New code routines for a secure system (e.g., a TPM) are stored in a memory (e.g., a flash memory) that is located external to the secure system. For example, a chip may include a TPM and an external flash memory may be connected to the chip. New routines for the TPM may then be stored in the flash. A function table may be used to determine whether a given function to be executed by the TPM is stored in on-chip memory (e.g., ROM) or in the flash. New function tables may be stored in the flash. For example, when a new set of functions is loaded into the flash, a new function table that references the new functions also may be loaded into the flash.