Linker --- linker

The linker

1. background

​ For regular use IDE The developer of the , Usually it's all right to click a button , Although this greatly simplifies the process , But for us C The language is very, very unfriendly to the underlying developers , A lot of details are masked , It's terrible not to know the inside details , therefore , It is very necessary to explore its principle .

2. Knowledge reserve

• take C The code generation executable object file needs to go through 4 A step
• We need to compile the tool chain to translate the code we write
• The compilation tool chain is just a program , It is just a program used to translate the code that we can read into machine code
  
• Preprocessing ： Use the preprocessor to put the header file 、 Macro definition , Expand directly in the current file intact , And generate .i file
• compile ： Use the compiler to .i File translated into assembly language file .s
• assembly ： Using the assembler to .s Translate into a relocatable target file （relocatable obj file）.o
• The linker ： The debut ！ This paper focuses on the next linker ld, Because linkers are very important , It needs to be used frequently , A set of .o The files are combined , Create an executable object file （executable object file）

3. What is a linker （Linker）

​ Borrow the explanation in the book "understanding computer systems in depth" ： Take a set of relocatable target files and command line parameters as input , Generate a fully linked 、 Executable target files that can be loaded and run as output . The input relocatable object file consists of a variety of different code and data sections （section） form , Each section is a continuous sequence of bytes . Instructions in a section , The initialized global variables are in another section , Uninitialized variables are in another section .

​ A section is actually a region

​ It's kind of similar Linux Archive tool under tar, Assemble multiple files , But it's just Linker One of the functions of , Another function is to reposition ！

4. What is in the relocatable object file ？

​ The above picture is the object file

• .text： Instructions （ Assembly instruction ,C Function after translation , A local variable is actually the operand of an instruction ） Put it in text paragraph
• .data： The initialized global variables and static variables are placed data paragraph
• .bss： Not initialized or initialized to 0 Global and static variables of
• .rodata： Use const Keyword modifies the variable
• .symtab： The symbol table ！ It stores the definitions in the program 、 Referenced function 、 Global symbol information . every last .o They all have a symbol table , Record their own definitions and global variable symbols 、 Self defined function symbols and symbols referring to others （ Variables and functions ）, Because of the existence of this table , The linker can put a large number of .o Files are linked to generate executable files .
• Others don't go into details

5. Symbols and symbol tables

​ Each relocatable target module m They all have a symbol table , It contains m Information about the symbols defined and quoted . In the context of the linker , There are three different symbols ：

• By module m Define global symbols that should be able to be referenced by other modules

• Defined by other modules and defined by modules m The global symbol of the reference . These symbols are called external symbols , Corresponding to the non static... Defined in other modules C Functions and global optimizations

• Only by module m Local symbols defined and referenced . They correspond to bands static Attribute C Functions and global variables The amount . These symbols are in the module m You can see it anywhere in the world , But it can't be referenced by other modules .

6. Symbol resolution

​ Symbol table Linker There are two kinds of information , One is the symbol that the current target file can provide to other target files , The other is the symbols that other files need to provide to the current target file . I have this information , The linker will be able to make symbolic decisions .

​ For those references defined in the same module （ Same target file ） A reference to a local symbol in , Symbolic parsing is very simple and clear . The compiler allows only one definition for each local symbol in each module . Static local variables also have local linker symbols , The compiler also ensures that they have unique names .

​ Reference resolution of global symbols is much more difficult , When the compiler encounters a symbol that is not defined in the current module , It is assumed that the symbol is defined in some other symbol , Generate a linker symbol table entry , And give it to the linker . If the linker cannot find the definition of the referenced symbol in any of its input modules , Just output an error message , And terminate .

7. Link static libraries

Not over and on ......

8. relocation

Not over and on ......