RTOS/TMS320F28335: Code Security Module(CSM), questions about thread

Can you show an example that allows me to write code to access secure memory from non-secure memory?

We do not have such an example. Once the security conditions are satisfied, it is no different from accessing one part of memory from another.

A call can be anywhere into secure zone memory. So, you can write functions in secure memory, and make calls to those functions from unsecure memory if desired. But you cannot read the secure code directly from code running in unsecure memory. It's not a mode of the CPU or anything, the security is just based off what address the code running resides at. So, when you ask about unsecure code "accessing" secure memory, if you are meaning how does one transfer execution from unsecure code to secure code, it is just done by changing the Program Counter (PC) value to an address in secure memory. This can be accomplished by a call or a branch instruction (or, in C, a call to a function located in secure memory). If you are asking how unsecure code can directly read variables located in secure memory, you cannot do this unless you unlock the CSM. Secure code (code running from secure memory) can always directly read/write both secure memory and unsecure memory. However, unsecure code (code running from unsecure memory) can only directly read/write unsecure memory while the CSM is locked.

There is no specific code that must be run. You just need to understand where you placed certain code. For instance, if a piece of code is copied to L4 RAM (unsecure memory), it will be considered unsecure code when run, and cannot directly access variables placed in L0 RAM (secure memory). However, code running from flash (secure memory), or copied to and run from L1 RAM (secure memory), would be able to access L0 RAM. Additionally, it would be able to access variables placed in unsecure memory (e.g., L5 RAM).

Hope this helps.

Regards,

Dave Foley