dynasm/dasm_x86.h

   1 /*
   2 ** DynASM x86 encoding engine.
   3 ** Copyright (C) 2005-2017 Mike Pall. All rights reserved.
   4 ** Released under the MIT license. See dynasm.lua for full copyright notice.
   5 */
   6
   7 #include <stddef.h>
   8 #include <stdarg.h>
   9 #include <string.h>
  10 #include <stdlib.h>
  11
  12 #define DASM_ARCH               "x86"
  13
  14 #ifndef DASM_EXTERN
  15 #define DASM_EXTERN(a,b,c,d)    0
  16 #endif
  17
  18 /* Action definitions. DASM_STOP must be 255. */
  19 enum {
  20   DASM_DISP = 233,
  21   DASM_IMM_S, DASM_IMM_B, DASM_IMM_W, DASM_IMM_D, DASM_IMM_WB, DASM_IMM_DB,
  22   DASM_VREG, DASM_SPACE, DASM_SETLABEL, DASM_REL_A, DASM_REL_LG, DASM_REL_PC,
  23   DASM_IMM_LG, DASM_IMM_PC, DASM_LABEL_LG, DASM_LABEL_PC, DASM_ALIGN,
  24   DASM_EXTERN, DASM_ESC, DASM_MARK, DASM_SECTION, DASM_STOP
  25 };
  26
  27 /* Maximum number of section buffer positions for a single dasm_put() call. */
  28 #define DASM_MAXSECPOS          25
  29
  30 /* DynASM encoder status codes. Action list offset or number are or'ed in. */
  31 #define DASM_S_OK               0x00000000
  32 #define DASM_S_NOMEM            0x01000000
  33 #define DASM_S_PHASE            0x02000000
  34 #define DASM_S_MATCH_SEC        0x03000000
  35 #define DASM_S_RANGE_I          0x11000000
  36 #define DASM_S_RANGE_SEC        0x12000000
  37 #define DASM_S_RANGE_LG         0x13000000
  38 #define DASM_S_RANGE_PC         0x14000000
  39 #define DASM_S_RANGE_VREG       0x15000000
  40 #define DASM_S_UNDEF_L          0x21000000
  41 #define DASM_S_UNDEF_PC         0x22000000
  42
  43 /* Macros to convert positions (8 bit section + 24 bit index). */
  44 #define DASM_POS2IDX(pos)       ((pos)&0x00ffffff)
  45 #define DASM_POS2BIAS(pos)      ((pos)&0xff000000)
  46 #define DASM_SEC2POS(sec)       ((sec)<<24)
  47 #define DASM_POS2SEC(pos)       ((pos)>>24)
  48 #define DASM_POS2PTR(D, pos)    (D->sections[DASM_POS2SEC(pos)].rbuf + (pos))
  49
  50 /* Action list type. */
  51 typedef const unsigned char *dasm_ActList;
  52
  53 /* Per-section structure. */
  54 typedef struct dasm_Section {
  55   int *rbuf;            /* Biased buffer pointer (negative section bias). */
  56   int *buf;             /* True buffer pointer. */
  57   size_t bsize;         /* Buffer size in bytes. */
  58   int pos;              /* Biased buffer position. */
  59   int epos;             /* End of biased buffer position - max single put. */
  60   int ofs;              /* Byte offset into section. */
  61 } dasm_Section;
  62
  63 /* Core structure holding the DynASM encoding state. */
  64 struct dasm_State {
  65   size_t psize;                 /* Allocated size of this structure. */
  66   dasm_ActList actionlist;      /* Current actionlist pointer. */
  67   int *lglabels;                /* Local/global chain/pos ptrs. */
  68   size_t lgsize;
  69   int *pclabels;                /* PC label chains/pos ptrs. */
  70   size_t pcsize;
  71   void **globals;               /* Array of globals (bias -10). */
  72   dasm_Section *section;        /* Pointer to active section. */
  73   size_t codesize;              /* Total size of all code sections. */
  74   int maxsection;               /* 0 <= sectionidx < maxsection. */
  75   int status;                   /* Status code. */
  76   dasm_Section sections[1];     /* All sections. Alloc-extended. */
  77 };
  78
  79 /* The size of the core structure depends on the max. number of sections. */
  80 #define DASM_PSZ(ms)    (sizeof(dasm_State)+(ms-1)*sizeof(dasm_Section))
  81
  82
  83 /* Initialize DynASM state. */
  84 void dasm_init(Dst_DECL, int maxsection)
  85 {
  86   dasm_State *D;
  87   size_t psz = 0;
  88   int i;
  89   Dst_REF = NULL;
  90   DASM_M_GROW(Dst, struct dasm_State, Dst_REF, psz, DASM_PSZ(maxsection));
  91   D = Dst_REF;
  92   D->psize = psz;
  93   D->lglabels = NULL;
  94   D->lgsize = 0;
  95   D->pclabels = NULL;
  96   D->pcsize = 0;
  97   D->globals = NULL;
  98   D->maxsection = maxsection;
  99   for (i = 0; i < maxsection; i++) {
 100     D->sections[i].buf = NULL;  /* Need this for pass3. */
 101     D->sections[i].rbuf = D->sections[i].buf - DASM_SEC2POS(i);
 102     D->sections[i].bsize = 0;
 103     D->sections[i].epos = 0;  /* Wrong, but is recalculated after resize. */
 104   }
 105 }
 106
 107 /* Free DynASM state. */
 108 void dasm_free(Dst_DECL)
 109 {
 110   dasm_State *D = Dst_REF;
 111   int i;
 112   for (i = 0; i < D->maxsection; i++)
 113     if (D->sections[i].buf)
 114       DASM_M_FREE(Dst, D->sections[i].buf, D->sections[i].bsize);
 115   if (D->pclabels) DASM_M_FREE(Dst, D->pclabels, D->pcsize);
 116   if (D->lglabels) DASM_M_FREE(Dst, D->lglabels, D->lgsize);
 117   DASM_M_FREE(Dst, D, D->psize);
 118 }
 119
 120 /* Setup global label array. Must be called before dasm_setup(). */
 121 void dasm_setupglobal(Dst_DECL, void **gl, unsigned int maxgl)
 122 {
 123   dasm_State *D = Dst_REF;
 124   D->globals = gl - 10;  /* Negative bias to compensate for locals. */
 125   DASM_M_GROW(Dst, int, D->lglabels, D->lgsize, (10+maxgl)*sizeof(int));
 126 }
 127
 128 /* Grow PC label array. Can be called after dasm_setup(), too. */
 129 void dasm_growpc(Dst_DECL, unsigned int maxpc)
 130 {
 131   dasm_State *D = Dst_REF;
 132   size_t osz = D->pcsize;
 133   DASM_M_GROW(Dst, int, D->pclabels, D->pcsize, maxpc*sizeof(int));
 134   memset((void *)(((unsigned char *)D->pclabels)+osz), 0, D->pcsize-osz);
 135 }
 136
 137 /* Setup encoder. */
 138 void dasm_setup(Dst_DECL, const void *actionlist)
 139 {
 140   dasm_State *D = Dst_REF;
 141   int i;
 142   D->actionlist = (dasm_ActList)actionlist;
 143   D->status = DASM_S_OK;
 144   D->section = &D->sections[0];
 145   memset((void *)D->lglabels, 0, D->lgsize);
 146   if (D->pclabels) memset((void *)D->pclabels, 0, D->pcsize);
 147   for (i = 0; i < D->maxsection; i++) {
 148     D->sections[i].pos = DASM_SEC2POS(i);
 149     D->sections[i].ofs = 0;
 150   }
 151 }
 152
 153
 154 #ifdef DASM_CHECKS
 155 #define CK(x, st) \
 156   do { if (!(x)) { \
 157     D->status = DASM_S_##st|(int)(p-D->actionlist-1); return; } } while (0)
 158 #define CKPL(kind, st) \
 159   do { if ((size_t)((char *)pl-(char *)D->kind##labels) >= D->kind##size) { \
 160     D->status=DASM_S_RANGE_##st|(int)(p-D->actionlist-1); return; } } while (0)
 161 #else
 162 #define CK(x, st)       ((void)0)
 163 #define CKPL(kind, st)  ((void)0)
 164 #endif
 165
 166 /* Pass 1: Store actions and args, link branches/labels, estimate offsets. */
 167 void dasm_put(Dst_DECL, int start, ...)
 168 {
 169   va_list ap;
 170   dasm_State *D = Dst_REF;
 171   dasm_ActList p = D->actionlist + start;
 172   dasm_Section *sec = D->section;
 173   int pos = sec->pos, ofs = sec->ofs, mrm = 4;
 174   int *b;
 175
 176   if (pos >= sec->epos) {
 177     DASM_M_GROW(Dst, int, sec->buf, sec->bsize,
 178       sec->bsize + 2*DASM_MAXSECPOS*sizeof(int));
 179     sec->rbuf = sec->buf - DASM_POS2BIAS(pos);
 180     sec->epos = (int)sec->bsize/sizeof(int) - DASM_MAXSECPOS+DASM_POS2BIAS(pos);
 181   }
 182
 183   b = sec->rbuf;
 184   b[pos++] = start;
 185
 186   va_start(ap, start);
 187   while (1) {
 188     int action = *p++;
 189     if (action < DASM_DISP) {
 190       ofs++;
 191     } else if (action <= DASM_REL_A) {
 192       int n = va_arg(ap, int);
 193       b[pos++] = n;
 194       switch (action) {
 195       case DASM_DISP:
 196         if (n == 0) { if ((mrm&7) == 4) mrm = p[-2]; if ((mrm&7) != 5) break; }
 197       case DASM_IMM_DB: if (((n+128)&-256) == 0) goto ob;
 198       case DASM_REL_A: /* Assumes ptrdiff_t is int. !x64 */
 199       case DASM_IMM_D: ofs += 4; break;
 200       case DASM_IMM_S: CK(((n+128)&-256) == 0, RANGE_I); goto ob;
 201       case DASM_IMM_B: CK((n&-256) == 0, RANGE_I); ob: ofs++; break;
 202       case DASM_IMM_WB: if (((n+128)&-256) == 0) goto ob;
 203       case DASM_IMM_W: CK((n&-65536) == 0, RANGE_I); ofs += 2; break;
 204       case DASM_SPACE: p++; ofs += n; break;
 205       case DASM_SETLABEL: b[pos-2] = -0x40000000; break;  /* Neg. label ofs. */
 206       case DASM_VREG: CK((n&-8) == 0 && (n != 4 || (*p&1) == 0), RANGE_VREG);
 207         if (*p++ == 1 && *p == DASM_DISP) mrm = n;
 208         continue;
 209       }
 210       mrm = 4;
 211     } else {
 212       int *pl, n;
 213       switch (action) {
 214       case DASM_REL_LG:
 215       case DASM_IMM_LG:
 216         n = *p++; pl = D->lglabels + n;
 217         /* Bkwd rel or global. */
 218         if (n <= 246) { CK(n>=10||*pl<0, RANGE_LG); CKPL(lg, LG); goto putrel; }
 219         pl -= 246; n = *pl;
 220         if (n < 0) n = 0;  /* Start new chain for fwd rel if label exists. */
 221         goto linkrel;
 222       case DASM_REL_PC:
 223       case DASM_IMM_PC: pl = D->pclabels + va_arg(ap, int); CKPL(pc, PC);
 224       putrel:
 225         n = *pl;
 226         if (n < 0) {  /* Label exists. Get label pos and store it. */
 227           b[pos] = -n;
 228         } else {
 229       linkrel:
 230           b[pos] = n;  /* Else link to rel chain, anchored at label. */
 231           *pl = pos;
 232         }
 233         pos++;
 234         ofs += 4;  /* Maximum offset needed. */
 235         if (action == DASM_REL_LG || action == DASM_REL_PC)
 236           b[pos++] = ofs;  /* Store pass1 offset estimate. */
 237         break;
 238       case DASM_LABEL_LG: pl = D->lglabels + *p++; CKPL(lg, LG); goto putlabel;
 239       case DASM_LABEL_PC: pl = D->pclabels + va_arg(ap, int); CKPL(pc, PC);
 240       putlabel:
 241         n = *pl;  /* n > 0: Collapse rel chain and replace with label pos. */
 242         while (n > 0) { int *pb = DASM_POS2PTR(D, n); n = *pb; *pb = pos; }
 243         *pl = -pos;  /* Label exists now. */
 244         b[pos++] = ofs;  /* Store pass1 offset estimate. */
 245         break;
 246       case DASM_ALIGN:
 247         ofs += *p++;  /* Maximum alignment needed (arg is 2**n-1). */
 248         b[pos++] = ofs;  /* Store pass1 offset estimate. */
 249         break;
 250       case DASM_EXTERN: p += 2; ofs += 4; break;
 251       case DASM_ESC: p++; ofs++; break;
 252       case DASM_MARK: mrm = p[-2]; break;
 253       case DASM_SECTION:
 254         n = *p; CK(n < D->maxsection, RANGE_SEC); D->section = &D->sections[n];
 255       case DASM_STOP: goto stop;
 256       }
 257     }
 258   }
 259 stop:
 260   va_end(ap);
 261   sec->pos = pos;
 262   sec->ofs = ofs;
 263 }
 264 #undef CK
 265
 266 /* Pass 2: Link sections, shrink branches/aligns, fix label offsets. */
 267 int dasm_link(Dst_DECL, size_t *szp)
 268 {
 269   dasm_State *D = Dst_REF;
 270   int secnum;
 271   int ofs = 0;
 272
 273 #ifdef DASM_CHECKS
 274   *szp = 0;
 275   if (D->status != DASM_S_OK) return D->status;
 276   {
 277     int pc;
 278     for (pc = 0; pc*sizeof(int) < D->pcsize; pc++)
 279       if (D->pclabels[pc] > 0) return DASM_S_UNDEF_PC|pc;
 280   }
 281 #endif
 282
 283   { /* Handle globals not defined in this translation unit. */
 284     int idx;
 285     for (idx = 10; idx*sizeof(int) < D->lgsize; idx++) {
 286       int n = D->lglabels[idx];
 287       /* Undefined label: Collapse rel chain and replace with marker (< 0). */
 288       while (n > 0) { int *pb = DASM_POS2PTR(D, n); n = *pb; *pb = -idx; }
 289     }
 290   }
 291
 292   /* Combine all code sections. No support for data sections (yet). */
 293   for (secnum = 0; secnum < D->maxsection; secnum++) {
 294     dasm_Section *sec = D->sections + secnum;
 295     int *b = sec->rbuf;
 296     int pos = DASM_SEC2POS(secnum);
 297     int lastpos = sec->pos;
 298
 299     while (pos != lastpos) {
 300       dasm_ActList p = D->actionlist + b[pos++];
 301       while (1) {
 302         int op, action = *p++;
 303         switch (action) {
 304         case DASM_REL_LG: p++; op = p[-3]; goto rel_pc;
 305         case DASM_REL_PC: op = p[-2]; rel_pc: {
 306           int shrink = op == 0xe9 ? 3 : ((op&0xf0) == 0x80 ? 4 : 0);
 307           if (shrink) {  /* Shrinkable branch opcode? */
 308             int lofs, lpos = b[pos];
 309             if (lpos < 0) goto noshrink;  /* Ext global? */
 310             lofs = *DASM_POS2PTR(D, lpos);
 311             if (lpos > pos) {  /* Fwd label: add cumulative section offsets. */
 312               int i;
 313               for (i = secnum; i < DASM_POS2SEC(lpos); i++)
 314                 lofs += D->sections[i].ofs;
 315             } else {
 316               lofs -= ofs;  /* Bkwd label: unfix offset. */
 317             }
 318             lofs -= b[pos+1];  /* Short branch ok? */
 319             if (lofs >= -128-shrink && lofs <= 127) ofs -= shrink;  /* Yes. */
 320             else { noshrink: shrink = 0; }  /* No, cannot shrink op. */
 321           }
 322           b[pos+1] = shrink;
 323           pos += 2;
 324           break;
 325         }
 326         case DASM_SPACE: case DASM_IMM_LG: case DASM_VREG: p++;
 327         case DASM_DISP: case DASM_IMM_S: case DASM_IMM_B: case DASM_IMM_W:
 328         case DASM_IMM_D: case DASM_IMM_WB: case DASM_IMM_DB:
 329         case DASM_SETLABEL: case DASM_REL_A: case DASM_IMM_PC: pos++; break;
 330         case DASM_LABEL_LG: p++;
 331         case DASM_LABEL_PC: b[pos++] += ofs; break; /* Fix label offset. */
 332         case DASM_ALIGN: ofs -= (b[pos++]+ofs)&*p++; break; /* Adjust ofs. */
 333         case DASM_EXTERN: p += 2; break;
 334         case DASM_ESC: p++; break;
 335         case DASM_MARK: break;
 336         case DASM_SECTION: case DASM_STOP: goto stop;
 337         }
 338       }
 339       stop: (void)0;
 340     }
 341     ofs += sec->ofs;  /* Next section starts right after current section. */
 342   }
 343
 344   D->codesize = ofs;  /* Total size of all code sections */
 345   *szp = ofs;
 346   return DASM_S_OK;
 347 }
 348
 349 #define dasmb(x)        *cp++ = (unsigned char)(x)
 350 #ifndef DASM_ALIGNED_WRITES
 351 #define dasmw(x) \
 352   do { *((unsigned short *)cp) = (unsigned short)(x); cp+=2; } while (0)
 353 #define dasmd(x) \
 354   do { *((unsigned int *)cp) = (unsigned int)(x); cp+=4; } while (0)
 355 #else
 356 #define dasmw(x)        do { dasmb(x); dasmb((x)>>8); } while (0)
 357 #define dasmd(x)        do { dasmw(x); dasmw((x)>>16); } while (0)
 358 #endif
 359
 360 /* Pass 3: Encode sections. */
 361 int dasm_encode(Dst_DECL, void *buffer)
 362 {
 363   dasm_State *D = Dst_REF;
 364   unsigned char *base = (unsigned char *)buffer;
 365   unsigned char *cp = base;
 366   int secnum;
 367
 368   /* Encode all code sections. No support for data sections (yet). */
 369   for (secnum = 0; secnum < D->maxsection; secnum++) {
 370     dasm_Section *sec = D->sections + secnum;
 371     int *b = sec->buf;
 372     int *endb = sec->rbuf + sec->pos;
 373
 374     while (b != endb) {
 375       dasm_ActList p = D->actionlist + *b++;
 376       unsigned char *mark = NULL;
 377       while (1) {
 378         int action = *p++;
 379         int n = (action >= DASM_DISP && action <= DASM_ALIGN) ? *b++ : 0;
 380         switch (action) {
 381         case DASM_DISP: if (!mark) mark = cp; {
 382           unsigned char *mm = mark;
 383           if (*p != DASM_IMM_DB && *p != DASM_IMM_WB) mark = NULL;
 384           if (n == 0) { int mrm = mm[-1]&7; if (mrm == 4) mrm = mm[0]&7;
 385             if (mrm != 5) { mm[-1] -= 0x80; break; } }
 386           if (((n+128) & -256) != 0) goto wd; else mm[-1] -= 0x40;
 387         }
 388         case DASM_IMM_S: case DASM_IMM_B: wb: dasmb(n); break;
 389         case DASM_IMM_DB: if (((n+128)&-256) == 0) {
 390             db: if (!mark) mark = cp; mark[-2] += 2; mark = NULL; goto wb;
 391           } else mark = NULL;
 392         case DASM_IMM_D: wd: dasmd(n); break;
 393         case DASM_IMM_WB: if (((n+128)&-256) == 0) goto db; else mark = NULL;
 394         case DASM_IMM_W: dasmw(n); break;
 395         case DASM_VREG: { int t = *p++; if (t >= 2) n<<=3; cp[-1] |= n; break; }
 396         case DASM_REL_LG: p++; if (n >= 0) goto rel_pc;
 397           b++; n = (int)(ptrdiff_t)D->globals[-n];
 398         case DASM_REL_A: rel_a:
 399           n -= (unsigned int)(ptrdiff_t)(cp+4); goto wd; /* !x64 */
 400         case DASM_REL_PC: rel_pc: {
 401           int shrink = *b++;
 402           int *pb = DASM_POS2PTR(D, n); if (*pb < 0) { n = pb[1]; goto rel_a; }
 403           n = *pb - ((int)(cp-base) + 4-shrink);
 404           if (shrink == 0) goto wd;
 405           if (shrink == 4) { cp--; cp[-1] = *cp-0x10; } else cp[-1] = 0xeb;
 406           goto wb;
 407         }
 408         case DASM_IMM_LG:
 409           p++; if (n < 0) { n = (int)(ptrdiff_t)D->globals[-n]; goto wd; }
 410         case DASM_IMM_PC: {
 411           int *pb = DASM_POS2PTR(D, n);
 412           n = *pb < 0 ? pb[1] : (*pb + (int)(ptrdiff_t)base);
 413           goto wd;
 414         }
 415         case DASM_LABEL_LG: {
 416           int idx = *p++;
 417           if (idx >= 10)
 418             D->globals[idx] = (void *)(base + (*p == DASM_SETLABEL ? *b : n));
 419           break;
 420         }
 421         case DASM_LABEL_PC: case DASM_SETLABEL: break;
 422         case DASM_SPACE: { int fill = *p++; while (n--) *cp++ = fill; break; }
 423         case DASM_ALIGN:
 424           n = *p++;
 425           while (((cp-base) & n)) *cp++ = 0x90; /* nop */
 426           break;
 427         case DASM_EXTERN: n = DASM_EXTERN(Dst, cp, p[1], *p); p += 2; goto wd;
 428         case DASM_MARK: mark = cp; break;
 429         case DASM_ESC: action = *p++;
 430         default: *cp++ = action; break;
 431         case DASM_SECTION: case DASM_STOP: goto stop;
 432         }
 433       }
 434       stop: (void)0;
 435     }
 436   }
 437
 438   if (base + D->codesize != cp)  /* Check for phase errors. */
 439     return DASM_S_PHASE;
 440   return DASM_S_OK;
 441 }
 442
 443 /* Get PC label offset. */
 444 int dasm_getpclabel(Dst_DECL, unsigned int pc)
 445 {
 446   dasm_State *D = Dst_REF;
 447   if (pc*sizeof(int) < D->pcsize) {
 448     int pos = D->pclabels[pc];
 449     if (pos < 0) return *DASM_POS2PTR(D, -pos);
 450     if (pos > 0) return -1;  /* Undefined. */
 451   }
 452   return -2;  /* Unused or out of range. */
 453 }
 454
 455 #ifdef DASM_CHECKS
 456 /* Optional sanity checker to call between isolated encoding steps. */
 457 int dasm_checkstep(Dst_DECL, int secmatch)
 458 {
 459   dasm_State *D = Dst_REF;
 460   if (D->status == DASM_S_OK) {
 461     int i;
 462     for (i = 1; i <= 9; i++) {
 463       if (D->lglabels[i] > 0) { D->status = DASM_S_UNDEF_L|i; break; }
 464       D->lglabels[i] = 0;
 465     }
 466   }
 467   if (D->status == DASM_S_OK && secmatch >= 0 &&
 468       D->section != &D->sections[secmatch])
 469     D->status = DASM_S_MATCH_SEC|(int)(D->section-D->sections);
 470   return D->status;
 471 }
 472 #endif
 473