Embedded Software Development Flashcards

Question 1

What defines an embedded system, and how does it differ from a general purpose computer?

Accepted Answer

An embedded system is a computer built into a larger device to perform a dedicated function. It typically has limited memory, modest processing power, and strict cost or energy constraints. A general purpose computer, by contrast, is designed to run many different applications that users choose at any time.

Question 2

Why do many embedded systems have no keyboard or screen for the user to operate them directly?

Accepted Answer

Many embedded systems interact with the world through sensors and actuators rather than through human input devices. A washing machine reads water level and door state, then drives motors and valves. Its control logic does the real work, so the interface can be a few buttons, a light, or nothing at all.

Question 3

What is the difference between hard real-time and soft real-time requirements?

Accepted Answer

A hard real-time requirement states that a missed deadline is a system failure. A soft real-time requirement tolerates occasional late responses, with degraded quality rather than failure, such as a video stream that drops a frame. Designers choose the category based on the consequences of lateness.

Question 4

Why is a watchdog timer commonly used in embedded systems?

Accepted Answer

A watchdog timer is a hardware counter that resets the processor unless the software periodically clears it. If the program hangs or enters an unexpected loop, the timer expires and forces a restart, which lets the device recover without human intervention. It is a common safety net in unattended systems.

Question 5

What does it mean for a microcontroller to be resource constrained?

Accepted Answer

Resource constrained means the device has only small amounts of RAM, flash storage, clock speed, and power compared with a desktop machine. Designers must count every byte and clock cycle, and they often avoid dynamic memory allocation and heavy abstraction layers. Efficient code and careful data type choices are therefore essential.

Question 6

What is the role of a board support package in embedded development?

Accepted Answer

A board support package is the layer of low level code that initializes a specific circuit board and exposes its peripherals to higher level software. It typically includes startup code, clock configuration, and drivers for memory and communication interfaces. It lets application code stay largely independent of the exact hardware revision.

Question 7

Why do embedded developers often use a cross compiler rather than compiling on the target?

Accepted Answer

A cross compiler runs on a host computer, such as a desktop, and produces machine code for a different processor architecture. Target devices often lack the memory, storage, or speed to host a full compiler and toolchain. Developers then flash or load the resulting binary onto the device for testing.

Question 8

What is a memory mapped peripheral register, and how does software use it?

Accepted Answer

A memory mapped peripheral register is a location that controls or reports the state of a hardware block, such as a timer or serial port. Reading or writing that address performs an action on the hardware. Software typically uses volatile pointers or structure definitions so the compiler does not optimize away accesses.

Question 9

What is the difference between a microcontroller and a microprocessor?

Accepted Answer

A microcontroller integrates a processor core with memory and peripheral interfaces on a single chip, making it a complete small computer. A microprocessor contains the central processing unit and relies on separate chips for memory and input output. Microcontrollers suit compact control tasks, while microprocessors suit systems needing external memory and higher performance.

Question 10

How does a general purpose input output pin differ from a dedicated peripheral pin?

Accepted Answer

A general purpose input output pin can be configured in software as either an input or an output, and it carries a single logic signal. A dedicated peripheral pin is wired to a specific function, such as a serial line, and it only works when that function is enabled.

Question 11

What is the purpose of a clock tree in a microcontroller?

Accepted Answer

A clock tree is the network of oscillators, dividers, and multiplexers that distributes timing signals to the processor core and its peripherals. Each branch can run at a different frequency, which lets designers trade speed against power consumption. Software selects clock sources and dividers at startup, and unused branches can be gated off.

Question 12

Why do designers add a brown out detector to a microcontroller board?

Accepted Answer

A brown out detector monitors the supply voltage and holds the processor in reset when the voltage falls below a safe threshold. Without it, the core may execute erratically during a sag, corrupting memory or leaving outputs in unsafe states. The detector lets the system restart cleanly once power returns to a stable level.

Question 13

What is an analog to digital converter used for in embedded sensing?

Accepted Answer

An analog to digital converter measures a continuously varying voltage and produces a number that software can process. Its resolution, expressed in bits, determines how finely the input range is divided. Sensors such as temperature probes and light detectors often connect to these converters, which allows firmware to convert raw readings into physical units.

Question 14

Why must firmware account for the warm up behavior of a sensor?

Accepted Answer

Many sensors drift or report unstable values until their internal temperature and bias settle after power is applied. Firmware that reads them immediately may treat these transient values as real measurements. Developers therefore discard early samples, wait a documented settling time, or average readings until they stabilize.

Question 15

What is pulse width modulation and what is it commonly used to control?

Accepted Answer

Pulse width modulation switches a digital output between high and low at a fixed frequency, varying the fraction of each cycle spent high. The average voltage delivered to the load then tracks the duty cycle. It is commonly used to control motor speed, LED brightness, and heater power without wasting energy in linear regulators.

Question 16

How does a direct memory access controller reduce processor workload?

Accepted Answer

A direct memory access controller moves blocks of data between peripherals and memory without the processor copying each byte. The core configures source, destination, and length, then starts the transfer and continues other work. The controller signals an interrupt when the transfer completes, which frees the processor for computation.

Question 17

What is a deadline in a real-time system, and how is it related to response time?

Accepted Answer

A deadline is the latest time by which a task must complete its work after an event occurs. Response time is the elapsed time from the event to the completed response. A task meets its deadline when its response time is less than or equal to the deadline value, and misses it otherwise.

Question 18

What is the difference between a periodic task and an aperiodic task?

Accepted Answer

A periodic task is released at fixed, regular intervals, such as sampling a sensor every ten milliseconds. An aperiodic task is triggered by irregular events, such as a button press or an incoming message. Aperiodic work is often queued or handled by a server to protect periodic deadlines.

Question 19

What does rate monotonic scheduling assign priorities based on?

Accepted Answer

Rate monotonic scheduling assigns fixed priorities according to task period, so tasks with shorter periods, meaning higher release rates, receive higher priority. It is a widely taught fixed priority method for periodic tasks with deadlines equal to their periods. Under its standard assumptions, it can guarantee schedulability using a utilization bound.

Question 20

What is priority inversion and why is it dangerous in a real-time system?

Accepted Answer

Priority inversion occurs when a high priority task waits for a resource held by a low priority task, while medium priority tasks delay the holder. The high priority task is blocked by lower priority work, risking missed deadlines. Protocols such as priority inheritance reduce this risk by temporarily raising the holder's priority.

Question 21

What is a context switch and what costs does it add to a real-time system?

Accepted Answer

A context switch saves the state of the running task, such as registers and program counter, and restores the state of another task. The switch itself consumes processor time, adds latency before the new task runs, and increases stack usage. Designers account for these overheads when estimating worst case response times.

Question 22

Why is worst case execution time analysis important for real-time design?

Accepted Answer

Worst case execution time analysis estimates the longest time a piece of code can take on a given processor. Schedulability tests rely on these bounds, so an underestimate can make an unsafe schedule look feasible. Average measured times alone cannot prove that deadlines will always be met.

Question 23

What is the role of a real-time operating system kernel in task management?

Accepted Answer

A real-time operating system kernel creates tasks, selects which ready task runs next according to its scheduling policy, and manages transitions between tasks. It provides services such as semaphores, message queues, and timers that tasks use to coordinate. Its design keeps overhead small and predictable so that timing analysis remains practical.

Question 24

Why might a designer choose a cyclic executive instead of a preemptive scheduler?

Accepted Answer

A cyclic executive runs a fixed table of tasks in a repeating frame, so the execution order is known in advance and is easy to verify. This suits simple systems with fixed well understood timing. The cost is reduced flexibility, because long tasks must be split across frames and sporadic events are awkwardly handled.

Question 25

What is an interrupt and how does it differ from polling?

Accepted Answer

An interrupt is a hardware signal that suspends normal execution to run a short service routine. Polling instead has software repeatedly check a status flag to see whether an event has occurred. Interrupts respond quickly and save processor time, while polling is simpler but wastes cycles when events are rare.

Question 26

What is a race condition in embedded software that uses interrupts?

Accepted Answer

A race condition occurs when the outcome depends on the timing of two or more execution contexts accessing shared data. An interrupt handler may modify a variable while main code is midway through reading it, producing a corrupted value. Protecting such data with brief critical sections or atomic operations prevents the problem.

Question 27

Why should an interrupt service routine be kept short?

Accepted Answer

An interrupt service routine blocks lower priority interrupts and delays the main program. Long handlers increase interrupt latency for other events and can cause missed deadlines. A common approach is to do only the urgent work in the handler, such as clearing the flag and copying data, then defer processing to a task.

Question 28

What is a volatile qualifier and when must it be used in embedded C?

Accepted Answer

The volatile qualifier tells the compiler that a variable can change outside the normal program flow, so every read and write must actually occur. It is required for memory mapped registers, variables shared with interrupt handlers, and some data modified by other hardware. It does not provide atomicity or mutual exclusion by itself.

Question 29

What is a stack overflow and why is it hard to detect in small systems?

Accepted Answer

A stack overflow occurs when the call stack grows into memory used for data or other stacks, often from deep recursion or large local arrays. Small systems often lack a memory protection unit, so the corruption may appear far from its cause. Designers mitigate the risk with stack sizing analysis, canary values, and watermark checks.

Question 30

What is a mutex and when should a task use one instead of disabling interrupts?

Accepted Answer

A mutex is a locking primitive that allows only one task at a time to hold a shared resource. Tasks use it for longer protected sections, such as updating a buffer, where disabling interrupts would add too much latency. Disabling interrupts suits only very short critical sections.

Question 31

What is a memory leak in embedded systems and why is it especially serious?

Accepted Answer

A memory leak happens when allocated memory is never released, so the free heap shrinks over time. Because embedded devices may run for years without restarting and have little RAM, even a slow leak can eventually cause allocation failures. For this reason many safety oriented designs forbid dynamic allocation after startup.

Question 32

How does a memory protection unit help an embedded system?

Accepted Answer

A memory protection unit restricts which code can read, write, or execute particular address regions, and it raises a fault when a rule is violated. This can stop a faulty task from corrupting the stacks or data of other tasks. Not every microcontroller includes one, so designers without one rely on careful coding and analysis.

Question 33

Why is testing embedded software on the target hardware important?

Accepted Answer

Host based tests can verify logic quickly, but they cannot reveal timing errors, register misuse, or problems caused by real electrical behavior. Target testing exercises the actual processor, clocks, peripherals, and power conditions that the product will experience. Both kinds are needed because each catches different risks.

Question 34

What is a hardware in the loop test and what does it accomplish?

Accepted Answer

A hardware in the loop test connects the embedded controller to a simulated model of its physical environment, such as a motor, vehicle, or plant. The model supplies realistic sensor inputs and records the controller's outputs. This allows engineers to test dangerous or costly scenarios safely, repeatedly, and early in development.

Question 35

How does a logic analyzer help when debugging timing problems?

Accepted Answer

A logic analyzer records the digital states of several pins against time, letting developers see the order and duration of signals without stopping the program. Triggers can capture a sequence around a specific event, such as a failed transaction on a bus. Comparing captured timing with the expected protocol often reveals the cause quickly.

Question 36

What is the purpose of a JTAG or SWD debug interface?

Accepted Answer

A JTAG or SWD interface gives a debugger access to the processor core, allowing it to halt execution, inspect registers and memory, and set breakpoints. These standard debug ports often need only a few pins. Manufacturers often disable them in production units to protect firmware.

Question 37

Why is version control important for embedded firmware projects?

Accepted Answer

Version control records every change to source code, configuration, and build scripts, so teams can review history and recover earlier working states. It also lets teams link a released firmware image to the exact source that produced it. This traceability is valuable when a field defect must be reproduced years after shipment.

Question 38

How do over the air updates introduce risk into embedded products?

Accepted Answer

An update can fail partway through, leaving the device with an unusable image or a mismatched bootloader. A malicious or corrupted package may also install unauthorized code if the device does not verify it. Robust designs use dual image layouts, rollback support, and cryptographic signature checks before any new firmware runs.

Question 39

What is regression testing and why does it matter after a firmware change?

Accepted Answer

Regression testing reruns previously passing tests after a change to confirm that existing behavior still works. Firmware changes can have unexpected side effects, such as altered timing, shared buffer conflicts, or changed power consumption. Automated test suites make this practical, especially when they run on target hardware or realistic simulators.

Question 40

What information should a good firmware crash log capture?

Accepted Answer

A useful crash log records the fault type, the program counter and key registers at the moment of failure, and the active task or interrupt. It should also store a firmware version identifier and a timestamp from a reliable clock source. Preserving this data across a reset lets engineers reconstruct the failure after the device restarts.

Embedded Software Development Flashcards

1Foundations of Embedded Systems

2Microcontroller Hardware and Peripherals

3Real-Time Concepts and Scheduling

4Interrupts, Concurrency and Memory

5Testing, Debugging and Maintenance

Learning Resources