nasm.h

   1 /* nasm.h   main header file for the Netwide Assembler: inter-module interface
   2  *
   3  * The Netwide Assembler is copyright (C) 1996 Simon Tatham and
   4  * Julian Hall. All rights reserved. The software is
   5  * redistributable under the licence given in the file "Licence"
   6  * distributed in the NASM archive.
   7  *
   8  * initial version: 27/iii/95 by Simon Tatham
   9  */
  10
  11 #ifndef NASM_H
  12 #define NASM_H
  13
  14 #define NASM_MAJOR_VER 0
  15 #define NASM_MINOR_VER 93
  16 #define NASM_VER "0.93"
  17
  18 #ifndef NULL
  19 #define NULL 0
  20 #endif
  21
  22 #ifndef FALSE
  23 #define FALSE 0                        /* comes in handy */
  24 #endif
  25 #ifndef TRUE
  26 #define TRUE 1
  27 #endif
  28
  29 #define NO_SEG -1L                     /* null segment value */
  30 #define SEG_ABS 0x40000000L            /* mask for far-absolute segments */
  31
  32 #ifndef FILENAME_MAX
  33 #define FILENAME_MAX 256
  34 #endif
  35
  36 /*
  37  * We must declare the existence of this structure type up here,
  38  * since we have to reference it before we define it...
  39  */
  40 struct ofmt;
  41
  42 /*
  43  * -------------------------
  44  * Error reporting functions
  45  * -------------------------
  46  */
  47
  48 /*
  49  * An error reporting function should look like this.
  50  */
  51 typedef void (*efunc) (int severity, char *fmt, ...);
  52
  53 /*
  54  * These are the error severity codes which get passed as the first
  55  * argument to an efunc.
  56  */
  57
  58 #define ERR_WARNING 0                  /* warn only: no further action */
  59 #define ERR_NONFATAL 1                 /* terminate assembly after phase */
  60 #define ERR_FATAL 2                    /* instantly fatal: exit with error */
  61 #define ERR_PANIC 3                    /* internal error: panic instantly
  62                                         * and dump core for reference */
  63 #define ERR_MASK 0x0F                  /* mask off the above codes */
  64 #define ERR_NOFILE 0x10                /* don't give source file name/line */
  65 #define ERR_USAGE 0x20                 /* print a usage message */
  66
  67 /*
  68  * -----------------------
  69  * Other function typedefs
  70  * -----------------------
  71  */
  72
  73 /*
  74  * A label-lookup function should look like this.
  75  */
  76 typedef int (*lfunc) (char *label, long *segment, long *offset);
  77
  78 /*
  79  * And a label-definition function like this.
  80  */
  81 typedef void (*ldfunc) (char *label, long segment, long offset,
  82                         struct ofmt *ofmt, efunc error);
  83
  84 /*
  85  * -----------------------------------------------------------
  86  * Format of the `insn' structure returned from `parser.c' and
  87  * passed into `assemble.c'
  88  * -----------------------------------------------------------
  89  */
  90
  91 /*
  92  * Here we define the operand types. These are implemented as bit
  93  * masks, since some are subsets of others; e.g. AX in a MOV
  94  * instruction is a special operand type, whereas AX in other
  95  * contexts is just another 16-bit register. (Also, consider CL in
  96  * shift instructions, DX in OUT, etc.)
  97  */
  98
  99 /* size, and other attributes, of the operand */
 100 #define BITS8     0x00000001L
 101 #define BITS16    0x00000002L
 102 #define BITS32    0x00000004L
 103 #define BITS64    0x00000008L          /* FPU only */
 104 #define BITS80    0x00000010L          /* FPU only */
 105 #define FAR       0x00000020L          /* grotty: this means 16:16 or */
 106                                        /* 16:32, like in CALL/JMP */
 107 #define NEAR      0x00000040L
 108 #define SHORT     0x00000080L          /* and this means what it says :) */
 109
 110 #define SIZE_MASK 0x000000FFL          /* all the size attributes */
 111 #define NON_SIZE  (~SIZE_MASK)
 112
 113 #define TO        0x00000100L          /* reverse effect in FADD, FSUB &c */
 114 #define COLON     0x00000200L          /* operand is followed by a colon */
 115
 116 /* type of operand: memory reference, register, etc. */
 117 #define MEMORY    0x00204000L
 118 #define REGISTER  0x00001000L          /* register number in 'basereg' */
 119 #define IMMEDIATE 0x00002000L
 120
 121 #define REGMEM    0x00200000L          /* for r/m, ie EA, operands */
 122 #define REGNORM   0x00201000L          /* 'normal' reg, qualifies as EA */
 123 #define REG8      0x00201001L
 124 #define REG16     0x00201002L
 125 #define REG32     0x00201004L
 126 #define FPUREG    0x01000000L          /* floating point stack registers */
 127 #define FPU0      0x01000800L          /* FPU stack register zero */
 128 #define MMXREG    0x00001008L          /* MMX registers */
 129
 130 /* special register operands: these may be treated differently */
 131 #define REG_SMASK 0x00070000L          /* a mask for the following */
 132 #define REG_ACCUM 0x00211000L          /* accumulator: AL, AX or EAX */
 133 #define REG_AL    0x00211001L          /* REG_ACCUM | BITSxx */
 134 #define REG_AX    0x00211002L          /* ditto */
 135 #define REG_EAX   0x00211004L          /* and again */
 136 #define REG_COUNT 0x00221000L          /* counter: CL, CX or ECX */
 137 #define REG_CL    0x00221001L          /* REG_COUNT | BITSxx */
 138 #define REG_CX    0x00221002L          /* ditto */
 139 #define REG_ECX   0x00221004L          /* another one */
 140 #define REG_DX    0x00241002L
 141 #define REG_SREG  0x00081002L          /* any segment register */
 142 #define REG_CS    0x01081002L          /* CS */
 143 #define REG_DESS  0x02081002L          /* DS, ES, SS (non-CS 86 registers) */
 144 #define REG_FSGS  0x04081002L          /* FS, GS (386 extended registers) */
 145 #define REG_CDT   0x00101004L          /* CRn, DRn and TRn */
 146 #define REG_CREG  0x08101004L          /* CRn */
 147 #define REG_CR4   0x08101404L          /* CR4 (Pentium only) */
 148 #define REG_DREG  0x10101004L          /* DRn */
 149 #define REG_TREG  0x20101004L          /* TRn */
 150
 151 /* special type of EA */
 152 #define MEM_OFFS  0x00604000L          /* simple [address] offset */
 153
 154 /* special type of immediate operand */
 155 #define UNITY     0x00802000L          /* for shift/rotate instructions */
 156
 157 /*
 158  * Next, the codes returned from the parser, for registers and
 159  * instructions.
 160  */
 161
 162 enum {                                 /* register names */
 163     R_AH = 1, R_AL, R_AX, R_BH, R_BL, R_BP, R_BX, R_CH, R_CL, R_CR0,
 164     R_CR2, R_CR3, R_CR4, R_CS, R_CX, R_DH, R_DI, R_DL, R_DR0, R_DR1,
 165     R_DR2, R_DR3, R_DR6, R_DR7, R_DS, R_DX, R_EAX, R_EBP, R_EBX,
 166     R_ECX, R_EDI, R_EDX, R_ES, R_ESI, R_ESP, R_FS, R_GS, R_MM0,
 167     R_MM1, R_MM2, R_MM3, R_MM4, R_MM5, R_MM6, R_MM7, R_SI, R_SP,
 168     R_SS, R_ST0, R_ST1, R_ST2, R_ST3, R_ST4, R_ST5, R_ST6, R_ST7,
 169     R_TR3, R_TR4, R_TR5, R_TR6, R_TR7, REG_ENUM_LIMIT
 170 };
 171
 172 enum {                                 /* instruction names */
 173     I_AAA, I_AAD, I_AAM, I_AAS, I_ADC, I_ADD, I_AND, I_ARPL,
 174     I_BOUND, I_BSF, I_BSR, I_BSWAP, I_BT, I_BTC, I_BTR, I_BTS,
 175     I_CALL, I_CBW, I_CDQ, I_CLC, I_CLD, I_CLI, I_CLTS, I_CMC, I_CMP,
 176     I_CMPSB, I_CMPSD, I_CMPSW, I_CMPXCHG, I_CMPXCHG8B, I_CPUID,
 177     I_CWD, I_CWDE, I_DAA, I_DAS, I_DB, I_DD, I_DEC, I_DIV, I_DQ,
 178     I_DT, I_DW, I_EMMS, I_ENTER, I_EQU, I_F2XM1, I_FABS, I_FADD,
 179     I_FADDP, I_FBLD, I_FBSTP, I_FCHS, I_FCLEX, I_FCMOVB, I_FCMOVBE,
 180     I_FCMOVE, I_FCMOVNB, I_FCMOVNBE, I_FCMOVNE, I_FCMOVNU, I_FCMOVU,
 181     I_FCOM, I_FCOMI, I_FCOMIP, I_FCOMP, I_FCOMPP, I_FCOS, I_FDECSTP,
 182     I_FDISI, I_FDIV, I_FDIVP, I_FDIVR, I_FDIVRP, I_FENI, I_FFREE,
 183     I_FIADD, I_FICOM, I_FICOMP, I_FIDIV, I_FIDIVR, I_FILD, I_FIMUL,
 184     I_FINCSTP, I_FINIT, I_FIST, I_FISTP, I_FISUB, I_FISUBR, I_FLD,
 185     I_FLD1, I_FLDCW, I_FLDENV, I_FLDL2E, I_FLDL2T, I_FLDLG2,
 186     I_FLDLN2, I_FLDPI, I_FLDZ, I_FMUL, I_FMULP, I_FNOP, I_FPATAN,
 187     I_FPREM, I_FPREM1, I_FPTAN, I_FRNDINT, I_FRSTOR, I_FSAVE,
 188     I_FSCALE, I_FSETPM, I_FSIN, I_FSINCOS, I_FSQRT, I_FST, I_FSTCW,
 189     I_FSTENV, I_FSTP, I_FSTSW, I_FSUB, I_FSUBP, I_FSUBR, I_FSUBRP,
 190     I_FTST, I_FUCOM, I_FUCOMI, I_FUCOMIP, I_FUCOMP, I_FUCOMPP,
 191     I_FXAM, I_FXCH, I_FXTRACT, I_FYL2X, I_FYL2XP1, I_HLT, I_ICEBP,
 192     I_IDIV, I_IMUL, I_IN, I_INC, I_INSB, I_INSD, I_INSW, I_INT,
 193     I_INT1, I_INT01, I_INT3, I_INTO, I_INVD, I_INVLPG, I_IRET,
 194     I_IRETD, I_IRETW, I_JCXZ, I_JECXZ, I_JMP, I_LAHF, I_LAR, I_LDS,
 195     I_LEA, I_LEAVE, I_LES, I_LFS, I_LGDT, I_LGS, I_LIDT, I_LLDT,
 196     I_LMSW, I_LOADALL, I_LODSB, I_LODSD, I_LODSW, I_LOOP, I_LOOPE,
 197     I_LOOPNE, I_LOOPNZ, I_LOOPZ, I_LSL, I_LSS, I_LTR, I_MOV, I_MOVD,
 198     I_MOVQ, I_MOVSB, I_MOVSD, I_MOVSW, I_MOVSX, I_MOVZX, I_MUL,
 199     I_NEG, I_NOP, I_NOT, I_OR, I_OUT, I_OUTSB, I_OUTSD, I_OUTSW,
 200     I_PACKSSDW, I_PACKSSWB, I_PACKUSWB, I_PADDB, I_PADDD, I_PADDSB,
 201     I_PADDSW, I_PADDUSB, I_PADDUSW, I_PADDW, I_PAND, I_PANDN,
 202     I_PCMPEQB, I_PCMPEQD, I_PCMPEQW, I_PCMPGTB, I_PCMPGTD,
 203     I_PCMPGTW, I_PMADDWD, I_PMULHW, I_PMULLW, I_POP, I_POPA,
 204     I_POPAD, I_POPAW, I_POPF, I_POPFD, I_POPFW, I_POR, I_PSLLD,
 205     I_PSLLQ, I_PSLLW, I_PSRAD, I_PSRAW, I_PSRLD, I_PSRLQ, I_PSRLW,
 206     I_PSUBB, I_PSUBD, I_PSUBSB, I_PSUBSW, I_PSUBUSB, I_PSUBUSW,
 207     I_PSUBW, I_PUNPCKHBW, I_PUNPCKHDQ, I_PUNPCKHWD, I_PUNPCKLBW,
 208     I_PUNPCKLDQ, I_PUNPCKLWD, I_PUSH, I_PUSHA, I_PUSHAD, I_PUSHAW,
 209     I_PUSHF, I_PUSHFD, I_PUSHFW, I_PXOR, I_RCL, I_RCR, I_RDMSR,
 210     I_RDPMC, I_RDTSC, I_RESB, I_RESD, I_RESQ, I_REST, I_RESW, I_RET,
 211     I_RETF, I_RETN, I_ROL, I_ROR, I_RSM, I_SAHF, I_SAL, I_SALC,
 212     I_SAR, I_SBB, I_SCASB, I_SCASD, I_SCASW, I_SGDT, I_SHL, I_SHLD,
 213     I_SHR, I_SHRD, I_SIDT, I_SLDT, I_SMSW, I_STC, I_STD, I_STI,
 214     I_STOSB, I_STOSD, I_STOSW, I_STR, I_SUB, I_TEST, I_UMOV, I_VERR,
 215     I_VERW, I_WAIT, I_WBINVD, I_WRMSR, I_XADD, I_XCHG, I_XLATB,
 216     I_XOR, I_CMOVcc, I_Jcc, I_SETcc
 217 };
 218
 219 enum {                                 /* condition code names */
 220     C_A, C_AE, C_B, C_BE, C_C, C_E, C_G, C_GE, C_L, C_LE, C_NA, C_NAE,
 221     C_NB, C_NBE, C_NC, C_NE, C_NG, C_NGE, C_NL, C_NLE, C_NO, C_NP,
 222     C_NS, C_NZ, C_O, C_P, C_PE, C_PO, C_S, C_Z
 223 };
 224
 225 /*
 226  * Note that because segment registers may be used as instruction
 227  * prefixes, we must ensure the enumerations for prefixes and
 228  * register names do not overlap.
 229  */
 230 enum {                                 /* instruction prefixes */
 231     PREFIX_ENUM_START = REG_ENUM_LIMIT,
 232     P_A16 = PREFIX_ENUM_START, P_A32, P_LOCK, P_O16, P_O32, P_REP, P_REPE,
 233     P_REPNE, P_REPNZ, P_REPZ, P_TIMES
 234 };
 235
 236 enum {                                 /* extended operand types */
 237     EOT_NOTHING, EOT_DB_STRING, EOT_DB_NUMBER
 238 };
 239
 240 typedef struct {                       /* operand to an instruction */
 241     long type;                         /* type of operand */
 242     int addr_size;                     /* 0 means default; 16; 32 */
 243     int basereg, indexreg, scale;      /* registers and scale involved */
 244     long segment;                      /* immediate segment, if needed */
 245     long offset;                       /* any immediate number */
 246     long wrt;                          /* segment base it's relative to */
 247 } operand;
 248
 249 typedef struct extop {                 /* extended operand */
 250     struct extop *next;                /* linked list */
 251     long type;                         /* defined above */
 252     char *stringval;                   /* if it's a string, then here it is */
 253     int stringlen;                     /* ... and here's how long it is */
 254     long segment;                      /* if it's a number/address, then... */
 255     long offset;                       /* ... it's given here ... */
 256     long wrt;                          /* ... and here */
 257 } extop;
 258
 259 #define MAXPREFIX 4
 260
 261 typedef struct {                       /* an instruction itself */
 262     char *label;                       /* the label defined, or NULL */
 263     int prefixes[MAXPREFIX];           /* instruction prefixes, if any */
 264     int nprefix;                       /* number of entries in above */
 265     int opcode;                        /* the opcode - not just the string */
 266     int condition;                     /* the condition code, if Jcc/SETcc */
 267     int operands;                      /* how many operands? 0-3 */
 268     operand oprs[3];                   /* the operands, defined as above */
 269     extop *eops;                       /* extended operands */
 270     int times;                         /* repeat count (TIMES prefix) */
 271     int forw_ref;                      /* is there a forward reference? */
 272 } insn;
 273
 274 /*
 275  * ------------------------------------------------------------
 276  * The data structure defining an output format driver, and the
 277  * interfaces to the functions therein.
 278  * ------------------------------------------------------------
 279  */
 280
 281 struct ofmt {
 282     /*
 283      * This is a short (one-liner) description of the type of
 284      * output generated by the driver.
 285      */
 286     char *fullname;
 287
 288     /*
 289      * This is a single keyword used to select the driver.
 290      */
 291     char *shortname;
 292
 293     /*
 294      * This procedure is called at the start of an output session.
 295      * It tells the output format what file it will be writing to,
 296      * what routine to report errors through, and how to interface
 297      * to the label manager if necessary. It also gives it a chance
 298      * to do other initialisation.
 299      */
 300     void (*init) (FILE *fp, efunc error, ldfunc ldef);
 301
 302     /*
 303      * This procedure is called by assemble() to write actual
 304      * generated code or data to the object file. Typically it
 305      * doesn't have to actually _write_ it, just store it for
 306      * later.
 307      *
 308      * The `type' argument specifies the type of output data, and
 309      * usually the size as well: its contents are described below.
 310      */
 311     void (*output) (long segto, void *data, unsigned long type,
 312                     long segment, long wrt);
 313
 314     /*
 315      * This procedure is called once for every symbol defined in
 316      * the module being assembled. It gives the name and value of
 317      * the symbol, in NASM's terms, and indicates whether it has
 318      * been declared to be global. Note that the parameter "name",
 319      * when passed, will point to a piece of static storage
 320      * allocated inside the label manager - it's safe to keep using
 321      * that pointer, because the label manager doesn't clean up
 322      * until after the output driver has.
 323      *
 324      * Values of `is_global' are: 0 means the symbol is local; 1
 325      * means the symbol is global; 2 means the symbol is common (in
 326      * which case `offset' holds the _size_ of the variable).
 327      * Anything else is available for the output driver to use
 328      * internally.
 329      */
 330     void (*symdef) (char *name, long segment, long offset, int is_global);
 331
 332     /*
 333      * This procedure is called when the source code requests a
 334      * segment change. It should return the corresponding segment
 335      * _number_ for the name, or NO_SEG if the name is not a valid
 336      * segment name.
 337      *
 338      * It may also be called with NULL, in which case it is to
 339      * return the _default_ section number for starting assembly in.
 340      *
 341      * It is allowed to modify the string it is given a pointer to.
 342      *
 343      * It is also allowed to specify a default instruction size for
 344      * the segment, by setting `*bits' to 16 or 32. Or, if it
 345      * doesn't wish to define a default, it can leave `bits' alone.
 346      */
 347     long (*section) (char *name, int pass, int *bits);
 348
 349     /*
 350      * This procedure is called to modify the segment base values
 351      * returned from the SEG operator. It is given a segment base
 352      * value (i.e. a segment value with the low bit set), and is
 353      * required to produce in return a segment value which may be
 354      * different. It can map segment bases to absolute numbers by
 355      * means of returning SEG_ABS types.
 356      */
 357     long (*segbase) (long segment);
 358
 359     /*
 360      * This procedure is called to allow the output driver to
 361      * process its own specific directives. When called, it has the
 362      * directive word in `directive' and the parameter string in
 363      * `value'. It is called in both assembly passes, and `pass'
 364      * will be either 1 or 2.
 365      *
 366      * This procedure should return zero if it does not _recognise_
 367      * the directive, so that the main program can report an error.
 368      * If it recognises the directive but then has its own errors,
 369      * it should report them itself and then return non-zero. It
 370      * should also return non-zero if it correctly processes the
 371      * directive.
 372      */
 373     int (*directive) (char *directive, char *value, int pass);
 374
 375     /*
 376      * This procedure is called before anything else - even before
 377      * the "init" routine - and is passed the name of the input
 378      * file from which this output file is being generated. It
 379      * should return its preferred name for the output file in
 380      * `outfunc'. Since it is called before the driver is properly
 381      * initialised, it has to be passed its error handler
 382      * separately.
 383      *
 384      * This procedure may also take its own copy of the input file
 385      * name for use in writing the output file: it is _guaranteed_
 386      * that it will be called before the "init" routine.
 387      *
 388      * The parameter `outname' points to an area of storage
 389      * guaranteed to be at least FILENAME_MAX in size.
 390      */
 391     void (*filename) (char *inname, char *outname, efunc error);
 392
 393     /*
 394      * This procedure is called after assembly finishes, to allow
 395      * the output driver to clean itself up and free its memory.
 396      * Typically, it will also be the point at which the object
 397      * file actually gets _written_.
 398      *
 399      * One thing the cleanup routine should always do is to close
 400      * the output file pointer.
 401      */
 402     void (*cleanup) (void);
 403 };
 404
 405 /*
 406  * values for the `type' parameter to an output function. Each one
 407  * must have the actual number of _bytes_ added to it.
 408  *
 409  * Exceptions are OUT_RELxADR, which denote an x-byte relocation
 410  * which will be a relative jump. For this we need to know the
 411  * distance in bytes from the start of the relocated record until
 412  * the end of the containing instruction. _This_ is what is stored
 413  * in the size part of the parameter, in this case.
 414  *
 415  * Also OUT_RESERVE denotes reservation of N bytes of BSS space,
 416  * and the contents of the "data" parameter is irrelevant.
 417  *
 418  * The "data" parameter for the output function points to a "long",
 419  * containing the address in question, unless the type is
 420  * OUT_RAWDATA, in which case it points to an "unsigned char"
 421  * array.
 422  */
 423 #define OUT_RAWDATA 0x00000000UL
 424 #define OUT_ADDRESS 0x10000000UL
 425 #define OUT_REL2ADR 0x20000000UL
 426 #define OUT_REL4ADR 0x30000000UL
 427 #define OUT_RESERVE 0x40000000UL
 428 #define OUT_TYPMASK 0xF0000000UL
 429 #define OUT_SIZMASK 0x0FFFFFFFUL
 430
 431 /*
 432  * -----
 433  * Other
 434  * -----
 435  */
 436
 437 /*
 438  * This is a useful #define which I keep meaning to use more often:
 439  * the number of elements of a statically defined array.
 440  */
 441
 442 #define elements(x)     ( sizeof(x) / sizeof(*(x)) )
 443
 444 #endif