Memory and stack related concepts

Memory
The smallest storage unit in a computer is bytes , Memory is no exception , Each byte acts as a storage unit . In this way, the memory is divided into hundreds of millions of memory units , To facilitate the management of these units , The computer has chosen to number each memory unit , When you want to use a memory unit , Just specify the memory unit number （ Memory address ） that will do . In order to number as many addresses as possible , The computer uses a 4 Number of bytes to address memory , This is good for 2^32 Block memory for addressing , Each block is 1byte,2^32 byte = 4Gb, The memory size that can be addressed is 4G size .

Stack
When the program runs, it will not immediately allocate all memory for the program , Will initialize a certain size of memory area to the program , Memory blocks in this memory space can be used by programs , Used to store data in the program . This space is used from memory blocks with large addresses , To see which address memory unit is used by the current stack , stay ESP The current memory location used is recorded in the register , If you add a data , Will be added at this location , then ESP Continue to move forward one position . therefore ESP The current position of the register record , Is the position of the next data insertion . This pointer is called the stack top pointer , The pointer is added with the data （push） And take out (pop) Move . alike ,EBP The address of the beginning of a stack is recorded in , Become a pointer at the bottom of the stack , The address usually does not change , Only in opening up a new stack space , For example, when executing a function call , When you need to re record a new function space for a function , Will change the position of the bottom of the stack . But when changing the pointer at the bottom of the stack , The original position at the bottom of the stack will be put on the stack and saved , It is convenient to restore to the original bottom of the stack in the future .

Stack elevation
We know , In advanced languages , When we execute a function call , Will open up a new stack space , At this point, stack lifting is required , You need to change ESP and EBP Bottom of stack and bottom of stack pointer recorded in register , In this way, the value generated in the function will not affect the original stack space ( The yellow part ) The data of . The right figure shows the result of stack elevation , here push The data will come from 1005 Start adding ,pop Only to 1005 The data of this address , Can't go down any further pop, This will not affect the contents of the metastack .

Stack balance
Stack balancing is actually restoring stack contents , For example, when performing a function operation , Stack up , Open up new space for functions , When function execution ends , The bottom and top pointers should be restored to their original positions .

The brief process is

1. take ESP Value in register , Change to EBP Value , In this way, the top of the stack is at the current bottom of the stack .ESP=1005
2. take EBP Get the value in the memory address pointed to by the register , namely 1000, take EBP The value of the register is changed to this value 1000. such EBP Back to 1000 The location of .

Thus, the stack recovery is completed , Stack balance is achieved

Data width
There are three data widths in assembly language , Respectively

BYTE： One byte wide , Occupy 8 position

WORD： word , Two bytes wide , Occupy 16 position

DOUBLEWORD： Two words , Four byte width , Occupy 32 position .

Large and small end mode
Memory is addressed from small to large , When storing a multi byte data , If the small address of memory stores the high order of multi byte data , It's the big end model . in other words , First get the high bit of the data for storage , Then take the low order storage . For example, a string "abcd",a Is the high order of the data ,d It's low , The memory will take out four consecutive memory spaces to store this string . If this continuous memory is stored in a small address a And then in turn bcd, It's the big end model . If the low order data is stored in the small address , Small end mode .

It is up to the compiler to store the data in big end or small end mode .

add , All registers contain addresses