Синтаксис команды i2c md

Если вы запускаете i2c в u-boot, это дает вам возможность использовать команду:

адрес микросхемы i2c md [.0, .1, .2] [количество объектов] — чтение с устройства I2C

Подробное объяснение дано в cmd_i2c.c в исходниках u-boot, папка u-boot/common:

"/*
 * I2C Functions similar to the standard memory functions.
 *
 * There are several parameters in many of the commands that bear further
 * explanations:
 *
 * {i2c_chip} is the I2C chip address (the first byte sent on the bus).
 *   Each I2C chip on the bus has a unique address.  On the I2C data bus,
 *   the address is the upper seven bits and the LSB is the "read/write"
 *   bit.  Note that the {i2c_chip} address specified on the command
 *   line is not shifted up: e.g. a typical EEPROM memory chip may have
 *   an I2C address of 0x50, but the data put on the bus will be 0xA0
 *   for write and 0xA1 for read.  This "non shifted" address notation
 *   matches at least half of the data sheets :-/.
 *
 * {addr} is the address (or offset) within the chip.  Small memory
 *   chips have 8 bit addresses.  Large memory chips have 16 bit
 *   addresses.  Other memory chips have 9, 10, or 11 bit addresses.
 *   Many non-memory chips have multiple registers and {addr} is used
 *   as the register index.  Some non-memory chips have only one register
 *   and therefore don't need any {addr} parameter.
 *
 *   The default {addr} parameter is one byte (.1) which works well for
 *   memories and registers with 8 bits of address space.
 *
 *   You can specify the length of the {addr} field with the optional .0,
 *   .1, or .2 modifier (similar to the .b, .w, .l modifier).  If you are
 *   manipulating a single register device which doesn't use an address
 *   field, use "0.0" for the address and the ".0" length field will
 *   suppress the address in the I2C data stream.  This also works for
 *   successive reads using the I2C auto-incrementing memory pointer.
 *
 *   If you are manipulating a large memory with 2-byte addresses, use
 *   the .2 address modifier, e.g. 210.2 addresses location 528 (decimal).
 *
 *   Then there are the unfortunate memory chips that spill the most
 *   significant 1, 2, or 3 bits of address into the chip address byte.
 *   This effectively makes one chip (logically) look like 2, 4, or
 *   8 chips.  This is handled (awkwardly) by #defining
 *   CONFIG_SYS_I2C_EEPROM_ADDR_OVERFLOW and using the .1 modifier on the
 *   {addr} field (since .1 is the default, it doesn't actually have to
 *   be specified).  Examples: given a memory chip at I2C chip address
 *   0x50, the following would happen...
 *     i2c md 50 0 10   display 16 bytes starting at 0x000
 *                      On the bus: <S> A0 00 <E> <S> A1 <rd> ... <rd>
 *     i2c md 50 100 10 display 16 bytes starting at 0x100
 *                      On the bus: <S> A2 00 <E> <S> A3 <rd> ... <rd>
 *     i2c md 50 210 10 display 16 bytes starting at 0x210
 *                      On the bus: <S> A4 10 <E> <S> A5 <rd> ... <rd>
 *   This is awfully ugly.  It would be nice if someone would think up
 *   a better way of handling this.
 *
 * Adapted from cmd_mem.c which is copyright Wolfgang Denk ([email protected]).
 */"

Для вашего примера:

i2c md 0x50 0.2 0x10

0x50 - адрес слейва

0.2 - отображение памяти начинается с адреса 0x000 (обозначается 0), .2 указывает длину адресного поля.

0x10 — показывает, что команда i2c md выведет 16 регистров.

Итак, команда i2c md 0x50 0.2 0x10 гласит: Показать 16 регистров, начиная с адреса 0x000 (длина адреса 2 байта) от подчиненного устройства, расположенного по адресу 0x50.