Home

Hitachi Microcomputer Support Software SH Series C Compiler USER S MANUAL

Contents

1. Define section R having the same size as section D in the output load module To reference the symbol allocated to section D copy the contents of section D to section R and reference to the symbol in section R Sections D and R are allocated to initialized data section in ROM and RAM respectively Allocate sections P C and D to internal ROM starting from address 0x400 and allocate sections R and B to RAM starting from address 0x9000000 62 2 2 Dynamic Area Allocation 2 2 1 Dynamic Areas Two types of dynamic areas are used Stack area Heap area used by the memory allocation library functions 2 2 2 Dynamic Area Size Calculation Stack Area The stack area used in C programs is allocated each time a function is called and is deallocated each time a function is returned The total stack area size is calculated based on the stack size used by each function and the nesting of function calls e Stack area used by each function The size of stack used by each function can be determined from the object list frame size output by the C compiler However note that this does not account for the size of parameters to be pushed onto the stack when a function is called Accordingly the parameter size must be added to stack area size The following example shows the object list stack allocation and stack size calculation method Example The following shows the object list and stack size calculation in a C program exter
2. Definition of static variables used in low level interface routines x KR KKK KKK KKK KKK KKK kok K kk KK OK k kk OK K kk KK OK OK KK OK KK KK KK char flmod FLMAX Open file mode specification area static union long dummy Dummy for 4 byte boundary char heap HEAPSIZE Declaration of the area managed by sbrk Jheap area static char brk char sheap area End address of area assigned by sbrk x 172 KR KKK KKK KKK KKK KKK kok K kk KK KK K kk OK KK OK OK OK KK OK KK KK KK open file open Return value File number Pass 1 Failure x7 KR KKK KKK KKK k kk KKK RK KK KKK K kk OK OK OK OK KK OK OK KK KKK int open char name File name int mode File mode i Check mode depending on file name and return file numbers if strcmp name stdin 0 Standard input file if mode amp O_RDONLY 0 return 1 flmod STDIN mode return STDIN else if strcmp name stdout 0 Standard output file if mode amp O_WRONLY 0 return 1 flmod STDOUT mode return STDOUT else if strcmp name stderr 0 f Standard error file if mode amp O_WRONLY 0 return 1 flmod STDERR mode return STDERR return 1 Error RR KKK KKK KKK KKK KKK kok K kk KK kk k kk K kk K KK KK OK OK OK KKK OK OK KK KK kk close File close Return value 0 Pass 1 Filure RR KKK KKK KKK KKK KKK k kk K kk K K
3. 2020 System identifier name redefined The name of the defined symbol is the same as that of the run time routine S Continues processing as a unique symbol P Define the symbol with a different name from that of the run time routine 2100 Multiple storage classes Two or more storage class specifiers are used in a declaration S Uses the first storage class specifier and ignores others P Specify the correct storage class specifier Error No Message 2101 Address of register Explanation The unary operator amp is used on a register variable S Assumes that the auto storage class is specified for the variable and continues processing P Modify the declaration so that the storage class of the variable is auto 2102 Illegal type combination A combination of type specifiers is illegal S Uses the first and longest legal combination of type specifiers and ignores the rest P Change the type specifiers to a legal combination 2103 Bad self reference structure A struct or union member has the same data type as its parent S Assumes the data type of the member is int P Declare the correct data type for the member 2104 Tllegal bit field width A constant expression indicating the width of a bit field is not an integer or it is negative S Ignores the bit field width specification and assumes that the member is not a bit field P Specify the c
4. 2147483648 to 2147483647 _4 bytes unsigned int _0 to 4294967295 _4 bytes long 2147483648 to 2147483647 _4 bytes unsigned long 0 to 4294967295 4 bytes 147 A 1 6 Floating Point Numbers Table A 7 Floating Point Number Specifications Item _C Compiler Specification Data that can be represented as The float double and long double are provided as floating floating point type and value point types Data converted from double or long See section A 3 Floating Point Number double to float Specifications for details on floating point Internal representation of floating point numbers internal representation conversion specifications and data operation specifications Table A 8 shows the limits on _tepresenting floating point numbers Table A 8 Limits on Floating Point Numbers Limit Item _Decimal 1 Internal Representation Maximum float 3 4028235677973364e 38f T 7 ffFF 3 4028234663852886e 38f i Positive minimum float 7 0064923216240862e 46f 00000001 i 1 4012984643248171e 45f Maximum double or long double 1 7976931348623158e 308 7fefffffffffffft 1 7976931348623157e 308 i Positive minimum double or long 4 9406564584124655e 324 0000000000000001 double 4 9406564584124654e 324 Note 1 Limits on decimal is non zero minimum value or maximum value not infinitive value Values within indicate theoritical values 148 A 1 7 Arrays and Pointers Table A 9 Array and Pointer Specifications Item Int
5. S Sy hats ae E NERA EVA VE eae ag V skt dete itech ae HAS dea VA AE u s E SH series simulator debugger interface routine x Only standard I O files stdin stdout stderr are supported a RR KKK KKK k kk k kk kok k kk K kk K kk kk k kk K kk K k k K KX KX KA KA AX AKA OK k k kk include lt string h gt file number define STDIN 0 Standard input console x define STDOUT 1 Standard output console 4 define STDERR 2 Standard error output console define FLMIN 0 Minimum file number x define FLMAX 3 Maximum number of files x file flag define O RDONLY 0x0001 Read only x define O WRONLY 0x0002 Write only x define O RDWR 0x0004 Both read and write A special character code define CR 0x0d Carriage return define LF 0x0a Line feed x size of area managed by sbrk define HEAPSIZE 1024 KR KKK KKK k kk KK KK kok K kk K OK KK KOK OK KK OK OK OK OK KK KK KK Declaration of reference function ir Reference of assembly program in which the simulator debugger input or output characters to the console KK KKK KKK KKK k kk KKK K kk K kk k kk k kk K kk OK KK OK KK KK OK OK KK KK extern void charput char One character input x extern char charget void One character output kok k kok k k kk k kk kok kok K kk K kk K kk k kk K kk K k k K KK OK KA KA AK AKA OK X k k kk Definition of static variable
6. identifier must not start witha _ A run time routine in a standard library must be used Check if a standard I O library library as the input library at linkage Create low level cation Part Il Progra mming 4 2 1 3 Execution routine in a standard library Part III System Installation 2 1 2 Part Ill System Installa function is used in the C program debug option must be specified interface routines for linking Specify debug option tion 4 6 Part I Overview and source level cannot be performed at both compilation and linkage A linkage editor of Ver 5 0 or at both compilation and linkage Use a linkage editor of higher must be used 54 Ver 5 0 or higher Operation 3 3 PART III SYSTEM INSTALLATION Section 1 Overview Part III describes how to install object programs generated by the C compiler on a SH system Before installation memory allocation and execution environment for the object program must be specified e Memory allocation Stack area heap area each section of a C compiler generated object program must be allocated in ROM or RAM on a SH system e Execution environment setting for a C compiler generated object program The execution environment can be specified by the register initialization processing memory area initialization and C program initiation processing These must be written by assembly language If C library function
7. A label is defined more than once in a function S Ignores redundant label definitions P Keep one label name and delete or modify the other 2303 Illegal continue A continue statement is specified outside a while for or do statement S Ignores the continue statement P Only use the continue statement in a while for or do statement 2304 Illegal break A break statement is specified outside a while for do or switch statement S Ignores the break statement P Only use the break statement in a while for do or switch statement 2305 Void function returns value A return statement specifies a return value for a function with a void return type S Ignores the return statement expression P For a function with a void return type do not specify an expression in a return statement or do not use the return statement 2306 Case label not constant A case label expression is not an integer constant expression S Ignores the case label P Use an integer constant expression for the case label 2307 Multiple case labels Two or more case labels with the same value are used in one switch statement S Ignores redundant case labels P Modify the switch statement so that each case label has a unique value 122 Error No Message 2308 Multiple default labels Explanation Two or more default labels are specified for one switch statement S Ignores redund
8. FILE NAME m0260 c Seq File Lin 0 1 2 3 4 5 1 m0260 c 1 include header h 4 m0260 c 2 5 m0260 c 3 int sum2 void 6 m0260 c 4 int j 7 m0260 c 5 8 m0260 c 6 ifdef SMALL 9 m0260 c 7 j SML INT 10 m0260 c 8 else 11 m0260 c 9 j LRG INT 12 m0260 c 10 endif 13 m0260 c 11 amp 14 m0260 c 12 return j continuel23456789012345678901234567 V W X 2345678901234567890 Figure 3 2 Source Listing Output for show include and expansion kk kkk k kkk SOURCE LISTING KKKKKKKKKKKK FILE NAME m0260 c m Seg File Lin 0 1 2 3 4 5 46 1 m0260 c 1 include header h 2 header h 1 define SML_INT 1 3 header h 2 define LRG_INT 100 4 m0260 c 2 5 m0260 c 3 int sum2 void 6 m0260 c 4 int j 7 m0260 c 5 8 m0260 c 6 ifdef SMALL 9 m0260 c 7 X j SML INT 10 m0260 c 8 Z felse 11 m0260 c 9 E j 100 12 m0260 c 10 endif 13 m0260 c 11 14 m0260 c 12 return j continuel23456789012345678901234567 Object Listing Figure 3 3 shows an example of an object listing 13 Description V W X Y Listing line number Source program file name or include file name Line number in source program or include file Source program lines resulting from an include file expansion when show include is specified Source program lines that are not to be compiled due to conditional directives such as ifdef and elif are marked with an X when show expansion is specified
9. 66 Listing file listfile lt listing file name gt 3 Output nolistfile No output Object file objectfile lt object file name gt Output Object code machine code Program in machine language is output progam format asmcode Assembly source progam is output Debug debug Output information nodebug No output Macro name define lt macro name gt lt name gt lt name gt is defined as lt macro name gt lt macro name gt lt constant gt lt constant gt is defined as lt macro name gt lt macro name gt 4 lt macro name gt is assumed to be defined Include file include lt path name gt 5 Include file destination path name is specified Multi specification is possible 4 Section name section program lt section name gt Program area section name is specified 5 const lt section name gt Constant area section name is specified data lt section name Initialized data area section name is specified bss lt section name gt Non initialized data area section name is specified Default p P c C d D b B Help message help Output Notes 1 show option is invalid when listfile is specified 2 The assignments of show width 0 or show length 0 are interpreted as below show width 0 No line feed is performed until line feed code is output show length 0 Maximum line number is not specified and p
10. An error occurs if the file is on an interactive device such as a console or printer the new offset value is negative or the new offset value exceeds the file size in the case of or above If Iseek correctly determines a new file position the new offset value is returned This value indicates the new read write position relative to the beginning of the file Otherwise the lseek routine returns a value of 1 87 f sbrk routine Purpose Allocates a memory area Interface char sbrk unsigned long size Parameters No Name Type Meaning 1 size unsigned long Size of the area to be allocated Return value Type Pointer to char Normal Start address of the allocated area Abnormal char 1 Explanation The size of the area to be allocated is given as a parameter Create the sbrk routine so that consecutive calls allocate consecutive areas beginning with the lowest available address An error will occur if there is insufficient memory If the routine ends normally it returns the start address of the allocated area Otherwise the routine returns char 1 88 PART IV ERROR MESSAGES Section 1 Error Messages Output by the C Compiler The C compiler checks C source programs for errors This section explains the format and meaning of error messages that may be generated during compile time and gives appropriate programmer responses 1 1 Error Message Format Error messages are output t
11. Lines containing a macro expansion due to define directives are marked with an E when show expansion is specified If a source program line is longer than the maximum listing line the continuation symbol is used to indicate that the source program line is extended over two or more listing lines 14 Figure 3 3 Object Listing KKKKKKKKKKKK SOURCE LISTING KKKKKKKKKKKK FILE NAME m0251 c M Seg File Lin 0 1 2 3 4 54 1 m0251 J extern int sum int 2 m0251 c 2 3 m0251 c 3 int 4 m0251 c 4 sum int x 5 m0251 c 5 6 m0251 c 6 Ine sks 7 m0251 c 7 int j 8 m0251 c 8 9 m0251 c 9 j 0 10 m0251 c 10 for i 0 i lt x i 11 m0251 c 11 j i 12 m0251 c 12 13 m0251 c 13 return j 14 m0251 c 14 KKKKKKKKKKKK OBJECT LISTING KKKKKKKKKKKK FILE NAME m0251 c SCT OFFSET CODE C LABE INSTRUCTION OPERAND COMMENT V W X Y P Firle m0251 A Line 4 block 00000000 _sum function sum frame size 8 X 00000000 7FF8 ADD 8 R15 File m025l c Line 5 block File m0251 c Line 9 expression statement 00000002 300 MOV 0 R3 00000004 2F32 MOV L R3 R15 File m0251 c Line 10 r for 00000006 300 MOV 0 R3 00000008 1F31 MOV L R3 4 R15 0000000A A009 BRA L104 0000000C 0009 NOP Description v Section attribute P C D B of each section w The offset indicates the offset address relative to the beginning of each section x Contents of the offset
12. The same type of interrupt specifications have been specified more than once S Ignore the interrupt function declaration P Delete one of the interrupt specifications 2703 Illegal pragma interrupt declaration 127 The specifications for the interrupt function declaration pragma interrupt are not correct S Ignores the interrupt function declaration P Specify correctly Error No Message Explanation 2704 Illegal reference to An interrupt function is illegally referenced interrupt function S Ignores the attempt to reference the interrupt function P An interrupt function cannot normally be referenced Define another function for referencing 2105 Illegal parameter in There are different parameter types in an interrupt function interrupt function S Ignores the interrupt function declaration P Specify correct parameter types 2706 Missing parameter The variables used in the option specification declaration in interrupt by the interrupt function are not specified function S Ignores the interrupt function declaration P Declare the variables before declaring the interrupt function declaration pragma interrupt 2707 Parameter out of range in Parameter tn in an interrupt function exceeds interrupt function 256 S Ignores the value of parameter tn P Modify the value of parameter tn so it does not exceed 256 2800 Illegal parameter number The number of parameters used in an intrins
13. default return break or continue P Correct the token sequence 2501 Division by zero An integer is divided by zero in a constant expression S Assumes a result value of zero and continues processing P Modify the constant expression so that an integer is not divided by zero 2600 character string An error message specified by string literal error is output to the list file if nolist option is not specified S Continues processing 126 Error No Message 2650 Invalid pointer reference Explanation The specified address does not match the required byte alignment S Uses the address with the lowest bit masked when accessing word data and the address with the lowest two bits masked when accessing long word data P Specify the address so as to match the byte alignment 2700 Function function name in pragma interrupt already A function already declared as a normal function has been specified with the interrupt declared function declaration pragma interrupt S Ignores the interrupt function declaration P Declare the function as an interrupt function before it is declared as a normal function 2701 Multiple interrupt for one function A function has been declared as an interrupt function with pragma interrupt more than once S Ignores the interrupt function declaration P Delete the declarations following the first one 2702 Multiple pragma interrupt options
14. fileno If the device such as a disk where a file is stored becomes full data less than the specified byte length is written to the file If zero is returned as the byte length of data actually written several times the routine assumes that the device is full and sends a return value of 1 The file read write position must be updated using the byte length of data actually written If the routine ends normally it returns the byte length of data actually written Otherwise the routine returns a value of 1 86 e Iseek routine Purpose Determines the next read write position in a file Interface long lseek int fileno long offset int base Parameters No Name Type Meaning fileno int File number of the target file 2 offset long Offset in bytes from specified point in the file 3 base int Base used for offset bytes Return value Type long Normal The offset bytes from the beginning of the file for the next read write position Abnormal 1 Explanation The Iseek routine determines the next read write position as an offset in bytes The next read write position is determined according to the third parameter base as follows Base 0 The second parameter gives the new offset relative to the beginning of the file Base 1 The second parameter is added to the current position to give the new offset Base 2 The second parameter is added to the file size to give the new offset
15. for i onexit count 1 i gt 0 i Sequencially executes functions defined by onexit onexit buf i _CLOSEALL Closes all files opened longjmp _init_env 1 Returns to the environment saved by the set jmp 170 Note To return to the environment before program execution create the callmain function and call the callmain function instead of calling the main function from the init routine as shown below include lt setjmp h gt jmp_buf _init_env int _exit_code void callmain Saves current environment by setjmp function and calls the main function Terminates C program if a termination code is returned from the exit function if setjmp init env exit code main D 3 Creating abort Routine To terminate the routine abnormally the program must be terminated by a abort routine prepared according to the user system specifications The following shows an example of abort routine in which an error message is output to the standard output device closes all files enters endless loop and wait for reset Example include lt stdio h gt extern void abort void extern void CLOSEALL void abort printf program is abort n Outputs message _CLOSEALL Closes all files while l Enters endless loop 171 Appendix E Examples of Low Level Interface Routine RR KKK KKK KKK k kk KKK kk K kk KK KKK KK KK OK KK OK KK KK KK KK ss lowsrc c
16. fread fwrite fseek ftell rewind perror 1200 Invalid radix An invalid radix was specified strtol atoi atol 1202 Number too long The specified number exceeds 17 strtod fscanf scanf digits sscanf atof 1204 Exponent too large The specified exponent exceeds strtod fscanf scanf three digits sscanf atof 1206 Normalized The exponent exceeds three digits strtod fscanf scanf exponent too large when the character string is normalized to the IEEE standard decimal format 141 sscanf atof Functions to Set for an in input output function using format Error No Message Explanation Error Numbers 1210 Overflow out of A float type decimal value is out of strtod fscanf scanf float range overflow sscanf atof 1220 Underflow out of A float type decimal value is out of strtod fscanf scanf float range underflow sscanf atof 1250 Overflow out of A double type decimal value is out strtod fscanf scanf double l of range overflow sscanf atof 1260 Underflow out of A double type decimal value is out strtod fscanf scanf double of range underflow sscanf atof 1270 Overflow out of A long double type decimal value is fscanf scanf long double out of range overflow 1280 Underflow out of A long double type decimal value is fscanf scanf long double out of range underflow 1300 File not open The file is not open fclose fflush setbuf setvbuf fprintf fscanf prin
17. point zero Division by the floating pomt number 0 01s carried out in the evaluation of a constant expression S Assumes the internal representation of the value corresponding to 00 or co depending on the sign of the operands P Specify the correct constant expression Invalid floating point operations such_as co oo or 0 0 0 0 are carried out in a constant S Assumes the internal representation of not a number to indicate the result of an ineffective operation P Correct the constant expression The same compiler option is specified more S Uses the last specified compiler option P Check that options are specified correctly 1201 Ineffective floating point operation expression t Command parameter specified twice than once 1301 Too many define options 97 The number of macro names specified as suboptions in the define option exceeds 16 S Uses the first 16 suboptions P Define the 17th and subsequent macro names using define directives at the beginning of the source program 2 Error Level Messages Error No Message 2000 Illegal preprocessor keyword Explanation An illegal keyword is used in a preprocessor directive S Ignores the line containing the preprocessor directive P Correct the keyword in the preprocessor directive 2001 Illegal preprocessor syntax There is an error in preprocessor directive or in a macro call spe
18. 2215 Illegal type for gt amp or its used ian expression that is not an integer expression S Assumes int as the result type and continues processing P Specify integer expressions as the operands The combination of operand types used with 2216 Illegal type for parameter the binary operator is egal S Assumes the result type is int and continues processing P Specify a correct type of operands Only the following type combinations are allowed for the binary operator Two arithmetic operands Pointer and integer void is specified for a function call parameter type 2217 Illegal type for 114 S Ignores the parameter type and continues processing P Specify a function call parameter type so that a value can be passed to the function The combination of operand types used with the binary operator is not allowed S Assumes that the result type is int and Error No Message Explanation 2218 continues processing P Specify a correct type combination of operands Only the following three combinations are allowed for the binary operator 1 Two arithmetic operands Scalar required 2 Two pointers assigned to the same data type 3 The first operand is a pointer and the second operand is an integer The first operand of the conditional operator is not a scalar 2219 S Assumes that the first operand is a scalar and continues proce
19. 999 An overflow or underflow occurs if a floating point number in decimal representation cannot be normalized If a floating point number in normalized decimal representation contains too many significant digits as a result of the conversion the lower digits are discarded In the above cases a warning level error occurs at compile time and the variable errno is set equal to the corresponding error number at run time To convert a floating point number from decimal representation to normalized decimal representation the length of the original ASCII string literal must be less than or equal to 511 characters Otherwise an error occurs at compile time and the variable errno is set equal to the corresponding error number at run time To convert a floating point number from internal representation to decimal representation the floating point number is first converted from internal representation to normalized decimal representation According to a specified format the result is then converted to an ASCII string literal 163 ii Conversion between Normalized Decimal Representation and Internal Representation If the exponent of a floating point number to be converted between decimal representation and internal representation is too large or too small a precise result cannot be obtained This section explains the range of exponents for precise conversion and the error that results from exceeding the range a b Range of Expone
20. A l Language Specifications of the C Compiler ocsccss cccsescscccsasensccetececssetsss cessetscceectensecenveeees 145 Bal Compilation SpecificatOnS sses paints e eee 145 A 1 2 Environmental Specifications eessseeseeseeeseeserseseresresseesresreesersresresseesrrsreesresee 145 AN3 Ident E S osineen a e a a ae shod sada SS 145 A14 CharaCtetSsccosiaoieieiesin sissioni iiai Es ee e RE EEan E EA ASE Ei as 146 AS TESE o ao ob ov E A G a a ee 147 AG Floating Point IOS a osedzdsskezue d mda as case demactoaareacunveearnyengeenceserecacnequede 148 PA Anas and BOMMETS 5 oiaa s eoickosnes eddaton so d n dasadtad oresar Eier niant 149 PNG KOPIS T eieiei accer rascn nenies su ce k adua ho duensacdetes ia ladeustaldebduabanaceeaeeebanna teas 149 A 1 9 Structure Union Enumeration and Bit Field Types 150 ALAO TOU seed alta adda ag obo c hen detect ain gt sce antecedent atch 150 AA LU Declarations sssrds iseni n a aa a ra N EE E S 151 A 1 12 Statement sincsen aa e e E a a E A A ESEESE 151 ATTI Preprocessor zhasne Se E kan en 152 A 2 C Library Function Specifications ac o cocecceecsenssticueeguatscaremuuderevaseaasdbceeebseauaeasemveudecaensiee 153 A21 SMOE saniser Eaa aE ea teas einai te ee EEEE 153 P22 T A E A a EEE E EE E E ET 153 223 CYPE Dares sacs seco eeepc ee ooo ooo 154 A2 math aise scence dem eae Sonia ae ee eo ee 154 AZI Sdi O Deae OCT REDE em AE M A AO PO Tn een M eae eee me ES 155 A26 SUI hh kisesi aa EE ER E EE AA EAER 156
21. Ox0F x sleep 49 0 1 2 4x 8x 12 Set the start address of table to GBR Set 10 to table cdata2 Set 100 to table idata2 tabl Referenc cdata2 AND 10 and table cdata2 and set it to table cdata2 OR H OF and table cdata2 and set it to table cdata2 Expanded to the sleep instruction x aha mf he 7 Effective use of intrinsic functions v Set the start address of a structure which is allocated to memory and frequently accessed in GBR and access its members by gbr_read_byte gbr_write_byte etc w In the case of v byte data frequently used in logical operations should be declared within 128 bytes from the start address of the structure 50 Section 4 Notes on Programming This section contains notes on coding programs for the C compiler and on troubleshooting when compiling or debugging programs 4 1 Coding Notes Functions with float Parameters For a function that declares float for parameters either a prototype must be declared or parameters must be declared as double Correct processing is not guaranteed if a function that has float parameters is called without a prototype declaration Example void f float ee ey V void f float x Since function f has a float parameter a prototype must be declared as shown at v Program Whose Evaluation Order is Not Regulated The effect of the execution is not guaranteed in a program wh
22. Parameters No Name Type Meaning fileno int File number of the file to be read 2 buf Pointer to Area to be used to store the read data char 3 count unsigned Byte length of data to be read int Return value Type int Normal Byte length of the data actually read Abnormal 1 Explanation The read routine loads data from the file indicated by the first parameter fileno into the area indicated by the second parameter buf The amount of data to be read is indicated by the third parameter count If an end of file is encountered during a read less than the specified number of bytes are read The file read write position is updated using the byte length of the data actually read If data is read normally the routine returns the number of bytes of the data read Otherwise the read routine returns a value of 1 85 d write routine Purpose Writes data to a file Interface int write int fileno char buf unsigned int count Parameters No Name Type Meaning fileno int File number 2 buf Pointer to char Area storing data to be written in the file 3 count unsigned int Byte length of the data to be written Return value Type int Normal Byte length of the data actually written Abnormal Explanation The write routine outputs data whose byte length is indicated by the third parameter count from the area indicated by the second parameter buf into the file indicated by the first parameter
23. The data type assigned to a pointer specified as a formal parameter in a prototype declaration differs from the data type assigned to a pointer used as the corresponding actual parameter in a function call S Uses the internal representation of the pointer used for the function call actual parameter P Use a cast operator for the function call actual parameter to convert the formal parameter to the type specified in the prototype declaration 1017 Return type mismatch The function return type and the expression type in a return statement are pointers but the data types assigned to these pointers are different S Uses the internal representation of the pointer specified in the return statement expression P Use a cast operator for the expression specified in the return statement expression to convert it to the type of the function return value 1019 Illegal constant expression 96 The operands of the relational operator lt gt lt or gt In a constant expression are pointers to different data types S Assumes 0 as the result value P Use an expression other than a constant expression to obtain the correct result Error No 4 Message expression of Explanation constant expression are pointers to different data types S Assumes 0 as the result value P Use an expression other than a constant expression to obtain the correct result t Division _by floating
24. __INIT Initializes registers and sequentially calls the initialization routine __INITLIB and file closing routine __CLOSEALL The coding example of _ INIT is shown below extern void _INITSECT void extern void _INITLIB void extern void _CLOSEALL void extern void main void void INIT void _INITSCT Calls section initialization routine _ _INIT Calls library initialization routine INI1 Calls C program main function Calls file close routine _ CLOSEALL Branches to endless loop after executing main function and waits for reset 75 4 3 Initializing Sections __ INITSCT Same as when the C library functions are not used For details refer to section 3 Setting Execution Environment 4 4 Initializing C Library Functions __INITLIB Initialization must be performed for related C library functions before being used The following description assumes the case when the initialization is performed in ___INITLIB in the program initiation routine To perform initialization the following must be considered 1 errno indicating the library error status must be initialized for all library functions 2 When using each function of lt stdio h gt and assert macro standard I O library function must be initialized 3 The user low level interface routine must be initialized according to the user low level initialization routine specification if required 4 When u
25. i Checks if the same function has been defined if onexit buf i f return NULL if _onexit_count 32 Checks if the No of defined functions exceed limit return NULL else _onexit_count t onexit buf onexit count f Defines the function address return 8 onexit buf onexit count 169 D 2 Creating exit Function This section describes how to create exit function that terminates program execution Note that the exit function must be created according to the user system specifications refereing to the following example because how to terminate a program differs depending on the user system The exit function terminates C program execution based on the termination code returned as a paramter and then returns to the environment at program initiation Returning to the environment at program initiation is achieved by the following two steps 1 Sets a termination code in an external variable 2 Returns to the environment that is saved by the setjmp function immediately before calling the main function An example of the exit function is shown below include lt setjmp h gt include lt stddef h gt typedef void onexit_t extern int _onexit_count extern onexit t _onexit_buf 32 void extern jmp buf _init_env extern int _exit_code extern void _CLOSEALL extern void exit int void exit code int code INE ig _exit_code code Sets return code to exit code
26. ifndef or elif mismatches directive corresponding to an elif directive S Ignores the elif directive P Insert the corresponding preprocessor directive or delete the elif directive 2009 Processor directive felse There is no if ifdef or ifndef directive mismatches corresponding to an else directive S Ignores the else directive P Insert the corresponding preprocessor directive or delete the else directive 2010 Macro parameters mismatch The number of macro call arguments is not equal to the number of macro definition arguments S Ignores the excess arguments if there are too many or assumes blank string literals if the number of arguments is insufficient P Specify the correct number of macro arguments 99 Error No Message 2011 Line too long Explanation After macro expansion a source program line exceeds the limit of 4095 characters for UNIX systems and 512 characters for PC systems S Ignores the 4096th and subsequent characters P Separate the line so that the length of each resulting line is within the limit after macro expansion 2012 Keyword as a macro name A preprocessor keyword is used as a macro name in a define or undef directive S Ignores the define or undef directive P Change the macro name 2013 Processor directive endif mismatches There is no if ifdef or ifndef directive corresponding to an endif directive S Ignores
27. section D ROM RAM for details refer to the following section 2 1 4 Initialized Data Area Allocation The initialized data area contains data with initial value Since the C language specifications allow the user to modify initialized data in programs the initialized data area is allocated to ROM and is copied to RAM before program execution Therefore the initialized data area must be allocated in both ROM and RAM However if the initialized data area contains only static variables that are not modified during program execution only a ROM area needs to be allocated 2 1 5 Example Memory Area Allocation and Address Specification at Program Linkage Each program section must be addressed by the option or subcommand of the linkage editor when the absolute load module is created as described below Figure 2 2 shows an example of allocating static areas 61 0x0000000 0x0000400 Program area section P Constant area Initialized data area section D 0x9000000 Initialized data area section R Noninitialized data area seciion B P C D B Default section name generated by the C compiler i R Section name specified by thROM Dynamic areg Internal RAN option of the linkage editor OxFFFF800 OxFFFFFFF Figure 2 2 Static Area Allocation Specify the following subcommands when allocating the static area as shown in figure 2 2 ROMA D R STARTAP C D 400 R B 9000000 A Description
28. 151 Preprocessor Specifications xc i3 seidhessvcccassasecisnaessacanessecasanecsacesbedsadenvenseedensdecdeanoesaceesess 152 stddef li Specifications 3 43 eba Beene Roe aha as 153 assert h Specifications ssrin r ati E E T E N a O od na 153 CPE A Spee a OMG a A E a Zde AES 154 Set of Characters that Returns True eseesseeseeesesressreeresresseserssressessrertessersresrersreseee 154 math Specifications inesike S A E A a A F 154 stdio h SpeciheatonS Aserre e e koa a S iso E RESE o ESES 155 Infinity and Not a Number eseeesenseesseesssseesseesseessessserssssessseesseesseesseresseeesseesseesseessees 156 string h Specifications sesse tesisi siiin ien AAE asi E eE y iiaa SES 156 Libraries Not Supported by the C Compiler eee 156 Types of Values Represented by Floating Point Numbers seeseesesereerereereereree 158 xi PART I OVERVIEW AND OPERATIONS Section 1 Overview The SH series C compiler inputs source programs written in C and outputs relocatable object programs or assembly source programs The C compiler supports the SH7000 series microcomputers referred to as SH Section 2 Developing Procedures Figure 2 1 shows the relationship between the C compiler package and other software for program development The C compiler package includes the software enclosed by the dotted line C source file creation Software included in User Standard the package include SH series C compiler ras so
29. O file names are specified These names are used by the low level interface routine open 2 If file could not be opened the file access disable flag is set 3 For equipment that can be used in interactive mode such as console the buffering disable flag is set Figure 4 2 FILE Type Data 4 4 2 Creating Initialization Routine INIT OTHERLIB for Other Library Function 78 Figure 4 1 Program Configuration When C Library Function Is Used 4 5 Closing Files __ CLOSEALL When a program ends normally all open files must be closed Usually the data destined for a file is stored in a memory buffer When the buffer becomes full data is output to an external storage device Therefore if the files are not closed data remaining in buffers is not output to external storage devices and may be lost When an program is installed in a device the program is not terminated normally However if the main function is terminated by a program error all open files must be closed The following shows an example of __CLOSEALL include lt stdio h gt void _CLOSEALL void Deletes an underline from symbol name in assembly routine TRE for i 0 i lt _NFILE i Checks that file is open if _iob i _ioflagl _IOREAD _IOWRITE Closes opened files fclose amp _iob i Example 79 4 6 Creating Low Level Interface Routines Low level interface routines must be supplied for C programs that use the
30. Parameter register 41 168 Return value storage register 42 Rules on changes in registers after a function call 35 Usage of registers 168 Reserved words 80 Result rounding 162 181 Return value 37 General rules concerning return values 37 Return value address 42 Return value setting location 42 Return value storage register 42 ROM linkage editor option 62 ROM and RAM allocation 61 ROM option 62 Run time routines 4 60 S sbrk routine low level interface routine 88 section option 9 Section 24 Constant area 22 Initialized data area 22 Non initialized data area 22 Program area 22 Section name 9 22 Section initialization 70 76 short 26 147 show option 9 SH series 3 Sign 157 Sign extension 30 source suboption 9 Source listing information C compiler list ing 12 13 Stack area 24 63 Higher addresses 34 Lower address 34 Stack frame 34 168 frame size 15 168 Stack pointer SP 34 66 68 Stack switching specification interrupt function 43 44 Standard include file 4 Standard library file 4 59 start linkage editor option 62 Static area size calculation 58 statistics suboption 9 Statistics information Compiler listings 12 16 Status register SR 44 47 Structure of object programs 24 Structure type 27 150 Suboption 9 Systems 5 System installation 57 Initialization 69 Program configuration 67 Section initialization 70 Vector table setting 68 T TRAPA instruction interrupt function 43 T
31. Table 2 6 lists the general parameter area allocation rules Stack Lower addresses sP gt Return value address R4 R5 Parameter area R6 R7 Parameter storage registers i Parameter area allocation Figure 2 2 Parameter Area Allocation Table 2 6 General Rules on Parameter Area Allocation 39 Allocation Rules Parameters Allocated to Registers Parameter Storage Registers Target Type Parameters Allocated to a Stack R4 R7 char unsigned char short v Parameters whose types are other than target unsigned short int types for register passing unsigned int long w Parameters of a function which has been unsigned long float and declared by a prototype declaration to have pointer variable number parameters x Other parameters are already allocated to R4 R7 Note If a function has been declared to have variable number parameters by a prototype definition parameters which do not have a corresponding type in the declaration and the immediately preceding parameter are allocated to a stack Example int f2 int int int int f2 a b c x y z i x y and z are allocated to a stack c Parameter allocation 40 i Allocation to parameter storage registers Following the order of their declaration in the source program parameters are allocated to the parameter storage registers starting with the smallest numbered register Figure 2 3 shows an example of parameter allocation t
32. a is int basic type and the number of types that qualify the basic type is zero ii char f f is a function type that returns pointer to char basic type The number of types that qualify the basic type is two A 1 12 Statement Table A 14 Statement Specifications item _C Compiler Specification The number of case label specified by a switch statement _Up to 511 labels can be specified 151 A 1 13 Preprocessor Table A 15 Preprocessor Specifications Item Correspondence between single character constant and execution environment characters in the conditional compilation Reading an include file Supporting an include file whose name is enclosed in a pair of double quotation marks Source file character string correspondence blank character in a character string after macro expansion pragma directive operation Value of DATE _ TIME _C Compiler Specification Character strings in the preprocessor statement match the execution environment characters The file within lt gt is read from a path specified by the include option Defalut The path _specified by environment variable SHC_LIB The C compiler supports include files whose names are delimited by double quotation marks The C compiler reads these include files from the current directory If the include files are not in the current directory the C compiler reads them from the directory specified in advance Strings of blan
33. assumed that the highest order bit 252 is 0 and a decimal point follows it Value represented by a denormalized number 1 lt sign gt x 2 1022 x lt mantissa gt x 2752 Example Bd Be Gi ijooooooaaaga Li1onadaonoonooooono nooo n ono nooo O ODO O ODO VOODOO V Sign Exponent 1022 Mantissa 0 111 2 0 875 Value 0 875 x 271022 Zero The sign bit is either 0 positive or 1 negative 1 e there are two distinct zero values 0 0 and 0 0 The exponent and mantissa are 0 Both 0 0 and 0 0 represent 0 0 See appendix A 3 4 Floating Point Operation Specifications for differences in each operation depending on the sign Infinity The sign bit is either 0 positive or 1 negative i e 00 and co can be represented The exponent is 2047 211 1 The mantissa is 0 Not a Number The exponent is 2047 211 1 and the mantissa is not equal to 0 Note The sign of a not a number is arbitrary and the value of the mantissa is not limited except that it may not be equal to 0 161 A 3 4 Floating point Operation Specifications This section explains the floating point arithmetic used in C language functions It also gives the specifications for converting between the decimal representation and the internal representation of floating point numbers generated during C compiler or standard library function processing Arithmetic Operation Specifications i Result Rounding If the prec
34. be assigned to the variable S Does not initialize the variable P Specify the correct type of initial value constant 2131 No type nor storage class A value other than a constant expression 1s specified as either the initial value of a struct union or array variable or as the initial value of a static variable S Does not initialize the variable P Specify a constant expression as the initial value 2132 No parameter name S ifi in an external data definition S Assumes int as the type specifier P Insert the storage class or type specifier 2133 Multiple parameter declarations 107 A parameter is declared even though the function parameter list is empty S Ignores the parameter declaration P Insert the parameter name in the function parameter list or delete the parameter declaration Either a parameter name is declared in a Error No Message Explanation function definition parameter list more than once or a parameter is declared inside and outside the function declarator S Uses the first declaration if a parameter is declared more than once in the function parameter list Uses the declaration inside the function declarator if a parameter is declared inside and outside the function declarator P Keep one of the declarations and delete the 2134 Initializer for parameter rest PRET CR er ee os Be Pear eee EE of an parameter S Does not use the initial value sp
35. codes 177 asmcode suboption 9 B Bit field 28 150 bss suboption 9 C C compiler listings 12 char type data 26 146 C library function 60 153 Error messages output for the C library functions 140 C library function execution environment setting 73 Closing files 79 Initialization 75 Initializing C library functions 76 Initializing sections 76 Low level interface routines 80 Program configuration 73 Vector table setting 74 close routine low level interface routine 84 code option 9 Coding notes 51 command line specification C compiler listings 17 Compiler option 9 C compiler options 5 6 C compiler option listings 9 Option combinations 11 const suboption 9 Constant area 24 cpu option 9 Creating low level interface routines 80 Creating termination function 169 Creating abort routine 171 Creating exit function 170 Creating library onexit function 169 D data suboption 9 debug option 9 Debug information 9 11 54 define option 9 Denormalized number 158 double 26 148 160 Dynamic area 24 63 Dynamic area allocation 63 Heap area 24 66 Stack area 24 63 Dynamic area size calculation 63 E errno 76 Error levels error message levels 92 Error level 92 Fatal level 92 Internal level 92 Warning level 92 Error messages 91 C compiler action and programmer response for each error level 92 Error levels 92 Error message format 91 Error message output for the C compiler library functions 140 Execution
36. depending on the sign Infinity The sign bit is either 0 positive or 1 negative i e 00 and co can be represented The exponent is 255 28 1 The mantissa is 0 Not a Number The exponent is 255 28 1 and the mantissa is not equal to 0 Note The sign of a not a number is arbitrary and the value of the mantissa is not limited except that it may not be equal to 0 A 3 3 double and long double A double or long double is represented as 1 sign bit 11 exponent bits and 52 mantissa bits Normalized Number The sign bit is either 0 positive or 1 negative The exponent is a number from 1 to 2046 211 2 From this value 1023 is subtracted and the result is used as the actual exponent The range of actual exponents is 1022 to 1023 The mantissa is a value from 0 to 252 1 For an actual mantissa it is assumed that the highest order bit 252 is 1 and a decimal point follows it Value represented by a normalized number 1 lt sign gt x 2 lt exponent gt 1023 x 1 lt mantissa gt x 2 52 Example i Bo 51 o o111111111141110000000000000000000000000000000000000000000000000 Sign Exponent 1111111111 1023 0 Mantissa 1 1110 1 875 Value 1 875 x 29 1 875 160 Denormalized Number The sign bit is either 0 positive or 1 negative The exponent is 0 which makes the actual exponent equal to 1022 The mantissa value is from 1 to 252 I For an actual mantissa it is
37. exceed 128 characters 4 Check that the disk has sufficient capacity for files late file i I T 3301 Cannot close internal file the C compiler cannot be closed S Terminates processing P 1 Check that there are no mistakes in the compiler installation procedure 2 Check that there are no abnormalities on the hard disk 3302 Cannot input internal file the C compiler cannot be read S Terminates processing P 1 Check that there are no mistakes in the compiler installation procedure 2 Check that there are no abnormalities on the hard disk An intermediate file internally generated by 137 Error No Message Explanation 3303 Cannot output internal the C compiler cannot be written file S Terminates processing P Increase the disk size er late file il 11 3304 Cannot delete internal the C compiler cannot be deleted file S Terminates processing P Check that the intermediate file generated by the C compiler is not being used An invalid compiler option is specified 3305 Invalid command parameter S Terminates processing option name P Specify the correct option An interrupt generated by a CNTL C 3306 Interrupt in compilation command from a standard input terminal is detected during compilation S Terminates processing P Input the compile command again FHe versions in the C compHler do not match 3307 Compiler version mismatch S Terminates proces
38. if freopen stdin r stdin NULL Opens standard input file stdin gt ioflagl 0xff Disables file access 2 stdin gt ioflagl _IOUNBUF No data buffering 3 1 if freopen stdout w stdout NULL Opens standard output file stdout gt _ioflagl Oxff stdout gt ioflagl _IOUNBUF 1 if freopen stderr w stderr NULL Opens standard error file stderr gt _ioflagl Oxff stderr gt ioflagl _IOUNBUF 77 Declares FILE type data in the C language define _NFILE 20 struct _iobuf unsigned char _bufptr Buffer pointer long _bufcnt Buffer counter unsigned char _bufbase Buffer base pointer long _buflen Buffer length char _ioflagl I O flag char _ioflag2 I O flag char _iofd 1 0 flag _iob _NFILE 4 4 1 Creating Initialization Routine _INIT_IOLIB for Standard I O Library Function The initialization routine for standard I O library function initializes FILE type data used to reference files and open the standard I O files The initialization must be performed before opening the standard I O files The following shows an example of _INIT_IOLIB include lt stddef h gt extern char _slptr extern void srand unsigned int void _INIT_OTHERLIB void srand 1 Sets initial value when rand function is used _slptr NULL Initializes the pointer used in the strtok function Example Notes 1 Standard I
39. if the byte data at the offset int offset unsigned char mask from GBR is 0 or not and sets the result in the T bit Special SLEEP instruction void sleep void Executes the SLEEP instruction instruc TAS instruction void tas char addr Executes TAS B addr tions TRAPA instruction int trapa int trap no Executes TRAPA trap_no 48 Warnings The offsets and masks shown in table 3 2 Intrinsic Functions must be constants Also the specification range for offsets is 255 bytes when the access size is shown in bytes 510 bytes when the access size is shown as a word and 1020 bytes when the access size is shown as a long word Masks which can be specified for performing logical operations AND OR XOR or TEST on a location relative to GBR global base register must be within the range of 0 to 255 As GBR is a control register whose contents are not preserved by all functions in this C compiler take care when changing GBR settings Example include lt machine h gt define CDATA1 0 define CDATA2 1 define CDATA3 2 define SDATA1 4 define IDATA1 8 define IDATA2 12 struct char cdatal offset char cdata2 offset char cdata3 offset char sdatal offset char idatal offset char idata2 offset stable void set_gbr stable gbr_write_byte CDATA2 10 gbr_write_long IDATA2 100 i if gbr read byte CDATA2 10 fx gbr_and_byte CDATA2 10 E gbr or byte CDATA2
40. name Explanation not a pointer to a struct or union S Assumes that the member is not referenced and continues processing Use an expression which deals with pointer to struct or union to the left of the gt operator according to the member 2206 Modifiable lvalue required for operator An undectared member name ts used to reference a struct or union S Assumes that the member is not referenced and continues processing P Specify the correct member name The operand for a unary prefix or suffix 2207 Scalar required for operator or has a left value that cannot be assigned a left value whose type is not array or const S Assumes that the expression with a left value that can be assigned is specified as an operand and continues processing Specify an expression whose left value can be assigned as an operand 2208 Pointer required for The unary operator is used on an expression that is not scalar S Assumes int as the type of the result and continues processing P Use a scalar expression as the operand 2209 Arithmetic type required for operator 112 The operand for the unary operator is an expression of pointer to void or is not an expression of pointer S Ignores P Use an operand that is an expression other than pointer to void The unary operator or is used on a non arithmetic expression Error No Messa
41. scalar type 116 Error No Message Explanation S Assumes that the result type is int and continues processing P Use a scalar type that is not a function or a pointer assigned to void as the operand The operand types for the assignment operator do not match S Assumes thatthe result type is int and continues processing P Specify a correct type combination of operands Only the following five type combinations are allowed for the operands of the assignment operator 1 Two arithmetic operands 2 Two pointers assigned to the same data type 3 The left operand is a pointer and the right operand is an integer constant whose value is 0 or another pointer that is assigned to void that was converted from an integer constant whose value is 0 4 A pointer and another pointer assigned to void 5 Two struct or union variables with the same data type An incomplete tag name is used for a struct or union in an expression 2223 Incomplete tag used in expression 2224 Illegal type for assign int and continues processing P Declare the tag name The operand types of the assignment operator or are illegal S Assumes that the result type is int and 117 Error No Message Explanation continues processing Specify a correct type combination of operands Only the following two type combinations are allowed as operands for the assignment operator or 1 Two arithm
42. standard input output or memory management library functions Table 4 1 shows the low level interface routines used by standard library functions Table 4 1 Low Level Interface Routines No Name Explanation 1 open Opens files 2 close Closes files 3 read Reads data from a file 4 write Writes data to a file 5 Iseek Sets the file read write address for data 6 sbrk Allocates a memory area Refer to the attached Standard Library Memory Stack Size Listing for details on low level interface routines required for each C library function Initialization of low level interface routines must be performed when the program is started For more information see the explanation concerning the INIT LOWLEVEL function in section 4 4 Initializing C Library Functions __INITLIB The rest of this section explains the basic concept of low level input and output and gives the specifications for each interface routine Refer to appendix E Examples of Low Level Interface Routines for details on the low level interface routines that run on the SH series simulator debugger Note The open close read write Iseek and sbrk are reserved words for low level interface routines Do not use these words in C programs 1 Concept of I O Operations Standard input output library functions manage files using the FILE type data Low level interface routines manage files using file numbers positive integers which correspond directly to ac
43. the area it has set with data control returns to the calling function usually with the RTS instruction The calling side then deallocates the area having an address higher than the return value address and the parameter area After function call and after control returns from a function 0 l Lower addresses SP gt Return value address Parameter area Higher addresses i Figure 2 1 Allocation and Deallocation of a Stack Frame 34 Registers Some registers change after a function call while some do not Table 2 5 shows how registers change according to the rules Table 2 5 Rules on Changes in Registers After a Function Call Item Registers Used in a Function Notes on Programming Guaranteed registers R0 R7 If registers used in a function contain valid data when a program calls the function the program must push the data onto the stack or register before calling the function Non guaranteed R8 R15 MACH MACL and PR The data in registers used in functions is pushed onto the stack or register before calling the function and popped from the stack or register only after control returns from the function The following examples show the rules governing register changes a A subroutine in an assembly program is called by a C program Assembly program called program EXPORT sub SECTION P CODE ALIGN 2 Ssub oes De daar ae Data in those registers neede
44. the characteristics of memory areas used for C source programs and standard library functions Section 2 2 Internal Data Representation This section explains the internal representation of data used by a C program This information is required when data is shared among C programs hardware and assembly programs Section 2 3 Linkage with Assembly Programs This section explains the rules for variable and function names that can be mutually referenced by multiple object programs This section also discusses how to use registers and how to transfer arguments and return values when a C program calls a function The above information is required for C program functions calling assembly program routines or assembly program routines calling C program functions Refer to respective hardware manuals for details on SH hardware 23 2 1 Structure of Object Programs This section explains the characteristics of memory areas used by a C program or standard library function in terms of the following items v Sections Composed of memory areas which are allocated statically by the C compiler Each section has a name and type A section name can be changed by the compiler option section w Write Operation Indicates whether write operations are enabled at program execution x Initial Value Shows whether there is an initial value when program execution starts y Alignment Restricts addresses to which data is allocated Table 2 1 shows the types an
45. 4 a DATA L 1 _b DATA L 1 fO END at b Example 2 An external identifier defined in a C program is referenced by an assembly program e Ina C program symbol names with no underscore character _ at the head are defined as external identifiers e In an assembly program external references to symbol names beginning with an underscore character _ are declared by an IMPORT directive C program definition Assembly program reference int a IMPORT _a SECTION P CODE ALIGN 2 MOV L A_a R1 MOV L R1 RO ADD 1 RO RTS MOV L RO R1 ALIGN 4 A a DATA L a 33 2 3 2 Function Call Interface When either a C program or an assembly program calls the other the assembly programs must be created using rules involving the following 1 Stack Pointer 2 Allocating and Deallocating Stack Frames 3 Registers 4 Setting and Referencing Parameters and Return Values Stack Pointer Valid data must not be stored in a stack area with an address lower than the stack pointer since the data may be destroyed by an interrupt process Allocating and Deallocating Stack Frames In a function call right after the JSR or the BSR instruction has been executed the stack pointer indicates the lowest address of the stack used by the calling function Allocating and setting data at addresses greater than this one is a role of the calling function After the called function deallocates
46. 4 C 4 256 256 256 260 D 4 0 0 0 4 isalnum isalpha ctype Note The standard library supplied by the C compiler includes C library functions based on C language specification and arithmetic routines reguired for C program execution The size required for run time routines must also be added to the memory area size in the same way as C library functions Run time routine names used by the C programs are output as external symbols in theassembly programs generated by the C compiler option code asmcode The user can see the run time routine names used in the C programs through the assembly program listing The following shows the example of C program and assembly program listings C program int a int b a b return a Assembly program output by the C compiler IMPORT divls An external reference definition for the run time routine EXPORT LE SECTION P CODE ALIGN 4 MOV R5 R1 MOV A_divls R2 JSR R2 MOV R4 RO RTS NOP DATA L __divls END An external reference definition MPORT beginning with __ indicates a run time routine In the above example divls is a run time routine used in the C program 60 2 1 3 ROM and RAM Allocation When allocating a program to memory static areas must be allocated to either ROM and RAM as shown below Program area section P ROM Constant area section C ROM Non initialized data area section B RAM Initialized data area
47. 5 Source listing 13 Statistics information 16 Structure of C compiler listings 12 long double 26 148 160 Low level interface routines 80 close routine 84 Examples of low level interface routine 172 Iseek routine 87 open routine 82 read routine 85 sbrk routine 88 write routine 86 Iseek routine low level interface routine 87 M machinecode suboption 9 Macro name definition 9 Mantissa 157 Memory allocation 57 sbrk routine 80 88 Memory area allocation 58 Area size calculation 58 63 Example Memory area allocation and address specification at program linkage 61 Dynamic area allocation 63 Initialized data area allocation 61 ROM and RAM allocation 61 Static area allocation 58 N Non initialized data area 24 Normalized number 158 Not a number 158 Notes on programming 51 Coding notes 51 Notes on programming development 54 O object suboption 9 objectfile option 9 Object information C compiler listings 12 15 onexit function termination routine 169 open routine low level interface routine 82 optimize option 9 Optimization level 9 Overflow 162 P Parameter 37 Parameter allocation example 165 Parameter area allocation 39 Passing parameters 37 Rules on type conversion 37 Stack parameter area 39 Storage registers 41 pragma interrupt 43 program suboption 9 Program area 24 Program configuration 67 73 Program development notes 54 PR register 36 Q R read routine 85 Register 149
48. 767 for UNIX systems and 16383 for PC systems S Terminates processing P Divide the file so that the number of physical lines does not exceed the limit 3031 Fhe size of an array or a structure exceeds Data size overflow 2147483647 S Terminates processing P Reduce the size of the array or the structure until it is less than or equal to 2147483647 The number of symbols used for debug 3033 Symbol table overflow information exceeds 30719 S Terminates processing P Divide the file so that the number of symbols does not exceed 30719 3201 The size of the object program exceeds Object size overflow 4 Gbytes S Terminates processing P Divide the program so that the size of the object program does not exceed 4 Gbytes An error has occurred in either one of the 136 Error No Message Explanation 3300 Cannot open internal file following cases 1 An intermediate file internally generated by the C compiler cannot be opened 2 A file having the same name as the intermediate file already exists 3 The path name for listing file specifications exceeds 128 characters 4 A file used internally by the C compiler cannot be opened S Terminates processing P 1 Check that the intermediate file generated by the C compiler is not being used 2 Do not use the intermediate file name for other files 3 Ensure that the path name for listing file specifications does not
49. 8 1 2147483647 float 4 4 Used 00 00 double 8 4 Used oo co long double Pointer 4 4 Unused 0 232 1 4294967295 26 Aggregate Type Data This part explains the internal representation of array structure and union data types Table 2 3 shows the internal data representation of aggregate type data Table 2 3 Internal Representation of Aggregate Type Data Data Type Alignment bytes Size bytes Data Allocation Example Array type Array element Number of array elements int a 10 alignment x Element size Alignment 4 bytes Size 40 bytes Structure type Maximum structure Total member size 1 struct member alignment int a b l Alignment 4 bytes Size 8 bytes Union type Maximum union Maximum value of member union member alignment size 2 int a b l Alignment 4 bytes Size 4 bytes Notes 1 When structure members are allocated unused area may be generated between structure members to align data types struct ates 4 bytes _ 4 bytes int b z za 1 byte Unused area If a structure has 4 byte alignment and the last member ends at an address indivisible by four the remaining bytes are included in this structure struct int a oba 4 bytes __ 4 byte gt C s 1 byte 27 2 When an union has 4 byte alignment and the maximum size of its members is not a multiple of four the remaining bytes up to a multiple of four are included in this union un
50. A27 Nat Supported Library escis na E EE 156 A 3 Floating Point Number SPSCill Cat ONS 2 cecesecczdeecusices eee cseteuneavaveveeacunssidenbeetutetenededacasueues 157 A 3 1 Internal Representation of Floating Point Numbers eee 157 AD Modlar a E ae ee eee 159 A535 doybl eand Lone doUbIo zaessutdsd do sodu zona sodou no st 160 A 3 4 Floating point Operation Specifications cccceseeceseeeeeeeeeceeeeeeeteceeneeeeneeeenas 162 Appendix B Parameter Allocation Example eee eeeeeeeeren n 165 Appendix C Usage of Registers and Stack Area een 168 Appendix D Creating Termination Functions L eeeeeeeeeeeeeeeenen ne 169 D 1 Creating Library onexit PONOR esses sce scar cpatecesaretieeient dans cacdei aad liete tedious 169 D2 Creating c Bunchon dada dodano EEE R RE ES EE sa 170 D 3 Creating abort ROUDE 33 55 csscciiasesnsdatossnanietsstcnasiadeedasnstelasasidanasauveaceaadeseedennssadeveueeiedasies 171 Appendix E Examples of Low Level Interface Routine eee 172 Appendix F ASCI COCs a a ixccctana sterner ia eean E 177 vi Figures Part I 2 1 Relationship between the C Compiler and Other Software eee 4 3 1 Source Listing Output for show noinclude and Noexpansion e eeee eee 13 3 2 Source Listing Output for show include and eXpansion eeeeeeeeeeeeeeeeeeenn 13 3 3 Object LISUDP io a does eh edu ezna e SO s oa os 15 3 4 Statistics Informationen oc catch DO oo dd eA S o oby ov ooo po Oo a n
51. EXPOR EXPORT EXPOR EXPORT EXPORT END Notes v Section names B and D must be the non initialized data area and initialized data area section names specified with the compiler option section w Section name R must be the section name in RAM area specified with the ROM option at linkage 71 If the above preparation is completed section initialization routine can be written in C as shown below Example extern int D ROM B BGN B END D BGN _D_END extern void _INITSCT void _INITSCT short p q Non initialized area is initialized to zeros for p _B_BGN p lt _B_END ptt p 0 Initialized data is copied from ROM to RAM for p D BGN q _D_ROM p lt _D_END ptt g p q 72 Section 4 Setting the C Library Function Execution Environment To use C library functions they must be initialized to set C program execution environment To use VO stdio h and memory management stdlib h functions low level I O and memory allocation routines must be created for each system This section describes how to set C program execution environment when C library functions are used Power on reset VEC_TBL Low level interface Table always required Routine always required Routine required when library is used Supplied by the C compiler Figure 4 1 shows a program configuration when C library functions are used 73 Eac
52. Explanation 1008 Identifier too long An identifier s length exceeds 31 characters S Uses the first 31 characters and ignores the rest P Use identifiers with 31 or less characters 1010 Character constant too The length of a character constant exceeds long four characters S Uses the first four characters and ignores the rest P Use character constant with four or less characters 1012 Floating point The value of a floating point constant exceeds constant overflow the limit S Assumes the internally represented value corresponding to 00 or co depending on the sign of the result P Specify floating point constants within their limits 1013 Integer constant The value of unsigned long integer constant overflow exceeds the limit S Ignores the overflow and uses the remaining bits P Specify integer constants within their limits 1014 Escape sequence The value of an escape sequence indicating a overflow bit pattern in a character constant or string literal exceeds 255 S Uses the low order byte P Change the value of the escape sequence to 255 or lower 95 Error No 1015 Message Floating point constant underflow Explanation The absolute value of a floating point constant is less than the lower limit S Assumes 0 0 as the value of the constant P Change the value of the constant to 0 0 or specify a constant whose value can be represented 1016 Argument mismatch
53. Hitachi Microcomputer Support Software SH Series C Compiler USER S MANUAL HITACHI ADE 702 095 HS0700CLCU1SE The Copyright Statement Preface This manual explains the facilities and operating procedures for the SH series C compiler Ver 2 0 The C compiler translates source programs written in C into relocatable object programs or assembly programs for Hitachi SH7000 series RISC microcomputers This manual consists of four parts and appendixes The information contained in each part is summarized below 1 PARTI OVERVIEW AND OPERATIONS The overview sections cover the following v C compiler functions w Developing procedures The operation sections cover the following x How to invoke the C compiler Yy Optional functions z Listings created by the C compiler 2 PART II PROGRAMMING This part explains the limitations of the C compiler and the special factors in object program execution which should be considered when creating a program 3 PART II SYSTEM INSTALLATION This part explains the requirements when installing an object program generated by the C compiler on a system They are the object program being written in ROM and memory allocation In addition specifications of the low level interface routine must be made by the user when using standard I O library and memory management library 4 PART IV ERROR MESSAGES This part explains the error messages corresponding to compilation errors and the stand
54. K kk k kk K kk K kk K KK OK KR OK KK KK OK KK KKK KK int close int fileno File number if fileno lt FLMIN FLMAX lt fileno File number range check return 1 flmod fileno 0 File mode reset return 0 173 RR KKK KKK KKK k kk KKK kk K kk KK kk K kk K kk KR OK KR OK KK OK KK KK KK read Data read Ef Return value Number of read characters Pass 1 Failure 8 RR KKK KKK k kk k kk k kok kk K kk K kk kk K kk K kk K k k K KK KX KA KA OK X KA KOK k kk int read int fileno File number char buf Destination buffer address unsigned int count Number of read characters unsigned int i Check mode according to file name and stores each character in buffer if flmod fileno amp O_RDONLY fImod fileno sO RDWR for i count i gt 0 i buf charget if buf CR Line feed character replacement buf LF buf return count RR KKK KKK KKK KK KKK KK K kk OK KKK kk KK OK OK OK KK OK OK KK KKK KK write Data write Return value Number of write characters Pass 1 Failure x KR KKK KKK k kk k kk k kok kok K kk K kk K kok K kk K kk K k k K K KK KX OK KA X AKA X KA OK KKK int write int fileno File number char buf Destination buffer address unsigned int count Number of write characters unsigned int i char c Check mode according to file name and output each character if flmod fileno amp O_WR
55. ONLY flmod fileno amp O_RDWR for i count i gt 0 i c buf charput c return count else return 1 174 KR KKK KKK k kk KKK KKK kok K kk KK kk k kk K OK KK OK OK KKK OK KK KK KK lseek Definition of file read write position VA Return value Offset from the top of file read write position Pass 1 Failure lseek is not supported in the console input output RR KKK KKK k kk KKK KKK RK K kk OK KKK KK KR KR OK OK KK KKK OK KK KKK KK long lseek int fileno File number long offset Read write potision x int base Origin of offset return 1 KR KKK KKK k kk KKK KKK KK kk K kk KK k kk K kk OK RK OK OK KK KR OK KK KK KK F sbrk Data write Af Return value Start address of the assigned area Pass 1 Failure x RR KKK kok k k kk k kk kok k kk k kk K kk kk k kk K kk X kk kk K kk K k X KX KA AK AKA KA X KA k kk char sbrk unsigned long size Assigned area size s char p if brk size gt heap area heap HEAPSIZE Empty area size x return char 1 p brk Area assignment brk size End address update a return p 175 lowlvl src SH series simulator debugger interface routine Input output one character EXPOR _charput EXPOR _charget SIM_IO EQU H 0080 Specifies TRAP_ADDRESS SECTION P CODE ALIGN 7 _charput One character output A C program interface charpu
56. Special Value Operations More about special value zero infinity and not a number operations The result is positive zero 1f positive zero and negative zero are added If zero is subtracted from zero and both zeros have the same sign the result is positive Zero The operation result is always a not a number if one or both operands are not a numbers Positive zero is equal to a negative zero for relational operations 162 If one or both operands are not a numbers in a relational or equivalence operation the result of is always true and all other results are false Conversion between Decimal Representation and Internal Representation This section explains the conversion between floating point constants in a source program and floating point constants in internal representation The conversion between decimal representation and internal representation of ASCII string literal floating point numbers by library functions is also explained i To convert a floating point number from decimal representation to internal representation the floating point number in decimal representation is first converted to a floating point number in normalized decimal representation A floating point number in normalized decimal representation is in the format M x 10 N The following ranges of M and N are used For normalized float 0 lt M lt 109 1 O lt N lt 99 For normalized double and long double 0 lt M lt 10 7 1 0 lt N lt
57. Standard File Extensions Used by the C Compiler eee 8 32 COMPS OU OM Sees va aeait causininiuacedpcececaeyel aacaseaseeanlecea ccsataasaiosacenacateessceuse AEE Eai 9 3 3 Macro Names Names and Constants Specified by the define Option 10 3 A Option Combinations 347 da as 0330 Aaa vee th de dan kase oa eos 11 3 5 Structure and Contents of C Compiler Listings ceeecessceceeececeeeeeesneeeeseeeeneeeeaees 12 Part II TSE Limitationot the CoC Onn p ler aio acne feu teal at ctedoea conus a nooo ogee asda 21 2 1 Memory Area Types and CharacteristiCs 2 9 cesd4y ciate edtniaeeeicin eens 24 2 2 Internal Representation of Scalar Type Data eee eeecceesscceceeeceesteeeenteeeeneeeeneeeeaees 26 2 3 Internal Representation of Aggregate Type Data eee 27 2 4 Bit Field Member Specifications sssessesssessseeeeseeessresseesseesseessseesseessersseeesseeesseesseesse 28 2 5 Rules on Changes in Registers After a Function Call eee 35 2 6 General Rules on Parameter Area Allocation eesesessesrseseerersersresreesetsresrresressereresresse 40 2 7 Return Value Type and Setting Location esseeseesesessesesseseresreesersresresseeseseresresseseresresse 42 Sql mtert pt Specifications ernn ole E alee EE E ede ees 43 3 2 NGC Functions nnno Be See a o o o o ooo 47 dals Troubl esho tint cenene n E oa a as dh A ea poh oto 54 Part II 2 17 stack Size Caleulation Example sid cassie dasdncguste seuss B wanes Wei ledu sesda nod k
58. a 16 3 3 Conimnand line speci heaton 33 eso eect Aloe 6d0 ko a de AA dne a ea Sob es 17 Part II 2 1 Allocation and Deallocation of a Stack Frame eee 34 22 Parameter Area Allocation Kesip n ea ea orade ripiani a 39 2 3 Example of Allocation to Parameter Registers sseseeeseseereeserssesreesresrrrerereesseseresresss 41 2 4 Return Value Setting Area Used When Return Value Is Written to Memory 42 3 1 Stack Processing by an Interrupt FUNCTION eee eee eee eee eee nenene 44 Part II zel Section Size Information dee on doe Maite hi nen ei alee E lee E AR RS 58 2 2 StaticArea Allocation boudy Sado Bessa Do KDO O Oo oo aang aa o oo 62 2 3 Nested Function Calls and Stack Size s b3iepz ad adeda li t ct ancadecasstvcacatesasedisivecaneceteiatts 65 3 1 Program Configuration No C Library Function is Used cece eeeeeeeeeeneeeeeeeseeeeees 67 4 1 Program Configuration When C Library Function Is Used ee 13 42 TE Py poe Data 00 erica sete ate as a A a ee a oun 78 Part IV 151 Error Messages Format UNIX Systems 42 00cieue nine daddy us eke eked 91 l 2 Error Messages Format PC Systems s cas e cecssccgessdesecasceesteceagasdhodacsenssedess oedvannahecaseeenteats 91 Appendix A 1 Structure for the Internal Representation of Floating Point Numbers 157 C 1 Usage of Registers and Stack Area i ras Sia ate aes cadens Sos deo cde sieads ontop SSV 168 vii Tables Part I 3 1
59. able 3 1 Table 3 1 Standard File Extensions Used by the C Compiler File Extensioner Description c Source program file written in C h Include file Ist lis Listing file obj Relocatable object program file src Assembly source program file lib Library file abs Absolute load module file rel Relocatable load module file map Linkage map listing file Note 1 The listing file extension is lis on UNIX systems and 1st on PC systems The general conventions for naming files depend on the host machine Refer to the manual of the host machine in use 3 3 Compiler Options Table 3 2 shows C compiler option formats abbreviations and defaults Characters underlined indicate the minimum valid abbreviation Bold characters indicate default assumption Table 3 2 C Compiler Options Item Format Suboption Specification Optimization optimize 0 Object without optimization is output level 1 Object with optimization is output Listings and show source nosource Source list yes no formats object noobject Object list yes no statistics nostatistics Statistics information yes no include noinclude List after include expansion yes no expansion noexpansion List after macro expansion yes no 2 width lt numeric value gt Maximum characters per line 0 80 132 2 length lt numeric value gt Maximum lines per page 0 40 255 Default w 132
60. address of each section y Assembly code corresponding to machine language z Comments indicating the C program structure only output when not optimized however labels are always output I Line information corresponding to the C program only output when not optimized Stack frame size in bytes always output Statistics Information Figure 3 4 shows an example of statistics information 15 Figure 3 4 Statistics Information kk kk k STATISTICS INFORMATION KKKKKKKKKK ERROR INFORMATION KKKKKKKKKKK NUMBER OF ERRORS 0 NUMBER OF WARNINGS 0 xxx SOURCE LINE INFORMATION T COMPILED SOURCE LINE 13 kkkkk SECTION SIZE INFORMATION PROGRAM SECTION CONSTANT SECTION DATA SECTION BSS SECTION P 0x00004A Byte s C 0x000000 Byte s D 0x000000 Byte s B 0x000000 Byte s TOTAL PROGRAM SIZE 0x00004A Byte s S KKKKKKKKKK LABEL INFORMATION KKKKKKKKKKK Y NUMBER OF EXTERNAL REFERENCE SYMBOLS 0 J NUMBER OF EXTERNAL DEFINITION SYMBOLS l Description v Total number of messages by the level w Number of compiled lines from the source file x Size of each section and total size of sections y Number of external reference symbols number of external definition symbols and total number of internal and external labels Note Section size information x and label information y are not output if an erro
61. age feed is not performed 3 If file name is not specified standard file extension is added to the source file name 4 Macro names specified by options are shown in table 3 3 Table 3 3 Macro Names Names and Constants Specified by the define Option Item Explanation Macro name A character string beginning with an alphabetic letter or an underscore followed by zero or more alphabetic letters underscores and numbers 0 to 9 Name A character string beginning with a letter or an underscore followed by zero or more alphabetic letters underscores and numbers Constant A character string of one or more numbers or a character string of one or more numbers followed by a period and zero or more numbers 5 Refer to descriptions in Preprocessor Specifications in Appendix A 1 for details on how to retrieve the include file 6 When the help option is specified all other options are invalid 3 4 Option Combinations 10 If a pair of conflicting options or suboptions are specified for a file only one of them is considered valid Table 3 4 shows such option combinations Table 3 4 Option Combinations Option Combinations Valid Option Invalid Options nolistfile show code asmcode debug and show object help All other options 3 5 C Compiler Listings 11 This section describes C compiler listings and their formats Structure of C Compiler Listings Table 3 5 shows the structure and content
62. ant default labels P Modify the switch statement so that it has only one default label 2309 No label for goto There is no label corresponding to the destination specified by a goto statement S Continues processing P Specify the correct label in the goto statement 2310 Scalar required The control expression that determines statement execution for a while for or do statement is not a scalar S Assumes that an int control expression is specified and continues processing P Use a scalar expression as the control expression for a while for or do statement 2311 Integer required The control expression that determines statement execution for a switch statement is not an integer S Assumes that an int control expression is specified and continues processing P Use an integer expression as the control expression for the switch statement 2312 Missing The control expression that determines statement execution does not follow a left parenthesis for an if while for do or switch statement S Assumes that the control expression follows a left parenthesis and continues processing P Specify the control expression for an if while for do or switch statement and enclose it in parentheses 123 Error No Message Explanation 2313 Missing A do statement is ended without a semicolon G S Assumes that the do statement ends with a semicolon and continu
63. ard library error messages corresponding to run time errors This manual corresponds to operating systems that function on UNIX MS DOS or IBM PC systems In this manual operating systems functioning on MS DOS or IBM PC systems are referred to as PC systems Notes on Symbols The following symbols are used in this manual Symbols Used in This Manual Symbol Explanation lt gt Indicates an item to be specified Indicates an item that can be omitted Indicates that the preceding item can be repeated A Indicates one or more blanks RET Indicates the carriage return key return key Indicates that one of the items must be selected CNTL Indicates that the control key should be held down while pressing the key that follows UNIX is an operating system administrated by the UNIX System Laboratories United States MS DOS is an operating system administrated by the Microsoft Corporation United States IBM PC is an personal computer system administrated by IBM United States Contents Part I OVERVIEW AND OPERATIONS eee eee 1 Section l VETY 10 itt uh en a ieoa iE niaaa toe on 3 Section 2 Developing Procedures s s 3s45ii5le sis es bc nose eedn ned 4 Section 3 C Compiler ExoGUGON x s5adn snese u ed un eke c naan nnaeaceasiteahtemearian 5 3 1 How to Invoke the C Compiler oiiccccscsssssccesssissastssuvesevesnsvcasesnvearsasnstensasnencatsdecsesassasenvasnesanees 5 22 Naming Files zein i y
64. ated by adding GBR and the offset specified Item Function Specification Description Global Referencing GBR unsigned long References long word data 32 base based long word gbr_read_long int offset bits at the address indicated by register adding GBR and the offset specified GBR Setting GBR based void gbr write byte Sets data 8bits at the address cont byte int offset unsigned char data indicated by adding GBR and the offset specified Setting GBR based void gbr write word Sets data 16 bits at the address word int offset unsigned short data indicated by adding GBR and the offset specified Setting GBR based void gbr write word Sets data 32 bits at the address long word int offset unsigned long data indicated by adding GBR and the offset specified AND of GBR base void gbr and byte ANDs mask with the byte data at int offset unsigned char mask the address indicated by adding GBR and the offset specified and then stores the result at the same address OR of GBR base void gbr or byte ORs mask with the byte data at the int offset unsigned char mask address indicated by adding GBR and the offset specified and then stores the result at the same address XOR of GBR base void gbr xor byte XORs mask with the byte data at the int offset unsigned char mask address indicated by adding GBR and the offset specified and then stores the result at the same address TEST of GBR base void gbr tst byte Checks
65. b 3 31 0 i gt LIL 2 3 w Consecutive bit field members having type specifiers of the same size are placed in the same area as much as possible Example struct bl y a 1 22 unsigned int b 3 gt y b SSS Jy Te 2 3 x Bit field members having type specifiers with different sizes are allocated to different areas 31 0 Example E struct bl CUNY et 3 5 int a 5 31 0 y 4 29 y If the number of remaining bits in the area is less than the next bit field size though type specifiers indicate the same size the remaining area is not used and the next bit field is allocated to the next area Example struct b2 31 24 16 seein gt va gt char b 4 5 E 5 4 v z If an anonymous bit field member or a bit field member with a bit field size of 0 is declared the next member is allocated to the next area Example struct b2 char a 5 char 0 a oe ue char c 3 e w 5 3 30 2 3 Linkage with Assembly Programs Because C is suitable for writing system programs it can be used to describe almost all processes in microcomputer application systems In particular the SH series C compiler supports operations such as access to the SH microcomputer registers as intrinsic functions Refer to section 3 2 Intrinsic Functions in Part II Programming for details on intrinsic functions Processes which cannot be written in C for example calculations like multiplication and addition performed by the MAC instru
66. cessing Immediately before the TRAPA instruction is issued Lower ee STK 0 STK 0 se pir gt STK 99 Higher ee Lower ze sp gt sp Previous PC Previous PC Previous PC Previous SR Previous SR Previous SR Higher s Note the following when using this function Table 3 2 Intrinsic Functions cont 44 Warnings v The storage class specifier of the interrupt function must be extern Even if storage class static is specified the storage class is handled as extern The function must return void data The return statement cannot have a return value If attempted an error is output Example pragma interrupt f1 SP 100 f2 void f1 SP 100 C9 creer rte ete eee a int 2 n B0060 Oe Oe ee ee a e b Description a is declared correctly b returns data that is not void thus b is declared incorrectly An error is output w A function declared as an interrupt function cannot be called within the program If attempted an error is output However if the function is called within a program which does not declare it to be an interrupt function an error is not output but correct program execution cannot be guaranteed Example An interrupt function is declared pragma interrupt f1 void td void int f2 f1 a Description Function f1 cannot be called in the program because it is declared as an interrupt funct
67. cification S Ignores the line containing the preprocessor directive or macro call If there is an error in a constant expression used in the preprocessor directive the system assumes that the constant expression is 0 P Specify the correct preprocessor directive or macro call 2002 Missing A comma is not used to delimit two arguments in a define directive S Assumes that there is a comma P Insert a comma 2003 Missing A right parenthesis does not follow a name in a defined expression The defined expression determines whether the name is defined by a define directive S Assumes that there is a right parenthesis P Insert a right parenthesis 2004 Missing gt 98 A right angle bracket gt does not follow a file name in an include directive S Assumes that there is a right angle bracket P Insert a right angle bracket Error No Message Explanation 2005 Cannot open include file The file specified by an include directive file name cannot be opened S Ignores the include directive P Specify the correct file name If the file name is correct check that the file does not have write only status 2006 Multiple define s The same macro name is redefined by define directives S Ignores the second define directive P Modify one of the macro names or delete one of the define directives 2008 Processor directive felif There is no if ifdef
68. ction must be written in assembly language and then linked with the C program This section explains two key items which must be considered when linking a C program to an assembly program e External identifier reference e Function call interface 31 2 3 1 External Identifier Reference Functions and variable names declared as external identifiers in a C program can be referenced or modified by both assembly programs and C programs The following are regarded as external identifiers by the C compiler e A global variable which has a storage class other than static e A variable name declared in a function with storage class extern e A function name whose storage class is other than static When variable or function names which are defined as external identifiers in C programs are used in assembly programs an underscore character _ must be added at the beginning of the variable or function name up to 31 characters without the leading underscore 32 Example 1 An external identifier defined in an assembly program is referenced by a C program e In an assembly program symbol names beginning with an underscore character _ are declared as external identifiers by an EXPORT directive e In aC program symbol names with no underscore character _ at the head are declared as external identifiers Assembly program definition C program reference EXPORT _a _b extern int a b SECTION D DATA ALIGN
69. d by the TRAPA instruction return lt constant gt Constant value trap vector number Explanation pragma interrupt declares an interrupt function An interrupt function will preserve register values before processing all registers used by the function are pushed onto and popped from the stack when entering and exiting the function The RTE instruction directs the function to return However if the trap instruction return is specified the TRAPA instruction is executed at the end of the function An interrupt function with no specifications is processed in the usual procedure The stack switching specification and the trap instruction return specification can be specified together 43 Example extern int STK 100 int ptr STK 100 pragma interrupt f sp ptr tn 10 V W Explanation v Stack switching specification ptr is set as the stack pointer used by interrupt function f w Trap instruction return specification After the interrupt function has completed its processing TRAPA 10 is executed The SP at the beginning of trap exception processing shown in the figure below After the previous PC and SR status register are popped from the stack by the RTE instruction in the trap routine control is returned from the interrupt function Figure 3 1 Stack Processing by an Interrupt Function bs Just after the interrupt function Immediately after interrug During interrupt function has completed processing pro
70. d by the called MOVs R13 R15 function is pushed onto the stack ADD 8 R15 Function processing Since data in registers RO to R7 is pushed onto a t stack by the calling C program the assembly ADD 8 R15 program can use them freely without having to save MOV L R15 R13 J them first RTS MOV L R15 R14 Register data is popped from the stack C program calling program extern void sub sub 35 b A subroutine in a C program is called by an assembly program C program called program void sub Assembly program calling program IMPORT STS L MOV L MOV MOV L JSR NOP LDS L A sub DATA L sub SECTION P CODE ALIGN 2 PR R15 R1 1 R15 R3 R12 A_sub RO RO R15 PR _sub 36 The called function is declared by the IMPORT control instruction with an underscore character _ at the beginning Store the PR register return address storage register when calling the function If registers RO and R7 contain valid data the data is pushed onto the stack or stored in unused registers The sub function is called The PR register is restored Address data of the sub function Setting and Referencing Parameters and Return Values This section explains how to set and reference parameters and return values The rules for parameters and return values differ depending on whether or not the type
71. d characteristics of those memory areas Table 2 1 Memory Area Types and Characteristics Memory Area Section Section Write Initial Name Name Type Operation Value Alignment Contents Program area P code Disabled Yes 4 bytes This area stores machine codes Constant area C data Disabled Yes 4 bytes This area stores const data Initialized data D data Enabled Yes 4 bytes This area stores data area whose initial values are specified Non initialized B data Enabled No 4 bytes This area stores data data area whose initial values are not specified Stack area Enabled No 4 bytes This area is allocated at run time and is required for C program execution Refer to section 2 2 Dynamic Area Allocation in PART Ill SYSTEM INSTALLATION Heap area Enabled No This area is used by a C library function malloc realloc or calloc Refer to section 2 2 Allocation to Dynamic Area in PART III SYSTEM INSTALLATION Note Section name shown is the default generated by the C compiler when a specific name is not specified by the compiler option section 24 Example This program example shows the relationship between a C program and the sections generated by the C compiler int a 1 Program area main A2 char b const int c 0 Constant area c main gt Initialized data area a Non initialized data area b C program Section generated by the C compil
72. d declaration A bit field is used in a union S Ignores the bit field 2141 Illegal bit field P Use the bit field in a struct member of a struct S Ignores the bit field P Specify the name of the bit field 2142 Illegal void type void is used illegally 109 Error No Message 2143 Illegal static function Explanation S Assumes that void is int P void can only be used in the following cases 1 To specify a type assigned to a pointer 2 To specify a function return value type 3 To explicitly specify that a function whose prototype is declared does not have a parameter 2144 Type mismatch A Static storage class function has no definition in the source program S Ignores the function declaration P Either delete the function declaration or define the function extern vartablesorfuneHons wththe sare 2200 Index not integer names are declared with different data types in different valid ranges S The currently declared variable or function type is valid within the range that can be referenced However when linked with another file the valid data type is determined as shown below 1 If there is a declaration that acts as a definition that data type is valid 2 If there is no declaration that acts as a definition The previously declared data type is valid when the current declaration is in the function The currently declared data type is vali
73. d from const attribute or if a program compiled separately uses a parameter of a different type the C compiler cannot detect the error Example V const char p Because the first parameter p in library function strcat is a pointer for char the area indicated by the parameter p x strcat p abc may change w file 1 const int i file 2 extern int i In file 2 parameter i is not declared as const therefore assignment to it in file 2 i 10 is not an error 4 2 Notes on Programming Development 53 Table 4 1 shows troubleshootings for developing programs at compilation or when debugging Table 4 1 Troubleshooting Trouble Error 314 cannot found section is output at linkage Error 105 undefined Check Points The section name which is output by the C compiler must be specified in capitals in start option of linkage editor If identifiers are mutually Solution Specify the correct section name Reference parameters References Part Il Programming 2 1 Part Il Programming external symbol is output at linkage Debugging at the C referenced by a C program and an assembly program an underscore must be attached to the symbol in the assembly program Check if the C program uses a with the correct para meters Specify a standard 2 3 1 Standard library specifi library function An undefined reference symbol
74. d when the current declaration is not in the function P Declare the same data types for extern variables or functions An array index expression type is not an 110 Error No Message 2201 Cannot convert parameter Explanation integer S Assumes that the type is int P Specify an integer expression for the array index 2202 Number of parameters mismatch The nth parameter of a function call cannot be converted to the type of parameter specified in the prototype declaration S Assumes that the correct parameter type is specified and continues processing P Specify an expression whose type corresponds to the one specified in the prototype declaration 2203 Illegal member reference for The number of parameters for a function calis not equal to the number of parameters specified in the prototype declaration S Assumes that the number of parameters for the function call is equal to the number of parameters specified in the prototype declaration and continues processing P Specify the correct number of parameters 2204 Illegal member reference for gt The expression to the left of the operator is not a struct or union S Assumes that the member is not referenced and continues processing P Use a struct or union expression to the left of the operator The expression to the left of the gt operator is Error No Message 2205 Undefined member
75. dd se etl d a 65 4 1 Low Level Interface ROuUnES 228 2s seda deda t ads Ga Ree Geet 80 Part IV 1 1 C Compiler Action and Programmer Response for Each Error Level 92 Appendix A 1 Compilation Specilicatioms 22 2 aii tenet Rai Pe eee Ee ae alae Saas 145 A 2 Environmental Specifications s4s cs chi vasa tessayessavineeewedy ass tao sunng ve lt 4 bess shnsavenssa nceoassd een teeseaes 145 A3 Identifier SPEC IG AM ONS BROS 145 A 4 Character Specifications isiin tinira E a NE EEE SR sess 146 ard integer Specilications z 5 nera ela A el eae ne ae 147 A 6 Integer Types and Their Corresponding Data Range 0 eee eeeceeeeeeeseeeeeeceeeeeeeeeeneees 147 A 7 Floating Point Number Specifications secca sisccdess sayecei ais cdcasececabial scadadeascanas ad ccascadeetbeds 148 viii A 8 A 9 A 10 A 11 A 12 A 13 A 14 A 15 A 16 A 17 A 18 A 19 A 20 A 21 A 22 A 23 A 24 A 25 Limits on Floating Point Numbers 1 size zi d sed dana deda idoly dodano a de av e e 148 Array and Pointer Specifications 46s eer candcveseesy ae bkda dd idu ed d nika duse eevee 149 IG BIGGS yee EC ACER T aa EE EE E hea eee 149 Specifications for Structure Union Enumeration and Bit Field Types 150 Modifier Specifications as za ena a a Ea an ob anu a dlou lol oddat S 150 Declaration Specifications ceskou dada koda e E E EEE ERE 151 Statement Specification S aniei E E E S oa AK p dh ka cena ao ak
76. decimal representation conversion in the fscan function the meaning of in If does not follow indicates the range between the fscanf function _the previous and following characters Value of errno specified by fgetpos and ftell The fgetpos function is not supported The ftell functions function does not specify the errno value The errno value is determined depending on the low level _interface routine Output format of messages generated by the perror See a below for the output message format function calloc malloc or realloc function operation when 0O byte area is allocated the size is 0 a Messages generated by a perror function follow this format lt string literal gt lt error message correpsonding to the error number indicated by errno gt b Table A 22 shows the format used to indicate infinity and not a number for floating point numbers when using the printf or fprintf function 155 Table A 22 Infinity and Not a Number Valle C__ Format Positive infinity tH HHH Negative infinity eee A 2 6 string h Table A 23 string h Specifications ltem _C Compiler Specification Error message returned by the strerror function See part IV section 2 Standard Library Error Messages A 2 7 Not Supported Library Table A 24 lists libraries in the C language specifications not supported by the C compiler Table A 24 Libraries Not Supported by the C Compiler Header File _Library Name si
77. dlib h gt main FILE fp fp fopen file w fp NULL fclose fp error occurred printf sXn strerror errno print error message A Description 1 An error occurs because the file pointer value NULL is passed to the fclose function as an actual argument In this case an error number is set in errno 2 If the error number is passed to the strerror function as an actual argument a pointer to the corresponding error message is returned Specifying the character string to be output in the printf function outputs the error message 140 C Library Function Error Messages Functions to Set Error No Message Explanation Error Numbers 1100 Data out of range An overflow occurs atan cos sin tan cosh sinh tanh exp fabs frexp Idexp modf ceil floor strtol atoi fscanf scanf sscanf atol 1101 Data out of domain Results for mathematical acos asin atan2 log parameters are not defined log10 sqrt fmod pow 1102 Division by zero Division by zero was performed divbs divws divls divbu divwu divlu 1104 Too long string The length of the character string strtol strtod atoi atol exceeds 512 characters atof 1106 Invalid file NULL pointer constant is specified fclose fflush freopen pointer as file pointer value setbuf setvbuf fprintf fscanf printf scanf sprintf sscanf viprintf vprintf vsprintf fgetc fgets fputc fputs ungetc
78. e status register w Setting and referencing the vector base register x I O functions using the global base register Y System instructions which do not compete with register sources in C Description include lt machine h gt must be specified when using intrinsic functions Intrinsic Function Specifications Table 3 2 lists intrinsic functions Table 3 2 Intrinsic Functions Item Function Specification Description Status Setting the status void set_cr int cr Sets cr 32 bits in the status register register register Referencing the int get_cr void References the status register status register Setting the interrupt void set_imask int mask Sets mask 4 bits in the interrupt mask mask 4 bits Referencing the int get_imask void References the interrupt mask interrupt mask 4 bits Vector Setting the vector void set_vbr void base Sets base 32 bits in VBR base base register register Referencing the int get_vbr void References VBR VBR vector base register Global Setting GBR void set_gbr void base Sets base 32 bits in GBR base Referencing GBR void get_gbr void References GBR register Referencing GBR unsigned char References byte data 8 bits at GBR based byte gbr_read_byte int offset the address indicated by adding GBR and the offset specified Referencing GBR based word unsigned word gbr_read_word int offset 47 References word data 16 bits at the address indic
79. ecification 2135 Multiple initialization P Delete the initial value specification Asvartable sattiahzedcuiere hattornees S Ignores the second and subsequent initialization directives 2136 Type mismatch P Delete any redundant directives declared more than once with different data types S Uses the type specified in the definition declaration where a definition is declared Otherwise the data type specified in the first declaration is used P Use the same data type in the declarations 2137 Null declaration for parameter An identifier is not specified in the function 108 Error No Message Explanation parameter declaration S Ignores the corresponding parameter declaration P Delete the parameter declaration or insert 2138 Too many initializers the correct parameter name EEA EEN SEN PEE fred struct or array is greater than the number of struct members or array elements This error also occurs if two or more initial values are specified when the first members of a union are scalar S Uses only the initial values corresponding to the number of struct members array elements or the first members of union The rest are ignored P Specify the correct number of initial 2139 No parameter type values A type is not specified in a function parameter declaration S Assumes int as the parameter declaration type P Specify the correct type for the parameter 2140 Illegal bit fiel
80. eger type required for array s maximum size size_t Conversion from pointer type data to integer type data Pointer type data size gt Integer type data size Conversion from pointer type data to integer type data Pointer type data size lt Integer type data size Conversion from integer type data to pointer type data Integer type data size gt Pointer type data size Conversion from integer type data to pointer type data Integer type data size lt Pointer type data size Integer type required for holding pointer difference between members in the same array ptrdiff_t A 1 8 Register Table A 10 Register Specifications Item The maximum number of register variables that can be allocated to registers Type of register variables that can be allocated to registers 149 _C Compiler Specification unsigned long The lower byte of pointer type data is used Extended with signs The lower byte of integer type data is used Extended with signs long _C Compiler Specification 7 char unsigned char short unsigned short int unsigned int long unsigned long float and _pointers A 1 9 Structure Union Enumeration and Bit Field Types Table A 11 Specifications for Structure Union Enumeration and Bit Field Types Item Effect of setting a union member and referencing a union member using another member whose data type is different Structure member alignment Sign of an int b
81. environment setting 67 Initialization 69 Program configuration 67 Section initialization 70 Vector table setting 68 exit function termination routine 170 expansion suboption 9 Exponent 157 Extended specifications 43 Interrupt functions 43 Intrinsic functions 47 External identifier 32 179 F int 26 147 Fatal level error message level 92 Initialization 67 69 74 75 File close 79 Initialized data area 24 File extension 8 Internal data representation 26 File I O operation 57 77 80 Internal errors 92 close routine 84 Internal representation 26 157 Iseek routine 87 Internal representation of scalar type data 26 open routine 82 Interrupt functions 43 read routine 85 Stack switching specification 43 44 sbrk routine 88 Trap instruction return specification 43 44 write routine 86 Intrinsic functions 47 File naming 8 Invalid operation 162 float 26 148 159 VO operation 57 Floating point number specifications 148 157 Concept of I O operations 80 Denormalized number 158 Low level interface routine 80 Exponent 157 Routine for standard library function 77 Infinity 158 Limits on floating point numbers 148 J Internal representation format 157 Mantissa 157 K Normalized number 158 Not a number 158 L Sign 157 Language specifications 145 Types of values 158 Arrays and pointers 149 Floating point operation specifications 162 assert h 153 Conversion between decimal represen Characters 146 tation and internal representation 163 C library f
82. er 25 2 2 Internal Data Representation This section explains the internal representation of C language data types The internal representation of data is determined according to the following four items v Size Shows the amount of memory needed to store the data w Alignment Restricts the addresses to which data is allocated There are three types of alignment 1 byte alignment in which data can be allocated to any address 2 byte alignment in which data is allocated to an even byte address and 4 byte alignment in which data is allocated to an address indivisible by four x Data range Shows the range of scalar type data y Data allocation example Shows how the elements of aggregate type data are allocated Scalar Type Data Table 2 2 shows the internal representation of scalar type data used in C Table 2 2 Internal Representation of Scalar Type Data Data Range Size Alignment Minimum Maximum Data Type bytes bytes Sign Bit Value Value char 1 1 Used 27 128 27 1 127 signed char 1 1 Used 27 128 27 1 127 unsigned char 1 1 Unused 0 28 1 255 short 2 2 Used 215 32768 215 _ 1 32767 unsigned short 2 2 Unused 0 a 1 65535 int 4 4 Used 231 2147483648 1 2147483647 unsigned int 4 4 Unused 0 1 4294967295 long 4 4 Used 231 2147483648 1 2147483647 unsigned long 4 4 Unused 0 1 4294967295 enum 4 4 Used 231 214748364
83. er how to resolve the error 93 1 Warning Level Messages Error No Message Explanation 1000 Illegal pointer A pointer is assigned to a pointer with a assignment different data type S Sets the left hand side to the internal representation of the right hand side pointer The resultant type is the same as the data type of the left pointer P Use the cast operator to specify explicit type conversion 1001 Illegal comparison in The operands of the binary operator or operator are a pointer and an integer other than 0 S Selects an internal representation for the operands P Specify the correct type for the operands 1002 Illegal pointer for The operands of the binary operator gt operator lt gt or lt are pointers assigned to different types S Assumes that the operands are pointers assigned to the same type P Use a cast operator so that the same operand type will be used 1005 Undefined escape An undefined escape sequence a character sequence following a backslash is used in a character constant or string literal S Ignores the backslash P Remove the backslash or specify the correct escape sequence 1007 Long character The length of a character constant is 2 constant characters S Uses the specified characters P Check that the correct character constant is specified 1020 Illegal constant The operands of the binary operator in a 94 Error No Message
84. erted Return value type conversion A return value is converted to the data type returned by the function 37 Example v long f long f float x return x The return value is converted to long w void p int EDI charo T P 1 0 c NOE is converted to int because a type is not declared for the Note When parameter types are not declared by a prototype declaration the correct specifications must be made by the calling and called functions so that parameters are correctly transferred Otherwise correct operation is not guaranteed Example f x f float x float x main main float x float x f x f x Incorrect specification Correct specification Since the parameter type belonging to function f is not declared by a prototype declaration in the incorrect specification above parameter x is converted to double when function main calls function f Function f cannot receive the parameter correctly because the parameter type is declared as float in function f Use the prototype declaration to declare the parameter type or make the parameter declaration double in function f The parameter type is declared by a prototype declaration in the correct specification above 38 b Parameter area allocation Parameters are allocated to registers or when this is impossible to a stack parameter area Figure 2 2 shows the parameter area allocation
85. es processing P Place a semicolon at the end of the do statement 2314 Scalar required A control expression that determines statement execution for an if statement is not a scalar S Assumes that an int control expression is specified and continues processing P Use a scalar expression as the control expression for if statement 2316 Illegal type for return An expression in a return statement cannot be value converted to the type of value expected to be returned by the function S Assumes that the expression in the return statement is the type expected to be returned by the function and continues processing P Convert the expression in the return statement so that it matches the type of value expected 2400 Illegal character character An illegal character is detected S Assumes that the character is a blank character and continues processing P Delete the illegal character 124 Error No Message Explanation 2401 Incomplete character constant An end of line indicator is detected in the middle of a character constant S Assumes that a quotation mark is placed before the end of line indicator and continues processing P Correct the character constant 2402 Incomplete string An end of line indicator is detected in the middle of a string literal S Assumes that a double quotation mark is placed before the end of line indicator and continues processing P Correct t
86. etic operands 2 The left operand is a pointer and the right operand is an integer An undeclared name is used in an expression Assumes that the name is declared as an int Lidentifi i 2225 Undeclared name S 2226 tar reguired fer operator P processing Either declare the name or modify it so that it corresponds with one of the declared names The binary operator amp amp or is used in a non scalar expression 118 continues processing P Use scalar expressions as operands Error No Message Explanation 2227 Illegal type for equality The combination of operand types for the equality operator or is not allowed S Assumes that the result type is int and continues processing P Specify a correct type combination of operands Only the following three combinations of operand types for the equality operator or are allowed 1 Two arithmetic operands 2 Two pointers assigned to the same data type 3 A pointer and an integer constant whose value is 0 or another pointer assigned to void 2228 Illegal type for The combination of operand types for the comparison relational operator gt lt gt or lt is not allowed S Assumes that the result type is int and continues processing P Specify a correct type combination of operands Only the following two combinations of operand types are allowed for a relational operator 1 Two arithmetic operands 2 Tw
87. exceed the limit The nesting level ofinacro expansion 3011 Macro expansion nest too performed by a define directive exceeds the deep limit of 32 for UNIX systems and 16 for PC systems S Terminates processing P Modify the program so that the nesting level of macro expansion never exceeds the limit Note that a macro may be defined recursively The number of function definitions exceeds the 132 Error No Message Explanation 3012 Too many function definitions limit of 512 for UNIX systems and 256 for PC systems S Terminates processing P Divide the program so that the number of function definitions is less than or equal to the limit in one compile unit 3013 Too many switches The number of switch statements exceeds the limit of 256 for UNIX systems and 128 for PC systems S Terminates processing P Divide the program so that the number of switch statements is less than or equal to the limit in one compile unit 3014 For nest too deep The nesting level of for statements exceeds the limit of 16 for UNIX systems and 15 for PC systems S Terminates processing P Ensure that the for nesting level does not exceed the limit 3015 Symbol table overflow The numberof symbols to be generated by the C compiler exceeds the limit of 8192 for UNIX systems and 1024 for PC systems S Terminates processing P Divide the file so that the number of symbols does no
88. ge 2210 Integer required for Explanation S Assumes that the operand type is int and continues processing P Change the expression to an arithmetic expression The unary operator is used on a non integral 2211 Illegal sizeof expression S Assumes that the result type is int and continues processing Change the expression to an integral expression A sizeof operator is used for a bit field member function void or array_with an _ 2212 Illegal cast undefined size S Assumes int as the operand type and continues processing A sizeof operator cannot be used to obtain the size of a bit field function void or array with an undefined size Use an appropriate operand Either array struct or union is specified in a 2213 Arithmetic type required for operator 113 cast operator or the operand of a cast operator is void struct or union and cannot be converted S Assumes that the result is int and continues processing Cast operation can only be performed on scalar data items Use appropriate operands The binary operator or is used in an Ty op expression that is not arithmetic Error No Message 2214 Integer required for operator Explanation S Assumes int as the result and continues processing P Specify arithmetic expressions as the operands The binary operator lt lt gt gt amp lt lt
89. gisters calls initialization routine sequentially and then calls main function The coding example of this routine is shown below Example extern void _INITSCT void extern void main void void _INIT INITSCT Calls section initialization routine INITSCT main Calls main routine _main for Branches to endless loop after executing main r A function and waits for reset 4 69 3 3 Section Initialization __INITSCT To set the C program execution environment clear the non initialized data area with zeros and copy the initialized data area in ROM to RAM To execute the __INITSCT function the following addresses must be known e Start address 1 of initialized data area in ROM e Start address 2 and end address 3 of initilalized data area in RAM e Start address 4 and end address 5 of non initialized data area Interrupt vector Program area section P Constant area section C Initialized data area section D Initialized data area section R Non initialized data are section B 70 To obtain the above addresses create the following assembly programs and link them together ECTION S SECTION S S ECTION ECTION DATA I start address of section D el DATA I start address of section R Li DATA L end address of section R DATA I start address of section B DATA I end address of section B
90. gnal h signal raise stdio h remove rename tmpfile tmpnam stdlib h _getenv system time h clock difftime time asctime ctime gmtime localtime 156 A 3 Floating Point Number Specifications A 3 1 Internal Representation of Floating Point Numbers The internal representation of floating point numbers follows the standard IEEE format This section explains this standard Internal Representation Format float is represented in IEEE single precision 32 bits double and long double are represented in IEEE double precision 64 bits Internal Representation Structure Figure A 1 shows the structure of float double and long double in internal representation float 31 30 23 22 0 Sign Exponent Mantissa 1 bit 8 bits 23 bits double and long double 63 62 5251 0 Sign Exponent Mantissa 1 bit 11 bits 52 bits Figure A 1 Structure for the Internal Representation of Floating Point Numbers The elements of the structure have the following meanings i Sign This indicates the sign of a floating point number Positive and negative are represented by 0 and 1 respectively ii Exponent This indicates the exponent of a floating point number as a power of two iii Mantissa This determines the significant digits of a floating point number 157 Types of Values Floating point numbers can represent infinity in addition to general real numbers The rest of this section explains the types of values that can be represen
91. h routine required to execute library functions as follows 1 Setting vector table VEC TBL Sets vector table to initiate register initialization program _ _INIT and set the stack pointer SP at power on reset 2 Initializing registers __INIT Initializes registers and sequentially calls the initialization routines 3 Initializing sections __INITSCT Clears non initialized dasta area with zeros and copies the initialized data area in ROM to RAM This routine is supplied as a standard library function 4 Initializing C library functions __INITLIB Initializes C library functions required to be initialized and prepares standard I O functions 5 Closing files __CLOSEALL Closes all files with open status 6 Low level interface routine Interfaces library functions and user system when standard I O and memory management library functions are used Creation of the above routines is described below Note When using the C library functions that terminates program execution such as exit onexit or abort the C library function must be created according to the user system For details refer to addpendix D Termination Processing Function Example In addition when using C library function assert macro the abort function must be supplied 4 1 Setting Vector Table VEC_TBL Same as when no C library function is used For details refer to section 3 Setting the Execution Environment 74 4 2 Initializing Registers
92. he string literal 2403 EOF in commment An end of file indicator is detected in the middle of a comment S Assumes that the program ends when the end of file indicator is reached and continues processing P End the comment with 2404 Illegal character code character code An illegal character code is detected S Assumes that the character code is a blank character and continues processing P Delete the illegal character code 2405 Null character constant 125 There are no characters in a character constant 1 e no characters are specified between two guotation marks S Assumes that 0 is specified and continues processing P Correct the character constant Error No Message Explanation 2406 Out of float The number of significant digits in a floating point constant exceeds 17 S Depending on the sign the system assumes 00 or oo P Ensure that the number of significant digits in a floating point constant is less than or equal to 17 2407 Incomplete logical line A backslash or a backslash followed by an end of line indicator ARET is specified as the last character in a non empty source file S Ignores the last logical line P Delete the backslash or continue the physical line 2500 Illegal token An illegal token sequence is used S Ignores data up to a semicolon left brace right brace comma or keyword if while for switch do case
93. ic in in line function function does not match the required number S Ignores the intrinsic function P Specify the correct number of parameters 2801 Illegal parameter type in There are different parameter types in an in line function intrinsic function S Ignores the intrinsic function P Specify the correct parameter types 2802 Parameter out of range in A parameter exceeds the range that can be 128 Error No Message Explanation in line function specified by an intrinsic function S Ignores the intrinsic function P Check the range that can be specified for the parameter and specify it correctly in line function intrinsic function S Ignores the intrinsic function P Check the intrinsic function specifications and specify it correctly 129 3 Fatal Level Messages Error No Message Explanation 3000 Statement nest too deep The nesting level of an if while for do and switch statements exceeds the limit of 32 for UNIX systems and 15 for PC systems S Terminates processing P Modify the program so that the nesting level is less than or equal to the limit 3001 Block nest too deep The nesting level of compound statements exceeds the limit of 32 for UNIX systems and 15 for PC systems S Terminates processing P Modify the program so that the nesting level is less than or equal to the limit 3002 if nest too deep The conditional compilation if ifdef ifndef elif and e
94. imultaneously 20 20 by the open function Notes 1 As the C compiler itself defines five macro names LINE __ FILE __ DATE __ 2 3 4 TIME __ and STDC ___ the user can define a maximum of 4091 macro names in UNIX systems and a maximum of 1019 macro names in PC systems As the C compiler itself defines two symbols the user can define a maximum of 4094 external identifiers in UNIX systems and a maximum of 509 external identifiers in PC systems An internal label is internally generated by the C compiler to indicate a static variable address case label address goto label address or a branch destination address generated by if switch while for and do statements The number of symbol table entries is determined by adding the following numbers Number of external identifiers Number of internal identifiers for each function Number of string literals Number of initial values for structures and arrays in compound statements Number of compound statements Number of case labels Number of goto labels 222 Section 2 Executing a C Program This section covers object programs which are generated by the C compiler In particular this section explains what items are required to link C programs with assembly programs and how to install programs on the SH system see PART III SYSTEM INSTALLATION This section consists of the following three parts Section 2 1 Structure of Object Programs This section discusses
95. ing 1 Parameter name Parameter Meaning of the parameter type Return value Type Type of return value Normal Return value for normal termination Abnormal Return value for abnormal termination 81 a open routine Purpose Opens a file Interface int open char name int mode Parameters No Name Type Meaning 1 name Pointer String literal indicating a file name to char 2 mode int Processing specification Return value Type int Normal File number of the file opened Abnormal 1 Explanation The open routine opens the file specified by the first parameter file name and returns a file number The open routine must determine the file device type console printer disk etc and assign this information to the file number The file type is referenced using the file number each time a read write operation is performed The second parameter mode gives processing specifications for the file The effect of each bit of this parameter is explained below 15 543210 mode O_RDONLY O_WRONLY O_RDWR O_CREAT O_TRUNC O_APPEND O_RDONLY bit 0 If this bit is 1 the file becomes read only O_WRONLY bit 1 If this bit is 1 the file becomes write only O_RDWR bit 2 If this bit is 1 the file becomes read write 82 O_CREAT bit 3 If this bit is 1 and the file indicated by the file name does not exist a new file is created O_TRUNC bit 4 If this bit is 1 and the file indicated by
96. ion int a 4 bytes 4 bytes 3 w a char b 7 w p 4 J w b 0 w b 1 w b 2 w b 3 w b 4 w b 5 w b 6 1 byte Bit Fields A bit field is a member of a structure This part explains how bit fields are allocated e Bit field members Table 2 4 shows the specifications of bit field members Table 2 4 Bit Field Member Specifications Item Specifications Type specifiers allowed for bit fields char unsigned char short unsigned short int unsigned int long and unsigned long How to treat a sign when data is A bit field with no sign unsigned type is specified Zero extension 2 expanded to the declared type 1 A bit field with a sign unsigned is not specified Sign extension 2 Notes 1 To use a member of a bit field data in the bit field is expanded to the declared type 2 Zero extension Zeros are written to the high order bits during extension Sign extension The most significant bit of a bit field is used as a sign and is written to all higher order bits generated during data extension Note One bit field data with a sign is interpreted as the sign and can only indicate 0 and 1 To indicate 0 and 1 bit field data must be declared with unsigned 28 Bit field allocation Bit field members are allocated according to the following five rules v Bit field members are placed in an area beginning from the left that is the most significant bit Example struct bl Unused area int a 2 X a int
97. ion An error is output at a Example An interrupt function is not declared int 2 fl j b Description Because function f1 is not declared as an interrupt function an object for extern int f1 is generated If function f1 is declared as an interrupt function not to be compiled in the same file as f2 correct program execution is not guaranteed 45 x A function declared as an interrupt function cannot be referenced in the same file Example pragma interrupt f1 main void a void a fl ccc cc etre eeee a Description Since the address of interrupt function f1 cannot be referenced at a an error is output If an interrupt function is referenced to set for example a vector table it must not be declared as an interrupt function in the same file Examples pragma interrupt f1 extern void fl void b main void f1 void void a void a f1 File with an interrupt function definition File referencing an interrupt function Description To reference the address of interrupt function f1 at b f1 is not declared as an interrupt function 46 3 2 Intrinsic Functions In this C compiler system control instructions of the SH microcomputer can be written in C as intrinsic functions The following describes the intrinsic functions provided Intrinsic Functions The following functions can be specified by intrinsic functions v Setting and referencing th
98. ion as a member 103 A member of a struct or union is declared as a function S Assumes int as the member type P Declare the correct member type Error No Message 2112 Illegal bit field Explanation The type specifier for a bit field is illegal char unsigned char short unsigned short int unsigned int long unsigned long or a combination of const or volatile with one of the above types is allowed as a type specifier for a bit field S Ignores the bit field specification and assumes that the member is not a bit field P Specify the correct type 2113 Bit field too wide The width of a bit field is greater than the size 8 16 or 32 bits indicated by its type specifier S Ignores the bit field specification and assumes that the member is not a bit field P Specify the correct bit field width 2114 Multiple variable declarations A variable name is declared more than once in the same scope S Uses the first declaration and ignores subsequent declarations P Keep one of the declarations and delete or modify the rest 2115 Multiple tag declarations A struct union or enum tag name is declared more than once in the same scope S Uses the first declaration and ignores subsequent declarations P Keep one of the tag name declarations and delete or modify the rest 2117 Empty source program 2118 There are no external definitions in the source program S Termina
99. ise result of a floating point operation exceeds the significant digits of the internally represented mantissa the result is rounded as follows The result is rounded to the nearest internally representable floating point number If the result is directly between the two nearest internally representable floating point numbers the result is rounded so that the lowest bit of the mantissa becomes 0 ii Overflow and Underflow Handling Invalid operations overflows and underflows resulting from numeric operations are handled as follows For an overflow positive or negative infinity is used depending on the sign of the result For an underflow positive or negative zero is used depending on the sign of the result An invalid operation is assumed when 1 infinity is added to infinity and each infinity has a different sign 11 infinity is subtracted from infinity and each infinity has the same sign iii zero is multiplied by infinity iv zero is divided by zero or v infinity is divided by infinity In each case the result is not a number In any case the variable errno is set to the error number corresponding to the error See part IV Error Messages section 2 C Library Error Messages for the error number Note Operations are performed with constant expressions at compile time If an overflow underflow or invalid operation is detected during these operations a warning level error occurs iii
100. it field Allocation order of bit fields in int area Result when a bit field has been allocated in an int area and the next bit field to be allocated is larger than the remaining int Type specifier allowed for bit field Integer describing enumeration _C Compiler Specification Reference is possible but the referred value is not guaranteed Structures consisting of char members are aligned in 1 byte units while structures consisting of short members are aligned in 2 byte units Structures consisting of any other members are _aligned in 4 byte units _Assumed to be signed int Beginning from the high order bit to low order bit The next bit field is allocated to the next int area 2 char unsigned char short unsigned short int unsigned int long and unsigned long int Notes 1 See section 2 2 2 Aggregate Data in part Il for details on structure member allocation 2 See section 2 2 3 Bit Fields in part II for details on bit field allocation A 1 10 Modifier Table A 12 Modifier Specifications Item volatile data access type 150 _C Compiler Specification _Not specified A 1 11 Declarations Table A 13 Declaration Specifications Item _C Compiler Specification Number of types that can qualify the basic types pointer Up to 16 types can be specified array and function a Example of counting the number of types that qualify the basic types Examples G inta
101. ks are expanded as one blank character _ pragma interrupt is supported Data depending on the host machine timer when the compilation starts Note 1 See section 3 1 Interrupt Functions in part II for details on pragma interrupt specifications 152 A 2 C Library Function Specifications This section explains the specifications for C library functions declared in standard include files Refer to the include file for the actual macro names defined in a standard include file A 2 1 stddef h Table A 16 stddef h Specifications Item C Compiler Specification Value of macro NULL The value 0 of pointer to void Contents of macro ptrdiff_t long A 2 2 assert h Table A 17 assert h Specifications Item C Compiler Specification Information output and terminal operation of See a for the format of output information The assert h program outputs information and then calls the abort function to stop the operation a The following message is output when the expression is 0 for assert expression Assertion Failed lt expression gt File lt file name gt Line lt line number gt 153 A 2 3 ctype h Table A 18 ctype h Specifications Item The character set for which the isalnum isalpha iscntrl islower isprint and isupper functions Table A 19 Set of Characters that Returns True Function Name isalnum _0 to 9 a to zZ a to isalpha JA to z a to z iscntrl _ 0 to 037 177 islower _a to z i
102. lse nesting level exceeds the limit of 32 for UNIX systems and 6 for PC systems S Terminates processing P Modify the program so that the nesting level is less than or equal to the limit 3003 Too many external a Y SATOR The number of external identifiers exceeds the limit of 4096 for UNIX systems and 511 for PC systems identifiers S Terminates processing P Divide the program so that the number of external identifiers is less than or equal to the limit The number of effective identifiers internal 130 Error No Message 3004 Too many local identifiers Explanation identifiers in one function exceeds the limit of 4096 for UNIX systems and 512 for PC systems S Terminates processing P Divide the compound statements so that the number of identifiers declared in one compound statement is less than or equal to the limit Fhe number ofinaere names_definedi na 3005 Too many macro identifiers define directive exceeds the limit of 4096 for UNIX systems and 1024 for PC systems S Terminates processing P Divide the program so that the number of macro names is less than or equal to the limit The number of parameters in either a function 3006 Too many parameters declaration or a function call exceeds the limit of 63 for UNIX systems and 31 for PC systems S Terminates processing P Divide the compound statements so that the number of identifiers declared i
103. mistake with respect W 1000to Yes Yes Check the list of error messages to to language specifica 1999 decide whether error recovery tions The compiler performed by the C compiler is has performed error correct If necessary modify and recovery recompile the source program 2 Error A mistake in language E 2000 to No Yes Correct the error and recompile the specifications 2999 source program 3 Fatal The source program F 3000 to No No Correct the error and recompile the exceeds the limit of the 3999 source program C compiler 4 Internal An error has occurred 4000 to No No Contact the sales office or represen in an internal process 4999 tative where the C compiler was of the C compiler 92 purchased 1 3 List of Error Messages This section gives lists of error messages in order of error number A list of error messages are provided for each level of errors Example Error Number Message Explanation v 2226 W Scalar required x Binary operator amp amp or is used in an for an operator expression that is not scalar Y S Assumes that the result is int and continues processing z P Specify a scalar expression as the operand Error Number Error Message This message 1s sent to the standard output device normally a terminal x Explanation This gives more details about the error y System Action This indicates the reaction of the C compiler to the error z Programmer Response This indicates to the programm
104. n a a ae Ges 8 33 Compiler Options s sess ersi eaa Ee Eae asto E EESE EREE SS 9 3 4 Option amma UNOS cds sean sens ancesuceseanedhasccnse toeoustendentexdsecvansencanseisadessn AEREE reaS A ENEE ASSE 11 3 3 C Compiler Listing Sissin rinier ri R EE A E E EAEE REE 12 Part IT PROGRAMMING cccccccseeseesesnestesesneeeesesneseeseseeneeeeaeenereeseeneeeeneaneneeess 19 Section 1 Limitations of the C Compiler j cccccccececseeseseeeseseeeeeeeesesesneneneaeeeeeeees 21 Section 2 Executing a C Program te yzss rezisid e bida einen 23 Zl Structure of Object Programs seinri r EE E E EEEE 24 2 2 Int r al Data Representation siiisiciiscie srein ainaani eee 26 2 3 Linkage with Assembly Programs sa cccssesnsqssasedeedensscieosnesuasbenasesesvacseandaa ituasdedenastatanceeunedes 31 2 3 1 External Identifier Rerercsies ssxciceiasisdecasadssicsastecedensssdasaaagvancesanadseaaiauaezerenietasoonrs 32 242 Bun hon Call Uitertaee 4 crdieciuhe Sieg aah awed elec 34 ee 43 SL Interrupt FUMCHONS sii eccessstevcasisnenadsarsucauaasiadansasduecesasnaveadaspedaasisbesaesaahuvendesaecanarabvenevasnezeds 43 3 2 ROCs FUNCHUONS ossiani n A E E eee 47 EE EEE 51 41 Coding MNO seiorn ia EE A RER 51 4 2 Notes on Programming Develo mic vss sccisiasteacss scdcaessiedeccsesdacessiocaacsiesseceaeastecasss aceite 54 Part III SYSTEM INSTALLATION eee eee eee een 55 ASPON 57 Section 2 Allocating Memory Areas c c ccccececscseseseessseseseeeeeeeeeeseseseeeseseeeeee
105. n int h char int double int h char a register int b double c char Ae d ga h d b c register int i i d return i 63 KKKKKKKKKKKK OBJECT LISTING KKKKKKKKKKKK E m0251 c OFFSE CODE BE INSTRUCTION OPERAND COMMI 00000000 function h _frame size 20 00000000 E MOV L R14 1 00000002 MOV L R13 Lower addresses R15 SP 0 pa a O RS function Parameter area For stack parameter 8 bytes Upper addresses The size of stack used by a function is determined by adding frame size and parameter area size for stack parameter Accordingly in the above example the stack size used by the function is calculated as follows 20 8 28 bytes For details on the size of parameters to be pushed onto the stack refer to the description of parameter and return value setting and referencing in section 2 3 2 of Part II 64 e Stack size calculation The following example shows a stack size calculation depending on the function call nesting Example Figure 2 3 illustrates the function call nestings and stack size Function Name Stack Size Bytes main 24 32 24 Figure 2 3 Nested Function Calls and Stack Size If funtion g is called via function f stack area size is calculated according to the formula listed in table 2 1 Table 2 1 Stack Size Calculation Example Function Calling Rou
106. n one compound statement is less than or equal to the limit 3007 Too many macro parameters 131 The number of parameters in a macro definition or a macro call exceeds the limit of 64 for UNIX systems and 31 for PC systems S Terminates processing P Modify the program so that the number of macro parameters is less than or equal to the limit After a macro expansion the length of a line Error No Message Explanation 3008 Line too long exceeds the limit of 4095 characters for UNIX systems and 512 characters for PC systems S Terminates processing P Divide the line so that its length does not exceed the limit after macro expansion Thetensth of string literals exceeds 512 3009 String literal too long characters The length of string literals is the byte number generated after the specified string is connected continuously The length of string literals in the source program is not the length of the source program in the string data This byte number is located in the string literal data with the expansion sign counted as one character S Terminates processing P Modify the program so that the total length of string literals does not exceeds 512 bytes The nesting level of the include directive 3010 Processor directive exceeds the limit of 8 for UNIX systems and 5 include nest too deep for PC systems S Terminates processing P Ensure that the file inclusion nesting level does not
107. neneeeseanes 58 ZL Static Area A MWGCAUON c caccesactsocaysthonatiussacenvesidvansesdasteseanasessastesadaalebeceeaadaseapessecassaasuarivasnesens 58 2 1 1 Data to be Allocated in Static Area ccc ececsscccrccsticeteiedacet eaters ccttoneecdcteetageeencce 58 2 12 Static Area Size Calctlatioi sssini aeeaiei sees 58 2 1 3 ROM and RAM Allocation sccicessccctzcssdcntancsdsonincesnctedenseotzcasssnleetasoeaisaceyscesdeseselvsdes 6l 2 1 4 Initialized Data Area Allocation eseeeeeeeseessessessseessetssstesseesserssererssessseessresse 61 2 1 5 Example Memory Area Allocation and Address Specification at Program De PRE ee 61 2 2 Dynamic Area Allocation ce Gii nkest a cddakns aa nao kAn s ee eatuansnaateaeeatavandosesteageeaaeadis 63 22 YY WANS AMC AS coescis is cxta seca accssexcacten dd isir reiser onitan ania KESSEN ENSE Esi eisin 63 222 D nameA saze Calculation sssrini unie o aan E EE EES 63 2 2 3 Rules for Allocating Dynamic Area ces caiciesscutlenieniidsereeaceatcereuleesahiddeawactenesamease 66 Www 67 Sb Vector Table Setting VEC BL sada dada ode ner dce din kareta o tak 68 32 Imualiz tio n INIT ciasne tiise ngia ra AE E A ASE a E ENESTE E 69 3 3 Section Initialization _INITSCT eeesseneeeseesssesoseeeesessssosroeesssssssseressesssssseereeessssssene 70 eee 73 4 1 Setting Vector Table VEC TBL Joseina a r 74 4 2 Initializing Registers AMY ssssesesesesssesessseessresssessesssseesssresseessersseeesseeessressee
108. ng unsigned long If the return value type is char or short perform sign float and Pointer extension before setting the return value in RO If the return value type is unsigned char or unsigned short perform zero extension before setting it in RO double long double structure and union Return value setting area memory Figure 2 4 Return Value Setting Area Used When Return Value Is Written to Memory Return value address area Return value setting area allocated by the calling side Parameter area 42 Section 3 Extended Specifications This section describes two C compiler extended specifications interrupt functions and intrinsic functions 3 1 Interrupt Functions A preprocessor directive pragma specifies an external hardware interrupt function The following section describes how to create an interrupt function Description pragma interrupt function name interrupt specifications function name interrupt specifications Table 3 1 lists interrupt specifications Table 3 1 Interrupt Specifications Item Form Options Specifications Stack switching sp lt variable gt The address of a new stack is specified with a amp lt variable gt variable or a constant lt constant gt lt variable gt Variable object type value amp lt variable gt Variable pointer type address lt constant gt Constant value Trap instruction tn lt constant gt Termination is specifie
109. nts for Precise Conversion Rounding as explained in the description Result Rounding in appendix A 3 4 Floating point Operation Specifications is performed precisely for floating point numbers whose exponents are in the following ranges For float 0 lt M lt 109 1 O lt N lt 13 For double and long double O lt M lt 10 7 1 0SN lt 27 An overflow or underflow will not occur if the exponent is within the proper ranges Conversion and Rounding Error The difference between i the error occurring when the exponent outside the proper range is converted and 11 the error occurring when the value is precisely rounded does not exceed the result of multiplying the lowest significant digit by 0 47 If an exponent outside the proper range is converted an overflow or underflow may occur In such a case a warning level error occurs at compile time and the variable errno is set equal to the corresponding error number at run time 164 Appendix B Parameter Allocation Example Example 1 Register parameters are allocated to registers from R4 to R7 depending on the order of declaration int f char short int float R4 Sign extension 1 2 3 4 0 R5 Sign extension R6 R7 Example 2 Parameters which could not be allocated to registers from R4 to R7 are allocated to the stack area as shown below If a char unsigned or short unsigned type parameter is allocated to a parameter area on a stack it is extended t
110. o a 4 byte area int f int short long float char 1 2 3 4 0 5 R5 Sign extension R6 R7 A Lower address Parameter area stack Sign extension y Upper address 165 Example 3 Parameters having a type that cannot be allocated to registers from R4 to R7 are allocated to the stack area struct s int x y a int f int struct s int f 1 a 3 A Lower address Parameter area stack y Upper address Example 4 If a function whose number of parameters changes is specified by prototype declaration parameters which do not have a corresponding type in the declaration and the immediately preceding paramters are allocated to a stack 1 0 2 3 4 A Lower address Parameter area stack y Upper address 166 Example 5 If no prototype is declared char and float types are extended to int and double types respectively a TT 4 Lower address Parameter area stack y Upper address Example 6 If a value returned by a function exceeds 4 bytes or is a structure type a return value is specified just before parameter area If structure size is not a multiple of four an unused area is generated A Lower address struct s char x y z a b double f struct s Return value address f a Unused a Parameter area stack y Upper address Return value setting area 167 Appendix C Usage of Registers and Stack Area This section describes how to use regi
111. o pointers assigned to the same data type 2230 Illegal function call An expression which is not a function type or a pointer assigned to a function type is used for a function call S Ignores the actual argument list and the parentheses which indicate this list P Specify a function type expression or pointer assigned to a function type correctly 119 Error No Message Explanation 2231 Address of bit field The unary operator amp is used on a bit field S Ignores the bit field assumes that the unary operator amp is correctly specified and continues processing P Correct the expression A bit field address cannot be used 2232 Illegal type for operator A type that is not a scalar or that is a pointer assigned to a function or void is specified as the operand for the prefix operator or S Assumes int as the result type and continues processing P Use an operand that is a scalar other than a pointer assigned to a function or void 2233 Illegal array reference An expression used as an array is not one of the following types Array Pointer assigned to a data type other than a function or void S Ignores the square brackets and the array subscript enclosed P When an array subscript is required specify the correct expression 2234 ELL 1t def xr Oak ad A typedef name is used as a variable in an reference expression S Ignores the expression P Use ty
112. o registers Figure 2 3 Example of Allocation to Parameter Registers f char a int b EO o sO 1 ii Allocation to a stack parameter area Parameters are allocated to the stack parameter area starting from lower addresses in the order that they are specified in the source program Note Regardless of the alignment determined by the structure type structure type or union type parameters are allocated using 4 byte alignment Also the area size for each parameter must be a multiple of four bytes This is because the SH stack pointer is incremented or decremented in 4 byte units Refer to appendix B Parameter Allocation Examples for examples of parameter allocation d Return value location 41 The return value is written to either a register or memory depending on its type Refer to table 2 7 for the relationship between the return value type and location When a function return value is to be written to memory the return value is written to the area indicated by the return value address The calling side must allocate this return value setting area in addition to the parameter area and must set the address of the former in the return value address area before calling the function The return value is not written if its type is void Table 2 7 Return Value Type and Setting Location Return Value Type Return Value Location char unsigned char short unsigned short RO 32 bits int unsigned int lo
113. o the standard output file normally a terminal Figures 1 1 and 1 2 show the formats used for error messages sample c line 23 2011 E Line too long V W X Y Z Figure 1 1 Error Messages Format UNIX Systems sample c 23 2011 E Line too long V W X Y Z Figure 1 2 Error Messages Format PC Systems Explanation v File name File name sample c of the source program in which the error was detected w Line number Line number 23 where the error was detected x Error number This number is unique to the error message See section 1 3 List of Error Messages for details on the errors and appropriate programmer responses y Message level The severity of the error See section 1 2 Message Levels for details z Message text This describes the error Note When an error not related to the source program has occurred e g an error internal to the compiler the file name is not output for the line number here O is output in UNIX systems and nothing is output in PC systems 91 1 2 C Compiler Action and Programmer Response for Each Error Level Error messages are classified into the following four levels according to their severity Table 1 1 shows C compiler action for each level of errors Table 1 1 C Compiler Action and Programmer Response for Each Error Level Object Error Error Error Program Processing No Level Meaning Symbol Number Output Continues User Response 1 Warning A
114. of each parameter or return value is explicitly declared in the function declaration A function prototype declaration is used to explicitly declare the type of each parameter or return value The rest of this section explains the general rules concerning parameters and return values how the parameter area is allocated and how areas are established for return values a General rules concerning parameters and return values i Passing parameters A function is called only after parameters have been copied to a parameter area in registers or on the stack Since the calling function does not reference the parameter area after control returns to it the calling function is not affected even if the called function modifies the parameters ii Rules on type conversion Type conversion may be performed automatically when parameters are transferred or a return value is returned This section explains the rules on type conversion Type Conversion of Parameters Whose Types are Declared Parameters whose types are declared by prototype declaration are converted to the declared types Type conversion of parameters for which types are not declared Parameters whose types are not declared by prototype declaration are converted according to the following rules e Parameters whose types are char unsigned char short or unsigned short are converted to int e Parameters whose types are float are converted to double e Other parameters are not conv
115. one another a request to allocate a new area may be rejected even if the net size of the free areas is sufficient To prevent this take note of the following If possible allocate the largest area first after program execution is started If possible specify data area size to be reused as a constant 2 2 3 Rules for Allocating Dynamic Area The dynamic area is allocated to RAM The stack area is determined by specifying the highest address of the stack to the vector table and refer to it as SP stack pointer The heap area is determined by the initial specification in the low level interface routine sbrk For details on stack and heap areas refer to section 3 1 Vector Table Setting VEC_TBL and section 4 6 Creating Low Level Interface Routine respectively 66 Section 3 Setting the Execution Environment This section describes the environment required for C program execution A C program environment specification program must be created according to the system specification because the C program execution environment differs depending on the user systems In this section basic C program execution specification where no C library function is used is described as an example Refer to section 4 Setting C Library Function Execution Environment for details on using C library functions Figure 3 1 shows an example of program configuration Required routine Power on Required table raset VEC_TBL _ _INITSCT User p
116. orrect width for the bit field 2105 Incomplete tag used in declaration An incomplete tag name declared with a struct or union or an undeclared tag name is used in a typedef declaration or in the declaration of a data type not assigned to a pointer or to a function return value S Assumes that the incomplete or undeclared tag name is an int P Declare the incomplete or undeclared tag name 102 Error No Message 2106 Extern variable initialized Explanation A compound statement specifies an initial value for an extern storage class variable S Ignores the initial value P Specify the initial value for the external definition of the variable 2107 Array of function An array with a function member type is specified S Ignores the function or array type P Specify the correct type 2108 Function returning array A function with an array return value type is specified S Ignores the function or array type P Specify the correct type 2109 Illegal function declaration A storage class other than extern is specified in the declaration of a function variable used in a compound statement S Assumes extern as the storage class P Specify the correct storage class 2110 Illegal storage class The storage class in an external definition is specified as auto or register S Assumes that the storage class is extern P Specify the correct storage class 2111 Funct
117. ose execution results differ depending on the evaluation order Example a i a i The value of i on the left side differs depending on whether the right side of the assignment expression is evaluated first sub i i The value of i for the second parameter differs depending on whether the first function parameter is evaluated first 51 Overflow Operation and Zero Division At run time if overflow operation or zero division is performed error messages will not be output However if an overflow operation or zero division is included in the operations for one or more constants error messages will be output at compilation Example main int ia int ib float fa float fb ib 32767 fb 3 4e 38f Compilation error messages are output when an overflow operation and zero division are included in operations for one or more constants ia 99999999999 W Detect integer constant overflow fa 3 5be 40f W Detect floating pointing constant overflow W W ia 1 0 E Detect division by zero fa 1 0 0 0 W Detect division by floating point zero No error message on overflow at execution is output ib ib 32767 Ignore integer constant overflow fb fb 3 4e 38f Ignore floating point constant overflow Assignment to const Variables Even if a variable is declared with const attribute if assignment 52 is done to a variable other than const converte
118. pedef correctly 2235 Illegal cast An attempt is made to cast a pointer with a floating point type S Ignores the attempt P Cast the pointer with an integer type then with a floating point type 120 Error No Message Explanation 2236 Illegal cast in constant An attempt is made to cast a pointer with a char or short S Ignores the cast operation P Use an expression other than a constant one 2237 Illegal constant expression In a constant expression a pointer constant is cast with an integer and the result is manipulated S Assumes that the conversion is not specified and continues processing Use an expression other than a constant expression 2238 Lvalue or function type required for a The unary operator amp is used on the left value or is used in an expression other than function type S Assumes that an expression with a left value is specified as the operand and continues processing Specify an expression that has a left value or a function type expression as the operand 2300 Case not in switch A case label is specified outside a switch statement S Ignores the case label P Specify the case label in a switch statement 2301 Default not in switch A default label is specified outside a switch statement S Ignores the default label P Specify the default label in a switch statement 121 Error No Message 2302 Multiple labels Explanation
119. r distinction in external identifiers used for external linkage C Compiler Specification Refer to part IV Error Messages C Compiler Specification Not specified Not specified C Compiler Specification The first 31 characters are valid The first 31 characters are valid Lowercase characters are distinguished from uppercase characters Note Two different identifiers with the same first 31 characters are considered to be identical Example a longnameabcdefghijklmnopgrstuvwx b longnameabcdefghijklmnopgrstuvwy Identifiers a and b are indistinguishable because the first 31 characters are the same 145 A 1 4 Characters Table A 4 Character Specifications Item Elements of character set and codes used during program execution Shift state used for encoding multiple byte characters The number of bits used to indicate a character sets during program execution Correspondence between the program compilation character set and the execution Extended representation that appears either in a character constant or a string literal and that is not defined in the language specifications Character constant or wide character constant of two or more characters locale specifications used to converting multiple byte character to wide character Simple char having normal the value range same as signed char 146 _C Compiler Specification ASCII character set Kanji used in host environment can be
120. r level error or a fatal level error has occurred when option noobject is specified In addition section size information x is not output when option code asmcode is specified command Line Specification The file names and options specified on the command line when the compiler is invoked are displayed Figure 3 5 shows an example of command line 16 specification information Figure 3 5 command Line Specification COMMAND PARAMETER listfile test c 17 PART IT PROGRAMMING Section 1 Limitations of the C compiler Table 1 1 shows the limits on source programs that can be handled by the C compiler Source programs must fall within these limits To edit and compile efficiently it is recommended to split the source program into smaller programs approximately 2 ksteps and compile them separately Table 1 1 Limitation of the C Compiler Limit Classification Item UNIX PC Invoking the Number of source programs that can be 16 1 C compiler compiled at one time Total number of macro names that can be 16 16 specified using the define option Length of file name characters 128 128 Source programs Length of one line characters 4096 512 Number of source program lines 32767 16383 Preprocessing Nesting level of files in an include directive 8 5 Total number of macro names that can be 4096 1024 specified in a define directive a Number of argument
121. rap instruction return specification interrupt function 43 44 Troubleshooting 54 U Underflow 162 Union type 27 150 unsigned 26 147 V Vector base register VBR 47 Vector table setting 67 68 74 W width suboption 9 write routine low level interface routine 86 X Y Z Zero extension 28 182
122. rogram Figure 3 1 Program Configuration No C Library Function is Used Each routine is described below Vector table setting VEC_TBL Sets vector table so as to initiate register initialization program _ _INIT and set the stack pointer SP by power on reset Initialization _ _INIT Initializes registers and sequentially calls initialization routines Section initialization __INITSCT Clears the non initialized data area with zeros and copies the initialized data area in ROM to RAM How to create the above routines are described as follows 67 3 1 Vector Table Setting VEC_TBL To call register initialization routine _ _INIT at power on reset specify the start address of function INIT at address 0 in the vector table Also to specify the SP specify the highest address of the stack to address H 4 When the user system executes interrupt handlings interrupt vector settings are also performed in the VEC_TBL routine The coding example of VEC_TBL is shown below Example ECTION VECT DATA LOCATE H 0000 H Assigns section VECT to address H 0 by the section directive Assigns the start address of INIT to addresses H 0x0 to H 0x3 Assigns the SP to addresses H 0x4 to H 0x7 a The highest address of the stack ORG H 00000100 DATA L _IRQO Assigns the start address of IRQO to addresses H 0x100 to END H 0x103 68 3 2 Initialization __INIT _ _INIT initializes re
123. s and 15 for PC systems S Terminates processing P Ensure that the switch nesting level does not exceed the limit The number of types pointer array and Error No Message Explanation 3023 Type nest too deep function that qualify the basic type exceeds 16 S Terminates processing P Ensure that the number of types is less than or equal to 16 3024 Array dimension too deep S Terminates processing P Ensure that arrays have no more than six dimensions A source file name is not specified in the 3025 Source file not found command line S Terminates processing P Specify a source file name ne Sines l 3026 Expression too complex S Terminates processing P Divide the expression into smaller units The nesting level of statements inthe program 3027 Source file too complex is too deep or an expression is too complex S Terminates processing P Reduce the nesting level of statements or divide the expression The last source line number exceeds the limit 3028 Source line number of 32767 for UNIX systems and 16383 for PC overflow systems S Terminates processing P Modify both the line count specified in the line directive and the source program so that the last source line number is less than or equal to the limit The number of physical lines including the 135 Error No Message Explanation 3029 Physical line overflow include files exceeds the limit of 32
124. s for I O are used library must be initialized according to the execution environment specification Specifically if I O function stdio h and memory allocation function stdlib h are used the user must create low level I O routines and memory allocation routines appropriate to the user system Section 2 describes how to allocate C programs in memory area and how to specify linkage editor s commands that actually allocate a program in memory area using examples Section 3 describes items to be specified in execution environment setting and execution environment specification programs Section 4 describes how to create C library function initialization and low level routines 57 Section 2 Allocating Memory Areas To install an object program generated by the C compiler on a system the size of each memory area must be determined then the areas must be appropriately allocated in memory Some memory areas such as the area used to store machine code and the area used to store data declared using external definitions are allocated statically Other memory areas such as the stack area are allocated dynamically This section describes how the size of each area is determined and how to allocate an area in memory 2 1 Static Area Allocation 2 1 1 Data to be Allocated in Static Area Sections of object programs such as program area constant area initialized data area and non initialized data area are allocated to the s
125. s of C compiler listings Table 3 5 Structure and Contents of C Compiler Listings List Structure Source listing Contents Listing consists of source programs Source program listing after include file and Option Specification Method Default show source Output show nosource show include 2 No output show expansion macro expansion show noinclude show noexpansion Object listing Machine language generated show object Output by the C compiler show noobject Statistics Total number of errors the show statistics Output number of source program show nostatistics command line specification lines length of each section byte and the number of symbols File names and options specified on the command Output line Notes 1 All options are valid when listfile is specified 2 The option enclosed in parentheses is only valid when show source is specified Source Listing The source listing can be output in two ways When show noinclude and show noexpansion is specified the unpreprocessed source program is output When show include 12 or show expansion is specified the preprocessed source program is output Figures 3 1 and 3 2 show these output formats respectively Bold characters in figure 3 2 show the differences Figure 3 1 Source Listing Output for show noinclude and noexpansion KKEKKKKKKKKKK SOURCE LISTING KKEKKKKKKKKKK
126. s that can be specified 63 31 using a macro definition or a macro call operation Depth of the recursive expansion of a macro name 32 16 Nesting level of if ifdef ifndef else or elif 32 6 directives Total number of operators and operands that can 512 210 be specified in an if or elif directive Declarations Number of function definitions 512 256 Number of external identifiers used for external 4096 511 linkage lt Number of internal identifiers that can be used 4096 512 in one function Number of internal labels 16384 2048 Number of symbol table entries 4 8192 1024 Total number of pointers arrays and functions 16 16 that qualify the basic type Array dimensions 6 6 211 Table 1 1 Limitation of the C Compiler cont Limit Classification Item UNIX PC Statements Nesting levels of compound statements 32 15 Levels of statement nesting in a combination of 32 15 repeat while do and for and select if and switch statements Number of goto labels that can be specified in 511 256 one function Number of switch statements 256 128 Nesting levels of switch statements 16 15 Number of case labels 511 255 Nesting levels of for statements 16 15 Expressions Number of arguments that can be specified using 63 31 a function definition or a function call operation Total number of operators and operands that can About 500 About 200 be specified in one expression C library functions Number of files that can be opened s
127. sing P Refer to the Install Guide for the installation procedure and reinstall the C compiler TI ili ificati is 256 3320 Command parameter buffer characters overflow S Terminates processing P Ensure that the command line does not exceed 256 characters An error has occurred in either of the 138 Error No Message Explanation 3321 Illegal environment following cases variable 1 The environment variable SHC_LIB is not specified 2 The file name does not satisfy file name specification rules or the path name exceeds 118 characters S Terminates processing P 1 Specify the environment variable SHC_LIB 2 Specify the file name according to file name specification rules 3 Ensure that the path name does not exceed 118 characters Atrinternal error oceurs during compilation 4000 Internal error S Terminates processing to P Report the error occurrence to your local 4999 Hitachi dealer 139 Section 2 Error Messages Output for the C Library Functions Some library functions set error numbers to macro errno defined by the header file lt stddef h gt in the C library function when an error occurs during the library function execution Error messages corresponding to error numbers have already been defined and can be output The following shows an example of a program which causes an error message output Example include lt stdio h gt include lt string h gt include lt st
128. sing the rand and strtok functions library functions other than I O must be initialized include lt stdlib h gt extern void INIT LOWLEVEL void extern void INIT IOLIB void extern void INIT OTHERLIB void void _INITLIB void Deletes an underline from symbol name used in the assembly routine errno 0 Initializes library functions commonly _INIT_LOWLEVEL Calls low level interface initialization routine _INIT_IOLIB Calls standard I O initialization routine INIT OTHERLIB Calls initialization routine other than that for standard I O Library function initialization program example is shown below Example The following shows examples of initialization routine _INIT_IOLIB for standard I O library function and initialization routine _INIT_OTHERLIB for other standard library function Initialization routine INIT LOWLEVED for low level interface routine must be created 76 according to the user low level interface routine s specifications include lt stdio h gt void INIT IOLIB void FILE fp Initializes FILE type data for fp _iob fp lt iob NFILE fp t gt _bufptr NULL Clears buffer pointer gt _bufcnt 0 Clears buffer counter gt _buflen 0 Clears buffer length gt _bufbase NULL Clears base pointer ioflagl 0 Clears I O flag _ioflag2 0 _iofd 0 Opens standard I O file 1
129. sprint _ 40 to 176 isupper _Ja to z A 2 4 math h Table A 20 math h Specifications _C Compiler Specification Table A 19 shows the character set that results in a true return value Characters That Become True Vo Note math h defines macro names EDOM and ERANGE that indicates a standard library error number Item Value returned by a mathematical function if an input parameter is out of the range Is errno set to the value of macro ERANGE if an underflow error occurs in a mathematical function Does a range error occur if the 2nd parameter in the fmod function is 0 154 _C Compiler Specification Returns a nonnumeric value Yes it is set A range error occurs A 2 4 stdio h Table A 21 stdio h Specifications Item _C Compiler Specification Is a return character indicating input text end Not specified Depends on the low level interface required routine specifications Is a blank character immediately before the carriage return read Number of NULL characters added to data written to binary file Initial value of file position specifier in addition mode Is a file data lost following text file output File bufferring specifications Is a file with file length 0 exists File name configuration rule Can the same files be opened simultaneously Output data representation of the p format Hexadecimal representation conversion in the fprintf function Input data representation of the p format Hexa
130. sseesseee 75 4 3 Initializing Sections _ _INITSCT scsscciisscsssscccsasscscstiavseadsaansccedassssanstanssaesaonsdedsoaceguateauvece 76 4 4 Initializing C Library Functions _INITLIB ccc cecsecssessrcssseccssssscesneccesnccssnsesesees 76 4 4 1 Creating Initialization Routine for Standard I O Library Function ME OTB soc ceeseaeiske colicin E aaa uo o ie E enaco o soba Slo eco ose 711 4 4 2 Creating Initialization Routine for Other Library Function INIT OTHERLIB sceatteedteeicy einen sree na eee eee 78 45 Closings Files COREA UL vsivenisaiutsnshcvsenianetcednatecanieseuiadsseutisbensasencetesvarantutesseciavins 79 4 6 Creating Low Level Interface Routines jssicidscsasssenssesesddaveanaancoetnsdadaveedvadsseasdceseesasadeseedsccee 80 Part IV ERROR MESSAGES cccececeeseeeseeseeseseeeeseseeneeeaeeneeeeneeneeteeeeneeees 89 Section 1 Error Messages Output by the C Compiler eee 91 t1 Error Messag PON z omd a asitka esbszssdasssdz oikdddd c ide ddteo eee snu peca es dogs ooo k n 91 1 2 C Compiler Action and Programmer Response for Each Error Level 92 13 Listof Error Messages ziedok s sides odassa taoedzoukodos b lacasb a az EEEE zu iden eee 93 Section 2 Error Messages Output for the C Library Functions 140 APPENDIX innuna ape okn 143 Appendix A Language and Standard Library Function Specifications of the C Compiler aio s ase oo sono na spe kua ooo an koa ak n aan 145
131. ssing P Specify a scalar expression as the first operand Type not compatible in Qen The types of the second and third operands of the conditional operator do not match with each other 115 Error No Message Explanation S Assumes that the result type is int and continues processing P Specify a correct type combination of operands Only one of the following six combinations is allowed for the second and third operands when using the operator 1 Two arithmetic operands 2 Two void operands 3 Two pointers assigned to the same data type 4 A pointer and an integer constant whose value is 0 or another pointer that is assigned to void that was converted from an integer constant whose value is 0 2220 Modifiable lvalue required 5 A pointer and another pointer assigned for operator to void 6 Two struct or union variables with the same data type An expression whose left value cannot be assigned a left value whose type is not array or const is used n operand of an assignment operator lt lt s gt gt amp s of be S Assumes that the left expression whose left value can be assigned is used and continues processing P Specify a left expression whose left value 2221 Illegal type for operator can be assigned The operand of the unary suffix operator or is function type a pointer assigned to void 2222 Type not compatible for or not
132. sters and stack area by the C compiler The user does not need to note how to use this area because registers and stack area used by a function are operated by the C compiler Figure C 1 shows the usage of registers and stack area For return value storage Stack area k Lower address Area used by the function Frame siz Stack frame Return value addres Parameter arec Stack arez i Upper address RO R14 For variable or temporary data storage R4 R7 For parameter storage indicated J Figure C 1 Usage of Registers and Stack Area 168 Appendix D Creating Termination Functions D 1 Creating Library onexit Function This section describes how to create library onexit function that defines termination routines The onexit function defines a function address which is passed as a parameter in the termination routine table If the number of defined functions exceeds the limit value assumed to be 32 in the following example or if the same function is defined twice or more NULL is returned Otherwise value other than NULL is returned In the following example an address in which a function is defined is returned An example of onexit routine is shown below Example include lt stdlib h gt typedef void onexit_t int _onexit_count 0 onexit_t _onexit_buf 32 void extern onexit_t onexit onexit_t void onexit_t onexit f onexit_t f void Int he for i 0 i lt onexit count
133. stfile and show When multiple options are specified separate them with a space A Also when multiple suboptions are specified separate them with a comma C Compiler Options PC shcA debugA listfileA show noobject expansion Atest c Insert a slash before the options debug listfile and show When multiple options are specified separate them with a space A Also when multiple suboptions are specified separate them with a comma and enclose them in parentheses Compiling Multiple Programs Several C source programs can be compiled by a single command on UNIX systems Example 1 Specifying multiple programs shcAtest1 cAtest2 c RET Example 2 Specifying options for all C source programs shcA listfileAtestl cAtest2 c RET The listfile option is valid for both test1 c and test2 c Example 3 Specifying options for particular C source programs shcAtest1 cAtest2 cA listfile RET The listfile option is valid for only test2 c Options specified for particular C source programs have priority over those specified for all C source programs Option List she RET Instead of compiling the C compiler outputs the standard command line format and option list 3 2 Naming Files A standard file extension is automatically added to the name of a file when omitted The standard file extensions used by the C compiler and related software are shown in t
134. t char F _charput MOV L A_PARM R1 MOV R4 RO Specifies data MOV B RO 3 R1 MOV H 21 RO Specifies function code MOV B RO R1 MOV L A FILENO RO Specifies file number MOV B RO RO MOV B RO 2 R1 MOV RI RO Specifies parameter block address TRAPA SIM_IO NOP RTS NOP eharget One character input C program interface char charget void 7 charget MOV L A PARM R1 MOV H 20 RO Specifies function code MOV B RO R1 MOV L A FILENO RO Specifies file number MOV B RO RO MOV B RO 2 R1 MOV R1 RO Specifies parameter block address TRAPA SIM_IO NOP MOV L A_PARM R1 MOV B 3 R1 RO References data RTS NOP ALIGN 4 A_PARM DATA L PARM Parameter block address A FILENO DATA L FILENO File number area address I O buffer definition SECTION B DATA ALIGN 4 PARM RES L I Parameter block area FILENO RES B 1 File number area END 176 Appendix F ASCII Codes M M a m hel 0 E m lt oc uw a jo a E gt m lt oc W o R A uN x Fa A D Z DI fa ja ja a ja SOH je e Je Bo s 3 DC2 177 Index Numbers Decimal and internal representation 163 A abort routine termination routine 171 Aggregate type data 27 Alignment 24 26 Area size calculation 58 63 Heap area 66 Stack area size calculation 58 Stack area 63 Array type 27 ASCII
135. t exceed the limit The number of internal labels to be generated 3016 Internal label overflow by the C compiler exceeds the limit of 16384 for UNIX systems and 2048 for PC systems S Terminates processing P Divide the file so that the number of internal labels does not exceed the limit The number of case labels in one switch Error No Message Explanation 2017 foo meny case Labels statement exceeds the limit of 511 for UNIX systems and 255 for PC systems S Terminates processing P Ensure that the number of case labels does not exceed the limit The number of goto tabets defined imone 3018 Too many gore labels function exceeds the limit of 511 for UNIX systems and 256 for PC systems S Terminates processing P Ensure that the number of goto labels defined in a function does not exceed the limit A source file cannot be opened 3019 Cannot open source file S Terminates processing file name P Specify the correct file name A source or include file cannot be read 3020 Source file input error S Terminates processing file name P Check that the file is not read protected The C compiler cannot allocate sufficient 3021 ponory Overt low memory to compile the program S Terminates processing P Divide the file so that less memory is needed for compilation The nesting level of switch statements exceeds 3022 Switch nest too deep 134 the limit of 16 for UNIX system
136. t is a number from 1 to 254 28 2 From this value 127 is subtracted and the result is used as the actual exponent The range of actual exponents is 126 to 127 The mantissa is a value from 0 to 223 1 For an actual mantissa it is assumed that the highest order bit 223 is 1 and a decimal point follows it Value represented by a normalized number 1 lt sign gt x 2 lt exponent gt 127 x 1 lt mantissa gt x 2725 Example 3130 aa de iji odd ddd LidggoooodododoododoOooD Sign Exponent 100000002 127 1 2 indicates decimal data throughout this manual Mantissa 1 11 1 75 Value 1 75 x 21 3 5 Denormalized Number The sign bit is either 0 positive or 1 negative The exponent is 0 which makes the actual exponent equal to 126 The mantissa is a value from 1 to 223 1 For an actual mantissa it is assumed that a highest order bit 223 is 0 and a decimal point follows it Value represented by a denormalized number 1 lt sign gt x 2 126 x lt mantissa gt x 2723 Example 7130 Ze ce n ojoooodoo 110000000 00000000000000 Sign Exponent 126 Mantissa 0 11 2 0 75 Value 0 75 x 27126 159 Zero The sign bit is either 0 positive or 1 negative 1 e there are two distinct zero values 0 0 and 0 0 The exponent and mantissa are 0 Both 0 0 and 0 0 represent 0 0 See appendix A 3 4 Floating Point Operation Specifications for differences in each operation
137. tatic area 2 1 2 Static Area Size Calculation The static area size is calculated by adding the size of C compiler generated object program and that of library functions used by the C program After object program linkage the static area size can be determined from each section size including library size output on a linkage map listing Before object program linkage the static area size can be approximately determined from the section size information on a compile listing Figure 2 1 shows an example of section size information kkkkkkk SECTION SIZE INFORMATION 993898 PROGRAM SECTION P 0x00004A Byte s CONSTANT SECTION C 0x000018 Byte s DATA SECTION D 0x000004 Byte s BSS SECTION B 0x000004 Byte s TOTAL PROGRAM SIZE 0x00006A Byte s Figure 2 1 Section Size Information 58 If the standard library is not used the static area size can be calculated by adding memory area size used by sections to the size shown in section size information However if the standard library is used the memory area used by the library functions must be added to the the memory area size of each section The standard library includes C library functions based on C language specifications and arithmetic routines required for C program execution Accordingly the standard library must be linked even if library functions are not used in the C source program For details on memory area size used by the standard library functions refer
138. te Total Stack Size main 24 gt f 32 gt g 24 80 bytes Maximum size of stack area main 24 gt g 24 48 bytes As can be seen from table 2 1 the maximum size of stack area required for the longest function calling route should be determined 80 bytes in this example and this size of memory should be allocated in RAM When using standard library functions the stack frame sizes for library functions must also be accounted for Refer to the Standard Library Memory Stack Size Listing included with the C compiler package Note If recursive calls are used in the C source program first determine the stack area required for a recursive call and then multiply with the maximum number of recursive calls 65 Heap Area The total heap area required is equal to the sum of the areas to be allocated by memory management library functions calloc malloc or realloc in the C program An additional 4 bytes must be summed because a 4 byte management area is used every time a memory management library function allocates an area An input output library function uses memory management library functions for internal processing The size of the area allocated in an input output is determined by the following formula 516 bytes x maximum number of simultaneously open files Note Areas released by the free function a memory management library function can be reused However since these areas are often fragmented separated from
139. ted by floating point numbers i Normalized Number The exponent is not 0 or the maximum A normalized number represents a general real number ii Denormalized Number The exponent is 0 and the mantissa is not 0 A denormalized number is a real number whose absolute value is very small iii Zero The exponent and mantissa are both 0 Zero represents the value 0 0 iv Infinity The exponent is the maximum and mantissa is 0 v Not a Number The exponent is the maximum and the mantissa is not 0 This is used to represent an operation result that is undefined such as 0 0 0 0 oo 00 co 00 Table A 25 shows the conditions used to determine values represented by floating point numbers Note A denormalized number represents a floating point number whose absolute value is so small that it cannot be represented as a normalized number Denormalized numbers have less significant digits than normalized numbers The significant digits of a result are not guaranteed if either the operation result or an intermediate result is a denormalized number Table A 25 Types of Values Represented by Floating Point Numbers i Exponent i Mantissa 0 _Other than 0 or Maximum _Maximum 0 0 Normalized number Infinity Other than 0 Denormalized number Not a number 158 A 3 2 float float is internally represented as 1 sign bit 8 exponent bits and 23 mantissa bits Normalized Number The sign bit is either 0 positive or 1 negative The exponen
140. tes processing P Specify and compile the correct source program Prototype mismatch 104 Error No Message Explanation A function type differs from the one specified in the declaration S Ignores the current declaration if the function prototype declaration is being processed Ignores the previous declaration if the declaration of an external function definition is being processed P Correct the declaration so that the function types match 2119 Nota parameter name An identifier not in the function parameter list is declared as a parameter S Ignores the parameter declaration P Check that the function parameter list matches all parameter declarations 2120 Flegal parameter storage A storage class other than register is specified in a function parameter declaration S Ignores the storage class specifier P Delete the storage class specifier 2121 Flegal tag name 2122 Bit field with 0 The combination of a tag name and struct union or enum differs from the declared combination S Assumes struct union or enum depending on the tag name type P Specify the correct combination of a tag name and a struct union or enum 2123 Undefined tag name 105 The width of a bit field which is a member of a struct or union is 0 S Ignores the bit field specification and assumes that the member is not a bit field P Delete the member name or specif
141. tf scanf sprintf sscanf vfprintf vprintf vsprintf fgetc fgets fputc fputs gets puts ungetc fread fwrite fseek ftell rewind perror freopen 1302 Bad file number An output function was issued for fprintf fscanf printf an input file or output function is scanf sprintf sscanf issued for input file vfprintf vprintf vsprintf fgetc fgets fputc fputs gets puts ungetc perror fread fwrite 1304 Error in format An erroneous format was specified fprintf fscanf printf scanf sprintf sscanf vfprintf vprintf vsprintf perror 142 APPENDIX Appendix A Language and Standard Library Function Specifications of the C Compiler This section shows the implementation dependent specifications of the C compiler that are not included in the C language specifications in ANSI standard for the C programming language A 1 Language Specifications of the C Compiler A 1 1 Compilation Specifications Table A 1 Compilation Specifications Item Error information when an error is detected A 1 2 Environmental Specifications Table A 2 Environmental Specifications Item Actual argument for the main function Interactive I O device configuration A 1 3 Identifiers Table A 3 Identifier Specifications Item Number of valid characters of internal identifiers not used for external linkage Number of valid characters of external identifiers used for external linkage Lowercase and uppercase characte
142. the endif directive P Check that the endif directive is used correctly 2014 Missing endif There is no endif directive corresponding to an if ifdef or ifndef directive and the end of file is detected S Assumes that there is an endif directive P Insert an endif directive 2016 Preprocessor constant expression too complex 100 The total number of operators and operands in a constant expression specified by an if or elif directive exceeds the limit of 512 for UNIX systems and 210 for PC systems S Assumes the value of the constant expression to be 0 P Correct the constant expression so that the number of operators and operands is less than or equal to the limit Error No Message Explanation Ww 2017 Missing A closing double quotation mark does not follow a file name in an include directive S Assumes that there is a closing double quotation mark P Insert a closing double quotation mark 2018 Illegal line The line count specified by a line directive exceeds the limit of 32767 for UNIX systems and 16383 for PC systems S Ignores the line directive P Modify the program so that the line count is less than or equal to the limit 2019 File name too long The length of a file name exceeds 128 characters S Uses the first 128 characters as the file name P Change the file name so that the length is less than or equal to 128 characters
143. the file name exists the file contents are discarded and the file size is set to zero O_APPEND bit 5 If this bit is 1 the read write position is set to the end of the file If this bit is 0 the read write position is set to the beginning of the file An error is assumed if the file processing specifications contradict with the actual characteristics of the file The open routine returns a file number positive integer which can be used by the read write Iseek and close routines provided the file opens normally The relationship between file numbers and actual files must be managed by the low level interface routines The open routine returns a value of 1 if the file fails to open properly 83 b close routine Purpose Closes a file Interface int close int fileno Parameters No Name Type Meaning 1 fileno int File number of the file to be closed Return value Type int Normal 0 Abnormal 1 Explanation The file number determined by the open routine is given as the parameter The area of memory allocated by the open routine for file management information is freed so that it can be reused If buffers are used the contents are output to their corresponding files Zero is returned if the file closes normally Otherwise 1 is returned 84 c read routine Purpose Reads data from a file Interface int read int fileno char buf unsigned int count
144. to the attached Standard Library Memory Stack Size Listing The following example shows how to calculate static area size based on the section size information shown in figure 2 1 Calculation Example lt ctype h gt Function Low Memory Size Bytes Stack Size Name Level Routine Library 1 SectionP SectionB SectionC SectionD Bytes isalnum None isalnum 32 0 256 0 16 _ctype isalpha None isalpha 32 0 256 0 16 _ctype Note 1 Library functions required for linkage The library functions include those used by the C program and the library function itself 1 isalnum function of lt ctype h gt is used Add 32 bytes to section P and 256 bytes to section C Size Bytes Section Name C Program Library Total P 74 32 106 B 24 0 24 C 4 256 260 D 4 0 4 2 isalnum and isalpha functions of lt ctype h gt are used When a library function is used by multiple functions memory size required for the library need not to be duplicated The following table shows memory size example when library function _ctype is used by multiple functions lt Library common routine gt Memory Size Bytes Section Name Section P Section B Section C Section D _ctype 0 0 256 0 59 Each section size is calculated by the following formula Note 1 Section size C program Library 1 Library 2 Duplicated library Size Byte Section Name C Program Library 1 Library 2 Duplicated Library Total 1 P 74 32 32 0 138 B 24 0 0 0 2
145. tual files 80 The open routine returns a file number for a given file name The open routine must determine the following so that other functions can access information about a file using the file number File device type console printer disk etc For a special device such as a console or printer file the user chooses a specific file name that can be recognized uniquely by the open routine Information such as the size and address of the buffer used for the file Fora disk file the offset in bytes from the beginning of the file to the next read write position The start position for read write operations is determined by the Iseek routine according to the information determined by the open routine If buffers are used the close routine outputs the contents to their corresponding files This allows the areas of memory allocated by the open routine to be reused 2 ww Low Level Interface Routine Specifications This section explains the specifications for creating low level interface routines gives examples of actual interfaces and explains their operations and notes on implementation The interface for each routine is shown using the format below Create each interface routine by assuming that the prototype declaration is made Example Routine name Purpose Purpose of the routine Interface Shows the interface as a C function declaration Parameters No Name Type Mean
146. unction specifications 153 Invalid operation 162 Compilation specifications 145 Overflow 162 ctype h 154 Result rounding 162 Declarations 151 Special value operations 162 Environmental specifications 145 Underflow 162 Floating point numbers 148 Frame size 15 64 168 Floating point number specifications 157 Function call interface 34 Identifiers 145 Integer 147 G Integer types and their corresponding data Global base register GBR 47 48 range 147 Limits on floating point numbers 148 H math h 154 Heap area 24 66 Modifier 150 help suboption 9 Not supported library 156 How to invoke the C compiler 5 Preprocessor 152 Register 149 I Statement 151 include option 9 stddef h 153 include suboption 9 stdio h 155 Include file 8 string h 156 Reading an include file 152 Structure union enumeration and bit Standard include file 4 field types 150 Infinity 158 length suboption 9 180 Library 4 C library function 4 60 Error messages output for the C library functions 140 Initializing C library functions 76 Low level interface routine 80 Not supported library 156 Run time routine 4 60 Standard library file 4 Limitation 21 22 Linkage with assembly programs 31 Allocating deallocating stack frames 34 External identifier reference 32 Function call interface 34 Registers 35 Setting and referencing parameters and return values 37 Stack pointer 34 listfile option 9 Listing 12 command line specification 17 Object information listing 1
147. urce SH series Relo cross assemble Standard library file User H series library linkage editoi Notes 1 Assembly source programs are output file based on chosen options 2 The standard include file defines C library functions and their macro H series object converte C gt names in order to use C library Load functions module 3 A functional group consisting of C library functions and run time routines S type is used as standard in the C program load module Refer to section 2 1 Note in Part II SH series System Installation simulator debugge __ is related software required during Target system program development Figure 2 1 Relationship between the C Compiler and Other Software Section 3 C Compiler Execution This section explains how to invoke the C compiler specify C compiler options and interpret C compiler listings 3 1 How to Invoke the C Compiler The format for the command line used to invoke the C compiler is as follows UNIX systems shc A lt option gt A lt file name gt A lt option gt PC systems shc A lt option gt JA lt file name gt The general operations of the C compiler are described below Compiling Programs shcAtest c RET The C source program test c is compiled C Compiler Options UNIX shcA debugA listfileA show noobject expansionAtest c RET Insert minus before options debug li
148. used for source program comment Shift state is not supported Eight bits are used for each character ASCIl is used for both Characters and extended representation other than that specified by the language are not _ supported The upper four characters of the character constant is valid and the upper two characters of the wide character is valid If a wide character of more than one character is specified a warning error message is output locale is not supported The same range as the signed char or unsigned char A 1 5 Integer Table A 5 Integer Specifications Item _C Compiler Specification Integer type data representation and value Table A 6 shows data representation and value Effect when an integer is too large to be converted into a signed The lower one or two bytes of integer type value or signed char the integer is used as the conversion result The result of bitwise operations on signed integers signed value Sign of the remainder for integer division Same as the sign of the dividend Effect of a right shift operation on the sign bit of signed integer type data The sign bit is unchanged by the _shift operation Table A 6 Integer Types and Their Corresponding Data Range Type _Range of Values _Data Size char 128 to 127 1 byte signed char 128 to 127 1 byte unsigned char _0 to 255 _1 byte short _ 32768 to 32767 _2 bytes unsigned short _0 to 65535 _2 bytes int
149. y the correct bit field width An undefined tag name is specified in an Error No Message 2124 Illegal enum value Explanation enum declaration S Ignores the declaration P Specify the correct tag name n j pa Ged 2125 Function returning as a value for an enum member S Ignores the value specification P Change the expression to an integer constant expression function 2126 Illegal array size A function with a function return value is specified S Ignores one of the function types P Specify the correct type 2127 Missing array size The value that specifies the number of elements in an array is other than an integer between 1 and 2147483647 S Assumes the number of array elements to be one P Specify a valid number of array elements 2128 Illegal pointer declaration for The number of elements in_an array_is not specified where it is required S Assumes that the number of array element is one P Specify the number of array elements A type specifier other than const or volatile is 106 Error No Message 2129 Illegal initializer type Explanation specified following an asterisk which indicates a pointer declaration S Ignores the type specifier following the asterisk P Specify the correct type specifier following the asterisk 2130 Initializer should be The initial value specified for a variable is not a type that can

Download Pdf Manuals

Related Search

Hitachi Microcomputer Support Software SH Series C Compiler USER S MANUAL

Related Contents

                MITEL DW9268 handbook    

Copyright © All rights reserved.