Nothing, just Mike and Johnny's notebook
Mike's articles written as if you have already known the basic and the knowledge only remind but not clearly for you to understand from zero
C++
Python
Robot Operating System
We are not familiar with web applications so our GitHub Webpage is for fun, our main target is knowledge
- Table of contents
- Some Words About This Notebook
- 1. C
- C++
- Python
- 2. Computer Architect
- 3. Embedded
- 4. Micro-controller
- 5. Operating System
- 6. RTOS
- Robot Operating System
- 7. Linux Embedded
- Available soon
| auto | double | int | struct |
|---|---|---|---|
| break | else | long | switch |
| case | enum | register | typedef |
| char | extern | return | union |
| continue | for | signed | void |
| do | if | static | while |
| default | goto | sizeof | volatile |
| const | float | short | unsigned |
type-qualifier(s) type-modifier data-type variable-name = initial-value;
type-qualifier: const, volatile, restrict
type-modifier: short, long, unsigned, signed
data-type: integer (int), floating point (double, float), enumerated, derived (array, struct, union, pointer), void
Memory layout from low to high addresses in C includes:
-
Text Segment(code segment): executable instructions - sharable & read-only -
initialized data: global variables, static variables initialized (declared
$\neq$ 0) - uninitialized data: global variables, static variables uninitialized or initialized to 0
- heap: dynamic memory allocation (forget deallocation cause Memory leak)
- stack: automatic variable storage, function frame (runout of stack memory cause Stack overflow)
Notice: memory layout in C is not memory layout in computer, this will be discussed in Operating Systems part
volatile: used to indicate to the compiler that a variable's value may change unexpectedly
volatile is usually used when you turn on compiler's optimization function
- Peripheral register mapping to memory
- Interrupt Service Routines (ISRs)
- Multi-thread applications const volatile: same as volatile but you can't change its value by the code
pointer is variables that hold address
pointer point to any type have same size
In C, we don't have reference parameter so we can use pointer to substitute
function pointer is a pointer to const because function address place on code segment which is read-only
syntax: data-type (*function-name)(parameter)
e.g. void (*pointer_function)(): pointer function point to a function with no parameter and return void
struct, union and enum are user-define data types
struct is used to group related variables (members) into one
struct in C need to use typedef to don't use struct keyword whenever declare a new struct variable. In C++, it is not necessary
union allows to store different data types to same memory
enum is usually used to assign name to integral constants
data structure alignment is the term that means compiler "padding" into space between 2 data "naturally aligned", this mechanism help compiler better performance. We should arrange data in struct reasonable to optimize memory and performance
extern is used to notify to the compiler that variable is memory allocated and don't need to allocate again, it is usually used in multi-source file projects
static have 2 meaning:
- If a variable is declared as global, that variable is only used in that file (multi-source file projects)
- Else if that variable is declared as local in a function (or a class in C++), that variable is memory allocated once until the program ends meaning that the variable will not be destroyed when the function call ends so it will use its last "state" when the function called again
Preprocessing: preprocessor will replace preprocessor directives (#)
Compiling: compiler will compile code to assembly code (.s/.asm)
Assembling: assembler will change assembly code to machine code in object code files(.o)
Linking: linker will link object code files together to shared library or executable file
A little bit about preprocessing and linking: preprocessing replaces header file (.h) but it usually only contains declarations not definitions and the linking link the definitions to these declarations
Static library is linked in compile time
Shared/Dynamic library is linked in run time
When you invoke a function, there are 2 important additional parts called prologue and epilogue
prologue is the concealed code run before the function is invoked, it is responsible to take the function's parameter to save to stack
epilogue is the concealed code run after the function is invoked, it is responsible for returning the function's result to the function invoker and removing the parameter saved to stack
function, prologue and epilogue take the same memory size no matter how many times the function is invoked which means the more you call the function, the more memory is saved
But there is a paradox in programming that the advantage in memory comes with the disadvantage in speed
The worst case is the function is too short that is shorter than prologue/epilogue
inline is used to define a macro-like function, meaning that the function's instructions will be replaced directly to its invoke in preprocessor step
This helps improve speed but paying by program size
As discuss in 1.3. Memory Layout there is a memory layer for dynamic memory allocation called heap. So how we manage dynamic memory allocation
There are some keyworks to manage dynamic memory
- malloc: dynamically allocates a single large block of memory, returns void pointer and has the default garbage value initially. Syntax:
ptr = (data-type*)malloc(byte-size) - calloc like malloc but has some differences: dynamically allocates the number of blocks of memory, blocks initialized with default value 0. Syntax:
ptr = (data-type*)calloc(size, element-size) - realloc: dynamically change the memory allocation of a previously allocated memory, maintains the already present value and new blocks will be initialized with the default garbage value. Syntax:
ptr = realloc(ptr, new byte-size) - free: dynamically de-allocate the memory. Syntax:
free(ptr)
Von Neumann is the computer architect that have intercommunity bus for program memory and data memory
Harvard is the opposite, so it have better performance
endian is a storage mechanism so it is opposite to our thinking
Little endian is from LSB to MSB and Big endian is the opposite
There are two memory types in computer: volatile and non-volatile
volatile with representation is RAM, also called memory, loses contents when power is off, a temporary workspace for data because it is high-speed
Some volatile memory:
- SRAM or static RAM: fast, small capacity, high energy consumption, used for caches
- DRAM or dynamic RAM: slower than SRAM, higher capacity, requires periodic refresh, used for main memory
non-volatile with representation is ROM, also called storage, preserves contents when power is off, used to store data
Some non-volatile memory (their name say all about them):
- PROM or programmable ROM and MROM or mask ROM
- EPROM or erasable programmable ROM
- EEPROM or electrically erasable programmable ROM
There are also some additional memory types:
- Cache: SRAM application
- Flash: erased a "block" at a time, limited number of program/erase cycles
- Bootloader: on power up, transfers data from non-volatile to volatile memory
- Disk
Memory component speed ordered from fastest to slowest: register, cache, main memory, non-volatile memory, hard-disk drives, optical disk, magnetic tapes
Our world is analog but computer's world is digital
analog refers to a continuously changing representation of a continuously variable quantity
digital refers to representing these variable quantities in terms of actual numbers, or digits
So we have to do something for computers to understand our world, there will be discuss in ADC & DAC part
Embedded system is computer system with a dedicated function embedded as a part of a complete device and have limited resources as memory, peripheral, speed, ...
They have some features as: small size, low per-unit cost, low power consumption, ...
This part will be added later
Sensors are devices that measure physical quantities
sensor can be seem as input of a system, it responsible to "read from physical world"
Actuators are devices that modify physical quantities
actuator can be seem as output of a system, it responsible to "write to physical world"
This part will be added later
Micro-processor:
Micro-controller
General-Purpose Input/Output Ports handles both incoming and outgoing digital signal
They can be INPUT or OUTPUT, LOW or HIGH
Also, they can be configurated for other functions, called alternate function
Pull-up resistor is pulled HIGH the GPIO when the button is not pressed and pulled LOW the GPIO when the button is pressed
Pull-down resistor is the opposite to pull-up
Open-drain make the GPIO can only be LOW or floating, so we have to use external pull-up resistor.
Analog to Digital Converter and Digital to Analog Converter are very important components in electronic equipments
ADC and DAC architect won't be mention here, let's go to their parameters
- Reference Voltage:
- Resolution:
- Quantization:
Pulse Width Modulation
A protocol usually come together with that protocol's interface so there are many people have ambiguous between protocol and interface. So, how are they different ?
protocol is a set of rules for devices to communicate with others as preamble, data length, conditions, crc, ... and they need to be agreed by all devices
interface is the way devices connect to others as wires, radio waves, ...
Synchronous and Asynchronous: these are two important term in communication, they imply to clock. Synchronous transmissions are synchronized by a clock and asynchronous transmissions are not
Wire and Wireless: just wire and wireless
Serial and Parallel: data transmission serial or parallel (e.g. one wire or multi wires)
Simplex, Half-duplex and Full-duplex: in simplex mode the signal is sent in one direction (only one device can sent data), in half-duplex the signal is sent in both directions but one at a time, in full-duplex signal is sent in both directions at the same time
Master and Slave: master will be the clock controller
Server and Client: client is requester and server serves
Universal Synchronous/Asynchronous Receiver-Transmitter protocol:
- Simplex, Half-duplex, Full-duplex
- Single Master - Single Slave (in UART there is not master and slave, both devices peer)
UART is asynchronous and USART is synchronous, their name said about it
In UART there are some definitions that must be the same on devices:
- Baudrate: transceive data rate
- Start bit
- Stop bit
- Data frame length: it can be 5, 6, 7 or 8 (even 9 when don't use parity bit)
- Parity bit: used to check data correction. When parity bit = 0 number of bit 1 must be even and when parity bit = 1 number of bit 1 must be odd unless the data is wrong
I have never used USART so I don't have information about it exclude it is synchronous
In addition, there is another mode called Multiprocessor UART, maybe I will add it later
Inter-Integrated Circuit, also called Two Wire Interface protocol:
- Half-duplex
- Multi Master - Multi Slave
- Synchronous
Serial Peripheral Interface protocol:
- Full-duplex
- Single Master - Multi Slave
- Synchronous
Controller Area Network protocol:
Direct Memory Access
Operating System is software or it can be seem as the intermediate between user and hardware that runs on a computing device and manages the hardware and software components that make up a functional computing system
Hardware management
Provide interface for users
Application installation
Connect hardware
Interaction between application and hardware
CPU scheduling
Process coordination and synchronization
Resource management
Access control and protect system
Integrity maintenance, error control and recovery
Base on processing:
- Uniprogramming OS
- Multiprogramming OS
- Time-sharing OS
- Parallel/multiprocessor/tightly-coupled OS
- Clustered/distributed/loosely-coupled OS
- RTOS
Process/thread management
Primary memory management
File management
I/O system management
Secondary memory management
Protect system
Command line interpreter system
Process is the active application instant, application become process when executable file loaded into memory (called process image)
Process layout is as same as Memory Layout In C Program
Process initialization step: allocate identifier, allocate memory for process load, initialize Process Control Block (PCB), setup necessary relation (e.g. arrange PCB to queue)
PCB is one of the most important data structure in OS to manage and control the execution of processes, it typically contains the following components:
- Process ID (PID): unique identifier assigned to each process by the operating system
- Program counter (PC): address of the next instruction to be executed by the process
- Process state: current state of the process, includes new, ready, running, waiting, terminated, suspended
- CPU registers: values of the processor registers that are being used by the process
- Memory management information: information about the process’s memory allocation, such as the base and limit registers
- I/O status information: information about the process's I/O operations, such as open files and network connections
- Accounting information: information about the resources used by the process, such as CPU time and memory usage
- CPU scheduling information: information about the process’s priority and scheduling requirements
Thread is a segment of a process or a lightweight process, threads share memory, data, resources together
There are 3 thread mapping model
- Multi user thread mapping to single kernel thread: any user thread blocked cause all user thread blocked, user threads can't run parallel because only one user thread can use kernel at one time
- Single user thread mapping to single kernel thread: create a user thread also create a kernel thread, better concurrency because any user thread blocked don't affect to other threads, limited number of threads
- Multi user thread mapping to multi kernel threads: solve the disavantages of the two below but hard to setting
Scheduling in OS is process scheduling with PCB: process scheduling is a crucial function in OS and PCB in OS plays a vital role in this function
Scheduler include several types:
- Long-term scheduler used to schedule job (or process), decide which program accepted to load into system to run, control system's multiprogramming, maintain both CPU-bound and I/O-bound process
- Short-term scheduler used to schedule CPU, decide which process in ready queue will take CPU to run next, trigger events: clock interrupt, I/O interrupt, OS call, signal, ...
- Mid-term scheduler (in time-sharing system) used to adjust system's multiprogramming (swap in - move process from memory to disk and swap out - move process from disk to memory) Scheduler will change the CPU control permit for the choosen process, includes:
- Context change
- User mode change
- Jump to the position in process to running proccess again Scheduling cause time consumption called dispatch latency, time that scheduler stop current process and start another process Scheduling criteria
- User-oriented: response time - time since process receive request to first request is responsed (time-sharing, interactive system) minimum, turnaround time: time since process loaded into system to process terminated minimum, waiting time: sum of process waiting time in ready queue minumum
- System-oriented: process utilization - schedule to make the CPU busiest maximum, fairness - all process treated peer, throughput - number of done process in a time maximum Scheduling features
- Selection function: used to choose which process in ready queue run usually based on priority, resources request, process
- Desicion mode: used to choose time to choose selection function for scheduling, there are two desicion mode: non-preemptive - when running process will run until end or blocked by I/O request and preemtive - running process can be interrupt to ready state, more cost but more better response time because no case that a process take CPU too long service time is the time process need CPU in a CPU-I/O period, processes have large service time are CPU-bound processes
These parts will be completed or distributed soon