| Andrea Paterniani | 814a8d5 | 2007-05-08 00:32:15 -0700 | [diff] [blame] | 1 | SPI devices have a limited userspace API, supporting basic half-duplex | 
|  | 2 | read() and write() access to SPI slave devices.  Using ioctl() requests, | 
|  | 3 | full duplex transfers and device I/O configuration are also available. | 
|  | 4 |  | 
|  | 5 | #include <fcntl.h> | 
|  | 6 | #include <unistd.h> | 
|  | 7 | #include <sys/ioctl.h> | 
|  | 8 | #include <linux/types.h> | 
|  | 9 | #include <linux/spi/spidev.h> | 
|  | 10 |  | 
|  | 11 | Some reasons you might want to use this programming interface include: | 
|  | 12 |  | 
|  | 13 | * Prototyping in an environment that's not crash-prone; stray pointers | 
|  | 14 | in userspace won't normally bring down any Linux system. | 
|  | 15 |  | 
|  | 16 | * Developing simple protocols used to talk to microcontrollers acting | 
|  | 17 | as SPI slaves, which you may need to change quite often. | 
|  | 18 |  | 
|  | 19 | Of course there are drivers that can never be written in userspace, because | 
|  | 20 | they need to access kernel interfaces (such as IRQ handlers or other layers | 
|  | 21 | of the driver stack) that are not accessible to userspace. | 
|  | 22 |  | 
|  | 23 |  | 
|  | 24 | DEVICE CREATION, DRIVER BINDING | 
|  | 25 | =============================== | 
|  | 26 | The simplest way to arrange to use this driver is to just list it in the | 
|  | 27 | spi_board_info for a device as the driver it should use:  the "modalias" | 
|  | 28 | entry is "spidev", matching the name of the driver exposing this API. | 
|  | 29 | Set up the other device characteristics (bits per word, SPI clocking, | 
|  | 30 | chipselect polarity, etc) as usual, so you won't always need to override | 
|  | 31 | them later. | 
|  | 32 |  | 
|  | 33 | (Sysfs also supports userspace driven binding/unbinding of drivers to | 
|  | 34 | devices.  That mechanism might be supported here in the future.) | 
|  | 35 |  | 
|  | 36 | When you do that, the sysfs node for the SPI device will include a child | 
|  | 37 | device node with a "dev" attribute that will be understood by udev or mdev. | 
|  | 38 | (Larger systems will have "udev".  Smaller ones may configure "mdev" into | 
|  | 39 | busybox; it's less featureful, but often enough.)  For a SPI device with | 
|  | 40 | chipselect C on bus B, you should see: | 
|  | 41 |  | 
|  | 42 | /dev/spidevB.C ... character special device, major number 153 with | 
|  | 43 | a dynamically chosen minor device number.  This is the node | 
|  | 44 | that userspace programs will open, created by "udev" or "mdev". | 
|  | 45 |  | 
|  | 46 | /sys/devices/.../spiB.C ... as usual, the SPI device node will | 
|  | 47 | be a child of its SPI master controller. | 
|  | 48 |  | 
|  | 49 | /sys/class/spidev/spidevB.C ... created when the "spidev" driver | 
|  | 50 | binds to that device.  (Directory or symlink, based on whether | 
|  | 51 | or not you enabled the "deprecated sysfs files" Kconfig option.) | 
|  | 52 |  | 
|  | 53 | Do not try to manage the /dev character device special file nodes by hand. | 
|  | 54 | That's error prone, and you'd need to pay careful attention to system | 
|  | 55 | security issues; udev/mdev should already be configured securely. | 
|  | 56 |  | 
|  | 57 | If you unbind the "spidev" driver from that device, those two "spidev" nodes | 
|  | 58 | (in sysfs and in /dev) should automatically be removed (respectively by the | 
|  | 59 | kernel and by udev/mdev).  You can unbind by removing the "spidev" driver | 
|  | 60 | module, which will affect all devices using this driver.  You can also unbind | 
|  | 61 | by having kernel code remove the SPI device, probably by removing the driver | 
|  | 62 | for its SPI controller (so its spi_master vanishes). | 
|  | 63 |  | 
|  | 64 | Since this is a standard Linux device driver -- even though it just happens | 
|  | 65 | to expose a low level API to userspace -- it can be associated with any number | 
|  | 66 | of devices at a time.  Just provide one spi_board_info record for each such | 
|  | 67 | SPI device, and you'll get a /dev device node for each device. | 
|  | 68 |  | 
|  | 69 |  | 
|  | 70 | BASIC CHARACTER DEVICE API | 
|  | 71 | ========================== | 
|  | 72 | Normal open() and close() operations on /dev/spidevB.D files work as you | 
|  | 73 | would expect. | 
|  | 74 |  | 
|  | 75 | Standard read() and write() operations are obviously only half-duplex, and | 
|  | 76 | the chipselect is deactivated between those operations.  Full-duplex access, | 
|  | 77 | and composite operation without chipselect de-activation, is available using | 
|  | 78 | the SPI_IOC_MESSAGE(N) request. | 
|  | 79 |  | 
|  | 80 | Several ioctl() requests let your driver read or override the device's current | 
|  | 81 | settings for data transfer parameters: | 
|  | 82 |  | 
|  | 83 | SPI_IOC_RD_MODE, SPI_IOC_WR_MODE ... pass a pointer to a byte which will | 
|  | 84 | return (RD) or assign (WR) the SPI transfer mode.  Use the constants | 
|  | 85 | SPI_MODE_0..SPI_MODE_3; or if you prefer you can combine SPI_CPOL | 
|  | 86 | (clock polarity, idle high iff this is set) or SPI_CPHA (clock phase, | 
|  | 87 | sample on trailing edge iff this is set) flags. | 
|  | 88 |  | 
|  | 89 | SPI_IOC_RD_LSB_FIRST, SPI_IOC_WR_LSB_FIRST ... pass a pointer to a byte | 
|  | 90 | which will return (RD) or assign (WR) the bit justification used to | 
|  | 91 | transfer SPI words.  Zero indicates MSB-first; other values indicate | 
|  | 92 | the less common LSB-first encoding.  In both cases the specified value | 
|  | 93 | is right-justified in each word, so that unused (TX) or undefined (RX) | 
|  | 94 | bits are in the MSBs. | 
|  | 95 |  | 
|  | 96 | SPI_IOC_RD_BITS_PER_WORD, SPI_IOC_WR_BITS_PER_WORD ... pass a pointer to | 
|  | 97 | a byte which will return (RD) or assign (WR) the number of bits in | 
|  | 98 | each SPI transfer word.  The value zero signifies eight bits. | 
|  | 99 |  | 
|  | 100 | SPI_IOC_RD_MAX_SPEED_HZ, SPI_IOC_WR_MAX_SPEED_HZ ... pass a pointer to a | 
|  | 101 | u32 which will return (RD) or assign (WR) the maximum SPI transfer | 
|  | 102 | speed, in Hz.  The controller can't necessarily assign that specific | 
|  | 103 | clock speed. | 
|  | 104 |  | 
|  | 105 | NOTES: | 
|  | 106 |  | 
|  | 107 | - At this time there is no async I/O support; everything is purely | 
|  | 108 | synchronous. | 
|  | 109 |  | 
|  | 110 | - There's currently no way to report the actual bit rate used to | 
|  | 111 | shift data to/from a given device. | 
|  | 112 |  | 
|  | 113 | - From userspace, you can't currently change the chip select polarity; | 
|  | 114 | that could corrupt transfers to other devices sharing the SPI bus. | 
|  | 115 | Each SPI device is deselected when it's not in active use, allowing | 
|  | 116 | other drivers to talk to other devices. | 
|  | 117 |  | 
|  | 118 | - There's a limit on the number of bytes each I/O request can transfer | 
|  | 119 | to the SPI device.  It defaults to one page, but that can be changed | 
|  | 120 | using a module parameter. | 
|  | 121 |  | 
|  | 122 | - Because SPI has no low-level transfer acknowledgement, you usually | 
|  | 123 | won't see any I/O errors when talking to a non-existent device. | 
|  | 124 |  | 
|  | 125 |  | 
|  | 126 | FULL DUPLEX CHARACTER DEVICE API | 
|  | 127 | ================================ | 
|  | 128 |  | 
| Randy Dunlap | 31a1629 | 2008-04-28 02:14:18 -0700 | [diff] [blame] | 129 | See the spidev_fdx.c sample program for one example showing the use of the | 
|  | 130 | full duplex programming interface.  (Although it doesn't perform a full duplex | 
| Andrea Paterniani | 814a8d5 | 2007-05-08 00:32:15 -0700 | [diff] [blame] | 131 | transfer.)  The model is the same as that used in the kernel spi_sync() | 
|  | 132 | request; the individual transfers offer the same capabilities as are | 
|  | 133 | available to kernel drivers (except that it's not asynchronous). | 
|  | 134 |  | 
|  | 135 | The example shows one half-duplex RPC-style request and response message. | 
|  | 136 | These requests commonly require that the chip not be deselected between | 
|  | 137 | the request and response.  Several such requests could be chained into | 
|  | 138 | a single kernel request, even allowing the chip to be deselected after | 
|  | 139 | each response.  (Other protocol options include changing the word size | 
|  | 140 | and bitrate for each transfer segment.) | 
|  | 141 |  | 
|  | 142 | To make a full duplex request, provide both rx_buf and tx_buf for the | 
|  | 143 | same transfer.  It's even OK if those are the same buffer. |