Merge pull request #10991 from RonEld/update_readme_for_optin_in_modules

Explaining how to opt in modules

Author: 0xc0170

features/cryptocell/FEATURE_CRYPTOCELL310/Readme.md

@@ -1,28 +1,37 @@
 # Guidelines for porting Cryptocell to Mbed OS
 
-Cryptocell (CC) 310 is a hardware accelerator that can be ported to several boards, assuming these boards have CC 310 embedded in their hardware.
+Cryptocell (CC) 310 is a hardware accelerator you can port to boards that have CC 310 embedded in their hardware.
 
 The CC 310 driver consists of three libraries:
 
-* A common core library(`libcc_core.a`).
-* A platform-specific TRNG library, containing TRNG-related information, for sampling sufficient entropy on the specific platform(`libcc_trng.a`).
-* A platform-specific library containing extra information, such as the CC register's base addresses on the specific board. (`libcc_ext.a`)
+* A common core library (`libcc_core.a`).
+* A platform-specific TRNG library containing TRNG-related information for sampling sufficient entropy on the specific platform (`libcc_trng.a`).
+* A platform-specific library containing extra information, such as the CC register's base addresses on the specific board (`libcc_ext.a`).
 
 Library version information:
 
-* The CC 310 libraries were built from version `arm_sw-cc310-1.1.0.1285`
+* The CC 310 libraries were built from version `arm_sw-cc310-1.1.0.1285`.
 * The `IAR` libraries were built using `IAR ANSI C/C++ Compiler V7.80.1.11864/W32 for ARM` with `--cpu Cortex-M4f`.
 * The `ARM` libraries were built using `ARM Compiler 5.06 update 4 (build 422)` with `--cpu cortex-m4`.
 * The `GCC_ARM` libraries were built using `arm-none-eabi-gcc 6.3.1 20170620 (release)` with `-mcpu=cortex-m4`.
 
-To port your CC 310 driver to Mbed OS on your specific target, do the following:
+To port your CC 310 driver to Mbed OS on your specific target:
 
-1. In `targets.json` add the following  to your target:
+1. In `targets.json`, add the following  to your target:
  	* `MBEDTLS_CONFIG_HW_SUPPORT` to `macros_add` key. This instructs Mbed TLS to look for an alternative cryptographic implementation.
  	* `CRYPTOCELL310` to `feature`. Use this in your common code that you need to remove from compilation in case CC exists in your board. Use `#if !defined(FEATURE_CRYPTOCELL310)` and  `#if defined(FEATURE_CRYPTOCELL310)`.
 1. In `objects.h`, include `objects_cryptocell.h`. You can use the `FEATURE_CRYPTOCELL310` precompilation check as defined above.
 1. In `features/cryptocell/FEATURE_CRYPTOCELL310/TARGET_<target name>`, add your platform-specific libraries for all toolchains in `TOOLCHAIN_ARM`, `TOOLCHAIN_GCC_ARM` and `TOOLCHAIN_IAR` respectively.
 1. Add your CC setup code:
-	* Implement `crypto_platform_setup()` and `crypto_platform_terminate()` to enable CC on your platform, in case you have board-specific setup functionality, required for CC setup. You MUST call 'SaSi_LibInit()` and 'SaSi_LibFini()' respectively in these functions.
+	* Implement `crypto_platform_setup()` and `crypto_platform_terminate()` to enable CC on your platform, in case you have board-specific setup functionality, required for CC setup. You MUST call `SaSi_LibInit()` and `SaSi_LibFini()` in these functions.
 	* Define `crypto_platform_ctx` in `crypto_device_platform.h` in a way that suits your implementation.
- 
+
+## Enabling optional alternative drivers
+
+Three additional modules that are not enabled by default have alternative implementation support. This allows backward compatability because these modules don't have full functionality and returns `MBEDTLS_ERR_PLATFORM_FEATURE_UNSUPPORTED` for some features. The modules are:
+
+* `AES`, which only supports 128 bit key size, as opposed to previous support for all key sizes.
+* `CMAC`, which only supports AES 128 bit key size, as opposed to previous support for DES and all key sizes.
+* `SHA512`, which only supports SHA512, as opposed to previous support for SHA384, as well.
+
+To enable these modules, define `MBEDTLS_AES_ALT`, `MBEDTLS_CMAC_ALT` and `MBEDTLS_SHA512_ALT`, either in `mbed_app.json` or in your `MBEDTLS_USER_CONFIG_FILE`, to have hardware accelerated module with reduced RAM size, on the expense of full functionality.