Finite State Machines (FSM) in Embedded Systems (Part 4) - Let 'em talk
No state machine is an island. State machines do not exist in a vacuum, they need to "talk" to their environment and each other to share information and provide synchronization to perform the system functions. In this conclusive article, you will find what kind of problems and which critical areas you need to pay attention to when designing a concurrent system. Although the focus is on state machines, the consideration applies to every system that involves more than one execution thread.
Getting Started With CUDA C on an Nvidia Jetson: A Meaningful Algorithm
In this blog post, I demonstrate a use case and corresponding GPU implementation where meaningful performance gains are realized and observed. Specifically, I implement a "blurring" algorithm on a large 1000x1000 pixel image. I show that the GPU-based implementation is 1000x faster than the CPU-based implementation.
Finite State Machines (FSM) in Embedded Systems (Part 3) - Unuglify C++ FSM with DSL
Domain Specific Languages (DSL) are an effective way to avoid boilerplate or repetitive code. Using DSLs lets the programmer focus on the problem domain, rather than the mechanisms used to solve it. Here I show how to design and implement a DSL using the C++ preprocessor, using the FSM library, and the examples I presented in my previous articles.
C++ Assertion? Well Yes, But Actually No.
Assertions are a double-edged sword - on one side you enforce program correctness catching bugs close to their origin, on the other your application is subject to run-time error, like any interpreted language. This article explores what C++ can offer to get the advantages of assertions, without risking the crashes by moving contract checking at compile time.
The volatile keyword
Although the C keyword volatile is very useful in embedded applications, care is needed to use it correctly and vigilance is required to ensure its correct implementation by compilers.
When a Mongoose met a MicroPython, part I
This is more a framework than an actual application, with it you can integrate MicroPython and Cesanta's Mongoose.
Mongoose runs when called by MicroPython and is able to run Python functions as callbacks for the events you decide in your event handler. The code is completely written in C, except for the example Python callback functions, of course. To try it, you can just build this example on a Linux machine, and, with just a small tweak, you can also run it on any ESP32 board.
Getting Started With CUDA C on an Nvidia Jetson: GPU Architecture
In the previous blog post (Getting Started With CUDA C on Jetson Nvidia: Hello CUDA World!) I showed how to develop applications targeted at a GPU on a Nvidia Jetson Nano. As we observed in that blog post, performing a calculation on a 1-D array on a GPU had no performance benefit compared to a traditional CPU implementation, even on an array with many elements. In this blog post, we will learn about the GPU architecture to better explain the behavior and to understand the applications where a GPU shines (hint: it has to do with graphics).
C to C++: Templates and Generics – Supercharging Type Flexibility
"C to C++: Templates and Generics – Supercharging Type Flexibility" illuminates the rigidity of C when managing multiple types and the confusion of code replication or macro complexity. In contrast, C++ offers templates, acting as type-agnostic blueprints for classes and functions, which allows for the creation of versatile and reusable code without redundancy. By using templates, developers can define operations like add once and apply them to any data type, simplifying codebases significantly. Generics further this concept, enabling a single code structure to handle diverse data types efficiently—a boon for embedded systems where operations must be performed on varying data, yet code efficiency is critical due to resource limitations. The blog walks through practical applications, showcasing how templates streamline processes and ensure type safety with static_assert, all while weighing the pros and cons of their use in embedded software, advocating for careful practice to harness their full potential.
Finite State Machines (FSM) in Embedded Systems (Part 2) - Simple C++ State Machine Engine
When implementing state machines in your project it is an advantage to rely on a tried and tested state machine engine. This component is reused for every kind of application and helps the developer focus on the domain part of the software. In this article, the design process that turns a custom C++ code into a finite-state machine engine is fully described with motivations and tradeoffs for each iteration.
Creating a GPIO HAL and Driver in C
Creating a GPIO Hardware Abstraction Layer (HAL) in C allows for flexible microcontroller interfacing, overcoming the challenge of variability across silicon vendors. This method involves reviewing datasheets, identifying features, designing interfaces, and iterative development, as detailed in the "Reusable Firmware" process. A simplified approach prioritizes essential functions like initialization and read/write operations, showcased through a minimal interface example. The post also highlights the use of AI to expedite HAL generation. A detailed GPIO HAL version is provided, featuring extended capabilities and facilitating driver connection through direct assignments or wrappers. The significance of a configuration table for adaptable peripheral setup is emphasized. Ultimately, the blog illustrates the ease and scalability of developing a GPIO HAL and driver in C, promoting hardware-independent and extensible code for various interfaces, such as SPI, I2C, PWM, and timers, underscoring the abstraction benefits.
C to C++: 3 Reasons to Migrate
Embedded C still powers most devices, but rising system complexity is revealing its limits. In this post Jacob Beningo kicks off a series on moving from C to C++, offering three practical reasons to start the migration now. He argues for an incremental approach that keeps low-level, hardware-dependent code in C while adopting C++ for higher-level, object-oriented application logic so teams can keep shipping during the transition.
Getting Started With Zephyr: DTS vs DTSI vs Overlays
Devicetrees can be daunting for traditional embedded software engineers that are new to Zephyr. In this blog post, I address these fears and show how navigating Devicetrees can be much easier if you understand that they represent the layered structure of the underlying hardware.
Working with Strings in Embedded C++
This article discusses the use of strings in embedded systems. It explains how the need for and use of strings in embedded systems has changed with the advent of cheaper, full graphic displays and the growth of the ‘Internet of Things’ (IoT). The article also covers character literals, C-Strings and string literals, and the difference in memory models between them. It also highlights the safety and security issues that arise from using strings in embedded systems. Finally, it explains how C++11 introduced a Raw string literal type that is useful for storing file paths or regular expressions.
The volatile keyword
Although the C keyword volatile is very useful in embedded applications, care is needed to use it correctly and vigilance is required to ensure its correct implementation by compilers.
Picowoose: The Raspberry Pi Pico-W meets Mongoose
This example application describes the way to adapt the George Robotics CYW43 driver, present in the Pico-SDK, to work with Cesanta's Mongoose. We are then able to use Mongoose internal TCP/IP stack (with TLS 1.3), instead of lwIP (and MbedTLS).
Creating a GPIO HAL and Driver in C
Creating a GPIO Hardware Abstraction Layer (HAL) in C allows for flexible microcontroller interfacing, overcoming the challenge of variability across silicon vendors. This method involves reviewing datasheets, identifying features, designing interfaces, and iterative development, as detailed in the "Reusable Firmware" process. A simplified approach prioritizes essential functions like initialization and read/write operations, showcased through a minimal interface example. The post also highlights the use of AI to expedite HAL generation. A detailed GPIO HAL version is provided, featuring extended capabilities and facilitating driver connection through direct assignments or wrappers. The significance of a configuration table for adaptable peripheral setup is emphasized. Ultimately, the blog illustrates the ease and scalability of developing a GPIO HAL and driver in C, promoting hardware-independent and extensible code for various interfaces, such as SPI, I2C, PWM, and timers, underscoring the abstraction benefits.
On hardware state machines: How to write a simple MAC controller using the RP2040 PIOs
Hardware state machines are nice, and the RP2040 has two blocks with up to four machines each. Their instruction set is limited, but powerful, and they can execute an instruction per cycle, pushing and popping from their FIFOs and shifting bytes in and out. The Raspberry Pi Pico does not have an Ethernet connection, but there are many PHY boards available… take a LAN8720 board and connect it to the Pico; you’re done. The firmware ? Introducing Mongoose…
Lightweight C++ Error-Codes Handling
The traditional C++ approach to error handling tends to distinguish the happy path from the unhappy path. This makes handling errors hard (or at least boring) to write and hard to read. In this post, I present a technique based on chaining operations that merges the happy and the unhappy paths. Thanks to C++ template and inlining the proposed technique is lightweight and can be used proficiently for embedded software.
Getting Started With Zephyr: Devicetree Bindings
This blog post shines some light on how devicetrees are used in The Zephyr Project. Specifically, we understand the mechanisms that enable us to use nodes in the devicetree in the C source files. We use a sample provided in the Zephyr repository itself and work our way through portions of the Zephyr codebase to get insight into the mechanisms that make this possible.
New book on Elliptic Curve Cryptography
New book on Elliptic Curve Cryptography now online. Deep discount for early purchase. Will really appreciate comments on how to improve the book because physical printing won't happen for a few more months. Check it out here: http://mng.bz/D9NA
C to C++: 5 Tips for Refactoring C Code into C++
The article titled "Simple Tips to Refactor C Code into C++: Improve Embedded Development" provides essential guidance for embedded developers transitioning from C to C++. The series covers fundamental details necessary for a seamless transition and emphasizes utilizing C++ as a better C rather than diving into complex language features. The article introduces five practical tips for refactoring C code into C++. Replace #define with constexpr and const: Discouraging the use of #define macros, the article advocates for safer alternatives like constexpr and const to improve type safety, debugging, namespaces, and compile-time computation. Use Namespaces: Demonstrating the benefits of organizing code into separate logical groupings through namespaces, the article explains how namespaces help avoid naming conflicts and improve code readability. Replace C-style Pointers with Smart Pointers and References: Emphasizing the significance of avoiding raw pointers, the article suggests replacing them with C++ smart pointers (unique_ptr, shared_ptr, weak_ptr) and using references
Review: Hands-On RTOS with Microcontrollers
Brian Amos's Hands-On RTOS with Microcontrollers delivers a practical path from bare-metal to full RTOS applications using FreeRTOS on an STM32 Nucleo-F767ZI board. The book combines clear explanations of concurrency, interrupts, and DMA with step-by-step toolchain setup and runnable examples that show building, debugging, monitoring, and scaling embedded systems for real projects and coursework.
Scorchers, Part 3: Bare-Metal Concurrency With Double-Buffering and the Revolving Fireplace
Jason Sachs presents a practical, low-overhead concurrency pattern for tiny bare-metal systems where an ISR (Speedy) must safely exchange data with a nonreal-time main loop (Poky). He describes the "revolving fireplace", a double-buffering variant that swaps ownership of two shared memory regions, and walks through C examples, atomic/volatile considerations, and testing strategies so you can implement it on RAM-constrained MCUs.
Finite State Machines (FSM) in Embedded Systems (Part 4) - Let 'em talk
No state machine is an island. State machines do not exist in a vacuum, they need to "talk" to their environment and each other to share information and provide synchronization to perform the system functions. In this conclusive article, you will find what kind of problems and which critical areas you need to pay attention to when designing a concurrent system. Although the focus is on state machines, the consideration applies to every system that involves more than one execution thread.
C to C++: 3 Proven Techniques for Embedded Systems Transformation
Jacob Beningo lays out a pragmatic, low-risk path for embedded teams to start using C++ without adding bloat or runtime cost. He recommends beginning by treating C++ as a cleaner C with namespaces, constexpr, and smart pointers, then adopting object-oriented design with composition, and finally introducing templates for static polymorphism where it makes sense. The post focuses on practical guardrails for resource-constrained firmware.
Lightweight C++ Error-Codes Handling
The traditional C++ approach to error handling tends to distinguish the happy path from the unhappy path. This makes handling errors hard (or at least boring) to write and hard to read. In this post, I present a technique based on chaining operations that merges the happy and the unhappy paths. Thanks to C++ template and inlining the proposed technique is lightweight and can be used proficiently for embedded software.
C to C++: Templates and Generics – Supercharging Type Flexibility
"C to C++: Templates and Generics – Supercharging Type Flexibility" illuminates the rigidity of C when managing multiple types and the confusion of code replication or macro complexity. In contrast, C++ offers templates, acting as type-agnostic blueprints for classes and functions, which allows for the creation of versatile and reusable code without redundancy. By using templates, developers can define operations like add once and apply them to any data type, simplifying codebases significantly. Generics further this concept, enabling a single code structure to handle diverse data types efficiently—a boon for embedded systems where operations must be performed on varying data, yet code efficiency is critical due to resource limitations. The blog walks through practical applications, showcasing how templates streamline processes and ensure type safety with static_assert, all while weighing the pros and cons of their use in embedded software, advocating for careful practice to harness their full potential.
C++ Assertion? Well Yes, But Actually No.
Assertions are a double-edged sword - on one side you enforce program correctness catching bugs close to their origin, on the other your application is subject to run-time error, like any interpreted language. This article explores what C++ can offer to get the advantages of assertions, without risking the crashes by moving contract checking at compile time.
NULL pointer protection with ARM Cortex-M MPU
This post explains how you can set up the ARM Cortex-M MPU (Memory Protection Unit) to protect thy code from dragons, demons, core dumps, and numberless other foul creatures awaiting thee after thou dereference the NULL pointer.
New book on Elliptic Curve Cryptography
New book on Elliptic Curve Cryptography now online. Deep discount for early purchase. Will really appreciate comments on how to improve the book because physical printing won't happen for a few more months. Check it out here: http://mng.bz/D9NA
