What is SPI?¶

The “Serial Peripheral Interface” (SPI) is a synchronous four wire seriallink used to connect microcontrollers to sensors, memory, and peripherals.It’s a simple “de facto” standard, not complicated enough to acquire astandardization body. SPI uses a master/slave configuration.

The three signal wires hold a clock (SCK, often on the order of 10 MHz),and parallel data lines with “Master Out, Slave In” (MOSI) or “Master In,Slave Out” (MISO) signals. (Other names are also used.) There are fourclocking modes through which data is exchanged; mode-0 and mode-3 are mostcommonly used. Each clock cycle shifts data out and data in; the clockdoesn’t cycle except when there is a data bit to shift. Not all data bitsare used though; not every protocol uses those full duplex capabilities.

SPI masters use a fourth “chip select” line to activate a given SPI slavedevice, so those three signal wires may be connected to several chipsin parallel. All SPI slaves support chipselects; they are usually activelow signals, labeled nCSx for slave ‘x’ (e.g. nCS0). Some devices haveother signals, often including an interrupt to the master.

Unlike serial busses like USB or SMBus, even low level protocols forSPI slave functions are usually not interoperable between vendors(except for commodities like SPI memory chips).

• SPI may be used for request/response style device protocols, as withtouchscreen sensors and memory chips.
• It may also be used to stream data in either direction (half duplex),or both of them at the same time (full duplex).
• Some devices may use eight bit words. Others may use different wordlengths, such as streams of 12-bit or 20-bit digital samples.
• Words are usually sent with their most significant bit (MSB) first,but sometimes the least significant bit (LSB) goes first instead.
• Sometimes SPI is used to daisy-chain devices, like shift registers.

In the same way, SPI slaves will only rarely support any kind of automaticdiscovery/enumeration protocol. The tree of slave devices accessible froma given SPI master will normally be set up manually, with configurationtables.

SPI is only one of the names used by such four-wire protocols, andmost controllers have no problem handling “MicroWire” (think of it ashalf-duplex SPI, for request/response protocols), SSP (“SynchronousSerial Protocol”), PSP (“Programmable Serial Protocol”), and otherrelated protocols.

Some chips eliminate a signal line by combining MOSI and MISO, andlimiting themselves to half-duplex at the hardware level. In factsome SPI chips have this signal mode as a strapping option. Thesecan be accessed using the same programming interface as SPI, but ofcourse they won’t handle full duplex transfers. You may find suchchips described as using “three wire” signaling: SCK, data, nCSx.(That data line is sometimes called MOMI or SISO.)

Microcontrollers often support both master and slave sides of the SPIprotocol. This document (and Linux) supports both the master and slavesides of SPI interactions.

Who uses it? On what kinds of systems?¶

Linux developers using SPI are probably writing device drivers for embeddedsystems boards. SPI is used to control external chips, and it is also aprotocol supported by every MMC or SD memory card. (The older “DataFlash”cards, predating MMC cards but using the same connectors and card shape,support only SPI.) Some PC hardware uses SPI flash for BIOS code.

SPI slave chips range from digital/analog converters used for analogsensors and codecs, to memory, to peripherals like USB controllersor Ethernet adapters; and more.

Most systems using SPI will integrate a few devices on a mainboard.Some provide SPI links on expansion connectors; in cases where nodedicated SPI controller exists, GPIO pins can be used to create alow speed “bitbanging” adapter. Very few systems will “hotplug” an SPIcontroller; the reasons to use SPI focus on low cost and simple operation,and if dynamic reconfiguration is important, USB will often be a moreappropriate low-pincount peripheral bus.

Many microcontrollers that can run Linux integrate one or more I/Ointerfaces with SPI modes. Given SPI support, they could use MMC or SDcards without needing a special purpose MMC/SD/SDIO controller.

I’m confused. What are these four SPI “clock modes”?¶

It’s easy to be confused here, and the vendor documentation you’llfind isn’t necessarily helpful. The four modes combine two mode bits:

• CPOL indicates the initial clock polarity. CPOL=0 means theclock starts low, so the first (leading) edge is rising, andthe second (trailing) edge is falling. CPOL=1 means the clockstarts high, so the first (leading) edge is falling.

• CPHA indicates the clock phase used to sample data; CPHA=0 sayssample on the leading edge, CPHA=1 means the trailing edge.

  Since the signal needs to stablize before it’s sampled, CPHA=0implies that its data is written half a clock before the firstclock edge. The chipselect may have made it become available.

Chip specs won’t always say “uses SPI mode X” in as many words,but their timing diagrams will make the CPOL and CPHA modes clear.

In the SPI mode number, CPOL is the high order bit and CPHA is thelow order bit. So when a chip’s timing diagram shows the clockstarting low (CPOL=0) and data stabilized for sampling during thetrailing clock edge (CPHA=1), that’s SPI mode 1.

Note that the clock mode is relevant as soon as the chipselect goesactive. So the master must set the clock to inactive before selectinga slave, and the slave can tell the chosen polarity by sampling theclock level when its select line goes active. That’s why many devicessupport for example both modes 0 and 3: they don’t care about polarity,and always clock data in/out on rising clock edges.

How do these driver programming interfaces work?¶

The <linux/spi/spi.h> header file includes kerneldoc, as does themain source code, and you should certainly read that chapter of thekernel API document. This is just an overview, so you get the bigpicture before those details.

SPI requests always go into I/O queues. Requests for a given SPI deviceare always executed in FIFO order, and complete asynchronously throughcompletion callbacks. There are also some simple synchronous wrappersfor those calls, including ones for common transaction types like writinga command and then reading its response.

There are two types of SPI driver, here called:

Controller drivers …

controllers may be built into System-On-Chipprocessors, and often support both Master and Slave roles.These drivers touch hardware registers and may use DMA.Or they can be PIO bitbangers, needing just GPIO pins.

Protocol drivers …

these pass messages through the controllerdriver to communicate with a Slave or Master device on theother side of an SPI link.

So for example one protocol driver might talk to the MTD layer to exportdata to filesystems stored on SPI flash like DataFlash; and others mightcontrol audio interfaces, present touchscreen sensors as input interfaces,or monitor temperature and voltage levels during industrial processing.And those might all be sharing the same controller driver.

A “struct spi_device” encapsulates the controller-side interface betweenthose two types of drivers.

There is a minimal core of SPI programming interfaces, focussing onusing the driver model to connect controller and protocol drivers usingdevice tables provided by board specific initialization code. SPIshows up in sysfs in several locations:

/sys/devices/.../CTLR ... physical node for a given SPI controller/sys/devices/.../CTLR/spiB.C ... spi_device on bus "B",     chipselect C, accessed through CTLR./sys/bus/spi/devices/spiB.C ... symlink to that physical     .../CTLR/spiB.C device/sys/devices/.../CTLR/spiB.C/modalias ... identifies the driver     that should be used with this device (for hotplug/coldplug)/sys/bus/spi/drivers/D ... driver for one or more spi*.* devices/sys/class/spi_master/spiB ... symlink (or actual device node) to     a logical node which could hold class related state for the SPI     master controller managing bus "B".  All spiB.* devices share one     physical SPI bus segment, with SCLK, MOSI, and MISO./sys/devices/.../CTLR/slave ... virtual file for (un)registering the     slave device for an SPI slave controller.     Writing the driver name of an SPI slave handler to this file     registers the slave device; writing "(null)" unregisters the slave     device.     Reading from this file shows the name of the slave device ("(null)"     if not registered)./sys/class/spi_slave/spiB ... symlink (or actual device node) to     a logical node which could hold class related state for the SPI     slave controller on bus "B".  When registered, a single spiB.*     device is present here, possible sharing the physical SPI bus     segment with other SPI slave devices.

Note that the actual location of the controller’s class state dependson whether you enabled CONFIG_SYSFS_DEPRECATED or not. At this time,the only class-specific state is the bus number (“B” in “spiB”), sothose /sys/class entries are only useful to quickly identify busses.

How does board-specific init code declare SPI devices?¶

Linux needs several kinds of information to properly configure SPI devices.That information is normally provided by board-specific code, even forchips that do support some of automated discovery/enumeration.

#include <mach/spi.h>   /* for mysoc_spi_data *//* if your mach-* infrastructure doesn't support kernels that can * run on multiple boards, pdata wouldn't benefit from "__init". */static struct mysoc_spi_data pdata __initdata = { ... };static __init board_init(void){        ...        /* this board only uses SPI controller #2 */        mysoc_register_spi(2, &pdata);        ...}

#include <mach/spi.h>static struct platform_device spi2 = { ... };void mysoc_register_spi(unsigned n, struct mysoc_spi_data *pdata){        struct mysoc_spi_data *pdata2;        pdata2 = kmalloc(sizeof *pdata2, GFP_KERNEL);        *pdata2 = pdata;        ...        if (n == 2) {                spi2->dev.platform_data = pdata2;                register_platform_device(&spi2);                /* also: set up pin modes so the spi2 signals are                 * visible on the relevant pins ... bootloaders on                 * production boards may already have done this, but                 * developer boards will often need Linux to do it.                 */        }        ...}

static struct ads7846_platform_data ads_info = {        .vref_delay_usecs       = 100,        .x_plate_ohms           = 580,        .y_plate_ohms           = 410,};static struct spi_board_info spi_board_info[] __initdata = {{        .modalias       = "ads7846",        .platform_data  = &ads_info,        .mode           = SPI_MODE_0,        .irq            = GPIO_IRQ(31),        .max_speed_hz   = 120000 /* max sample rate at 3V */ * 16,        .bus_num        = 1,        .chip_select    = 0,},};

spi_register_board_info(spi_board_info, ARRAY_SIZE(spi_board_info));

How do I write an “SPI Protocol Driver”?¶

Most SPI drivers are currently kernel drivers, but there’s also supportfor userspace drivers. Here we talk only about kernel drivers.

SPI protocol drivers somewhat resemble platform device drivers:

static struct spi_driver CHIP_driver = {        .driver = {                .name           = "CHIP",                .owner          = THIS_MODULE,                .pm             = &CHIP_pm_ops,        },        .probe          = CHIP_probe,        .remove         = CHIP_remove,};

The driver core will automatically attempt to bind this driver to any SPIdevice whose board_info gave a modalias of “CHIP”. Your probe() codemight look like this unless you’re creating a device which is managinga bus (appearing under /sys/class/spi_master).

static int CHIP_probe(struct spi_device *spi){        struct CHIP                     *chip;        struct CHIP_platform_data       *pdata;        /* assuming the driver requires board-specific data: */        pdata = &spi->dev.platform_data;        if (!pdata)                return -ENODEV;        /* get memory for driver's per-chip state */        chip = kzalloc(sizeof *chip, GFP_KERNEL);        if (!chip)                return -ENOMEM;        spi_set_drvdata(spi, chip);        ... etc        return 0;}

As soon as it enters probe(), the driver may issue I/O requests tothe SPI device using “struct spi_message”. When remove() returns,or after probe() fails, the driver guarantees that it won’t submitany more such messages.

• An spi_message is a sequence of protocol operations, executedas one atomic sequence. SPI driver controls include:

  • when bidirectional reads and writes start … by how itssequence of spi_transfer requests is arranged;
  • which I/O buffers are used … each spi_transfer wraps abuffer for each transfer direction, supporting full duplex(two pointers, maybe the same one in both cases) and halfduplex (one pointer is NULL) transfers;
  • optionally defining short delays after transfers … usingthe spi_transfer.delay_usecs setting (this delay can be theonly protocol effect, if the buffer length is zero);
  • whether the chipselect becomes inactive after a transfer andany delay … by using the spi_transfer.cs_change flag;
  • hinting whether the next message is likely to go to this samedevice … using the spi_transfer.cs_change flag on the lasttransfer in that atomic group, and potentially saving costsfor chip deselect and select operations.

• Follow standard kernel rules, and provide DMA-safe buffers inyour messages. That way controller drivers using DMA aren’t forcedto make extra copies unless the hardware requires it (e.g. workingaround hardware errata that force the use of bounce buffering).

  If standard dma_map_single() handling of these buffers is inappropriate,you can use spi_message.is_dma_mapped to tell the controller driverthat you’ve already provided the relevant DMA addresses.

• The basic I/O primitive isspi_async(). Async requests may beissued in any context (irq handler, task, etc) and completionis reported using a callback provided with the message.After any detected error, the chip is deselected and processingof that spi_message is aborted.

• There are also synchronous wrappers likespi_sync(), and wrapperslikespi_read(),spi_write(), andspi_write_then_read(). Thesemay be issued only in contexts that may sleep, and they’re allclean (and small, and “optional”) layers overspi_async().

• Thespi_write_then_read() call, and convenience wrappers aroundit, should only be used with small amounts of data where thecost of an extra copy may be ignored. It’s designed to supportcommon RPC-style requests, such as writing an eight bit commandand reading a sixteen bit response –spi_w8r16() being one itswrappers, doing exactly that.

Some drivers may need to modify spi_device characteristics like thetransfer mode, wordsize, or clock rate. This is done withspi_setup(),which would normally be called from probe() before the first I/O isdone to the device. However, that can also be called at any timethat no message is pending for that device.

While “spi_device” would be the bottom boundary of the driver, theupper boundaries might include sysfs (especially for sensor readings),the input layer, ALSA, networking, MTD, the character device framework,or other Linux subsystems.

Note that there are two types of memory your driver must manage as partof interacting with SPI devices.

• I/O buffers use the usual Linux rules, and must be DMA-safe.You’d normally allocate them from the heap or free page pool.Don’t use the stack, or anything that’s declared “static”.
• The spi_message and spi_transfer metadata used to glue thoseI/O buffers into a group of protocol transactions. These canbe allocated anywhere it’s convenient, including as part ofother allocate-once driver data structures. Zero-init these.

If you like, spi_message_alloc() and spi_message_free() convenienceroutines are available to allocate and zero-initialize an spi_messagewith several transfers.

How do I write an “SPI Master Controller Driver”?¶

An SPI controller will probably be registered on the platform_bus; writea driver to bind to the device, whichever bus is involved.

The main task of this type of driver is to provide an “spi_master”.Use spi_alloc_master() to allocate the master, and spi_master_get_devdata()to get the driver-private data allocated for that device.

struct spi_master       *master;struct CONTROLLER       *c;master = spi_alloc_master(dev, sizeof *c);if (!master)        return -ENODEV;c = spi_master_get_devdata(master);

The driver will initialize the fields of that spi_master, including thebus number (maybe the same as the platform device ID) and three methodsused to interact with the SPI core and SPI protocol drivers. It willalso initialize its own internal state. (See below about bus numberingand those methods.)

After you initialize the spi_master, then use spi_register_master() topublish it to the rest of the system. At that time, device nodes for thecontroller and any predeclared spi devices will be made available, andthe driver model core will take care of binding them to drivers.

If you need to remove your SPI controller driver, spi_unregister_master()will reverse the effect of spi_register_master().

THANKS TO¶

Contributors to Linux-SPI discussions include (in alphabetical order,by last name):

• Mark Brown
• David Brownell
• Russell King
• Grant Likely
• Dmitry Pervushin
• Stephen Street
• Mark Underwood
• Andrew Victor
• Linus Walleij
• Vitaly Wool