_S32_SPI_INTERRUPT*/
static const struct spi_driver_api spi_nxp_s32_driver_api = {
.transceive = spi_nxp_s32_transceive,
#ifdef CONFIG_SPI_ASYNC
.transceive_async = spi_nxp_s32_transceive_async,
#endif
#ifdef CONFIG_SPI_RTIO
.iodev_submit = spi_rtio_iodev_default_submit,
#endif
.release = spi_nxp_s32_release,
};
#define SPI_NXP_S32_HW_INSTANCE_CHECK(i, n) \
((DT_INST_REG_ADDR(n) == IP_SPI_##i##_BASE) ? i : 0)
#define SPI_NXP_S32_HW_INSTANCE(n) \
LISTIFY(__DEBRACKET SPI_INSTANCE_COUNT, SPI_NXP_S32_HW_INSTANCE_CHECK, (|), n)
#define SPI_NXP_S32_NUM_CS(n) DT_INST_PROP(n, num_cs)
#define SPI_NXP_S32_IS_MASTER(n) !DT_INST_PROP(n, slave)
#ifdef CONFIG_SPI_SLAVE
#define SPI_NXP_S32_SET_SLAVE(n) .SlaveMode = DT_INST_PROP(n, slave),
#else
#define SPI_NXP_S32_SET_SLAVE(n)
#endif
#ifdef CONFIG_NXP_S32_SPI_INTERRUPT
#define SPI_NXP_S32_CONFIG_INTERRUPT_FUNC(n) \
.irq_config_func = spi_nxp_s32_config_func_##n,
#define SPI_NXP_S32_INTERRUPT_DEFINE(n) \
static void spi_nxp_s32_config_func_##n(const struct device *dev) \
{ \
IRQ_CONNECT(DT_INST_IRQN(n), DT_INST_IRQ(n, priority), \
spi_nxp_s32_isr, DEVICE_DT_INST_GET(n), \
DT_INST_IRQ(n, flags)); \
irq_enable(DT_INST_IRQN(n)); \
}
#define SPI_NXP_S32_CONFIG_CALLBACK_FUNC(n) \
.cb = spi_nxp_s32_##n##_callback,
#define SPI_NXP_S32_CALLBACK_DEFINE(n) \
static void spi_nxp_s32_##n##_callback(uint8 instance, Spi_Ip_EventType event) \
{ \
ARG_UNUSED(instance); \
const struct device *dev = DEVICE_DT_INST_GET(n); \
\
spi_nxp_s32_transfer_callback(dev, event); \
}
#else
#define SPI_NXP_S32_CONFIG_INTERRUPT_FUNC(n)
#define SPI_NXP_S32_INTERRUPT_DEFINE(n)
#define SPI_NXP_S32_CONFIG_CALLBACK_FUNC(n)
#define SPI_NXP_S32_CALLBACK_DEFINE(n)
#endif /*CONFIG_NXP_S32_SPI_INTERRUPT*/
/*
* Declare the default configuration for SPI driver, no DMA
* support, all inner module Chip Selects are active low.
*/
#define SPI_NXP_S32_INSTANCE_CONFIG(n) \
static const Spi_Ip_ConfigType spi_nxp_s32_default_config_##n = { \
.Instance = SPI_NXP_S32_HW_INSTANCE(n), \
.Mcr = (SPI_MCR_MSTR(SPI_NXP_S32_IS_MASTER(n)) | \
SPI_MCR_CONT_SCKE(0U) | SPI_MCR_FRZ(0U) | \
SPI_MCR_MTFE(0U) | SPI_MCR_SMPL_PT(0U) | \
SPI_MCR_PCSIS(BIT_MASK(SPI_NXP_S32_NUM_CS(n))) | \
SPI_MCR_MDIS(0U) | SPI_MCR_XSPI(1U) | SPI_MCR_HALT(1U)), \
.TransferMode = SPI_IP_POLLING, \
.StateIndex = n, \
SPI_NXP_S32_SET_SLAVE(n) \
}
#define SPI_NXP_S32_TRANSFER_CONFIG(n) \
.transfer_cfg = { \
.Instance = SPI_NXP_S32_HW_INSTANCE(n), \
.Ctare = SPI_CTARE_FMSZE(0U) | SPI_CTARE_DTCP(1U), \
}
#define SPI_NXP_S32_DEVICE(n) \
PINCTRL_DT_INST_DEFINE(n); \
SPI_NXP_S32_CALLBACK_DEFINE(n) \
SPI_NXP_S32_INTERRUPT_DEFINE(n) \
SPI_NXP_S32_INSTANCE_CONFIG(n); \
static const struct spi_nxp_s32_config spi_nxp_s32_config_##n = { \
.num_cs = SPI_NXP_S32_NUM_CS(n), \
.clock_dev = DEVICE_DT_GET(DT_INST_CLOCKS_CTLR(n)), \
.clock_subsys = (clock_control_subsys_t)DT_INST_CLOCKS_CELL(n, name), \
.sck_cs_delay = DT_INST_PROP_OR(n, spi_sck_cs_delay, 0U), \
.cs_sck_delay = DT_INST_PROP_OR(n, spi_cs_sck_delay, 0U), \
.cs_cs_delay = DT_INST_PROP_OR(n, spi_cs_cs_delay, 0U), \
.spi_hw_cfg = (Spi_Ip_ConfigType *)&spi_nxp_s32_default_config_##n, \
.pincfg = PINCTRL_DT_INST_DEV_CONFIG_GET(n), \
SPI_NXP_S32_CONFIG_CALLBACK_FUNC(n) \
SPI_NXP_S32_CONFIG_INTERRUPT_FUNC(n) \
}; \
static struct spi_nxp_s32_data spi_nxp_s32_data_##n = { \
SPI_NXP_S32_TRANSFER_CONFIG(n), \
SPI_CONTEXT_INIT_LOCK(spi_nxp_s32_data_##n, ctx), \
SPI_CONTEXT_INIT_SYNC(spi_nxp_s32_data_##n, ctx), \
SPI_CONTEXT_CS_GPIOS_INITIALIZE(DT_DRV_INST(n), ctx) \
}; \
DEVICE_DT_INST_DEFINE(n, \
spi_nxp_s32_init, NULL, \
&spi_nxp_s32_data_##n, &spi_nxp_s32_config_##n, \
POST_KERNEL, CONFIG_SPI_INIT_PRIORITY, \
&spi_nxp_s32_driver_api);
DT_INST_FOREACH_STATUS_OKAY(SPI_NXP_S32_DEVICE)