src/lib/lttng-ctl/filter-visitor-generate-bytecode.c

   1 /*
   2  * filter-visitor-generate-bytecode.c
   3  *
   4  * LTTng filter bytecode generation
   5  *
   6  * Copyright 2012 - Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
   7  *
   8  * This library is free software; you can redistribute it and/or modify it
   9  * under the terms of the GNU Lesser General Public License, version 2.1 only,
  10  * as published by the Free Software Foundation.
  11  *
  12  * This library 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 GNU
  15  * Lesser General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Lesser General Public License
  18  * along with this library; if not, write to the Free Software Foundation,
  19  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20  */
  21
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include <errno.h>
  25 #include "align.h"
  26 #include "filter-bytecode.h"
  27 #include "filter-ir.h"
  28 #include "filter-ast.h"
  29
  30 #ifndef max_t
  31 #define max_t(type, a, b)       ((type) ((a) > (b) ? (a) : (b)))
  32 #endif
  33
  34 //#define INIT_ALLOC_SIZE               PAGE_SIZE
  35 #define INIT_ALLOC_SIZE         4
  36
  37 static
  38 int recursive_visit_gen_bytecode(struct filter_parser_ctx *ctx,
  39                 struct ir_op *node);
  40
  41 static
  42 int bytecode_init(struct lttng_filter_bytecode_alloc **fb)
  43 {
  44         *fb = calloc(sizeof(struct lttng_filter_bytecode_alloc) + INIT_ALLOC_SIZE, 1);
  45         if (!*fb) {
  46                 return -ENOMEM;
  47         } else {
  48                 (*fb)->alloc_len = INIT_ALLOC_SIZE;
  49                 return 0;
  50         }
  51 }
  52
  53 static
  54 int32_t bytecode_reserve(struct lttng_filter_bytecode_alloc **fb, uint32_t align, uint32_t len)
  55 {
  56         int32_t ret;
  57         uint32_t padding = offset_align((*fb)->b.len, align);
  58
  59         if ((*fb)->b.len + padding + len > (*fb)->alloc_len) {
  60                 uint32_t new_len =
  61                         max_t(uint32_t, (*fb)->b.len + padding + len,
  62                                 (*fb)->alloc_len << 1);
  63                 uint32_t old_len = (*fb)->alloc_len;
  64
  65                 if (new_len > 0xFFFF)
  66                         return -EINVAL;
  67                 *fb = realloc(*fb, sizeof(struct lttng_filter_bytecode_alloc) + new_len);
  68                 if (!*fb)
  69                         return -ENOMEM;
  70                 memset(&(*fb)->b.data[old_len], 0, new_len - old_len);
  71                 (*fb)->alloc_len = new_len;
  72         }
  73         (*fb)->b.len += padding;
  74         ret = (*fb)->b.len;
  75         (*fb)->b.len += len;
  76         return ret;
  77 }
  78
  79 static
  80 int bytecode_push(struct lttng_filter_bytecode_alloc **fb, const void *data,
  81                 uint32_t align, uint32_t len)
  82 {
  83         int32_t offset;
  84
  85         offset = bytecode_reserve(fb, align, len);
  86         if (offset < 0)
  87                 return offset;
  88         memcpy(&(*fb)->b.data[offset], data, len);
  89         return 0;
  90 }
  91
  92 static
  93 int bytecode_push_logical(struct lttng_filter_bytecode_alloc **fb,
  94                 struct logical_op *data,
  95                 uint32_t align, uint32_t len,
  96                 uint16_t *skip_offset)
  97 {
  98         int32_t offset;
  99
 100         offset = bytecode_reserve(fb, align, len);
 101         if (offset < 0)
 102                 return offset;
 103         memcpy(&(*fb)->b.data[offset], data, len);
 104         *skip_offset =
 105                 (void *) &((struct logical_op *) &(*fb)->b.data[offset])->skip_offset
 106                         - (void *) &(*fb)->b.data[0];
 107         return 0;
 108 }
 109
 110 static
 111 int bytecode_patch(struct lttng_filter_bytecode_alloc **fb,
 112                 const void *data,
 113                 uint16_t offset,
 114                 uint32_t len)
 115 {
 116         if (offset >= (*fb)->b.len) {
 117                 return -EINVAL;
 118         }
 119         memcpy(&(*fb)->b.data[offset], data, len);
 120         return 0;
 121 }
 122
 123 static
 124 int visit_node_root(struct filter_parser_ctx *ctx, struct ir_op *node)
 125 {
 126         int ret;
 127         struct return_op insn;
 128
 129         /* Visit child */
 130         ret = recursive_visit_gen_bytecode(ctx, node->u.root.child);
 131         if (ret)
 132                 return ret;
 133
 134         /* Generate end of bytecode instruction */
 135         insn.op = FILTER_OP_RETURN;
 136         return bytecode_push(&ctx->bytecode, &insn, 1, sizeof(insn));
 137 }
 138
 139 static
 140 enum filter_register reg_sel(struct ir_op *node)
 141 {
 142         switch (node->side) {
 143         case IR_SIDE_UNKNOWN:
 144         default:
 145                 fprintf(stderr, "[error] Unknown node side in %s\n",
 146                         __func__);
 147                 return REG_ERROR;
 148         case IR_LEFT:
 149                 return REG_R0;
 150         case IR_RIGHT:
 151                 return REG_R1;
 152         }
 153 }
 154
 155 static
 156 int visit_node_load(struct filter_parser_ctx *ctx, struct ir_op *node)
 157 {
 158         int ret;
 159
 160         switch (node->data_type) {
 161         case IR_DATA_UNKNOWN:
 162         default:
 163                 fprintf(stderr, "[error] Unknown data type in %s\n",
 164                         __func__);
 165                 return -EINVAL;
 166
 167         case IR_DATA_STRING:
 168         {
 169                 struct load_op *insn;
 170                 uint32_t insn_len = sizeof(struct load_op)
 171                         + strlen(node->u.load.u.string) + 1;
 172
 173                 insn = calloc(insn_len, 1);
 174                 if (!insn)
 175                         return -ENOMEM;
 176                 insn->op = FILTER_OP_LOAD_STRING;
 177                 insn->reg = reg_sel(node);
 178                 if (insn->reg == REG_ERROR)
 179                         return -EINVAL;
 180                 strcpy(insn->data, node->u.load.u.string);
 181                 ret = bytecode_push(&ctx->bytecode, insn, 1, insn_len);
 182                 free(insn);
 183                 return ret;
 184         }
 185         case IR_DATA_NUMERIC:
 186         {
 187                 struct load_op *insn;
 188                 uint32_t insn_len = sizeof(struct load_op)
 189                         + sizeof(struct literal_numeric);
 190
 191                 insn = calloc(insn_len, 1);
 192                 if (!insn)
 193                         return -ENOMEM;
 194                 insn->op = FILTER_OP_LOAD_S64;
 195                 insn->reg = reg_sel(node);
 196                 if (insn->reg == REG_ERROR)
 197                         return -EINVAL;
 198                 *(int64_t *) insn->data = node->u.load.u.num;
 199                 ret = bytecode_push(&ctx->bytecode, insn, 1, insn_len);
 200                 free(insn);
 201                 return ret;
 202         }
 203         case IR_DATA_FLOAT:
 204         {
 205                 struct load_op *insn;
 206                 uint32_t insn_len = sizeof(struct load_op)
 207                         + sizeof(struct literal_double);
 208
 209                 insn = calloc(insn_len, 1);
 210                 if (!insn)
 211                         return -ENOMEM;
 212                 insn->op = FILTER_OP_LOAD_DOUBLE;
 213                 insn->reg = reg_sel(node);
 214                 if (insn->reg == REG_ERROR)
 215                         return -EINVAL;
 216                 *(double *) insn->data = node->u.load.u.flt;
 217                 ret = bytecode_push(&ctx->bytecode, insn, 1, insn_len);
 218                 free(insn);
 219                 return ret;
 220         }
 221         case IR_DATA_FIELD_REF:
 222         {
 223                 struct load_op *insn;
 224                 uint32_t insn_len = sizeof(struct load_op)
 225                         + sizeof(struct field_ref);
 226                 struct field_ref ref_offset;
 227                 uint16_t reloc_offset;
 228
 229                 insn = calloc(insn_len, 1);
 230                 if (!insn)
 231                         return -ENOMEM;
 232                 insn->op = FILTER_OP_LOAD_FIELD_REF;
 233                 insn->reg = reg_sel(node);
 234                 ref_offset.offset = (uint16_t) -1U;
 235                 memcpy(insn->data, &ref_offset, sizeof(ref_offset));
 236                 if (insn->reg == REG_ERROR)
 237                         return -EINVAL;
 238                 /* reloc_offset points to struct field_ref */
 239                 reloc_offset = bytecode_get_len(&ctx->bytecode->b);
 240                 reloc_offset += sizeof(struct load_op);
 241                 ret = bytecode_push(&ctx->bytecode, insn, 1, insn_len);
 242                 if (ret) {
 243                         free(insn);
 244                         return ret;
 245                 }
 246                 /* append reloc */
 247                 ret = bytecode_push(&ctx->bytecode_reloc, &reloc_offset,
 248                                         1, sizeof(reloc_offset));
 249                 if (ret) {
 250                         free(insn);
 251                         return ret;
 252                 }
 253                 ret = bytecode_push(&ctx->bytecode_reloc, node->u.load.u.ref,
 254                                         1, strlen(node->u.load.u.ref) + 1);
 255                 free(insn);
 256                 return ret;
 257         }
 258         }
 259 }
 260
 261 static
 262 int visit_node_unary(struct filter_parser_ctx *ctx, struct ir_op *node)
 263 {
 264         int ret;
 265         struct unary_op insn;
 266
 267         /* Visit child */
 268         ret = recursive_visit_gen_bytecode(ctx, node->u.unary.child);
 269         if (ret)
 270                 return ret;
 271
 272         /* Generate end of bytecode instruction */
 273         switch (node->u.unary.type) {
 274         case AST_UNARY_UNKNOWN:
 275         default:
 276                 fprintf(stderr, "[error] Unknown unary node type in %s\n",
 277                         __func__);
 278                 return -EINVAL;
 279         case AST_UNARY_PLUS:
 280                 /* Nothing to do. */
 281                 return 0;
 282         case AST_UNARY_MINUS:
 283                 insn.op = FILTER_OP_UNARY_MINUS;
 284                 insn.reg = reg_sel(node);
 285                 if (insn.reg == REG_ERROR)
 286                         return -EINVAL;
 287                 return bytecode_push(&ctx->bytecode, &insn, 1, sizeof(insn));
 288         case AST_UNARY_NOT:
 289                 insn.op = FILTER_OP_UNARY_NOT;
 290                 insn.reg = reg_sel(node);
 291                 if (insn.reg == REG_ERROR)
 292                         return -EINVAL;
 293                 return bytecode_push(&ctx->bytecode, &insn, 1, sizeof(insn));
 294         }
 295 }
 296
 297 /*
 298  * Binary comparator nesting is disallowed. This allows fitting into
 299  * only 2 registers.
 300  */
 301 static
 302 int visit_node_binary(struct filter_parser_ctx *ctx, struct ir_op *node)
 303 {
 304         int ret;
 305         struct binary_op insn;
 306
 307         /* Visit child */
 308         ret = recursive_visit_gen_bytecode(ctx, node->u.binary.left);
 309         if (ret)
 310                 return ret;
 311         ret = recursive_visit_gen_bytecode(ctx, node->u.binary.right);
 312         if (ret)
 313                 return ret;
 314
 315         switch (node->u.binary.type) {
 316         case AST_OP_UNKNOWN:
 317         default:
 318                 fprintf(stderr, "[error] Unknown unary node type in %s\n",
 319                         __func__);
 320                 return -EINVAL;
 321
 322         case AST_OP_AND:
 323         case AST_OP_OR:
 324                 fprintf(stderr, "[error] Unexpected logical node type in %s\n",
 325                         __func__);
 326                 return -EINVAL;
 327
 328         case AST_OP_MUL:
 329                 insn.op = FILTER_OP_MUL;
 330                 break;
 331         case AST_OP_DIV:
 332                 insn.op = FILTER_OP_DIV;
 333                 break;
 334         case AST_OP_MOD:
 335                 insn.op = FILTER_OP_MOD;
 336                 break;
 337         case AST_OP_PLUS:
 338                 insn.op = FILTER_OP_PLUS;
 339                 break;
 340         case AST_OP_MINUS:
 341                 insn.op = FILTER_OP_MINUS;
 342                 break;
 343         case AST_OP_RSHIFT:
 344                 insn.op = FILTER_OP_RSHIFT;
 345                 break;
 346         case AST_OP_LSHIFT:
 347                 insn.op = FILTER_OP_LSHIFT;
 348                 break;
 349         case AST_OP_BIN_AND:
 350                 insn.op = FILTER_OP_BIN_AND;
 351                 break;
 352         case AST_OP_BIN_OR:
 353                 insn.op = FILTER_OP_BIN_OR;
 354                 break;
 355         case AST_OP_BIN_XOR:
 356                 insn.op = FILTER_OP_BIN_XOR;
 357                 break;
 358
 359         case AST_OP_EQ:
 360                 insn.op = FILTER_OP_EQ;
 361                 break;
 362         case AST_OP_NE:
 363                 insn.op = FILTER_OP_NE;
 364                 break;
 365         case AST_OP_GT:
 366                 insn.op = FILTER_OP_GT;
 367                 break;
 368         case AST_OP_LT:
 369                 insn.op = FILTER_OP_LT;
 370                 break;
 371         case AST_OP_GE:
 372                 insn.op = FILTER_OP_GE;
 373                 break;
 374         case AST_OP_LE:
 375                 insn.op = FILTER_OP_LE;
 376                 break;
 377         }
 378         return bytecode_push(&ctx->bytecode, &insn, 1, sizeof(insn));
 379 }
 380
 381 static
 382 int visit_node_logical(struct filter_parser_ctx *ctx, struct ir_op *node)
 383 {
 384         int ret;
 385         struct logical_op insn;
 386         uint16_t skip_offset_loc;
 387         uint16_t target_loc;
 388
 389         /* Visit left child */
 390         ret = recursive_visit_gen_bytecode(ctx, node->u.binary.left);
 391         if (ret)
 392                 return ret;
 393         switch (node->u.logical.type) {
 394         default:
 395                 fprintf(stderr, "[error] Unknown node type in %s\n",
 396                         __func__);
 397                 return -EINVAL;
 398
 399         case AST_OP_AND:
 400                 insn.op = FILTER_OP_AND;
 401                 break;
 402         case AST_OP_OR:
 403                 insn.op = FILTER_OP_OR;
 404                 break;
 405         }
 406         insn.skip_offset = (uint16_t) -1UL;     /* Temporary */
 407         ret = bytecode_push_logical(&ctx->bytecode, &insn, 1, sizeof(insn),
 408                         &skip_offset_loc);
 409         if (ret)
 410                 return ret;
 411         /* Visit right child */
 412         ret = recursive_visit_gen_bytecode(ctx, node->u.binary.right);
 413         if (ret)
 414                 return ret;
 415         /* We now know where the logical op can skip. */
 416         target_loc = (uint16_t) bytecode_get_len(&ctx->bytecode->b);
 417         ret = bytecode_patch(&ctx->bytecode,
 418                         &target_loc,                    /* Offset to jump to */
 419                         skip_offset_loc,                /* Where to patch */
 420                         sizeof(uint16_t));
 421         return ret;
 422 }
 423
 424 /*
 425  * Postorder traversal of the tree. We need the children result before
 426  * we can evaluate the parent.
 427  */
 428 static
 429 int recursive_visit_gen_bytecode(struct filter_parser_ctx *ctx,
 430                 struct ir_op *node)
 431 {
 432         switch (node->op) {
 433         case IR_OP_UNKNOWN:
 434         default:
 435                 fprintf(stderr, "[error] Unknown node type in %s\n",
 436                         __func__);
 437                 return -EINVAL;
 438
 439         case IR_OP_ROOT:
 440                 return visit_node_root(ctx, node);
 441         case IR_OP_LOAD:
 442                 return visit_node_load(ctx, node);
 443         case IR_OP_UNARY:
 444                 return visit_node_unary(ctx, node);
 445         case IR_OP_BINARY:
 446                 return visit_node_binary(ctx, node);
 447         case IR_OP_LOGICAL:
 448                 return visit_node_logical(ctx, node);
 449         }
 450 }
 451
 452 void filter_bytecode_free(struct filter_parser_ctx *ctx)
 453 {
 454         free(ctx->bytecode);
 455         ctx->bytecode = NULL;
 456         free(ctx->bytecode_reloc);
 457         ctx->bytecode_reloc = NULL;
 458 }
 459
 460 int filter_visitor_bytecode_generate(struct filter_parser_ctx *ctx)
 461 {
 462         int ret;
 463
 464         ret = bytecode_init(&ctx->bytecode);
 465         if (ret)
 466                 return ret;
 467         ret = bytecode_init(&ctx->bytecode_reloc);
 468         if (ret)
 469                 goto error;
 470         ret = recursive_visit_gen_bytecode(ctx, ctx->ir_root);
 471         if (ret)
 472                 goto error;
 473
 474         /* Finally, append symbol table to bytecode */
 475         ctx->bytecode->b.reloc_table_offset = bytecode_get_len(&ctx->bytecode->b);
 476         return bytecode_push(&ctx->bytecode, ctx->bytecode_reloc->b.data,
 477                         1, bytecode_get_len(&ctx->bytecode_reloc->b));
 478
 479 error:
 480         filter_bytecode_free(ctx);
 481         return ret;
 482 }