The simple answer is no, the Cortex-M0+ does not have a dedicated flash patch mechanism like the Flash Patch and Breakpoint (FPB) unit available in Cortex-M4. The FPB unit allows for dynamic patching of flash memory at runtime, without needing to halt or reprogram the chip. This is useful for applying bug fixes or implementing runtime security patches. The Cortex-M0+ lacks this dedicated hardware unit and does not support dynamic flash patching like the Cortex-M4.
Overview of Cortex-M0+
The Cortex-M0+ is ARM’s ultra low power microcontroller targeted for simple embedded applications. It is an extremely compact and efficient 32-bit RISC core, intended for cost-sensitive and power-constrained devices. The M0+ focuses on minimizing silicon area and maximizing energy efficiency.
As a result, the M0+ has a simplified design and reduced feature set compared to higher-end Cortex-M cores. It lacks certain capabilities like memory protection units, floating point support, and advanced debug features. This includes omitting the dedicated flash patch hardware present in Cortex-M3 and M4 chips.
Lack of Flash Patch and Breakpoint Unit
The key reason Cortex-M0+ does not have flash patch capabilities is the absence of the Flash Patch and Breakpoint (FPB) unit. This is a specialized piece of hardware included in the Cortex-M3 and M4 processors to enable dynamic self-patching of flash memory at runtime.
The FPB unit works by remapping flash regions to RAM whenever a breakpoint matches or a patch needs to be applied. This happens automatically without stopping the processor, allowing for seamless patching. The modified flash contents in RAM are then used until the next reset. This enables bug fixes and security updates to be applied on-the-fly.
However, the silicon area and power cost of the FPB hardware is too high for the low-end Cortex-M0+ design goals. So ARM opted to remove the FPB unit entirely from the M0+ to save on die space and power consumption. This was likely an acceptable tradeoff given the simpler applications targeted by M0+ chips.
Alternative Options for Cortex-M0+
While the Cortex-M0+ lacks a dedicated flash patching mechanism, there are still some options to dynamically modify flash at runtime:
- RAM-Based Patching: Patch code can be loaded into RAM at runtime and executed from there instead of flash. This allows bug fixes but is less persistent.
- External Flash: An external flash chip can be used to dynamically update application firmware.
- Partial Reprogramming: Small sections of internal flash can be reprogrammed without full erase.
However, these options are not as seamless as the FPB unit in Cortex-M4 and may require halting the processor temporarily. They also have limitations in flexibility, performance, and ease-of-use compared to dedicated flash patch hardware.
Cortex-M4 Flash Patch and Breakpoint Unit
Unlike the M0+, the Cortex-M3 and Cortex-M4 processors do include the Flash Patch and Breakpoint unit to support runtime flash patching efficiently. Some key capabilities of the FPB include:
- Remapping flash regions to RAM when a programmed address matches.
- Breakpoint matches can trigger remap to perform dynamic patching.
- Up to 8 simultaneous flash patch remaps.
- Dedicated remap address bus to overlay flash with RAM contents.
- Remapping is automatic and transparent to software.
With the FPB, critical patches can be applied immediately without halting operation. The flash contents are dynamically synchronized to RAM only when a remap occurs. This minimal remapping approach maximizes performance.
The Cortex-M4 FPB capabilities enable use cases like:
- Runtime bug fixes for embedded firmware.
- Security patches to embedded applications.
- Dynamically applying tuning parameters or sensor calibration data.
- Runtime profiling and instrumentation of flash code execution.
Overall, the Cortex-M4 FPB unit provides great flexibility and performance for dynamic flash self-modification at runtime.
Summary – Key Differences in Flash Patching Capability
In summary, there are some key differences between Cortex-M0+ and Cortex-M4 when it comes to flash patching:
- Cortex-M0+ lacks any dedicated flash patch hardware unit.
- Cortex-M4 has a Flash Patch and Breakpoint (FPB) unit allowing for dynamic patching.
- FPB enables seamless runtime flash remaps to RAM without halting processor.
- Cortex-M0+ requires alternative techniques like RAM patching or partial reflashing.
- FPB provides greater flexibility, performance, and ease-of-use for flash patching.
- Absence of FPB unit on M0+ is an architectural limitation by design.
In conclusion, while the Cortex-M0+ does not have dedicated flash patch capabilities, the Cortex-M4 FPB unit does enable advanced real-time self modification of flash memory at runtime. This difference stems directly from the architectural design goals and tradeoffs between the low-power M0+ and more fully-featured M4 cores.
