Introduction to Keil C

The use of C language to program microcontrollers is becoming too common. And most of the time its not easy to buld an application in assembly which instead you can make easily in C. So Its important that you know C language for microcontroller which is commonly known as Embedded C. As we are going to use Keil C51 Compiler, hence we also call it Keil C.


Keil C51 compiler adds few more keywords to the scope C Language:

_at_ far sbit
alien idata sfr
bdata interrupt sfr16
bit large small
code pdata _task_
compact _priority_ using
data reentrant xdata

Description: The variable will be stored in internal data memory of controller. example:

unsigned char data x;
unsigned char idata y;

Description: The variable will be stored in bit addressable memory of controller. example:

unsigned char bdata x;
//each bit of the variable x can be accessed as follows
x ^ 1 = 1; //1st bit of variable x is set
x ^ 0 = 0; //0th bit of variable x is cleared

Description: The variable will be stored in external RAM memory of controller. example:

unsigned char xdata x;

Description: This keyword is used to store a constant variable in code memory. Lets say you have a big string which is not going to change anywhere in program. Wasting ram for such string will be foolish thing. So instead we will make use of the keyword "code" as shown in example below. example:

unsigned char code str="this is a constant string";

Description: This keyword will store the variable in paged data memory. This keyword is used occasionally. example:

unsigned char pdata x;

Description: This keyword is used to store a variable on a defined location in ram. example:

unsigned char idata x _at_ 0x30;
// variable x will be stored at location 0x30
// in internal data memory

Description: This keyword is used to define a special bit from SFR (special function register) memory. example:

sbit Port0_0 = 0x80;
// Special bit with name Port0_0 is defined at address 0x80

Description: sfr is used to define an 8-bit special function register from sfr memory. example:

sfr Port1 = 0x90;
// Special function register with name Port1 defined at addrress 0x90

Description: This keyword is used to define a two sequential 8-bit registers in SFR memory. example:

sfr16 DPTR = 0x82;
// 16-bit special function register starting at 0x82
// DPL at 0x82, DPH at 0x83

Description: This keyword is used to define register bank for a function. User can specify register bank 0 to 3. example:

void function () using 2
	// code
// Funtion named "function" uses register bank 2 while executing its code

Description: This keyword will tells the compiler that function described is an interrupt service routine. C51 compiler supports interrupt functions for 32 interrupts (0-31). Use the interrupt vector address in the following table to determine the interrupt number.


Vector Address Locations


void External_Int0() interrupt 0

Memory Models

There are three kind of memory models available for the user:

All variables in internal data memory.
Variables in one page, maximum 256 variables (limited due to addressing scheme, memory accessed indirectly using r0 and r1 registers);
All variables in external ram. variables are accessed using DPTR.

Depending on our hardware configuration we can specify the momory models as shown below:

//For Small Memory model
#pragma small
//For Compact memory model
#pragma compact
//For large memory model
#pragma large

Pointers in Keil C

Pointers in keil C is are similar to that of standard C and can perform all the operations that are available in standard C. In addition, keil C extends the operatability of pointers to match with the 8051 Controller architecture. Keil C provides two different types of pointers:

  • Generic Pointers
  • Memory-Specific Pointers

Generic Pointers

Generic Pointers are declared same as standard C Pointers as shown below:

char *ptr;	//Character Pointer
int *num;	//Integer Pointer

Generic pointers are always stored using three bytes. The first byte is the memory type, the second byte is the high-order byte of the offset, and the third byte is the low-order byte of the offset. Generic pointers maybe used to access any variable regardless of its location.

Memory-Specific Pointers

Memory specific pointers are defined along with memory type to which the pointer refers to, for example:

char data *c;
//Pointer to character stored in Data memory
char xdata *c1;
//Pointer to character stored in External Data Memory.
char code *c2;
//Pointer to character stored in Code memory

As Memory-Specific pointers are defined with a memory type at compile time, so memory type byte as required for generic pointers is not needed. Memory-Specific pointers can be stored using 1 byte (for idata, data, bdata and pdata pointers) or 2 bytes (for code and xdata pointers).

The Code generated by keil C compiler for memory-specific pointer executes mroe quickly than the equivalent code generated for a generic pointer. This is because the memory area accessed by the pointer is known at the compile time rather at run-time. The compiler can use this information to optimize memory access. So If execution speed is your priority then it is recommended to use memory-specific pointers. Generic pointers and Memory-Specific pointers can be declared with memory area in which they are to be stored. For example:

//Generic Pointer
char * idata ptr;
//character pointer stored in data memory
int * xdata ptr1;
//Integer pointer stored in external data memory
//Memory Specific pointer
char idata * xdata ptr2;
//Pointer to character stored in Internal Data memory
//and pointer is going to be stored in External data memory
int xdata * data ptr3;
//Pointer to character stored in External Data memory
//and pointer is going to be stored in data memory

Functions in Keil C

Keil C compiler provides number of extensions for standarad C function declerations. These extensions allows you to:

  • Specify a function as an interrupt procedure
  • Choose the register bank used
  • Select memory model

Function Declaration

[Return_type] Fucntion_name ( [Arguments] ) [Memory_model] [reentrant] [interrupt n] [using n]

The type of value returned from the function. If return type of a function is not specified, int is assumed by default.
Name of function
Arguments passed to function
Optional Stuff

These are options that you can specify along with function declaration. Memory_model: explicit memory model (Large, Compact, Small) for the function. Example:

int add_number (int a, int b) Large
To indicate if the function is reentrant or recursive. This option is explained later in the tutorial.
Indicates that function is an interrupt service routine. This option is explained later in the tutorial.
Specify register bank to be used during function execution. We have three register banks in 8051 architecture. These register banks are specified using number 0 for Bank 0 to 3 for Bank 3 as shown in example
void function_name () using 2
	//function uses Bank 2
	//function code
Interrupt Service Routines

A function can be specified as an interrupt service routine using the keyword interrupt and interrupt number. The interrupt number indicates the interrupt for which the function is declared as service routine.

Following table describes the default interrupts:


8051 Interrupt vector

As 8051 vendors create new parts, more interrupts are added. Keil C51 compiler supports interrupt functions for 32 interrupts (0-31). Use the interrupt vector address in the following table to determine the interrupt number.


Interrupt vector

The interrupt function can be declared as follows:

void isr_name (void) interrupt 2
	// Interrupt routine code

Please make sure that interrupt service routines should not have any arguments or return type except void.

Reentrant Functions

In ANSI C we have recursive function, to meet the same requirement in embedded C, we have reentrant function. These functions can be called recursively and can be called simultaneously by two or more processes. Now you might be thinking, why special definition for recursive functions?

Well you must know how these functions work when they are called recursively. when a function is running there is some runtime data associated with it, like local variables associated with it etc. when the same function called recursively or two process calls same function, CPU has to maintain the state of function along with its local variables. Reentrant functions can be defined as follows:

void function_name (int argument) reentrant
	//function code

Each reentrant function has reentrant stack associated with it, which is defined by startup.A51 file. Reentrant stack area is simulated internal or external memory depending upon the memory model used:

  • Small model reentrant functions simulate reentrant stack in idata memory.
  • Compant model reentrant functions simulate reentrant stack in pdata memory.
  • Large model reentrant functions simulate reentrant stack in xdata memory.
Real-time Function Tasks

Keil or C51 provides support for real-time operating system (RTOS) RTX51 Full and RTX51 Tiny. Real-time function task are declared using _task_ and _priority_ keywords. The _task_ defines a function as real-time task. The _priority_ keyword specify the priority of task.

Fucntions are declared as follows:

void func (void) _task_ Number _priority_ Priority


is task ID from 0 to 255 for RTX51 Full and 0 to 15 for RTX51 Tiny.
is priority of task.

Real-time task functions must be declared with void return type and void argument list (say no arguments passed to task function).

Writing First C program in Keil

Basic of a C program

As we already discussed, Keil C is not much different from a normal C program. If you know assembly, writing a C program is not a problem, only thing you have to keep in mind is forget your controller has general purpose registers, accumulators or whatever. But do not forget about Ports and other on chip peripherals and related registers to them.

In basic C, all programs have atleast one function which is entry point for your application that function is named as \"main\" function. Similarly in keil, we will have a main function, in which all your application specific work will be defined. Lets move further deep into the working of applications and programs.

When you run your C programs in your PC or computer, you run them as a child program or process to your Operating System so when you exit your programs (exits main function of program) you come back to operating system. Whereas in case of embedded C, you do not have any operating system running in there. So you have to make sure that your program or main file should never exit. This can be done with the help of simple while(1) or for(;;) loop as they are going to run infinitely. Following layout provides a skeleton of Basic C program.

void main()
	//Your one time initialization code will come here
	while (1) {
		//while 1 loop
		//This loop will have all your application code
		//which will run infinitely

When we are working on controller specific code, then we need to add header file for that controller. I am considering you have already gone through "Keil Microvision" tutorial. After project is created, add the C file to project. Now first thing you have to do is adding the header file. All you have to do is right click in editor window, it will show you correct header file for your project.

Figure below shows the windows context for adding header file to your c file.


Include Header file in Keil

Writing Hardware specific code

In harware specific code, we use hardware peripherals like ports, timers and uart etc. Do not forget to add header file for controller you are using, otherwise you will not be able to access registers related to peripherals. Lets write a simple code to Blink LED on Port1, Pin1.

#include <REGx51.h>
//header file for 89C51
void main()
	//main function starts
	unsigned int i;

	//Initializing Port1 pin1
	P1_1 = 0; //Make Pin1 o/p

	while (1) {
		//Infinite loop main application
		//comes here
			; //delay loop

		P1_1 = ~P1_1;
		//complement Port1.1
		//this will blink LED connected on Port1.1

You can now try out more programs. "Practice makes a man perfect".

Writing C and Assembly together

Interfacing C program to Assembler

You can easily interface your programs to routines written in 8051 Assembler. All you need to do is follow few programming rules, you can call assembly routines from C and vice-versa. Public variables declared in assembly modules are available to your C program.

There maybe several reasons to call an assembly routine like faster execution of program, accessing SFRs directly using assembly etc. In this part of tutorial we will discuss how to write assembly progarms that can be directly interfaced with C programs.

For any assembly routine to be called from C program, you must know how to pass parameters or arguements to fucntion and get return values from a function.

Segment naming

C51 compiler generates objects for every program like program code, program data and constant data. These objects are stored in segments which are units of code or data memory. Segment naming is standard for C51 compiler, so every assembly program need to follow this convention.

Segment names include module_name which is the name of the source file in which the object is declared. Each segment has a prefix that corresponds to memory type used for the segment. Prefix is enclosed in question marks (?). The following is the list of the standard segment name prefixes:


C51 module prefix

Data Objects

Data objects are the variables and constants you declare in your C programs. The C51 compiler generates a saperate segment for each memory type for which variable is declared. The following table lists the segment names generated for different variable data objects. Data objects segment prefix


Data Segment Prefix

Program Objects

Program onjects includes code generated for C programs functions by C51 compiler. Each function in a source module is assigned a separate code segment using the ?PR?function_name?module_name naming convention. For example, for a function name send_char in file name uart.c will have a segment name of ?PR?SEND_CHAR?UART.

C51 compiler creates saperate segments for local variables that are declared within the body of a function. Segment naming conventions for different memory models are given in following tables:


Small model segment naming convention


Compact model segment naming convention


Large model segment naming convention

Function names are modified slightly depending on type of function (functions without arguments, functions with arguments and reentrant functions). Following tables explains the segment names:


function segment naming convention

Advanced C programming

Function Parameters

C51 make use of registers and memory locations for passing parameters. By default C function pass up to three parameters in registers and further parameters are passed in fixed memory locations. You can disable parameter passing in register using NOREGPARMS keyword. Parameters are passed in fixed memory location if parameter passing in register is disabled or if there are too many parameters to fit in registers.

Parameter passing in registers

C functions may pass parameter in registers and fixed memory locations. Following table gives an idea how registers are user for parameter passing.


parameter passing to functions

Following example explains a little more clearly the parameter passing technique:


example parameter passing to functions

Parameter passing in Fixed Memory Locations

Parameters passed to assembly routines in fixed memory lcoation use segments named

All except bit parameters are defined in this segment.
Bit parameters are defined in this segment.

All parameters are assigned in this space even if they are passed using registers. Parameters are stored in the order in which they are declared in each respective segment.

The fixed memory locations used for parameters passing may be in internal data memory or external data memory depending upon the memory model used. The SMALL memory model is the most efficient and uses internal data memory for parameter segment. The COMPACT and LARGE models use external data memory for the parameter passing segments.

Fucntion Return Values

Function return values are always passed using CPU registers. The following table lists the possible return types and the registers used for each.


function return values


Following example shows how these segment and function decleration is done in assembler.

	;Assembly program example which is compatible
;and called from any C program
;lets say asm_test.asm is file name
name asm_test
;We are going to write a function
;add which can be used in c programs as
; unsigned long add(unsigned long, unsigned long);
; as we are passing arguments to function
;so function name is prefixed with '_' (underscore)
;code segment for function "add"
?PR?_add?asm_test segment code
;data segment for function "add"
?DT?_add?asm_test segment data
;let other function use this data space for passing variables
public ?_add?BYTE
;make function public or accessible to everyone
public _add
;define the data segment for function add
rseg ?DT?_add?asm_test
parm1:	DS 4	;First Parameter
parm2:	ds 4	;Second Parameter
;either you can use parm1 for reading passed value as shown below
;or directly use registers used to pass the value.
rseg ?PR?_add?asm_test
;reading first argument
mov parm1+3,r7
mov parm1+2,r6
mov parm1+1,r5
mov parm1,r4
;param2 is stored in fixed location given by param2
;now adding two variables
mov a,parm2+3
add a,parm1+3
;after addition of LSB, move it to r7(LSB return register for Long)
mov r7,a
mov a,parm2+2
addc a,parm1+2
;store second LSB
mov r6,a
mov a,parm2+1
addc a,parm1+1
;store second MSB
mov r5,a
mov a,parm2
addc a,parm1
;store MSB of result and return
;keil will automatically store it to
;varable reading the resturn value
mov r4,a

Now calling this above function from a C program is very simple. We make function call as normal function as shown below:

extern unsigned long add(unsigned long, unsigned long);
void main()
	unsigned long a;
	a = add(10,30);
	//a will have 40 after execution

Powered by MediaWiki
This page was last modified on 6 April 2015, at 10:54.
ArrayContent is available under Creative Commons Attribution Non-Commercial Share Alike unless otherwise noted.