The eFuse Manager library is designed to structure access to eFuse bits and make using these easy. This library operates eFuse bits by a structure name which is assigned in eFuse table. This sections introduces some concepts used by eFuse Manager.
The ESP32 has a number of eFuses which can store system and user parameters. Each eFuse is a one-bit field which can be programmed to 1 after which it cannot be reverted back to 0.
Some of system parameters are using these eFuse bits directly by hardware modules and have special place (for example EFUSE_BLK0). For more details see `ESP32 Technical Reference Manual <https://www.espressif.com/sites/default/files/documentation/esp32_technical_reference_manual_en.pdf>`_ in part 20 eFuse controller. Some eFuse bits are available for user applications.
The component has API functions for reading and writing fields. Access to the fields is carried out through the structures that describe the location of the eFuse bits in the blocks. The component provides the ability to form fields of any length and from any number of individual bits. The description of the fields is made in a CSV file in a table form. To generate from a tabular form (CSV file) in the C-source uses the tool `efuse_table_gen.py`. The tool checks the CSV file for uniqueness of field names and bit intersection, in case of using a `custom` file from the user's project directory, the utility will check with the `common` CSV file.
* common (`esp_efuse_table.csv`) - contains eFuse fields which are used inside the IDF. C-source generation should be done manually when changing this file (run command ``idf.py efuse_common_table``). Note that changes in this file can lead to incorrect operation.
* custom - (optional and can be enabled by :envvar:`CONFIG_EFUSE_CUSTOM_TABLE`) contains eFuse fields that are used by the user in their application. C-source generation should be done manually when changing this file and running ``idf.py efuse_custom_table``.
Name of field. The prefix `ESP_EFUSE_` will be added to the name, and this field name will be available in the code. This name will be used to access the fields. The name must be unique for all fields. If the line has an empty name, then this line is combined with the previous field. This allows you to set an arbitrary order of bits in the field, and expand the field as well (see ``MAC_FACTORY`` field in the common table).
Start bit number (0..255). The bit_start field can be omitted. In this case, it will be set to bit_start + bit_count from the previous record, if it has the same efuse_block. Otherwise (if efuse_block is different, or this is the first entry), an error will be generated.
The number of bits to use in this field (1..-). This parameter can not be omitted. This field also may be ``MAX_BLK_LEN`` in this case, the field length will have the maximum block length, taking into account the coding scheme (applicable for ``ESP_EFUSE_SECURE_BOOT_KEY`` and ``ESP_EFUSE_ENCRYPT_FLASH_KEY`` fields). The value ``MAX_BLK_LEN`` depends on :envvar:`CONFIG_EFUSE_MAX_BLK_LEN`, will be replaced with "None" - 256, "3/4" - 192, "REPEAT" - 128.
This param is using for comment field, it also move to C-header file. The comment field can be omitted.
If a non-sequential bit order is required to describe a field, then the field description in the following lines should be continued without specifying a name, this will indicate that it belongs to one field. For example two fields MAC_FACTORY and MAC_FACTORY_CRC:
::
# Factory MAC address #
#######################
MAC_FACTORY, EFUSE_BLK0, 72, 8, Factory MAC addr [0]
, EFUSE_BLK0, 64, 8, Factory MAC addr [1]
, EFUSE_BLK0, 56, 8, Factory MAC addr [2]
, EFUSE_BLK0, 48, 8, Factory MAC addr [3]
, EFUSE_BLK0, 40, 8, Factory MAC addr [4]
, EFUSE_BLK0, 32, 8, Factory MAC addr [5]
MAC_FACTORY_CRC, EFUSE_BLK0, 80, 8, CRC8 for factory MAC address
This field will available in code as ESP_EFUSE_MAC_FACTORY and ESP_EFUSE_MAC_FACTORY_CRC.
efuse_table_gen.py tool
-----------------------
The tool is designed to generate C-source files from CSV file and validate fields. First of all, the check is carried out on the uniqueness of the names and overlaps of the field bits. If an additional `custom` file is used, it will be checked with the existing `common` file (esp_efuse_table.csv). In case of errors, a message will be displayed and the string that caused the error. C-source files contain structures of type `esp_efuse_desc_t`.
The coding scheme affects only EFUSE_BLK1, EFUSE_BLK2 and EFUSE_BLK3 blocks. EUSE_BLK0 block always has a coding scheme ``None``.
Coding changes the number of bits that can be written into a block, the block length is constant 256, some of these bits are used for encoding and are not used.
* from ``esptool`` utility logs (during flashing).
* calling the function in the code :cpp:func:`esp_efuse_get_coding_scheme` for the EFUSE_BLK3 block.
eFuse tables must always comply with the coding scheme in the chip. There is an :envvar:`EFUSE_CODE_SCHEME_SELECTOR` option to select the coding type for tables in a Kconfig. When generating source files, if your tables do not follow the coding scheme, an error message will be displayed. Adjust the length or offset fields.
If your program was compiled with ``None`` encoding and ``3/4`` is used in the chip, then the ``ESP_ERR_CODING`` error may occur when calling the eFuse API (the field is outside the block boundaries). If the field matches the new block boundaries, then the API will work without errors.
Also, 3/4 coding scheme imposes restrictions on writing bits belonging to one coding unit. The whole block with a length of 256 bits is divided into 4 coding units, and in each coding unit there are 6 bytes of useful data and 2 service bytes. These 2 service bytes contain the checksum of the previous 6 data bytes.
It turns out that only one field can be written into one coding unit. Repeated rewriting in one coding unit is prohibited. But if the record was made in advance or through a :cpp:func:`esp_efuse_write_block` function, then reading the fields belonging to one coding unit is possible.
In case ``3/4`` coding scheme, the writing process is divided into the coding units and we can not use the usual mode of writing some fields. We can prepare all the data for writing and burn it in one time. You can also use this mode for ``None`` coding scheme but it is not necessary. It is important for ``3/4`` coding scheme.
To write some fields in one time need to use ``the batch writing mode``. Firstly set this mode through :cpp:func:`esp_efuse_batch_write_begin` function then write some fields as usual use the ``esp_efuse_write_...`` functions. At the end to burn they, need to call the :cpp:func:`esp_efuse_batch_write_commit` function. It burns prepared data to the efuse blocks and disable the ``batch recording mode``.
The Kconfig option :envvar:`CONFIG_EFUSE_VIRTUAL` will virtualize eFuse values inside the eFuse Manager, so writes are emulated and no eFuse values are permanently changed. This can be useful for debugging app and unit tests.
espefuse.py
^^^^^^^^^^^
esptool includes a useful tool for reading/writing ESP32 eFuse bits - `espefuse.py <https://github.com/espressif/esptool/wiki/espefuse>`_.
