mbed library sources. Supersedes mbed-src.

Dependents:   Nucleo_Hello_Encoder BLE_iBeaconScan AM1805_DEMO DISCO-F429ZI_ExportTemplate1 ... more

Revision:
187:0387e8f68319
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/targets/TARGET_NUVOTON/TARGET_M2351/spi_api.c	Thu Sep 06 13:40:20 2018 +0100
@@ -0,0 +1,848 @@
+/* mbed Microcontroller Library
+ * Copyright (c) 2015-2016 Nuvoton
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include "spi_api.h"
+
+#if DEVICE_SPI
+
+#include "cmsis.h"
+#include "mbed_error.h"
+#include "pinmap.h"
+#include "PeripheralPins.h"
+#include "nu_modutil.h"
+#include "nu_miscutil.h"
+#include "nu_bitutil.h"
+
+#if DEVICE_SPI_ASYNCH
+#include "dma_api.h"
+#include "dma.h"
+#endif
+
+#define NU_SPI_FRAME_MIN    8
+#define NU_SPI_FRAME_MAX    32
+
+struct nu_spi_var {
+#if DEVICE_SPI_ASYNCH
+    uint8_t     pdma_perp_tx;
+    uint8_t     pdma_perp_rx;
+#endif
+};
+
+static struct nu_spi_var spi0_var = {
+#if DEVICE_SPI_ASYNCH
+    .pdma_perp_tx       =   PDMA_SPI0_TX,
+    .pdma_perp_rx       =   PDMA_SPI0_RX
+#endif
+};
+static struct nu_spi_var spi1_var = {
+#if DEVICE_SPI_ASYNCH
+    .pdma_perp_tx       =   PDMA_SPI1_TX,
+    .pdma_perp_rx       =   PDMA_SPI1_RX
+#endif
+};
+static struct nu_spi_var spi2_var = {
+#if DEVICE_SPI_ASYNCH
+    .pdma_perp_tx       =   PDMA_SPI2_TX,
+    .pdma_perp_rx       =   PDMA_SPI2_RX
+#endif
+};
+static struct nu_spi_var spi3_var = {
+#if DEVICE_SPI_ASYNCH
+    .pdma_perp_tx       =   PDMA_SPI3_TX,
+    .pdma_perp_rx       =   PDMA_SPI3_RX
+#endif
+};
+static struct nu_spi_var spi5_var = {
+#if DEVICE_SPI_ASYNCH
+    .pdma_perp_tx       =   PDMA_SPI5_TX,
+    .pdma_perp_rx       =   PDMA_SPI5_RX
+#endif
+};
+
+/* Synchronous version of SPI_ENABLE()/SPI_DISABLE() macros
+ *
+ * The SPI peripheral clock is asynchronous with the system clock. In order to make sure the SPI
+ * control logic is enabled/disabled, this bit indicates the real status of SPI controller.
+ *
+ * NOTE: All configurations shall be ready before calling SPI_ENABLE_SYNC().
+ * NOTE: Before changing the configurations of SPIx_CTL, SPIx_CLKDIV, SPIx_SSCTL and SPIx_FIFOCTL registers,
+ *       user shall clear the SPIEN (SPIx_CTL[0]) and confirm the SPIENSTS (SPIx_STATUS[15]) is 0
+ *       (by SPI_DISABLE_SYNC here).
+ */
+__STATIC_INLINE void SPI_ENABLE_SYNC(SPI_T *spi_base)
+{
+    if (! (spi_base->CTL & SPI_CTL_SPIEN_Msk)) {
+        SPI_ENABLE(spi_base);
+    }
+    while (! (spi_base->STATUS & SPI_STATUS_SPIENSTS_Msk));
+}
+__STATIC_INLINE void SPI_DISABLE_SYNC(SPI_T *spi_base)
+{
+    if (spi_base->CTL & SPI_CTL_SPIEN_Msk) {
+        // NOTE: SPI H/W may get out of state without the busy check.
+        while (SPI_IS_BUSY(spi_base));
+    
+        SPI_DISABLE(spi_base);
+    }
+    while (spi_base->STATUS & SPI_STATUS_SPIENSTS_Msk);
+}
+
+#if DEVICE_SPI_ASYNCH
+static void spi_enable_vector_interrupt(spi_t *obj, uint32_t handler, uint8_t enable);
+static void spi_master_enable_interrupt(spi_t *obj, uint8_t enable);
+static uint32_t spi_master_write_asynch(spi_t *obj, uint32_t tx_limit);
+static uint32_t spi_master_read_asynch(spi_t *obj);
+static uint32_t spi_event_check(spi_t *obj);
+static void spi_enable_event(spi_t *obj, uint32_t event, uint8_t enable);
+static void spi_buffer_set(spi_t *obj, const void *tx, size_t tx_length, void *rx, size_t rx_length);
+static void spi_check_dma_usage(DMAUsage *dma_usage, int *dma_ch_tx, int *dma_ch_rx);
+static uint8_t spi_get_data_width(spi_t *obj);
+static int spi_is_tx_complete(spi_t *obj);
+static int spi_is_rx_complete(spi_t *obj);
+static int spi_writeable(spi_t * obj);
+static int spi_readable(spi_t * obj);
+static void spi_dma_handler_tx(uint32_t id, uint32_t event_dma);
+static void spi_dma_handler_rx(uint32_t id, uint32_t event_dma);
+static uint32_t spi_fifo_depth(spi_t *obj);
+#endif
+
+static uint32_t spi_modinit_mask = 0;
+
+static const struct nu_modinit_s spi_modinit_tab[] = {
+    {SPI_0, SPI0_MODULE, CLK_CLKSEL2_SPI0SEL_PCLK1, MODULE_NoMsk, SPI0_RST, SPI0_IRQn, &spi0_var},
+    {SPI_1, SPI1_MODULE, CLK_CLKSEL2_SPI1SEL_PCLK0, MODULE_NoMsk, SPI1_RST, SPI1_IRQn, &spi1_var},
+    {SPI_2, SPI2_MODULE, CLK_CLKSEL2_SPI2SEL_PCLK1, MODULE_NoMsk, SPI2_RST, SPI2_IRQn, &spi2_var},
+    {SPI_3, SPI3_MODULE, CLK_CLKSEL2_SPI3SEL_PCLK0, MODULE_NoMsk, SPI3_RST, SPI3_IRQn, &spi3_var},
+    {SPI_5, SPI5_MODULE, CLK_CLKSEL2_SPI5SEL_PCLK1, MODULE_NoMsk, SPI5_RST, SPI5_IRQn, &spi5_var},
+
+    {NC, 0, 0, 0, 0, (IRQn_Type) 0, NULL}
+};
+
+void spi_init(spi_t *obj, PinName mosi, PinName miso, PinName sclk, PinName ssel)
+{
+    // Determine which SPI_x the pins are used for
+    uint32_t spi_mosi = pinmap_peripheral(mosi, PinMap_SPI_MOSI);
+    uint32_t spi_miso = pinmap_peripheral(miso, PinMap_SPI_MISO);
+    uint32_t spi_sclk = pinmap_peripheral(sclk, PinMap_SPI_SCLK);
+    uint32_t spi_ssel = pinmap_peripheral(ssel, PinMap_SPI_SSEL);
+    uint32_t spi_data = pinmap_merge(spi_mosi, spi_miso);
+    uint32_t spi_cntl = pinmap_merge(spi_sclk, spi_ssel);
+    obj->spi.spi = (SPIName) pinmap_merge(spi_data, spi_cntl);
+    MBED_ASSERT((int)obj->spi.spi != NC);
+
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    /* Reset module
+     *
+     * NOTE: We must call secure version (from non-secure domain) because SYS/CLK regions are secure.
+     */
+    SYS_ResetModule_S(modinit->rsetidx);
+
+    /* Select IP clock source
+     *
+     * NOTE: We must call secure version (from non-secure domain) because SYS/CLK regions are secure.
+     */
+    CLK_SetModuleClock_S(modinit->clkidx, modinit->clksrc, modinit->clkdiv);
+
+    /* Enable IP clock
+     *
+     * NOTE: We must call secure version (from non-secure domain) because SYS/CLK regions are secure.
+     */
+    CLK_EnableModuleClock_S(modinit->clkidx);
+
+    pinmap_pinout(mosi, PinMap_SPI_MOSI);
+    pinmap_pinout(miso, PinMap_SPI_MISO);
+    pinmap_pinout(sclk, PinMap_SPI_SCLK);
+    pinmap_pinout(ssel, PinMap_SPI_SSEL);
+
+    obj->spi.pin_mosi = mosi;
+    obj->spi.pin_miso = miso;
+    obj->spi.pin_sclk = sclk;
+    obj->spi.pin_ssel = ssel;
+
+
+#if DEVICE_SPI_ASYNCH
+    obj->spi.dma_usage = DMA_USAGE_NEVER;
+    obj->spi.event = 0;
+    obj->spi.dma_chn_id_tx = DMA_ERROR_OUT_OF_CHANNELS;
+    obj->spi.dma_chn_id_rx = DMA_ERROR_OUT_OF_CHANNELS;
+    
+    /* NOTE: We use vector to judge if asynchronous transfer is on-going (spi_active).
+     *       At initial time, asynchronous transfer is not on-going and so vector must
+     *       be cleared to zero for correct judgement. */
+    NVIC_SetVector(modinit->irq_n, 0);
+#endif
+
+    // Mark this module to be inited.
+    int i = modinit - spi_modinit_tab;
+    spi_modinit_mask |= 1 << i;
+}
+
+void spi_free(spi_t *obj)
+{
+#if DEVICE_SPI_ASYNCH
+    if (obj->spi.dma_chn_id_tx != DMA_ERROR_OUT_OF_CHANNELS) {
+        dma_channel_free(obj->spi.dma_chn_id_tx);
+        obj->spi.dma_chn_id_tx = DMA_ERROR_OUT_OF_CHANNELS;
+    }
+    if (obj->spi.dma_chn_id_rx != DMA_ERROR_OUT_OF_CHANNELS) {
+        dma_channel_free(obj->spi.dma_chn_id_rx);
+        obj->spi.dma_chn_id_rx = DMA_ERROR_OUT_OF_CHANNELS;
+    }
+#endif
+
+    SPI_Close((SPI_T *) NU_MODBASE(obj->spi.spi));
+
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    SPI_DisableInt(((SPI_T *) NU_MODBASE(obj->spi.spi)), (SPI_FIFO_RXOV_INT_MASK | SPI_FIFO_RXTH_INT_MASK | SPI_FIFO_TXTH_INT_MASK));
+    NVIC_DisableIRQ(modinit->irq_n);
+
+    /* Disable IP clock
+     *
+     * NOTE: We must call secure version (from non-secure domain) because SYS/CLK regions are secure.
+     */
+    CLK_DisableModuleClock_S(modinit->clkidx);
+
+    // Mark this module to be deinited.
+    int i = modinit - spi_modinit_tab;
+    spi_modinit_mask &= ~(1 << i);
+}
+
+void spi_format(spi_t *obj, int bits, int mode, int slave)
+{
+    MBED_ASSERT(bits >= NU_SPI_FRAME_MIN && bits <= NU_SPI_FRAME_MAX);
+
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    SPI_DISABLE_SYNC(spi_base);
+
+    SPI_Open(spi_base,
+             slave ? SPI_SLAVE : SPI_MASTER,
+             (mode == 0) ? SPI_MODE_0 : (mode == 1) ? SPI_MODE_1 : (mode == 2) ? SPI_MODE_2 : SPI_MODE_3,
+             bits,
+             SPI_GetBusClock(spi_base));
+    // NOTE: Hardcode to be MSB first.
+    SPI_SET_MSB_FIRST(spi_base);
+
+    if (! slave) {
+        // Master
+        if (obj->spi.pin_ssel != NC) {
+            // Configure SS as low active.
+            SPI_EnableAutoSS(spi_base, SPI_SS, SPI_SS_ACTIVE_LOW);
+        } else {
+            SPI_DisableAutoSS(spi_base);
+        }
+    } else {
+        // Slave
+        // Configure SS as low active.
+        spi_base->SSCTL &= ~SPI_SSCTL_SSACTPOL_Msk;
+    }
+
+    /* NOTE: M451's/M480's/M2351's SPI_Open() will enable SPI transfer (SPI_CTL_SPIEN_Msk).
+     *       We cannot use SPI_CTL_SPIEN_Msk for judgement of spi_active().
+     *       Judge with vector instead. */
+}
+
+void spi_frequency(spi_t *obj, int hz)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    SPI_DISABLE_SYNC(spi_base);
+
+    SPI_SetBusClock((SPI_T *) NU_MODBASE(obj->spi.spi), hz);
+}
+
+
+int spi_master_write(spi_t *obj, int value)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    // NOTE: Data in receive FIFO can be read out via ICE.
+    SPI_ENABLE_SYNC(spi_base);
+
+    // Wait for tx buffer empty
+    while(! spi_writeable(obj));
+    SPI_WRITE_TX(spi_base, value);
+
+    // Wait for rx buffer full
+    while (! spi_readable(obj));
+    int value2 = SPI_READ_RX(spi_base);
+
+    /* We don't call SPI_DISABLE_SYNC here for performance. */
+
+    return value2;
+}
+
+int spi_master_block_write(spi_t *obj, const char *tx_buffer, int tx_length,
+                           char *rx_buffer, int rx_length, char write_fill) {
+    int total = (tx_length > rx_length) ? tx_length : rx_length;
+
+    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;
+        }
+    }
+
+    return total;
+}
+
+#if DEVICE_SPISLAVE
+int spi_slave_receive(spi_t *obj)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    SPI_ENABLE_SYNC(spi_base);
+
+    return spi_readable(obj);
+};
+
+int spi_slave_read(spi_t *obj)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    SPI_ENABLE_SYNC(spi_base);
+
+    // Wait for rx buffer full
+    while (! spi_readable(obj));
+    int value = SPI_READ_RX(spi_base);
+    return value;
+}
+
+void spi_slave_write(spi_t *obj, int value)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    SPI_ENABLE_SYNC(spi_base);
+
+    // Wait for tx buffer empty
+    while(! spi_writeable(obj));
+    SPI_WRITE_TX(spi_base, value);
+}
+#endif
+
+#if DEVICE_SPI_ASYNCH
+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)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+    SPI_SET_DATA_WIDTH(spi_base, bit_width);
+
+    obj->spi.dma_usage = hint;
+    spi_check_dma_usage(&obj->spi.dma_usage, &obj->spi.dma_chn_id_tx, &obj->spi.dma_chn_id_rx);
+    uint32_t data_width = spi_get_data_width(obj);
+    // Conditions to go DMA way:
+    // (1) No DMA support for non-8 multiple data width.
+    // (2) tx length >= rx length. Otherwise, as tx DMA is done, no bus activity for remaining rx.
+    if ((data_width % 8) ||
+            (tx_length < rx_length)) {
+        obj->spi.dma_usage = DMA_USAGE_NEVER;
+        dma_channel_free(obj->spi.dma_chn_id_tx);
+        obj->spi.dma_chn_id_tx = DMA_ERROR_OUT_OF_CHANNELS;
+        dma_channel_free(obj->spi.dma_chn_id_rx);
+        obj->spi.dma_chn_id_rx = DMA_ERROR_OUT_OF_CHANNELS;
+    }
+
+    // SPI IRQ is necessary for both interrupt way and DMA way
+    spi_enable_event(obj, event, 1);
+    spi_buffer_set(obj, tx, tx_length, rx, rx_length);
+
+    SPI_ENABLE_SYNC(spi_base);
+
+    if (obj->spi.dma_usage == DMA_USAGE_NEVER) {
+        // Interrupt way
+        spi_master_write_asynch(obj, spi_fifo_depth(obj) / 2);
+        spi_enable_vector_interrupt(obj, handler, 1);
+        spi_master_enable_interrupt(obj, 1);
+    } else {
+        // DMA way
+        const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+        MBED_ASSERT(modinit != NULL);
+        MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+        PDMA_T *pdma_base = dma_modbase();
+
+        // Configure tx DMA
+        pdma_base->CHCTL |= 1 << obj->spi.dma_chn_id_tx;    // Enable this DMA channel
+        PDMA_SetTransferMode(pdma_base,
+                             obj->spi.dma_chn_id_tx,
+                             ((struct nu_spi_var *) modinit->var)->pdma_perp_tx,    // Peripheral connected to this PDMA
+                             0,                                                     // Scatter-gather disabled
+                             0);                                                    // Scatter-gather descriptor address
+        PDMA_SetTransferCnt(pdma_base,
+                            obj->spi.dma_chn_id_tx,
+                            (data_width == 8) ? PDMA_WIDTH_8 : (data_width == 16) ? PDMA_WIDTH_16 : PDMA_WIDTH_32,
+                            tx_length);
+        PDMA_SetTransferAddr(pdma_base,
+                             obj->spi.dma_chn_id_tx,
+                             (uint32_t) tx, // NOTE:
+                                            // NUC472: End of source address
+                                            // M451/M480/M2351: Start of source address
+                             PDMA_SAR_INC,  // Source address incremental
+                             (uint32_t) &spi_base->TX,  // Destination address
+                             PDMA_DAR_FIX); // Destination address fixed
+        PDMA_SetBurstType(pdma_base,
+                          obj->spi.dma_chn_id_tx,
+                          PDMA_REQ_SINGLE,  // Single mode
+                          0);               // Burst size
+        PDMA_EnableInt(pdma_base,
+                       obj->spi.dma_chn_id_tx,
+                       PDMA_INT_TRANS_DONE);    // Interrupt type
+        // Register DMA event handler
+        dma_set_handler(obj->spi.dma_chn_id_tx, (uint32_t) spi_dma_handler_tx, (uint32_t) obj, DMA_EVENT_ALL);
+
+        // Configure rx DMA
+        pdma_base->CHCTL |= 1 << obj->spi.dma_chn_id_rx;  // Enable this DMA channel
+        PDMA_SetTransferMode(pdma_base,
+                             obj->spi.dma_chn_id_rx,
+                             ((struct nu_spi_var *) modinit->var)->pdma_perp_rx,    // Peripheral connected to this PDMA
+                             0,  // Scatter-gather disabled
+                             0); // Scatter-gather descriptor address
+        PDMA_SetTransferCnt(pdma_base,
+                            obj->spi.dma_chn_id_rx,
+                            (data_width == 8) ? PDMA_WIDTH_8 : (data_width == 16) ? PDMA_WIDTH_16 : PDMA_WIDTH_32,
+                            rx_length);
+        PDMA_SetTransferAddr(pdma_base,
+                             obj->spi.dma_chn_id_rx,
+                             (uint32_t) &spi_base->RX,   // Source address
+                             PDMA_SAR_FIX,   // Source address fixed
+                             (uint32_t) rx,  // NOTE:
+                             // NUC472: End of destination address
+                             // M451/M480/M2351: Start of destination address
+                             PDMA_DAR_INC);  // Destination address incremental
+        PDMA_SetBurstType(pdma_base,
+                          obj->spi.dma_chn_id_rx,
+                          PDMA_REQ_SINGLE,    // Single mode
+                          0); // Burst size
+        PDMA_EnableInt(pdma_base,
+                       obj->spi.dma_chn_id_rx,
+                       PDMA_INT_TRANS_DONE);   // Interrupt type
+        // Register DMA event handler
+        dma_set_handler(obj->spi.dma_chn_id_rx, (uint32_t) spi_dma_handler_rx, (uint32_t) obj, DMA_EVENT_ALL);
+
+        /* Start tx/rx DMA transfer
+         *
+         * If we have both PDMA and SPI interrupts enabled and PDMA priority is lower than SPI priority,
+         * we would trap in SPI interrupt handler endlessly with the sequence:
+         *
+         * 1. PDMA TX transfer done interrupt occurs and is well handled.
+         * 2. SPI RX FIFO threshold interrupt occurs. Trap here because PDMA RX transfer done interrupt doesn't get handled.
+         * 3. PDMA RX transfer done interrupt occurs but it cannot be handled due to above.
+         *
+         * To fix it, we don't enable SPI TX/RX threshold interrupts but keep SPI vector handler set to be called
+         * in PDMA TX/RX transfer done interrupt handlers (spi_dma_handler_tx/spi_dma_handler_rx).
+         */
+        NVIC_SetVector(modinit->irq_n, handler);
+
+        /* Order to enable PDMA TX/RX functions
+         *
+         * H/W spec: In SPI Master mode with full duplex transfer, if both TX and RX PDMA functions are
+         *           enabled, RX PDMA function cannot be enabled prior to TX PDMA function. User can enable
+         *           TX PDMA function firstly or enable both functions simultaneously.
+         * Per real test, it is safer to start RX PDMA first and then TX PDMA. Otherwise, receive FIFO is 
+         * subject to overflow by TX DMA.
+         *
+         * With the above conflicts, we enable PDMA TX/RX functions simultaneously.
+         */
+        spi_base->PDMACTL |= (SPI_PDMACTL_TXPDMAEN_Msk | SPI_PDMACTL_RXPDMAEN_Msk);
+
+        /* Don't enable SPI TX/RX threshold interrupts as commented above */
+    }
+}
+
+/**
+ * Abort an SPI transfer
+ * This is a helper function for event handling. When any of the events listed occurs, the HAL will abort any ongoing
+ * transfers
+ * @param[in] obj The SPI peripheral to stop
+ */
+void spi_abort_asynch(spi_t *obj)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+    PDMA_T *pdma_base = dma_modbase();
+
+    if (obj->spi.dma_usage != DMA_USAGE_NEVER) {
+        // Receive FIFO Overrun in case of tx length > rx length on DMA way
+        if (spi_base->STATUS & SPI_STATUS_RXOVIF_Msk) {
+            spi_base->STATUS = SPI_STATUS_RXOVIF_Msk;
+        }
+
+        if (obj->spi.dma_chn_id_tx != DMA_ERROR_OUT_OF_CHANNELS) {
+            PDMA_DisableInt(pdma_base, obj->spi.dma_chn_id_tx, PDMA_INT_TRANS_DONE);
+            // NOTE: On NUC472, next PDMA transfer will fail with PDMA_STOP() called. Cause is unknown.
+            pdma_base->CHCTL &= ~(1 << obj->spi.dma_chn_id_tx);
+        }
+        SPI_DISABLE_TX_PDMA(((SPI_T *) NU_MODBASE(obj->spi.spi)));
+
+        if (obj->spi.dma_chn_id_rx != DMA_ERROR_OUT_OF_CHANNELS) {
+            PDMA_DisableInt(pdma_base, obj->spi.dma_chn_id_rx, PDMA_INT_TRANS_DONE);
+            // NOTE: On NUC472, next PDMA transfer will fail with PDMA_STOP() called. Cause is unknown.
+            pdma_base->CHCTL &= ~(1 << obj->spi.dma_chn_id_rx);
+        }
+        SPI_DISABLE_RX_PDMA(((SPI_T *) NU_MODBASE(obj->spi.spi)));
+    }
+
+    // Necessary for both interrupt way and DMA way
+    spi_enable_vector_interrupt(obj, 0, 0);
+    spi_master_enable_interrupt(obj, 0);
+
+    /* Necessary for accessing FIFOCTL below */
+    SPI_DISABLE_SYNC(spi_base);
+
+    SPI_ClearRxFIFO(spi_base);
+    SPI_ClearTxFIFO(spi_base);
+}
+
+/**
+ * Handle the SPI interrupt
+ * Read frames until the RX FIFO is empty.  Write at most as many frames as were read.  This way,
+ * it is unlikely that the RX FIFO will overflow.
+ * @param[in] obj The SPI peripheral that generated the interrupt
+ * @return
+ */
+uint32_t spi_irq_handler_asynch(spi_t *obj)
+{
+    // Check for SPI events
+    uint32_t event = spi_event_check(obj);
+    if (event) {
+        spi_abort_asynch(obj);
+    }
+
+    return (obj->spi.event & event) | ((event & SPI_EVENT_COMPLETE) ? SPI_EVENT_INTERNAL_TRANSFER_COMPLETE : 0);
+}
+
+uint8_t spi_active(spi_t *obj)
+{
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    /* Vector will be cleared when asynchronous transfer is finished or aborted.
+       Use it to judge if asynchronous transfer is on-going. */
+    uint32_t vec = NVIC_GetVector(modinit->irq_n);
+    return vec ? 1 : 0;
+}
+
+static int spi_writeable(spi_t * obj)
+{
+    // Receive FIFO must not be full to avoid receive FIFO overflow on next transmit/receive
+    return (! SPI_GET_TX_FIFO_FULL_FLAG(((SPI_T *) NU_MODBASE(obj->spi.spi))));
+}
+
+static int spi_readable(spi_t * obj)
+{
+    return ! SPI_GET_RX_FIFO_EMPTY_FLAG(((SPI_T *) NU_MODBASE(obj->spi.spi)));
+}
+
+static void spi_enable_event(spi_t *obj, uint32_t event, uint8_t enable)
+{
+    obj->spi.event &= ~SPI_EVENT_ALL;
+    obj->spi.event |= (event & SPI_EVENT_ALL);
+    if (event & SPI_EVENT_RX_OVERFLOW) {
+        SPI_EnableInt((SPI_T *) NU_MODBASE(obj->spi.spi), SPI_FIFO_RXOV_INT_MASK);
+    }
+}
+
+static void spi_enable_vector_interrupt(spi_t *obj, uint32_t handler, uint8_t enable)
+{
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    if (enable) {
+        NVIC_SetVector(modinit->irq_n, handler);
+        NVIC_EnableIRQ(modinit->irq_n);
+    } else {
+        NVIC_DisableIRQ(modinit->irq_n);
+        NVIC_SetVector(modinit->irq_n, 0);
+    }
+}
+
+static void spi_master_enable_interrupt(spi_t *obj, uint8_t enable)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    if (enable) {
+        uint32_t fifo_depth = spi_fifo_depth(obj);
+        SPI_SetFIFO(spi_base, fifo_depth / 2, fifo_depth / 2);
+        // Enable tx/rx FIFO threshold interrupt
+        SPI_EnableInt(spi_base, SPI_FIFO_RXTH_INT_MASK | SPI_FIFO_TXTH_INT_MASK);
+    } else {
+        SPI_DisableInt(spi_base, SPI_FIFO_RXTH_INT_MASK | SPI_FIFO_TXTH_INT_MASK);
+    }
+}
+
+static uint32_t spi_event_check(spi_t *obj)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+    uint32_t event = 0;
+
+    if (obj->spi.dma_usage == DMA_USAGE_NEVER) {
+        uint32_t n_rec = spi_master_read_asynch(obj);
+        spi_master_write_asynch(obj, n_rec);
+    }
+
+    if (spi_is_tx_complete(obj) && spi_is_rx_complete(obj)) {
+        event |= SPI_EVENT_COMPLETE;
+    }
+
+    // Receive FIFO Overrun
+    if (spi_base->STATUS & SPI_STATUS_RXOVIF_Msk) {
+        spi_base->STATUS = SPI_STATUS_RXOVIF_Msk;
+        // In case of tx length > rx length on DMA way
+        if (obj->spi.dma_usage == DMA_USAGE_NEVER) {
+            event |= SPI_EVENT_RX_OVERFLOW;
+        }
+    }
+
+    // Receive Time-Out
+    if (spi_base->STATUS & SPI_STATUS_RXTOIF_Msk) {
+        spi_base->STATUS = SPI_STATUS_RXTOIF_Msk;
+        // Not using this IF. Just clear it.
+    }
+    // Transmit FIFO Under-Run
+    if (spi_base->STATUS & SPI_STATUS_TXUFIF_Msk) {
+        spi_base->STATUS = SPI_STATUS_TXUFIF_Msk;
+        event |= SPI_EVENT_ERROR;
+    }
+
+    return event;
+}
+
+/**
+ * Send words from the SPI TX buffer until the send limit is reached or the TX FIFO is full
+ * tx_limit is provided to ensure that the number of SPI frames (words) in flight can be managed.
+ * @param[in] obj       The SPI object on which to operate
+ * @param[in] tx_limit  The maximum number of words to send
+ * @return The number of SPI words that have been transfered
+ */
+static uint32_t spi_master_write_asynch(spi_t *obj, uint32_t tx_limit)
+{
+    uint32_t n_words = 0;
+    uint32_t tx_rmn = obj->tx_buff.length - obj->tx_buff.pos;
+    uint32_t rx_rmn = obj->rx_buff.length - obj->rx_buff.pos;
+    uint32_t max_tx = NU_MAX(tx_rmn, rx_rmn);
+    max_tx = NU_MIN(max_tx, tx_limit);
+    uint8_t data_width = spi_get_data_width(obj);
+    uint8_t bytes_per_word = (data_width + 7) / 8;
+    uint8_t *tx = (uint8_t *)(obj->tx_buff.buffer) + bytes_per_word * obj->tx_buff.pos;
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    while ((n_words < max_tx) && spi_writeable(obj)) {
+        if (spi_is_tx_complete(obj)) {
+            // Transmit dummy as transmit buffer is empty
+            SPI_WRITE_TX(spi_base, 0);
+        } else {
+            switch (bytes_per_word) {
+            case 4:
+                SPI_WRITE_TX(spi_base, nu_get32_le(tx));
+                tx += 4;
+                break;
+            case 2:
+                SPI_WRITE_TX(spi_base, nu_get16_le(tx));
+                tx += 2;
+                break;
+            case 1:
+                SPI_WRITE_TX(spi_base, *((uint8_t *) tx));
+                tx += 1;
+                break;
+            }
+
+            obj->tx_buff.pos ++;
+        }
+        n_words ++;
+    }
+
+    //Return the number of words that have been sent
+    return n_words;
+}
+
+/**
+ * Read SPI words out of the RX FIFO
+ * Continues reading words out of the RX FIFO until the following condition is met:
+ * o There are no more words in the FIFO
+ * OR BOTH OF:
+ * o At least as many words as the TX buffer have been received
+ * o At least as many words as the RX buffer have been received
+ * This way, RX overflows are not generated when the TX buffer size exceeds the RX buffer size
+ * @param[in] obj The SPI object on which to operate
+ * @return Returns the number of words extracted from the RX FIFO
+ */
+static uint32_t spi_master_read_asynch(spi_t *obj)
+{
+    uint32_t n_words = 0;
+    uint32_t tx_rmn = obj->tx_buff.length - obj->tx_buff.pos;
+    uint32_t rx_rmn = obj->rx_buff.length - obj->rx_buff.pos;
+    uint32_t max_rx = NU_MAX(tx_rmn, rx_rmn);
+    uint8_t data_width = spi_get_data_width(obj);
+    uint8_t bytes_per_word = (data_width + 7) / 8;
+    uint8_t *rx = (uint8_t *)(obj->rx_buff.buffer) + bytes_per_word * obj->rx_buff.pos;
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    while ((n_words < max_rx) && spi_readable(obj)) {
+        if (spi_is_rx_complete(obj)) {
+            // Disregard as receive buffer is full
+            SPI_READ_RX(spi_base);
+        } else {
+            switch (bytes_per_word) {
+            case 4: {
+                uint32_t val = SPI_READ_RX(spi_base);
+                nu_set32_le(rx, val);
+                rx += 4;
+                break;
+            }
+            case 2: {
+                uint16_t val = SPI_READ_RX(spi_base);
+                nu_set16_le(rx, val);
+                rx += 2;
+                break;
+            }
+            case 1:
+                *rx ++ = SPI_READ_RX(spi_base);
+                break;
+            }
+
+            obj->rx_buff.pos ++;
+        }
+        n_words ++;
+    }
+
+    // Return the number of words received
+    return n_words;
+}
+
+static void spi_buffer_set(spi_t *obj, const void *tx, size_t tx_length, void *rx, size_t rx_length)
+{
+    obj->tx_buff.buffer = (void *) tx;
+    obj->tx_buff.length = tx_length;
+    obj->tx_buff.pos = 0;
+    obj->tx_buff.width = spi_get_data_width(obj);
+    obj->rx_buff.buffer = rx;
+    obj->rx_buff.length = rx_length;
+    obj->rx_buff.pos = 0;
+    obj->rx_buff.width = spi_get_data_width(obj);
+}
+
+static void spi_check_dma_usage(DMAUsage *dma_usage, int *dma_ch_tx, int *dma_ch_rx)
+{
+    if (*dma_usage != DMA_USAGE_NEVER) {
+        if (*dma_ch_tx == DMA_ERROR_OUT_OF_CHANNELS) {
+            *dma_ch_tx = dma_channel_allocate(DMA_CAP_NONE);
+        }
+        if (*dma_ch_rx == DMA_ERROR_OUT_OF_CHANNELS) {
+            *dma_ch_rx = dma_channel_allocate(DMA_CAP_NONE);
+        }
+
+        if (*dma_ch_tx == DMA_ERROR_OUT_OF_CHANNELS || *dma_ch_rx == DMA_ERROR_OUT_OF_CHANNELS) {
+            *dma_usage = DMA_USAGE_NEVER;
+        }
+    }
+
+    if (*dma_usage == DMA_USAGE_NEVER) {
+        dma_channel_free(*dma_ch_tx);
+        *dma_ch_tx = DMA_ERROR_OUT_OF_CHANNELS;
+        dma_channel_free(*dma_ch_rx);
+        *dma_ch_rx = DMA_ERROR_OUT_OF_CHANNELS;
+    }
+}
+
+static uint8_t spi_get_data_width(spi_t *obj)
+{
+    SPI_T *spi_base = (SPI_T *) NU_MODBASE(obj->spi.spi);
+
+    uint32_t data_width = ((spi_base->CTL & SPI_CTL_DWIDTH_Msk) >> SPI_CTL_DWIDTH_Pos);
+    if (data_width == 0) {
+        data_width = 32;
+    }
+
+    return data_width;
+}
+
+static int spi_is_tx_complete(spi_t *obj)
+{
+    return (obj->tx_buff.pos == obj->tx_buff.length);
+}
+
+static int spi_is_rx_complete(spi_t *obj)
+{
+    return (obj->rx_buff.pos == obj->rx_buff.length);
+}
+
+static void spi_dma_handler_tx(uint32_t id, uint32_t event_dma)
+{
+    spi_t *obj = (spi_t *) id;
+
+    // FIXME: Pass this error to caller
+    if (event_dma & DMA_EVENT_ABORT) {
+    }
+    // Expect SPI IRQ will catch this transfer done event
+    if (event_dma & DMA_EVENT_TRANSFER_DONE) {
+        obj->tx_buff.pos = obj->tx_buff.length;
+    }
+    // FIXME: Pass this error to caller
+    if (event_dma & DMA_EVENT_TIMEOUT) {
+    }
+
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    void (*vec)(void) = (void (*)(void)) NVIC_GetVector(modinit->irq_n);
+    vec();
+}
+
+static void spi_dma_handler_rx(uint32_t id, uint32_t event_dma)
+{
+    spi_t *obj = (spi_t *) id;
+
+    // FIXME: Pass this error to caller
+    if (event_dma & DMA_EVENT_ABORT) {
+    }
+    // Expect SPI IRQ will catch this transfer done event
+    if (event_dma & DMA_EVENT_TRANSFER_DONE) {
+        obj->rx_buff.pos = obj->rx_buff.length;
+    }
+    // FIXME: Pass this error to caller
+    if (event_dma & DMA_EVENT_TIMEOUT) {
+    }
+
+    const struct nu_modinit_s *modinit = get_modinit(obj->spi.spi, spi_modinit_tab);
+    MBED_ASSERT(modinit != NULL);
+    MBED_ASSERT(modinit->modname == (int) obj->spi.spi);
+
+    void (*vec)(void) = (void (*)(void)) NVIC_GetVector(modinit->irq_n);
+    vec();
+}
+
+/** Return FIFO depth of the SPI peripheral
+ *
+ * @details
+ *  M2351
+ *      SPI0/1/2/3      8 if data width <=16; 4 otherwise
+ */
+static uint32_t spi_fifo_depth(spi_t *obj)
+{
+    return (spi_get_data_width(obj) <= 16) ? 8 : 4;
+}
+
+#endif
+
+#endif