The ESP32 has four SPI peripheral devices, called SPI0, SPI1, HSPI and VSPI. SPI0 is entirely dedicated to the flash cache the ESP32 uses to map the SPI flash device it is connected to into memory. SPI1 is connected to the same hardware lines as SPI0 and is used to write to the flash chip. HSPI and VSPI are free to use. SPI1, HSPI and VSPI all have three chip select lines, allowing them to drive up to three SPI devices each as a master.
The spi_master driver allows easy communicating with SPI slave devices, even in a multithreaded environment. It fully transparently handles DMA transfers to read and write data and automatically takes care of multiplexing between different SPI slaves on the same master.
Note
Notes about thread safety
The SPI driver API is thread safe when multiple SPI devices on the same bus are accessed from different tasks. However, the driver is not thread safe if the same SPI device is accessed from multiple tasks.
In this case, it is recommended to either refactor your application so only a single task accesses each SPI device, or to add mutex locking around access of the shared device.
The spi_master driver uses the following terms:
- Host: The SPI peripheral inside the ESP32 initiating the SPI transmissions. One of SPI, HSPI or VSPI. (For now, only HSPI or VSPI are actually supported in the driver; it will support all 3 peripherals somewhere in the future.)
- Bus: The SPI bus, common to all SPI devices connected to one host. In general the bus consists of the
miso, mosi, sclk and optionally quadwp and quadhd signals. The SPI slaves are connected to these
signals in parallel.
- miso - Also known as q, this is the input of the serial stream into the ESP32
- mosi - Also known as d, this is the output of the serial stream from the ESP32
- sclk - Clock signal. Each data bit is clocked out or in on the positive or negative edge of this signal
- quadwp - Write Protect signal. Only used for 4-bit (qio/qout) transactions.
- quadhd - Hold signal. Only used for 4-bit (qio/qout) transactions.
- Device: A SPI slave. Each SPI slave has its own chip select (CS) line, which is made active when a transmission to/from the SPI slave occurs.
- Transaction: One instance of CS going active, data transfer from and/or to a device happening, and CS going inactive again. Transactions are atomic, as in they will never be interrupted by another transaction.
A transaction on the SPI bus consists of five phases, any of which may be skipped:
- The command phase. In this phase, a command (0-16 bit) is clocked out.
- The address phase. In this phase, an address (0-64 bit) is clocked out.
- The write phase. The master sends data to the slave.
- The dummy phase. The phase is configurable, used to meet the timing requirements.
- The read phase. The slave sends data to the master.
In full duplex mode, the read and write phases are combined, and the SPI host reads and
writes data simultaneously. The total transaction length is decided by
command_bits + address_bits + trans_conf.length
, while the trans_conf.rx_length
only determins length of data received into the buffer.
While in half duplex mode, the host have independent write and read phases. The length of write phase and read phase are
decided by trans_conf.length
and trans_conf.rx_length
respectively.
The command and address phase are optional in that not every SPI device will need to be sent a command
and/or address. This is reflected in the device configuration: when the command_bits
or address_bits
fields are set to zero, no command or address phase is done.
Something similar is true for the read and write phase: not every transaction needs both data to be written
as well as data to be read. When rx_buffer
is NULL (and SPI_USE_RXDATA) is not set) the read phase
is skipped. When tx_buffer
is NULL (and SPI_USE_TXDATA) is not set) the write phase is skipped.
The driver offers two different kinds of transactions: the interrupt transactions and the polling transactions. Each device can choose one kind of transaction to send. See :ref:`mixed_transactions` if your device do require both kinds of transactions.
The interrupt transactions use an interrupt-driven logic when the transactions are in-flight. The routine will get blocked, allowing the CPU to run other tasks, while it is waiting for a transaction to be finished.
Interrupt transactions can be queued into a device, the driver automatically send them one-by-one in the ISR. A task can queue several transactions, and then do something else before the transactions are finished.
The polling transactions don't rely on the interrupt, the routine keeps polling the status bit of the SPI peripheral until the transaction is done.
All the tasks that do interrupt transactions may get blocked by the queue, at which point they need to wait for the ISR to run twice before the transaction is done. Polling transactions save the time spent on queue handling and context switching, resulting in a smaller transaction interval smaller. The disadvantage is that the the CPU is busy while these transactions are in flight.
The spi_device_polling_end
routine spends at least 1us overhead to
unblock other tasks when the transaction is done. It is strongly recommended
to wrap a series of polling transactions inside of spi_device_acquire_bus
and spi_device_release_bus
to avoid the overhead. (See
:ref:`bus_acquiring`)
During the command and address phases, cmd
and addr
field in the
spi_transaction_t
struct are sent to the bus, while nothing is read at the
same time. The default length of command and address phase are set in the
spi_device_interface_config_t
and by spi_bus_add_device
. When the the
flag SPI_TRANS_VARIABLE_CMD
and SPI_TRANS_VARIABLE_ADDR
are not set in
the spi_transaction_t
,the driver automatically set the length of these
phases to the default value as set when the device is initialized respectively.
If the length of command and address phases needs to be variable, declare a
spi_transaction_ext_t
descriptor, set the flag SPI_TRANS_VARIABLE_CMD
or/and SPI_TRANS_VARIABLE_ADDR
in the flags
of base
member and
configure the rest part of base
as usual. Then the length of each phases
will be command_bits
and address_bits
set in the spi_transaction_ext_t
.
Normally, data to be transferred to or from a device will be read from or written to a chunk of memory
indicated by the rx_buffer
and tx_buffer
members of the transaction structure.
When DMA is enabled for transfers, these buffers are highly recommended to meet the requirements as below:
- allocated in DMA-capable memory using
pvPortMallocCaps(size, MALLOC_CAP_DMA)
;- 32-bit aligned (start from the boundary and have length of multiples of 4 bytes).
If these requirements are not satisfied, efficiency of the transaction will suffer due to the allocation and memcpy of temporary buffers.
Note
Half duplex transactions with both read and write phases are not supported when using DMA. See :ref:`spi_known_issues` for details and workarounds.
Sometimes you may want to send spi transactions exclusively, continuously, to
make it as fast as possible. You may use spi_device_acquire_bus
and
spi_device_release_bus
to realize this. When the bus is acquired,
transactions to other devices (no matter polling or interrupt) are pending
until the bus is released.
Initialize a SPI bus by calling
spi_bus_initialize
. Make sure to set the correct IO pins in thebus_config
struct. Take care to set signals that are not needed to -1.Tell the driver about a SPI slave device connected to the bus by calling spi_bus_add_device. Make sure to configure any timing requirements the device has in the
dev_config
structure. You should now have a handle for the device, to be used when sending it a transaction.To interact with the device, fill one or more spi_transaction_t structure with any transaction parameters you need. Then send them either in a polling way or the interrupt way:
- :ref:`Interrupt <interrupt_transactions>`
- Either queue all transactions by calling
spi_device_queue_trans
, and at a later time query the result usingspi_device_get_trans_result
, or handle all requests synchroneously by feeding them intospi_device_transmit
.
- :ref:`Polling <polling_transactions>`
- Call the
spi_device_polling_transmit
to send polling transactions. Alternatively, you can send a polling transaction byspi_device_polling_start
andspi_device_polling_end
if you want to insert something between them.
Optional: to do back-to-back transactions to a device, call
spi_device_acquire_bus
before andspi_device_release_bus
after the transactions.Optional: to unload the driver for a device, call
spi_bus_remove_device
with the device handle as an argumentOptional: to remove the driver for a bus, make sure no more drivers are attached and call
spi_bus_free
.
- Transactions with small amount of data:
Sometimes, the amount of data is very small making it less than optimal allocating a separate buffer for it. If the data to be transferred is 32 bits or less, it can be stored in the transaction struct itself. For transmitted data, use the
tx_data
member for this and set theSPI_USE_TXDATA
flag on the transmission. For received data, userx_data
and setSPI_USE_RXDATA
. In both cases, do not touch thetx_buffer
orrx_buffer
members, because they use the same memory locations astx_data
andrx_data
.
- Transactions with integers other than uint8_t
The SPI peripheral reads and writes the memory byte-by-byte. By default, the SPI works at MSB first mode, each bytes are sent or received from the MSB to the LSB. However, if you want to send data with length which is not multiples of 8 bits, unused bits are sent.
E.g. you write
uint8_t data = 0x15
(00010101B), and set length to only 5 bits, the sent data is00010B
rather than expected10101B
.Moreover, ESP32 is a little-endian chip whose lowest byte is stored at the very beginning address for uint16_t and uint32_t variables. Hence if a uint16_t is stored in the memory, it's bit 7 is first sent, then bit 6 to 0, then comes its bit 15 to bit 8.
To send data other than uint8_t arrays, macros
SPI_SWAP_DATA_TX
is provided to shift your data to the MSB and swap the MSB to the lowest address; whileSPI_SWAP_DATA_RX
can be used to swap received data from the MSB to it's correct place.
Most peripheral signals in ESP32 can connect directly to a specific GPIO, which is called its IOMUX pin. When a peripheral signal is routed to a pin other than its IOMUX pin, ESP32 uses the less direct GPIO matrix to make this connection.
If the driver is configured with all SPI signals set to their specific IOMUX pins (or left unconnected), it will bypass the GPIO matrix. If any SPI signal is configured to a pin other than its IOMUx pin, the driver will automatically route all the signals via the GPIO Matrix. The GPIO matrix samples all signals at 80MHz and sends them between the GPIO and the peripheral.
When the GPIO matrix is used, signals faster than 40MHz cannot propagate and the setup time of MISO is more easily violated, since the input delay of MISO signal is increased. The maximum clock frequency with GPIO Matrix is 40MHz or less, whereas using all IOMUX pins allows 80MHz.
Note
More details about influence of input delay on the maximum clock frequency, see :ref:`timing_considerations` below.
IOMUX pins for SPI controllers are as below:
Pin Name | HSPI | VSPI |
---|---|---|
GPIO Number | ||
CS0* | 15 | 5 |
SCLK | 14 | 18 |
MISO | 12 | 19 |
MOSI | 13 | 23 |
QUADWP | 2 | 22 |
QUADHD | 4 | 21 |
note * Only the first device attaching to the bus can use CS0 pin.
Though we suggest to send only one type (interrupt or polling) of transactions to one device to reduce coding complexity, it is supported to send both interrupt and polling transactions alternately. Notes below is to help you do this.
The polling transactions should be started when all the other transactions are finished, no matter they are polling or interrupt.
An unfinished polling transaction forbid other transactions from being sent.
Always call spi_device_polling_end
after spi_device_polling_start
to
allow other device using the bus, or allow other transactions to be started
to the same device. You can use spi_device_polling_transmit
to simplify
this if you don't need to do something during your polling transaction.
An in-flight polling transaction would get disturbed by the ISR operation
caused by interrupt transactions. Always make sure all the interrupt
transactions sent to the ISR are finished before you call
spi_device_polling_start
. To do that, you can call
spi_device_get_trans_result
until all the transactions are returned.
It is strongly recommended to send mixed transactions to the same device in only one task to control the calling sequence of functions.
There're three factors limiting the transferring speed: (1) The transaction interval, (2) The SPI clock frequency used. (3) The cache miss of SPI functions including callbacks. When large transactions are used, the clock frequency determines the transferring speed; while the interval effects the speed a lot if small transactions are used.
Transaction interval: It takes time for the software to setup spi peripheral registers as well as copy data to FIFOs, or setup DMA links. When the interrupt transactions are used, an extra overhead is appended, from the cost of FreeRTOS queues and the time switching between tasks and the ISR.
- For interrupt transactions, the CPU can switched to other tasks when the transaction is in flight. This save the cpu time but increase the interval (See :ref:`interrupt_transactions`). For polling transactions, it does not block the task but do polling when the transaction is in flight. (See :ref:`polling_transactions`).
- When the DMA is enabled, it needs about 2us per transaction to setup the linked list. When the master is transferring, it automatically read data from the linked list. If the DMA is not enabled, CPU has to write/read each byte to/from the FIFO by itself. Usually this is faster than 2us, but the transaction length is limited to 64 bytes for both write and read.
Typical transaction interval with one byte data is as below:
Typical Transaction Time (us)
Interrupt
Polling
DMA
24
8
No DMA
22
7
SPI clock frequency: Each byte transferred takes 8 times of the clock period 8/fspi. If the clock frequency is too high, some functions may be limited to use. See :ref:`timing_considerations`.
The cache miss: the default config puts only the ISR into the IRAM. Other SPI related functions including the driver itself and the callback may suffer from the cache miss and wait for some time while reading code from the flash. Select :ref:`CONFIG_SPI_MASTER_IN_IRAM` to put the whole SPI driver into IRAM, and put the entire callback(s) and its callee functions into IRAM to prevent this.
For an interrupt transaction, the overall cost is 20+8n/Fspi[MHz] [us] for n bytes tranferred in one transaction. Hence the transferring speed is : n/(20+8n/Fspi). Example of transferring speed under 8MHz clock speed:
Frequency (MHz) |
Transaction Interval (us) |
Transaction Length (bytes) |
Total Time (us) |
Total Speed (kBps) |
---|---|---|---|---|
8 | 25 | 1 | 26 | 38.5 |
8 | 25 | 8 | 33 | 242.4 |
8 | 25 | 16 | 41 | 490.2 |
8 | 25 | 64 | 89 | 719.1 |
8 | 25 | 128 | 153 | 836.6 |
When the length of transaction is short, the cost of transaction interval is really high. Please try to squash data into one transaction if possible to get higher transfer speed.
BTW, the ISR is disabled during flash operation by default. To keep sending
transactions during flash operations, enable
:ref:`CONFIG_SPI_MASTER_ISR_IN_IRAM` and set :cpp:class:`ESP_INTR_FLAG_IRAM`
in the intr_flags
member of :cpp:class:`spi_bus_config_t`. Then all the
transactions queued before the flash operations will be handled by the ISR
continuously during flash operation. Note that the callback of each devices,
and their callee functions, should be in the IRAM in this case, or your
callback will crash due to cache miss.
As shown in the figure below, there is a delay on the MISO signal after SCLK launch edge and before it's latched by the internal register. As a result, the MISO pin setup time is the limiting factor for SPI clock speed. When the delay is too large, setup slack is < 0 and the setup timing requirement is violated, leads to the failure of reading correctly.
The maximum frequency allowed is related to the input delay (maximum valid
time after SCLK on the MISO bus), as well as the usage of GPIO matrix. The
maximum frequency allowed is reduced to about 33~77% (related to existing
input delay) when the GPIO matrix is used. To work at higher frequency, you
have to use the IOMUX pins or the dummy bit workaround. You can get the
maximum reading frequency of the master by spi_get_freq_limit
.
Dummy bit workaround: We can insert dummy clocks (during which the host does not read data) before the read phase actually begins. The slave still sees the dummy clocks and gives out data, but the host does not read until the read phase. This compensates the lack of setup time of MISO required by the host, allowing the host reading at higher frequency.
In the ideal case (the slave is so fast that the input delay is shorter than an apb clock, 12.5ns), the maximum frequency host can read (or read and write) under different conditions is as below:
Frequency Limit (MHz) | Dummy Bits Used By Driver | Comments | |
---|---|---|---|
GPIO matrix | IOMUX pins | ||
26.6 | 80 | No | |
40 | -- | Yes | Half Duplex, no DMA allowed |
And if the host only writes, the dummy bit workaround is not used and the frequency limit is as below:
GPIO matrix (MHz) | IOMUX pins (MHz) |
---|---|
40 | 80 |
The spi master driver can work even if the input delay in the spi_device_interface_config_t
is set to 0.
However, setting a accurate value helps to: (1) calculate the frequency limit in full duplex mode, and (2) compensate
the timing correctly by dummy bits in half duplex mode. You may find the maximum data valid time after the launch edge
of SPI clocks in the AC characteristics chapter of the device specifications, or measure the time on a oscilloscope or
logic analyzer.
As shown in the figure above, the input delay is usually:
[input delay] = [sample delay] + [slave output delay]
- The sample delay is the maximum random delay due to the asynchronization of SCLK and peripheral clock of the slave. It's usually 1 slave peripheral clock if the clock is asynchronize with SCLK, or 0 if the slave just use the SCLK to latch the SCLK and launch MISO data. e.g. for ESP32 slaves, the delay is 12.5ns (1 apb clock), while it is reduced to 0 if the slave is in the same chip as the master.
- The slave output delay is the time for the MOSI to be stable after the launch edge. e.g. for ESP32 slaves, the output delay is 37.5ns (3 apb clocks) when IOMUX pins in the slave is used, or 62.5ns (5 apb clocks) if through the GPIO matrix.
Some typical delays are shown in the following table:
Device | Input delay (ns) |
---|---|
Ideal device | 0 |
ESP32 slave IOMUX* | 50 |
ESP32 slave GPIO* | 75 |
ESP32 slave is on an independent chip, 12.5ns sample delay included. |
The MISO path delay(tv), consists of slave input delay and master GPIO matrix delay, finally determines the frequency limit, above which the full duplex mode will not work, or dummy bits are used in the half duplex mode. The frequency limit is:
Freq limit[MHz] = 80 / (floor(MISO delay[ns]/12.5) + 1)
The figure below shows the relations of frequency limit against the input delay. 2 extra apb clocks should be counted into the MISO delay if the GPIO matrix in the master is used.
Corresponding frequency limit for different devices with different input delay are shown in the following table:
Master | Input delay (ns) | MISO path delay (ns) | Freq. limit (MHz) |
---|---|---|---|
IOMUX (0ns) | 0 | 0 | 80 |
50 | 50 | 16 | |
75 | 75 | 11.43 | |
GPIO (25ns) | 0 | 25 | 26.67 |
50 | 75 | 11.43 | |
75 | 100 | 8.89 |
Half duplex mode is not compatible with DMA when both writing and reading phases exist.
If such transactions are required, you have to use one of the alternative solutions:
use full-duplex mode instead.
disable the DMA by setting the last parameter to 0 in bus initialization function just as below:
ret=spi_bus_initialize(VSPI_HOST, &buscfg, 0);
this may prohibit you from transmitting and receiving data longer than 64 bytes.
try to use command and address field to replace the write phase.
Full duplex mode is not compatible with the dummy bit workaround, hence the frequency is limited. See :ref:`dummy bit speed-up workaround <dummy_bit_workaround>`.
cs_ena_pretrans
is not compatible with command, address phases in full duplex mode.
Display graphics on the 320x240 LCD of WROVER-Kits: :example:`peripherals/spi_master`.