Several implementation methods of print function in ARM

1Use the C library function printf.

step:

1) First you need to include the header file stdio.h.

2) Then define the file handle, which is actually an int variable encapsulated in a structure.

struct__FILE{inthandle;};

3) Define FILE__stdout; FILE is __FILE, which is defined by the stdio.h macro.

4) Implement the function

intfputc(intch,FILE*f){chartempch=ch;sendchar(tempch);returnch;}

5) Implement the function

intsendchar(intch){

if(ch=='n'){

while(!(console_tty_f->lsr&UART_LSR_THRE));

console_tty_f->dll_fifo=0x0d;

}

while(!(console_tty_f->lsr&UART_LSR_THRE));

return(console_tty_f->dll_fifo=ch);

}

As can be seen from the above code, printf is a blocking function. Waiting for the message to be sent may affect other processes. If you write a non-waiting print function, you can use the second method.

2 Use the C library function vsprintf and variable parameter function.

step:

1) Include the header files stdio.h and stdarg.h.

2) Write a variable parameter function.

voidprint(constchar*lpszFormat,...){charszBuffer[PRINT_BUF]={0};

va_listargs; enter; va_start(args, lpszFormat);

ret=vsprintf(szBuffer,lpszFormat,args);

va_end(args);

uart[console_uart].put_2_ring(console_uart,szBuffer,ret);

}

It can be seen that the library function vsprintf is used to format the input string and then send it when idle.

3. Complete parameter extraction and formatting by yourself.

step:

1) Define variable parameter list typedef char*va_list;

2) Define address alignment macro

#define_AUPBND(sizeof(int)-1)

#define_ADNBND(sizeof(int)-1)

#define_bnd(X,bnd)(((sizeof(X))+(bnd))&(~(bnd)))

3) Define variable length parameter extraction macro

#defineva_start(ap,A)(void)((ap)=(((char*)&(A))+(_bnd(A,_AUPBND))))

#defineva_arg(ap,T)(*(T*)(((ap)+=(_bnd(T,_AUPBND)))-(_bnd(T,_ADNBND)))

#defineva_end(ap)(void)0

4) Write a variable parameter function, as in step 2 of method 2.

5) Implement the vsprintf function. There are many implementation source codes, such as Linux, etc. But there is no support for floating points.