SPI Slave Half Duplex

[中文]

Introduction

The Half Duplex (HD) Mode is a special mode provided by ESP SPI Slave peripheral. Under this mode, the hardware provides more services than the Full Duplex (FD) Mode (the mode for general-purpose SPI transactions, see SPI Slave Driver). These services reduce the CPU load and the response time of SPI Slave. However, it is important to note that the communication format is determined by the hardware and is always in a half-duplex configuration, allowing only one-way data transfer at any given time. Hence, the mode is named Half Duplex Mode due to this characteristic.

When conducting an SPI transaction, transactions can be classified into several types based on the command phase of the transaction. Each transaction may consist of the following phases: command, address, dummy, and data. The command phase is mandatory, while the other phases may be determined by the command field. During the command, address, and dummy phases, the bus is always controlled by the master (usually the host), while the direction of the data phase depends on the command. The data phase can be either an input phase, where the master writes data to the slave (e.g., the host sends data to the slave), or an output phase, where the master reads data from the slave (e.g., the host receives data from the slave).

Protocol

About the details of how master should communicate with the SPI Slave, see ESP SPI Slave HD (Half Duplex) Mode Protocol.

Through these different transactions, the slave provides these services to the master:

  • A DMA channel for the master to write a great amount of data to the slave.

  • A DMA channel for the master to read a great amount of data from the slave.

  • Several general purpose registers, shared between the master and the slave.

  • Several general purpose interrupts, for the master to interrupt the SW of the slave.

Terminology

  • Transaction

  • Channel

  • Sending

  • Receiving

  • Data Descriptor

Driver Feature

  • Transaction read/write by master in segments

  • Queues for data to send and received

Driver Usage

Slave Initialization

Call spi_slave_hd_init() to initialize the SPI bus as well as the peripheral and the driver. The SPI Slave exclusively uses the SPI peripheral, pins of the bus before it is deinitialized, which means other devices are unable to use the above resources during initialization. Thus, to ensure SPI resources are correctly occupied and the connections work properly, most configurations of the slave should be done as soon as the slave is initialized.

The spi_bus_config_t specifies how the bus should be initialized, while spi_slave_hd_slot_config_t specifies how the SPI Slave driver should work.

Deinitialization (Optional)

Call spi_slave_hd_deinit() to uninstall the driver. The resources, including the pins, SPI peripheral, internal memory used by the driver, and interrupt sources, are released by the deinit() function.

Send/Receive Data by DMA Channels

To send data to the master through the sending DMA channel, the application should properly wrap the data in an spi_slave_hd_data_t descriptor structure before calling spi_slave_hd_queue_trans() with the data descriptor and the channel argument of SPI_SLAVE_CHAN_TX. The pointers to descriptors are stored in the queue, and the data is sent to the master in the same order they are enqueued using spi_slave_hd_queue_trans(), upon receiving the master's Rd_DMA command.

The application should check the result of data sending by calling spi_slave_hd_get_trans_res() with the channel set as SPI_SLAVE_CHAN_TX. This function blocks until the transaction with the command Rd_DMA from the master successfully completes (or timeout). The out_trans argument of the function outputs the pointer of the data descriptor which is just finished, providing information about the sending.

Receiving data from the master through the receiving DMA channel is quite similar. The application calls spi_slave_hd_queue_trans() with proper data descriptor and the channel argument of SPI_SLAVE_CHAN_RX. And the application calls the spi_slave_hd_get_trans_res() later to get the descriptor to the receiving buffer before it handles the data in the receiving buffer.

Note

This driver itself does not have an internal buffer for the data to send or just received. The application should provide data buffer for driver via data descriptors to send to the master, or to receive data from the master.

The application has to properly keep the data descriptor as well as the buffer it points, after the descriptor is successfully sent into the driver internal queue by spi_slave_hd_queue_trans(), and before returned by spi_slave_hd_get_trans_res(). During this period, the hardware as well as the driver may read or write to the buffer and the descriptor when required at any time.

Please note that, when using this driver for data transfer, the buffer does not have to be fully sent or filled before it is terminated. For example, in the segment transaction mode, the master has to send CMD7 to terminate a Wr_DMA transaction or send CMD8 to terminate an Rd_DMA transaction (in segments), no matter whether the send (receive) buffer is used up (full) or not.

Using Data Descriptor with Customized User Arguments

Sometimes you may have initiator (sending data descriptor) and closure (handling returned descriptors) functions in different places. When you get the returned data descriptor in the closure, you may need some extra information when handling the finished data descriptor. For example, you may want to know which round it is for the returned descriptor when you send the same piece of data several times.

Set the arg member in the data descriptor to a variable indicating the transaction by force casting, or point it to a structure that wraps all the information you may need when handling the sending/receiving data. Then you can get what you need in your closure.

Using Callbacks

Note

These callbacks are called in the ISR, so the required operations need to be processed quickly and returned as soon as possible to ensure that the system is functioning properly. You may need to be very careful to write the code in the ISR.

Since the interrupt handling is executed concurrently with the application, long delays or blocking may cause the system to respond slower or lead to unpredictable behavior. Therefore, when writing callback functions, avoid using operations that may cause delays or blocking, e.g., waiting, sleeping, resource locking, etc.

The spi_slave_hd_callback_config_t member in the spi_slave_hd_slot_config_t configuration structure passed when initializing the SPI Slave HD driver, allows you to have callbacks for each event you may concern.

The corresponding interrupt for each callback that is not NULL is enabled, so that the callbacks can be called immediately when the events happen. You do not need to provide callbacks for the unconcerned events.

The arg member in the configuration structure can help you pass some context to the callback or indicate the specific SPI Slave instance when using the same callbacks for multiple SPI Slave peripherals. You can set the arg member to a variable that indicates the SPI Slave instance by performing a forced type casting or point it to a context structure. All the callbacks are called with this arg argument you set when the callbacks are initialized.

There are two other arguments: the event and the awoken.

  • The event passes the information of the current event to the callback. The spi_slave_hd_event_t type contains the information of the event, for example, event type, the data descriptor just finished (The data argument is very useful in this case!).

  • The awoken argument serves as an output parameter. It informs the ISR that tasks have been awakened after the callback function, and the ISR should call portYIELD_FROM_ISR() to schedule these tasks. Simply pass the awoken argument to all FreeRTOS APIs that may unblock tasks, and the value of awoken will be returned to the ISR.

Writing/Reading Shared Registers

Call spi_slave_hd_write_buffer() to write the shared buffer, and spi_slave_hd_read_buffer() to read the shared buffer.

Note

On ESP32-P4, the shared registers are read/written in words by the application but read/written in bytes by the master. There is no guarantee four continuous bytes read from the master are from the same word written by the slave's application. It is also possible that if the slave reads a word while the master is writing bytes of the word, the slave may get one word with half of them just written by the master, and the other half has not been written into.

The master can confirm that the word is not in transition by reading the word twice and comparing the values.

For the slave, it is more difficult to ensure the word is not in transition because the process of master writing four bytes can be very long (32 SPI clocks). You can put some CRC in the last (largest address) byte of a word so that when the byte is written, the word is sure to be all written.

Due to the conflicts that may be among read/write from SW (worse if there are multi-cores) and master, it is suggested that a word is only used in one direction (only written by the master or only written by the slave).

Receiving General Purpose Interrupts from the Master

When the master sends CMD8, CMD9 or CMDA, the slave corresponding is triggered. Currently the CMD8 is permanently used to indicate the termination of Rd_DMA segments. To receive general-purpose interrupts, register callbacks for CMD9 and CMDA when the slave is initialized, see Using Callbacks.

Sleep Retention

ESP32-P4 supports to retain the SPI register context before entering light sleep and restore them after waking up. This means you don't have to re-init the SPI driver after the light sleep.

This feature can be enabled by setting the flag SPICOMMON_BUSFLAG_SLP_ALLOW_PD. It will allow the system to power down the SPI in light sleep, meanwhile save the register context. It can help to save more power consumption with some extra cost of the memory.

Notice that when GPSPI is working as a slave, it is not support to enter sleep when any transaction (including TX and RX) is not finished.

Application Examples

The code example for Device/Host communication can be found in the peripherals/spi_slave_hd directory of ESP-IDF examples.

  • example

    peripherals/spi_slave_hd/append_mode demonstrates how to use the SPI Slave HD driver and ESSL driver to communicate (ESSL driver is an encapsulated layer based on SPI Master driver to communicate with halfduplex mode SPI Slave).

  • example

    peripherals/spi_slave_hd/segment_mode demonstrate two ways to use the SPI Slave Halfduplex Segment Mode: Using the SPI Slave Halfduplex driver with two tasks repeating transactions with the SPI Master, and using the ESP Serial Slave Link APIs for multiple exchanges with the slave.

API Reference

Header File

  • components/esp_driver_spi/include/driver/spi_slave_hd.h

  • This header file can be included with:

    #include "driver/spi_slave_hd.h"

  • This header file is a part of the API provided by the esp_driver_spi component. To declare that your component depends on esp_driver_spi, add the following to your CMakeLists.txt:

    REQUIRES esp_driver_spi

    or

    PRIV_REQUIRES esp_driver_spi

Functions

esp_err_t spi_slave_hd_init(spi_host_device_t host_id, const spi_bus_config_t *bus_config, const spi_slave_hd_slot_config_t *config)

Initialize the SPI Slave HD driver.

Parameters

  • host_id -- The host to use

  • bus_config -- Bus configuration for the bus used

  • config -- Configuration for the SPI Slave HD driver

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: invalid argument given

  • ESP_ERR_INVALID_STATE: function called in invalid state, may be some resources are already in use

  • ESP_ERR_NOT_FOUND if there is no available DMA channel

  • ESP_ERR_NO_MEM: memory allocation failed

  • or other return value from esp_intr_alloc

esp_err_t spi_slave_hd_deinit(spi_host_device_t host_id)

Deinitialize the SPI Slave HD driver.

Parameters

host_id -- The host to deinitialize the driver

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: if the host_id is not correct

esp_err_t spi_slave_hd_queue_trans(spi_host_device_t host_id, spi_slave_chan_t chan, spi_slave_hd_data_t *trans, TickType_t timeout)

Queue transactions (segment mode)

Parameters

  • host_id -- Host to queue the transaction

  • chan -- SPI_SLAVE_CHAN_TX or SPI_SLAVE_CHAN_RX

  • trans -- Transaction descriptors

  • timeout -- Timeout before the data is queued

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: The input argument is invalid. Can be the following reason:

    • The buffer given is not DMA capable

    • The length of data is invalid (not larger than 0, or exceed the max transfer length)

    • The transaction direction is invalid

  • ESP_ERR_TIMEOUT: Cannot queue the data before timeout. Master is still processing previous transaction.

  • ESP_ERR_INVALID_STATE: Function called in invalid state. This API should be called under segment mode.

esp_err_t spi_slave_hd_get_trans_res(spi_host_device_t host_id, spi_slave_chan_t chan, spi_slave_hd_data_t **out_trans, TickType_t timeout)

Get the result of a data transaction (segment mode)

Note

This API should be called successfully the same times as the spi_slave_hd_queue_trans.

Parameters

  • host_id -- Host to queue the transaction

  • chan -- Channel to get the result, SPI_SLAVE_CHAN_TX or SPI_SLAVE_CHAN_RX

  • out_trans -- [out] Pointer to the transaction descriptor (spi_slave_hd_data_t) passed to the driver before. Hardware has finished this transaction. Member trans_len indicates the actual number of bytes of received data, it's meaningless for TX.

  • timeout -- Timeout before the result is got

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: Function is not valid

  • ESP_ERR_TIMEOUT: There's no transaction done before timeout

  • ESP_ERR_INVALID_STATE: Function called in invalid state. This API should be called under segment mode.

void spi_slave_hd_read_buffer(spi_host_device_t host_id, int addr, uint8_t *out_data, size_t len)

Read the shared registers.

Parameters

  • host_id -- Host to read the shared registers

  • addr -- Address of register to read, 0 to SOC_SPI_MAXIMUM_BUFFER_SIZE-1

  • out_data -- [out] Output buffer to store the read data

  • len -- Length to read, not larger than SOC_SPI_MAXIMUM_BUFFER_SIZE-addr

void spi_slave_hd_write_buffer(spi_host_device_t host_id, int addr, uint8_t *data, size_t len)

Write the shared registers.

Parameters

  • host_id -- Host to write the shared registers

  • addr -- Address of register to write, 0 to SOC_SPI_MAXIMUM_BUFFER_SIZE-1

  • data -- Buffer holding the data to write

  • len -- Length to write, SOC_SPI_MAXIMUM_BUFFER_SIZE-addr

esp_err_t spi_slave_hd_append_trans(spi_host_device_t host_id, spi_slave_chan_t chan, spi_slave_hd_data_t *trans, TickType_t timeout)

Load transactions (append mode)

Note

In this mode, user transaction descriptors will be appended to the DMA and the DMA will keep processing the data without stopping

Parameters

  • host_id -- Host to load transactions

  • chan -- SPI_SLAVE_CHAN_TX or SPI_SLAVE_CHAN_RX

  • trans -- Transaction descriptor

  • timeout -- Timeout before the transaction is loaded

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: The input argument is invalid. Can be the following reason:

    • The buffer given is not DMA capable

    • The length of data is invalid (not larger than 0, or exceed the max transfer length)

    • The transaction direction is invalid

  • ESP_ERR_TIMEOUT: Master is still processing previous transaction. There is no available transaction for slave to load

  • ESP_ERR_INVALID_STATE: Function called in invalid state. This API should be called under append mode.

esp_err_t spi_slave_hd_get_append_trans_res(spi_host_device_t host_id, spi_slave_chan_t chan, spi_slave_hd_data_t **out_trans, TickType_t timeout)

Get the result of a data transaction (append mode)

Note

This API should be called the same times as the spi_slave_hd_append_trans

Parameters

  • host_id -- Host to load the transaction

  • chan -- SPI_SLAVE_CHAN_TX or SPI_SLAVE_CHAN_RX

  • out_trans -- [out] Pointer to the transaction descriptor (spi_slave_hd_data_t) passed to the driver before. Hardware has finished this transaction. Member trans_len indicates the actual number of bytes of received data, it's meaningless for TX.

  • timeout -- Timeout before the result is got

Returns

  • ESP_OK: on success

  • ESP_ERR_INVALID_ARG: Function is not valid

  • ESP_ERR_TIMEOUT: There's no transaction done before timeout

  • ESP_ERR_INVALID_STATE: Function called in invalid state. This API should be called under append mode.

Structures

struct spi_slave_hd_data_t

Descriptor of data to send/receive.

Public Members

uint8_t *data

Buffer to send, must be DMA capable.

size_t len

Len of data to send/receive. For receiving the buffer length should be multiples of 4 bytes, otherwise the extra part will be truncated.

size_t trans_len

For RX direction, it indicates the data actually received. For TX direction, it is meaningless.

uint32_t flags

Bitwise OR of SPI_SLAVE_HD_TRANS_* flags.

void *arg

Extra argument indicating this data.

struct spi_slave_hd_event_t

Information of SPI Slave HD event.

Public Members

spi_event_t event

Event type.

spi_slave_hd_data_t *trans

Corresponding transaction for SPI_EV_SEND and SPI_EV_RECV events.

struct spi_slave_hd_callback_config_t

Callback configuration structure for SPI Slave HD.

Public Members

slave_cb_t cb_buffer_tx

Callback when master reads from shared buffer.

slave_cb_t cb_buffer_rx

Callback when master writes to shared buffer.

slave_cb_t cb_send_dma_ready

Callback when TX data buffer is loaded to the hardware (DMA)

slave_cb_t cb_sent

Callback when data are sent.

slave_cb_t cb_recv_dma_ready

Callback when RX data buffer is loaded to the hardware (DMA)

slave_cb_t cb_recv

Callback when data are received.

slave_cb_t cb_cmd9

Callback when CMD9 received.

slave_cb_t cb_cmdA

Callback when CMDA received.

void *arg

Argument indicating this SPI Slave HD peripheral instance.

struct spi_slave_hd_slot_config_t

Configuration structure for the SPI Slave HD driver.

Public Members

uint8_t mode

SPI mode, representing a pair of (CPOL, CPHA) configuration:

  • 0: (0, 0)

  • 1: (0, 1)

  • 2: (1, 0)

  • 3: (1, 1)

uint32_t spics_io_num

CS GPIO pin for this device.

uint32_t flags

Bitwise OR of SPI_SLAVE_HD_* flags.

uint32_t command_bits

command field bits, multiples of 8 and at least 8.

uint32_t address_bits

address field bits, multiples of 8 and at least 8.

uint32_t dummy_bits

dummy field bits, multiples of 8 and at least 8.

uint32_t queue_size

Transaction queue size. This sets how many transactions can be 'in the air' (queued using spi_slave_hd_queue_trans but not yet finished using spi_slave_hd_get_trans_result) at the same time.

spi_dma_chan_t dma_chan

DMA channel to used.

spi_slave_hd_callback_config_t cb_config

Callback configuration.

Macros

SPI_SLAVE_HD_TRANS_DMA_BUFFER_ALIGN_AUTO

Automatically re-malloc dma buffer if user buffer doesn't meet hardware alignment or dma_capable, this process may lose some memory and performance.

SPI_SLAVE_HD_TXBIT_LSBFIRST

Transmit command/address/data LSB first instead of the default MSB first.

SPI_SLAVE_HD_RXBIT_LSBFIRST

Receive data LSB first instead of the default MSB first.

SPI_SLAVE_HD_BIT_LSBFIRST

Transmit and receive LSB first.

SPI_SLAVE_HD_APPEND_MODE

Adopt DMA append mode for transactions. In this mode, users can load(append) DMA descriptors without stopping the DMA.

Type Definitions

typedef bool (*slave_cb_t)(void *arg, spi_slave_hd_event_t *event, BaseType_t *awoken)

Callback for SPI Slave HD.

Enumerations

enum spi_slave_chan_t

Channel of SPI Slave HD to do data transaction.

Values:

enumerator SPI_SLAVE_CHAN_TX

The output channel (RDDMA)

enumerator SPI_SLAVE_CHAN_RX

The input channel (WRDMA)

Was this page helpful?