1

/*

     2

 *  matiec - a compiler for the programming languages defined in IEC 61131-3

     3

 *

     4

 *  Copyright (C) 2009-2011  Mario de Sousa (msousa@fe.up.pt)

     5

 *

     6

 *  This program is free software: you can redistribute it and/or modify

     7

 *  it under the terms of the GNU General Public License as published by

     8

 *  the Free Software Foundation, either version 3 of the License, or

     9

 *  (at your option) any later version.

    10

 *

    11

 *  This program is distributed in the hope that it will be useful,

    12

 *  but WITHOUT ANY WARRANTY; without even the implied warranty of

    13

 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

    14

 *  GNU General Public License for more details.

    15

 *

    16

 *  You should have received a copy of the GNU General Public License

    17

 *  along with this program.  If not, see <http://www.gnu.org/licenses/>.

    18

 *

    19

 *

    20

 * This code is made available on the understanding that it will not be

    21

 * used in safety-critical situations without a full and competent review.

    22

 */

    23

    24

/*

    25

 * An IEC 61131-3 compiler.

    26

 *

    27

 * Based on the

    28

 * FINAL DRAFT - IEC 61131-3, 2nd Ed. (2001-12-10)

    29

 *

    30

 */

    31

    32

/* Determine the size, in bits, of the data type.

    33

 * 

    34

 * NOTE: Currently, only elementary data types with well defined sizes (in the standard) are supported.

    35

 *       - derived data types are not supported, and these will return 0

    36

 *       - TIME, DATE, TIME_OF_DAY, and DATE_AND_TIME are not supported, and will return 0

    37

 *       - STRING and WSTRING are not supported, and the standard merely defines bit per character,

    38

 *              and not the maximum number of characters, so these will return 0

    39

 *

    40

 *       We also support the 'Numeric Literals' Data types.

    41

 *       i.e., numeric literals are considerd basic data types

    42

 *       as their data type is undefined (e.g. the datat type of '30'

    43

 *       could be 'INT' or 'SINT' or 'LINT' or 'USINT' or ...

    44

 *       NOTE: for base 10 numeric literals, any number taking up more than 64 bits

    45

 *             will only return a bitsize of 1024!

    46

 *

    47

 *       For numeric literals, we return the minimum number of bits

    48

 *       required to store the value.

    49

 *

    50

 * E.g. TYPE new_int_t : INT; END_TYPE;

    51

 *      TYPE new_int2_t : INT = 2; END_TYPE;

    52

 *      TYPE new_subr_t : INT (4..5); END_TYPE;

    53

 *

    54

 *    sizeof(SINT) ->  8

    55

 *    sizeof(INT)  -> 16

    56

 *    sizeof(DINT) -> 32

    57

 *    sizeof(LINT) -> 64

    58

 *

    59

 *    sizeof('1')       ->  1

    60

 *    sizeof('015')     ->  4    # Leading zeros are ignored!

    61

 *    sizeof('0')       ->  1    # This is a special case! Even the value 0 needs at least 1 bit to store!

    62

 *    sizeof('16')      ->  5

    63

 *    sizeof('2#00101') ->  3

    64

 *    sizeof('8#334')   ->  9

    65

 *    sizeof('16#2A')   ->  8

    66

 *

    67

 *    sizeof('7.4')     ->  32   # all real literals return 32 bits, the size of a 'REAL'

    68

 *                               # TODO: study IEC 60559 for the range of values that may be

    69

 *                               #       stored in a REAL (basic single width floating point format)

    70

 *                               #       and in a LREAL (basic double width floating point format)

    71

 *                               #       and see if some real literals need to return 64 instead!

    72

 */

    73

    74

#include "get_sizeof_datatype.hh"

    75

    76

#include <stdlib.h>

    77

#include <string.h>

    78

#include <limits.h>  // get definition of ULLONG_MAX

    79

/* tell stdint.h we want the definition of UINT64_MAX */

    80

#define __STDC_LIMIT_MACROS

    81

#include <stdint.h>  // get definition of uint64_t and UINT64_MAX

    82

    83

    84

#define ERROR error_exit(__FILE__,__LINE__)

    85

/* function defined in main.cc */

    86

extern void error_exit(const char *file_name, int line_no);

    87

    88

    89

/* This class is a singleton.

    90

 * So we need a pointer to the singe instance...

    91

 */

    92

get_sizeof_datatype_c *get_sizeof_datatype_c::singleton = NULL;

    93

    94

    95

#define _encode_int(value)   ((void *)(((char *)NULL) + value))

    96

#define _decode_int(ptr)     (((char *)ptr) - ((char *)NULL))

    97

    98

    99

   100

   101

/* divide a base 10 literal in a string by 2 */

   102

/* returns remainder of division (0 or 1)    */

   103

static int strdivby2(char **strptr) {

   104

  char *str = *strptr;

   105

  int carry = 0;

   106

   107

  while (*str != '\0') {

   108

    /* Assumes ASCII */

   109

    int newcarry;

   110

//     newcarry = ((*str-'0') mod 2);

   111

    newcarry = ((*str-'0') - ((*str-'0')/2)*2);

   112

    *str = (((*str-'0') + 10*carry)/2) + '0';

   113

    carry = newcarry;

   114

    str++;

   115

  }

   116

   117

  /* ignore leading zeros in result... */

   118

  while (**strptr == '0') 

   119

    (*strptr)++;

   120

   121

  return carry;

   122

}

   123

   124

   125

/* Constructor for the singleton class */

   126

int get_sizeof_datatype_c::getsize(symbol_c *data_type_symbol) {

   127

      if (NULL == singleton) {

   128

        singleton = new get_sizeof_datatype_c;

   129

        if (NULL == singleton)

   130

          ERROR;

   131

      }

   132

      return _decode_int(data_type_symbol->accept(*singleton));

   133

    }

   134

   135

/* Destructor for the singleton class */

   136

get_sizeof_datatype_c::~get_sizeof_datatype_c(void) {

   137

      if (NULL != singleton) delete singleton;

   138

      singleton = NULL;

   139

    }

   140

   141

   142

/*********************/

   143

/* B 1.2 - Constants */

   144

/*********************/

   145

   146

/******************************/

   147

/* B 1.2.1 - Numeric Literals */

   148

/******************************/

   149

 /* Numeric literals without any explicit type cast have unknown data type, 

   150

  * so we continue considering them as their own basic data types until

   151

  * they can be resolved (for example, when using '30+x' where 'x' is a LINT variable, the

   152

  * numeric literal '30' must then be considered a LINT so the ADD function may be called

   153

  * with all inputs of the same data type.

   154

  * If 'x' were a SINT, then the '30' would have to be a SINT too!

   155

  */

   156

   157

/* NOTE: all integer_c and real_c tokens will always be positive (i.e. no leading '-')

   158

 * due to the way the source code is parsed by iec.flex.

   159

 */

   160

void *get_sizeof_datatype_c::visit(real_c *symbol) {

   161

  return _encode_int(32);

   162

}

   163

   164

void *get_sizeof_datatype_c::visit(neg_real_c *symbol) {

   165

  return symbol->exp->accept(*this);

   166

}

   167

   168

   169

/* NOTE: all integer_c and real_c literal tokens will always be positive (i.e. no leading '-')

   170

 * due to the way the source code is parsed by iec.flex.

   171

 */

   172

void *get_sizeof_datatype_c::visit(integer_c *symbol) {

   173

  int bitsize = 0;

   174

   175

  if (NULL ==   symbol->value ) ERROR;

   176

  if ('\0' == *(symbol->value)) ERROR;

   177

   178

#if 0

   179

  char *endptr;

   180

  /* Convert the string to an unsigned 64 bit integer */

   181

  /* We can use strtoull(), but we are not guaranteed that an unsigned long long int

   182

   * is 64 bits wide. First make sure that it is...

   183

   *

   184

   * We could also use the strtouq() instead, which converts

   185

   * to a quad word (64 bits). However, this requires either GCC or BSD extensions.

   186

   */

   187

  #ifdef strtoull // this ifdef does not work!!

   188

    /* we have long long int, use it... */

   189

    #define ival_MAX ULLONG_MAX

   190

    unsigned long long int ival = 0;

   191

    ival = strtoull (symbol->value, &endptr, 10 /* base */);

   192

  #else

   193

    /* use long int ... */

   194

    #define ival_MAX ULONG_MAX

   195

    unsigned long int ival = 0;

   196

    ival = strtoul (symbol->value, &endptr, 10 /* base */);

   197

  #endif

   198

   199

  #if (ival_MAX < UINT64_MAX)

   200

  #error Largest strtoint() conversion function converts to an int less than 64 bits wide!

   201

  #endif

   202

   203

  if (NULL ==  endptr) ERROR;

   204

  if ('\0' != *endptr) ERROR;

   205

  // TODO: return _encode_int(1024) if value takes up more than 64 bits! 

   206

   207

  /* determine the number of bits used... */

   208

  for (bitsize = 0; ival > 0; ival /= 2, bitsize++);

   209

   210

  /* special case... if (value == 0) <=> (bitsize == 0), return bit size of 1 ! */

   211

  if (0 == bitsize) bitsize = 1;

   212

   213

  return _encode_int(bitsize);

   214

#endif

   215

  /* We could first convert from string to some kind of integer, and then

   216

   * determine the the bitsize using integer aritmetic.

   217

   * However, we are then limited to the bit size of the widest available integer

   218

   * (usually 64 bits), which is not good at all!

   219

   */

   220

  /* Let's try to determine bitsize by converting directly from the string!! */

   221

  char *sval = strdup(symbol->value);

   222

  char *oval = sval;

   223

  if (NULL ==  sval) ERROR;

   224

  if ('\0' == *sval) ERROR;

   225

   226

  for (bitsize = 0; *sval != '\0'; strdivby2(&sval), bitsize ++);

   227

   228

  /* Even for (value == 0), the above loop will return bitsize == 1!, 

   229

   * so we don't need to handle the special case...

   230

   */

   231

  /* special case... if (value == 0) <=> (bitsize == 0), return bit size of 1 ! */

   232

  // if (0 == bitsize) bitsize = 1;

   233

   234

  free(oval);

   235

  return _encode_int(bitsize);

   236

}

   237

   238

   239

void *get_sizeof_datatype_c::visit(neg_integer_c *symbol) {

   240

  return symbol->exp->accept(*this);

   241

}

   242

   243

   244

/* NOTE: all binary_integer_c tokens will always be positive (i.e. no leading '-')

   245

 * due to the syntax definition of IEC 61131-3.

   246

 */

   247

void *get_sizeof_datatype_c::visit(binary_integer_c *symbol) {

   248

  const char *sval = symbol->value;

   249

  int bitsize = 0;

   250

   251

  /* first 2 characters had better be "2#" ! */

   252

  if (NULL ==  sval) ERROR;

   253

  if ('\0' == *sval) ERROR;

   254

  if ( '2' != *sval) ERROR;

   255

  sval++;

   256

  if ('\0' == *sval) ERROR;

   257

  if ( '#' != *sval) ERROR;

   258

  sval++;

   259

  if ('\0' == *sval) ERROR;

   260

   261

  /* determine the number of bits used... */

   262

  for (bitsize = 0; '\0' != *sval; sval++, bitsize++) {

   263

    /* consistency check: make sure we only have binary digits! */

   264

    if (('0' != *sval) && ('1' != *sval))

   265

      ERROR;

   266

  }

   267

   268

  /* special case... if (value == 0) <=> (bitsize == 0), return bit size of 1 ! */

   269

  if (0 == bitsize) bitsize = 1;

   270

   271

  return _encode_int(bitsize);

   272

}

   273

   274

   275

/* NOTE: all octal_integer_c tokens will always be positive (i.e. no leading '-')

   276

 * due to the syntax definition of IEC 61131-3.

   277

 */

   278

void *get_sizeof_datatype_c::visit(octal_integer_c *symbol) {

   279

  const char *sval = symbol->value;

   280

  int bitsize = 0;

   281

   282

  /* first 2 characters had better be "8#" ! */

   283

  if (NULL ==  sval) ERROR;

   284

  if ('\0' == *sval) ERROR;

   285

  if ( '8' != *sval) ERROR;

   286

  sval++;

   287

  if ('\0' == *sval) ERROR;

   288

  if ( '#' != *sval) ERROR;

   289

  sval++;

   290

  if ('\0' == *sval) ERROR;

   291

   292

  /* determine the number of bits used... */

   293

  for (bitsize = 0; '\0' != *sval; sval++, bitsize += 3 /* 3 bits per octal digit */) {

   294

    /* consistency check: make sure we only have octal digits! */

   295

    /* Assumes ASCII */

   296

    if (('0' > *sval) || ('7' < *sval))

   297

      ERROR;

   298

  }

   299

   300

  /* special case... if (value == 0) <=> (bitsize == 0), return bit size of 1 ! */

   301

  if (0 == bitsize) bitsize = 1;

   302

   303

  return _encode_int(bitsize);

   304

}

   305

   306

   307

/* NOTE: all hex_integer_c tokens will always be positive (i.e. no leading '-')

   308

 * due to the syntax definition of IEC 61131-3.

   309

 */

   310

void *get_sizeof_datatype_c::visit(hex_integer_c *symbol) {

   311

  const char *sval = symbol->value;

   312

  int bitsize = 0;

   313

   314

  /* first 3 characters had better be "16#" ! */

   315

  if (NULL ==  sval) ERROR;

   316

  if ('\0' == *sval) ERROR;

   317

  if ( '1' != *sval) ERROR;

   318

  sval++;

   319

  if ('\0' == *sval) ERROR;

   320

  if ( '6' != *sval) ERROR;

   321

  sval++;

   322

  if ('\0' == *sval) ERROR;

   323

  if ( '#' != *sval) ERROR;

   324

  sval++;

   325

  if ('\0' == *sval) ERROR;

   326

   327

  /* determine the number of bits used... */

   328

  for (bitsize = 0; '\0' != *sval; sval++, bitsize += 4 /* 4 bits per hex digit */) {

   329

    /* consistency check: make sure we only have hex digits! */

   330

    /* Assumes ASCII */

   331

    if (!(('0' <= *sval) && ('9' >= *sval)) && 

   332

        !(('A' <= *sval) && ('F' >= *sval)) &&

   333

        !(('a' <= *sval) && ('b' >= *sval)))

   334

      ERROR;

   335

  }

   336

   337

  /* special case... if (value == 0) <=> (bitsize == 0), return bit size of 1 ! */

   338

  if (0 == bitsize) bitsize = 1;

   339

   340

  return _encode_int(bitsize);

   341

}

   342

   343

   344

/***********************************/

   345

/* B 1.3.1 - Elementary Data Types */

   346

/***********************************/

   347

// void *get_sizeof_datatype_c::visit(time_type_name_c *symbol) {return _encode_int(0); }

   348

void *get_sizeof_datatype_c::visit(bool_type_name_c *symbol)    {return _encode_int(1); }

   349

void *get_sizeof_datatype_c::visit(sint_type_name_c *symbol)    {return _encode_int(8); }

   350

void *get_sizeof_datatype_c::visit(int_type_name_c *symbol)     {return _encode_int(16);}

   351

void *get_sizeof_datatype_c::visit(dint_type_name_c *symbol)    {return _encode_int(32);}

   352

void *get_sizeof_datatype_c::visit(lint_type_name_c *symbol)    {return _encode_int(64);}

   353

void *get_sizeof_datatype_c::visit(usint_type_name_c *symbol)   {return _encode_int(8); }

   354

void *get_sizeof_datatype_c::visit(uint_type_name_c *symbol)    {return _encode_int(16);}

   355

void *get_sizeof_datatype_c::visit(udint_type_name_c *symbol)   {return _encode_int(32);}

   356

void *get_sizeof_datatype_c::visit(ulint_type_name_c *symbol)   {return _encode_int(64);}

   357

void *get_sizeof_datatype_c::visit(real_type_name_c *symbol)    {return _encode_int(32);}

   358

void *get_sizeof_datatype_c::visit(lreal_type_name_c *symbol)   {return _encode_int(64);}

   359

// void *get_sizeof_datatype_c::visit(date_type_name_c *symbol) {return _encode_int(0); }

   360

// void *get_sizeof_datatype_c::visit(tod_type_name_c *symbol)  {return _encode_int(0); }

   361

// void *get_sizeof_datatype_c::visit(dt_type_name_c *symbol)   {return _encode_int(0); }

   362

void *get_sizeof_datatype_c::visit(byte_type_name_c *symbol)    {return _encode_int(8); }

   363

void *get_sizeof_datatype_c::visit(word_type_name_c *symbol)    {return _encode_int(16);}

   364

void *get_sizeof_datatype_c::visit(dword_type_name_c *symbol)   {return _encode_int(32);}

   365

void *get_sizeof_datatype_c::visit(lword_type_name_c *symbol)   {return _encode_int(64);}

   366

// void *get_sizeof_datatype_c::visit(string_type_name_c *symbol)    {return _encode_int(0); }

   367

// void *get_sizeof_datatype_c::visit(wstring_type_name_c *symbol)   {return _encode_int(0); }

   368

/******************************************************/

   369

/* Extensions to the base standard as defined in      */

   370

/* "Safety Software Technical Specification,          */

   371

/*  Part 1: Concepts and Function Blocks,             */

   372

/*  Version 1.0 – Official Release"                   */

   373

/* by PLCopen - Technical Committee 5 - 2006-01-31    */

   374

/******************************************************/

   375

//     void *get_sizeof_datatype_c::visit(safetime_type_name_c *symbol); {return _encode_int(0); }

   376

void *get_sizeof_datatype_c::visit(safebool_type_name_c *symbol)         {return _encode_int(1); }

   377

void *get_sizeof_datatype_c::visit(safesint_type_name_c *symbol)         {return _encode_int(8); }

   378

void *get_sizeof_datatype_c::visit(safeint_type_name_c *symbol)          {return _encode_int(16);}

   379

void *get_sizeof_datatype_c::visit(safedint_type_name_c *symbol)         {return _encode_int(32);}

   380

void *get_sizeof_datatype_c::visit(safelint_type_name_c *symbol)         {return _encode_int(64);}

   381

void *get_sizeof_datatype_c::visit(safeusint_type_name_c *symbol)        {return _encode_int(8); }

   382

void *get_sizeof_datatype_c::visit(safeuint_type_name_c *symbol)         {return _encode_int(16);}

   383

void *get_sizeof_datatype_c::visit(safeudint_type_name_c *symbol)        {return _encode_int(32);}

   384

void *get_sizeof_datatype_c::visit(safeulint_type_name_c *symbol)        {return _encode_int(64);}

   385

void *get_sizeof_datatype_c::visit(safereal_type_name_c *symbol)         {return _encode_int(32);}

   386

void *get_sizeof_datatype_c::visit(safelreal_type_name_c *symbol)        {return _encode_int(64);}

   387

//     void *get_sizeof_datatype_c::visit(safedate_type_name_c *symbol); {return _encode_int(0); }

   388

//     void *get_sizeof_datatype_c::visit(safetod_type_name_c *symbol);  {return _encode_int(0); }

   389

//     void *get_sizeof_datatype_c::visit(safedt_type_name_c *symbol);   {return _encode_int(0); }

   390

void *get_sizeof_datatype_c::visit(safebyte_type_name_c *symbol)         {return _encode_int(8); }

   391

void *get_sizeof_datatype_c::visit(safeword_type_name_c *symbol)         {return _encode_int(16);}

   392

void *get_sizeof_datatype_c::visit(safedword_type_name_c *symbol)        {return _encode_int(32);}

   393

void *get_sizeof_datatype_c::visit(safelword_type_name_c *symbol)        {return _encode_int(64);}

   394

//     void *get_sizeof_datatype_c::visit(safestring_type_name_c *symbol);  {return _encode_int(0); }

   395

//     void *get_sizeof_datatype_c::visit(safewstring_type_name_c *symbol); {return _encode_int(0); }

   396

   397

   398

/********************************/

   399

/* B 1.3.3 - Derived data types */

   400

/********************************/

   401

// Not yet supported...