Wolfram Sang | 7c60375 | 2015-03-23 09:26:37 +0100 | [diff] [blame] | 1 | Linux I2C slave interface description |
| 2 | ===================================== |
| 3 | |
| 4 | by Wolfram Sang <wsa@sang-engineering.com> in 2014-15 |
| 5 | |
Wolfram Sang | 976cf20 | 2015-05-14 14:40:04 +0200 | [diff] [blame] | 6 | Linux can also be an I2C slave if the I2C controller in use has slave |
| 7 | functionality. For that to work, one needs slave support in the bus driver plus |
| 8 | a hardware independent software backend providing the actual functionality. An |
| 9 | example for the latter is the slave-eeprom driver, which acts as a dual memory |
| 10 | driver. While another I2C master on the bus can access it like a regular |
| 11 | EEPROM, the Linux I2C slave can access the content via sysfs and handle data as |
| 12 | needed. The backend driver and the I2C bus driver communicate via events. Here |
| 13 | is a small graph visualizing the data flow and the means by which data is |
| 14 | transported. The dotted line marks only one example. The backend could also |
| 15 | use a character device, be in-kernel only, or something completely different: |
Wolfram Sang | 7c60375 | 2015-03-23 09:26:37 +0100 | [diff] [blame] | 16 | |
| 17 | |
| 18 | e.g. sysfs I2C slave events I/O registers |
| 19 | +-----------+ v +---------+ v +--------+ v +------------+ |
| 20 | | Userspace +........+ Backend +-----------+ Driver +-----+ Controller | |
| 21 | +-----------+ +---------+ +--------+ +------------+ |
| 22 | | | |
| 23 | ----------------------------------------------------------------+-- I2C |
| 24 | --------------------------------------------------------------+---- Bus |
| 25 | |
| 26 | Note: Technically, there is also the I2C core between the backend and the |
| 27 | driver. However, at this time of writing, the layer is transparent. |
| 28 | |
| 29 | |
| 30 | User manual |
| 31 | =========== |
| 32 | |
| 33 | I2C slave backends behave like standard I2C clients. So, you can instantiate |
Wolfram Sang | 8366610 | 2015-06-15 19:51:46 +0200 | [diff] [blame] | 34 | them as described in the document 'instantiating-devices'. A quick example for |
| 35 | instantiating the slave-eeprom driver from userspace at address 0x64 on bus 1: |
Wolfram Sang | 7c60375 | 2015-03-23 09:26:37 +0100 | [diff] [blame] | 36 | |
Wolfram Sang | 8366610 | 2015-06-15 19:51:46 +0200 | [diff] [blame] | 37 | # echo slave-24c02 0x64 > /sys/bus/i2c/devices/i2c-1/new_device |
Wolfram Sang | 7c60375 | 2015-03-23 09:26:37 +0100 | [diff] [blame] | 38 | |
| 39 | Each backend should come with separate documentation to describe its specific |
| 40 | behaviour and setup. |
| 41 | |
| 42 | |
| 43 | Developer manual |
| 44 | ================ |
| 45 | |
Wolfram Sang | 976cf20 | 2015-05-14 14:40:04 +0200 | [diff] [blame] | 46 | First, the events which are used by the bus driver and the backend will be |
| 47 | described in detail. After that, some implementation hints for extending bus |
| 48 | drivers and writing backends will be given. |
| 49 | |
| 50 | |
Wolfram Sang | 7c60375 | 2015-03-23 09:26:37 +0100 | [diff] [blame] | 51 | I2C slave events |
| 52 | ---------------- |
| 53 | |
| 54 | The bus driver sends an event to the backend using the following function: |
| 55 | |
| 56 | ret = i2c_slave_event(client, event, &val) |
| 57 | |
| 58 | 'client' describes the i2c slave device. 'event' is one of the special event |
| 59 | types described hereafter. 'val' holds an u8 value for the data byte to be |
| 60 | read/written and is thus bidirectional. The pointer to val must always be |
| 61 | provided even if val is not used for an event, i.e. don't use NULL here. 'ret' |
| 62 | is the return value from the backend. Mandatory events must be provided by the |
| 63 | bus drivers and must be checked for by backend drivers. |
| 64 | |
| 65 | Event types: |
| 66 | |
| 67 | * I2C_SLAVE_WRITE_REQUESTED (mandatory) |
| 68 | |
| 69 | 'val': unused |
| 70 | 'ret': always 0 |
| 71 | |
| 72 | Another I2C master wants to write data to us. This event should be sent once |
| 73 | our own address and the write bit was detected. The data did not arrive yet, so |
| 74 | there is nothing to process or return. Wakeup or initialization probably needs |
| 75 | to be done, though. |
| 76 | |
| 77 | * I2C_SLAVE_READ_REQUESTED (mandatory) |
| 78 | |
| 79 | 'val': backend returns first byte to be sent |
| 80 | 'ret': always 0 |
| 81 | |
| 82 | Another I2C master wants to read data from us. This event should be sent once |
| 83 | our own address and the read bit was detected. After returning, the bus driver |
| 84 | should transmit the first byte. |
| 85 | |
| 86 | * I2C_SLAVE_WRITE_RECEIVED (mandatory) |
| 87 | |
| 88 | 'val': bus driver delivers received byte |
| 89 | 'ret': 0 if the byte should be acked, some errno if the byte should be nacked |
| 90 | |
| 91 | Another I2C master has sent a byte to us which needs to be set in 'val'. If 'ret' |
| 92 | is zero, the bus driver should ack this byte. If 'ret' is an errno, then the byte |
| 93 | should be nacked. |
| 94 | |
| 95 | * I2C_SLAVE_READ_PROCESSED (mandatory) |
| 96 | |
| 97 | 'val': backend returns next byte to be sent |
| 98 | 'ret': always 0 |
| 99 | |
| 100 | The bus driver requests the next byte to be sent to another I2C master in |
| 101 | 'val'. Important: This does not mean that the previous byte has been acked, it |
| 102 | only means that the previous byte is shifted out to the bus! To ensure seamless |
| 103 | transmission, most hardware requests the next byte when the previous one is |
| 104 | still shifted out. If the master sends NACK and stops reading after the byte |
| 105 | currently shifted out, this byte requested here is never used. It very likely |
| 106 | needs to be sent again on the next I2C_SLAVE_READ_REQUEST, depending a bit on |
| 107 | your backend, though. |
| 108 | |
| 109 | * I2C_SLAVE_STOP (mandatory) |
| 110 | |
| 111 | 'val': unused |
| 112 | 'ret': always 0 |
| 113 | |
| 114 | A stop condition was received. This can happen anytime and the backend should |
| 115 | reset its state machine for I2C transfers to be able to receive new requests. |
| 116 | |
| 117 | |
| 118 | Software backends |
| 119 | ----------------- |
| 120 | |
| 121 | If you want to write a software backend: |
| 122 | |
| 123 | * use a standard i2c_driver and its matching mechanisms |
| 124 | * write the slave_callback which handles the above slave events |
| 125 | (best using a state machine) |
| 126 | * register this callback via i2c_slave_register() |
| 127 | |
| 128 | Check the i2c-slave-eeprom driver as an example. |
| 129 | |
| 130 | |
| 131 | Bus driver support |
| 132 | ------------------ |
| 133 | |
| 134 | If you want to add slave support to the bus driver: |
| 135 | |
| 136 | * implement calls to register/unregister the slave and add those to the |
| 137 | struct i2c_algorithm. When registering, you probably need to set the i2c |
| 138 | slave address and enable slave specific interrupts. If you use runtime pm, you |
| 139 | should use pm_runtime_forbid() because your device usually needs to be powered |
| 140 | on always to be able to detect its slave address. When unregistering, do the |
| 141 | inverse of the above. |
| 142 | |
| 143 | * Catch the slave interrupts and send appropriate i2c_slave_events to the backend. |
| 144 | |
| 145 | Check the i2c-rcar driver as an example. |
| 146 | |
| 147 | |
| 148 | About ACK/NACK |
| 149 | -------------- |
| 150 | |
| 151 | It is good behaviour to always ACK the address phase, so the master knows if a |
| 152 | device is basically present or if it mysteriously disappeared. Using NACK to |
| 153 | state being busy is troublesome. SMBus demands to always ACK the address phase, |
| 154 | while the I2C specification is more loose on that. Most I2C controllers also |
| 155 | automatically ACK when detecting their slave addresses, so there is no option |
| 156 | to NACK them. For those reasons, this API does not support NACK in the address |
| 157 | phase. |
| 158 | |
| 159 | Currently, there is no slave event to report if the master did ACK or NACK a |
| 160 | byte when it reads from us. We could make this an optional event if the need |
| 161 | arises. However, cases should be extremely rare because the master is expected |
| 162 | to send STOP after that and we have an event for that. Also, keep in mind not |
| 163 | all I2C controllers have the possibility to report that event. |
| 164 | |
| 165 | |
| 166 | About buffers |
| 167 | ------------- |
| 168 | |
| 169 | During development of this API, the question of using buffers instead of just |
| 170 | bytes came up. Such an extension might be possible, usefulness is unclear at |
| 171 | this time of writing. Some points to keep in mind when using buffers: |
| 172 | |
| 173 | * Buffers should be opt-in and slave drivers will always have to support |
| 174 | byte-based transactions as the ultimate fallback because this is how the |
| 175 | majority of HW works. |
| 176 | |
| 177 | * For backends simulating hardware registers, buffers are not helpful because |
| 178 | on writes an action should be immediately triggered. For reads, the data in |
| 179 | the buffer might get stale. |
| 180 | |
| 181 | * A master can send STOP at any time. For partially transferred buffers, this |
| 182 | means additional code to handle this exception. Such code tends to be |
| 183 | error-prone. |
| 184 | |