1 ===================================== 2 Linux I2C slave interface description 3 ===================================== 4 5 by Wolfram Sang <wsa@sang-engineering.com> in 2014-15 6 7 Linux can also be an I2C slave if the I2C controller in use has slave 8 functionality. For that to work, one needs slave support in the bus driver plus 9 a hardware independent software backend providing the actual functionality. An 10 example for the latter is the slave-eeprom driver, which acts as a dual memory 11 driver. While another I2C master on the bus can access it like a regular 12 EEPROM, the Linux I2C slave can access the content via sysfs and handle data as 13 needed. The backend driver and the I2C bus driver communicate via events. Here 14 is a small graph visualizing the data flow and the means by which data is 15 transported. The dotted line marks only one example. The backend could also 16 use a character device, be in-kernel only, or something completely different:: 17 18 19 e.g. sysfs I2C slave events I/O registers 20 +-----------+ v +---------+ v +--------+ v +------------+ 21 | Userspace +........+ Backend +-----------+ Driver +-----+ Controller | 22 +-----------+ +---------+ +--------+ +------------+ 23 | | 24 ----------------------------------------------------------------+-- I2C 25 --------------------------------------------------------------+---- Bus 26 27 Note: Technically, there is also the I2C core between the backend and the 28 driver. However, at this time of writing, the layer is transparent. 29 30 31 User manual 32 =========== 33 34 I2C slave backends behave like standard I2C clients. So, you can instantiate 35 them as described in the document instantiating-devices.rst. The only 36 difference is that i2c slave backends have their own address space. So, you 37 have to add 0x1000 to the address you would originally request. An example for 38 instantiating the slave-eeprom driver from userspace at the 7 bit address 0x64 39 on bus 1:: 40 41 # echo slave-24c02 0x1064 > /sys/bus/i2c/devices/i2c-1/new_device 42 43 Each backend should come with separate documentation to describe its specific 44 behaviour and setup. 45 46 47 Developer manual 48 ================ 49 50 First, the events which are used by the bus driver and the backend will be 51 described in detail. After that, some implementation hints for extending bus 52 drivers and writing backends will be given. 53 54 55 I2C slave events 56 ---------------- 57 58 The bus driver sends an event to the backend using the following function:: 59 60 ret = i2c_slave_event(client, event, &val) 61 62 'client' describes the I2C slave device. 'event' is one of the special event 63 types described hereafter. 'val' holds an u8 value for the data byte to be 64 read/written and is thus bidirectional. The pointer to val must always be 65 provided even if val is not used for an event, i.e. don't use NULL here. 'ret' 66 is the return value from the backend. Mandatory events must be provided by the 67 bus drivers and must be checked for by backend drivers. 68 69 Event types: 70 71 * I2C_SLAVE_WRITE_REQUESTED (mandatory) 72 73 'val': unused 74 75 'ret': 0 if the backend is ready, otherwise some errno 76 77 Another I2C master wants to write data to us. This event should be sent once 78 our own address and the write bit was detected. The data did not arrive yet, so 79 there is nothing to process or return. After returning, the bus driver must 80 always ack the address phase. If 'ret' is zero, backend initialization or 81 wakeup is done and further data may be received. If 'ret' is an errno, the bus 82 driver should nack all incoming bytes until the next stop condition to enforce 83 a retry of the transmission. 84 85 * I2C_SLAVE_READ_REQUESTED (mandatory) 86 87 'val': backend returns first byte to be sent 88 89 'ret': always 0 90 91 Another I2C master wants to read data from us. This event should be sent once 92 our own address and the read bit was detected. After returning, the bus driver 93 should transmit the first byte. 94 95 * I2C_SLAVE_WRITE_RECEIVED (mandatory) 96 97 'val': bus driver delivers received byte 98 99 'ret': 0 if the byte should be acked, some errno if the byte should be nacked 100 101 Another I2C master has sent a byte to us which needs to be set in 'val'. If 'ret' 102 is zero, the bus driver should ack this byte. If 'ret' is an errno, then the byte 103 should be nacked. 104 105 * I2C_SLAVE_READ_PROCESSED (mandatory) 106 107 'val': backend returns next byte to be sent 108 109 'ret': always 0 110 111 The bus driver requests the next byte to be sent to another I2C master in 112 'val'. Important: This does not mean that the previous byte has been acked, it 113 only means that the previous byte is shifted out to the bus! To ensure seamless 114 transmission, most hardware requests the next byte when the previous one is 115 still shifted out. If the master sends NACK and stops reading after the byte 116 currently shifted out, this byte requested here is never used. It very likely 117 needs to be sent again on the next I2C_SLAVE_READ_REQUEST, depending a bit on 118 your backend, though. 119 120 * I2C_SLAVE_STOP (mandatory) 121 122 'val': unused 123 124 'ret': always 0 125 126 A stop condition was received. This can happen anytime and the backend should 127 reset its state machine for I2C transfers to be able to receive new requests. 128 129 130 Software backends 131 ----------------- 132 133 If you want to write a software backend: 134 135 * use a standard i2c_driver and its matching mechanisms 136 * write the slave_callback which handles the above slave events 137 (best using a state machine) 138 * register this callback via i2c_slave_register() 139 140 Check the i2c-slave-eeprom driver as an example. 141 142 143 Bus driver support 144 ------------------ 145 146 If you want to add slave support to the bus driver: 147 148 * implement calls to register/unregister the slave and add those to the 149 struct i2c_algorithm. When registering, you probably need to set the I2C 150 slave address and enable slave specific interrupts. If you use runtime pm, you 151 should use pm_runtime_get_sync() because your device usually needs to be 152 powered on always to be able to detect its slave address. When unregistering, 153 do the inverse of the above. 154 155 * Catch the slave interrupts and send appropriate i2c_slave_events to the backend. 156 157 Note that most hardware supports being master _and_ slave on the same bus. So, 158 if you extend a bus driver, please make sure that the driver supports that as 159 well. In almost all cases, slave support does not need to disable the master 160 functionality. 161 162 Check the i2c-rcar driver as an example. 163 164 165 About ACK/NACK 166 -------------- 167 168 It is good behaviour to always ACK the address phase, so the master knows if a 169 device is basically present or if it mysteriously disappeared. Using NACK to 170 state being busy is troublesome. SMBus demands to always ACK the address phase, 171 while the I2C specification is more loose on that. Most I2C controllers also 172 automatically ACK when detecting their slave addresses, so there is no option 173 to NACK them. For those reasons, this API does not support NACK in the address 174 phase. 175 176 Currently, there is no slave event to report if the master did ACK or NACK a 177 byte when it reads from us. We could make this an optional event if the need 178 arises. However, cases should be extremely rare because the master is expected 179 to send STOP after that and we have an event for that. Also, keep in mind not 180 all I2C controllers have the possibility to report that event. 181 182 183 About buffers 184 ------------- 185 186 During development of this API, the question of using buffers instead of just 187 bytes came up. Such an extension might be possible, usefulness is unclear at 188 this time of writing. Some points to keep in mind when using buffers: 189 190 * Buffers should be opt-in and backend drivers will always have to support 191 byte-based transactions as the ultimate fallback anyhow because this is how 192 the majority of HW works. 193 194 * For backends simulating hardware registers, buffers are largely not helpful 195 because after each byte written an action should be immediately triggered. 196 For reads, the data kept in the buffer might get stale if the backend just 197 updated a register because of internal processing. 198 199 * A master can send STOP at any time. For partially transferred buffers, this 200 means additional code to handle this exception. Such code tends to be 201 error-prone.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.