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

 *  Copyright (C) 2007-2011  Laurent Bessard and Edouard Tisserant

     6

 *

     7

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

     8

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

     9

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

    10

 *  (at your option) any later version.

    11

 *

    12

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

    13

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

    14

 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

    15

 *  GNU General Public License for more details.

    16

 *

    17

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

    18

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

    19

 *

    20

 *

    21

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

    22

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

    23

 */

    24

    25

/*

    26

 * An IEC 61131-3 compiler.

    27

 *

    28

 * Based on the

    29

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

    30

 *

    31

 */

    32

    33

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

    34

 * 

    35

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

    36

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

    37

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

    38

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

    39

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

    40

 *

    41

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

    42

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

    43

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

    44

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

    45

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

    46

 *             will only return a bitsize of 1024!

    47

 *

    48

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

    49

 *       required to store the value.

    50

 *

    51

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

    52

 *      TYPE new_int2_t : INT = 2; END_TYPE;

    53

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

    54

 *

    55

 *    sizeof(SINT) ->  8

    56

 *    sizeof(INT)  -> 16

    57

 *    sizeof(DINT) -> 32

    58

 *    sizeof(LINT) -> 64

    59

 *

    60

 *    sizeof('1')       ->  1

    61

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

    62

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

    63

 *    sizeof('16')      ->  5

    64

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

    65

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

    66

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

    67

 *

    68

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

    69

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

    70

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

    71

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

    72

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

    73

 */

    74

    75

#include "get_sizeof_datatype.hh"

    76

    77

#include <stdlib.h>

    78

#include <string.h>

    79

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

    80

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

    81

#define __STDC_LIMIT_MACROS

    82

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

    83

    84

    85

#define ERROR error_exit(__FILE__,__LINE__)

    86

/* function defined in main.cc */

    87

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

    88

    89

    90

/* This class is a singleton.

    91

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

    92

 */

    93

get_sizeof_datatype_c *get_sizeof_datatype_c::singleton = NULL;

    94

    95

    96

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

    97

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

    98

    99

   100

   101

   102

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

   103

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

   104

static int strdivby2(char **strptr) {

   105

  char *str = *strptr;

   106

  int carry = 0;

   107

   108

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

   109

    /* Assumes ASCII */

   110

    int newcarry;

   111

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

   112

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

   113

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

   114

    carry = newcarry;

   115

    str++;

   116

  }

   117

   118

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

   119

  while (**strptr == '0') 

   120

    (*strptr)++;

   121

   122

  return carry;

   123

}

   124

   125

   126

/* Constructor for the singleton class */

   127

int get_sizeof_datatype_c::getsize(symbol_c *data_type_symbol) {

   128

      if (NULL == singleton) {

   129

        singleton = new get_sizeof_datatype_c;

   130

        if (NULL == singleton)

   131

          ERROR;

   132

      }

   133

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

   134

    }

   135

   136

/* Destructor for the singleton class */

   137

get_sizeof_datatype_c::~get_sizeof_datatype_c(void) {

   138

      if (NULL != singleton) delete singleton;

   139

      singleton = NULL;

   140

    }

   141

   142

   143

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

   144

/* B 1.2 - Constants */

   145

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

   146

   147

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

   148

/* B 1.2.1 - Numeric Literals */

   149

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

   150

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

   151

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

   152

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

   153

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

   154

  * with all inputs of the same data type.

   155

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

   156

  */

   157

   158

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

   159

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

   160

 */

   161

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

   162

  return _encode_int(32);

   163

}

   164

   165

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

   166

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

   167

}

   168

   169

   170

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

   171

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

   172

 */

   173

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

   174

  int bitsize = 0;

   175

   176

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

   177

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

   178

   179

#if 0

   180

  char *endptr;

   181

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

   182

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

   183

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

   184

   *

   185

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

   186

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

   187

   */

   188

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

   189

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

   190

    #define ival_MAX ULLONG_MAX

   191

    unsigned long long int ival = 0;

   192

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

   193

  #else

   194

    /* use long int ... */

   195

    #define ival_MAX ULONG_MAX

   196

    unsigned long int ival = 0;

   197

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

   198

  #endif

   199

   200

  #if (ival_MAX < UINT64_MAX)

   201

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

   202

  #endif

   203

   204

  if (NULL ==  endptr) ERROR;

   205

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

   206

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

   207

   208

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

   209

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

   210

   211

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

   212

  if (0 == bitsize) bitsize = 1;

   213

   214

  return _encode_int(bitsize);

   215

#endif

   216

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

   217

   * determine the the bitsize using integer aritmetic.

   218

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

   219

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

   220

   */

   221

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

   222

  char *sval, *oval;

   223

  const char *pval;

   224

  oval = sval = (char *)malloc(strlen(symbol->value)+1);

   225

  if (NULL ==  sval) ERROR;

   226

   227

  /* copy the original string, but leave out any underscores... */

   228

  for (pval = symbol->value, sval = oval; *pval != '\0'; pval++) {

   229

    if ('_' != *pval) {*sval = *pval; sval++;}

   230

  }  

   231

  *sval = '\0';  

   232

   233

  sval = oval;

   234

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

   235

   236

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

   237

   238

/*

   239

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

   240

  char *oval = sval;

   241

  if (NULL ==  sval) ERROR;

   242

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

   243

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

   244

*/

   245

   246

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

   247

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

   248

   */

   249

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

   250

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

   251

   252

  free(oval);

   253

  return _encode_int(bitsize);

   254

}

   255

   256

   257

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

   258

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

   259

}

   260

   261

   262

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

   263

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

   264

 */

   265

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

   266

  const char *sval = symbol->value;

   267

  int bitsize = 0;

   268

   269

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

   270

  if (NULL ==  sval) ERROR;

   271

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

   272

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

   273

  sval++;

   274

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

   275

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

   276

  sval++;

   277

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

   278

   279

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

   280

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

   281

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

   282

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

   283

      ERROR;

   284

   285

    if ('_' != *sval) bitsize ++; /* 1 bits per binary digit */

   286

  }

   287

   288

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

   289

  if (0 == bitsize) bitsize = 1;

   290

   291

  return _encode_int(bitsize);

   292

}

   293

   294

   295

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

   296

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

   297

 */

   298

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

   299

  const char *sval = symbol->value;

   300

  int bitsize = 0;

   301

   302

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

   303

  if (NULL ==  sval) ERROR;

   304

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

   305

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

   306

  sval++;

   307

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

   308

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

   309

  sval++;

   310

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

   311

   312

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

   313

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

   314

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

   315

    /* Assumes ASCII */

   316

    if ((('0' > *sval) || ('7' < *sval)) && ('_' != *sval))

   317

      ERROR;

   318

   319

    if ('_' != *sval) bitsize += 3; /* 3 bits per octal digit */

   320

  }

   321

   322

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

   323

  if (0 == bitsize) bitsize = 1;

   324

   325

  return _encode_int(bitsize);

   326

}

   327

   328

   329

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

   330

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

   331

 */

   332

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

   333

  const char *sval = symbol->value;

   334

  int bitsize = 0;

   335

   336

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

   337

  if (NULL ==  sval) ERROR;

   338

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

   339

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

   340

  sval++;

   341

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

   342

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

   343

  sval++;

   344

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

   345

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

   346

  sval++;

   347

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

   348

   349

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

   350

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

   351

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

   352

    /* Assumes ASCII */

   353

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

   354

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

   355

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

   356

        ! ('_' == *sval))

   357

      ERROR;

   358

   359

    if ('_' != *sval) bitsize += 4; /* 4 bits per hex digit */

   360

  }

   361

   362

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

   363

  if (0 == bitsize) bitsize = 1;

   364

   365

  return _encode_int(bitsize);

   366

}

   367

   368

   369

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

   370

/* B 1.3.1 - Elementary Data Types */

   371

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

   372

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

   373

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

   374

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

   375

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

   376

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

   377

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

   378

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

   379

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

   380

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

   381

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

   382

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

   383

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

   384

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

   385

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

   386

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

   387

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

   388

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

   389

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

   390

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

   391

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

   392

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

   393

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

   394

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

   395

/* "Safety Software Technical Specification,          */

   396

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

   397

/*  Version 1.0 – Official Release"                   */

   398

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

   399

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

   400

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

   401

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

   402

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

   403

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

   404

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

   405

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

   406

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

   407

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

   408

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

   409

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

   410

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

   411

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

   412

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

   413

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

   414

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

   415

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

   416

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

   417

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

   418

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

   419

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

   420

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

   421

   422

   423

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

   424

/* B 1.3.3 - Derived data types */

   425

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

   426

// Not yet supported...