smatch_struct_assignment.c

   1 /*
   2  * Copyright (C) 2014 Oracle.
   3  *
   4  * This program is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU General Public License
   6  * as published by the Free Software Foundation; either version 2
   7  * of the License, or (at your option) any later version.
   8  *
   9  * This program is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12  * GNU General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU General Public License
  15  * along with this program; if not, see http://www.gnu.org/copyleft/gpl.txt
  16  */
  17
  18 /*
  19  * This file started out by saying that if you have:
  20  *
  21  *      struct foo one, two;
  22  *      ...
  23  *      one = two;
  24  *
  25  * That's equivalent to saying:
  26  *
  27  *      one.x = two.x;
  28  *      one.y = two.y;
  29  *
  30  * Turning an assignment like that into a bunch of small fake assignments is
  31  * really useful.
  32  *
  33  * The call to memcpy(&one, &two, sizeof(foo)); is the same as "one = two;" so
  34  * we can re-use the code.  And we may as well use it for memset() too.
  35  * Assigning pointers is almost the same:
  36  *
  37  *      p1 = p2;
  38  *
  39  * Is the same as:
  40  *
  41  *      p1->x = p2->x;
  42  *      p1->y = p2->y;
  43  *
  44  * The problem is that you can go a bit crazy with pointers to pointers.
  45  *
  46  *      p1->x->y->z->one->two->three = p2->x->y->z->one->two->three;
  47  *
  48  * We probably want to distinguish between between shallow and deep copies so
  49  * that if we memset(p1, 0, sizeof(*p1)) then it just sets p1->x to zero and not
  50  * p1->x->y.
  51  *
  52  * I don't have a proper solution for this problem right now.  I just copy one
  53  * level and don't nest.  It should handle limitted nesting but intelligently.
  54  *
  55  * The other thing is that you end up with a lot of garbage assignments where
  56  * we record "x could be anything. x->y could be anything. x->y->z->a->b->c
  57  * could *also* be anything!".  There should be a better way to filter this
  58  * useless information.
  59  *
  60  */
  61
  62 #include "smatch.h"
  63 #include "smatch_slist.h"
  64 #include "smatch_extra.h"
  65
  66 enum {
  67         COPY_NORMAL,
  68         COPY_UNKNOWN,
  69         COPY_ZERO,
  70 };
  71
  72 static struct symbol *get_struct_type(struct expression *expr)
  73 {
  74         struct symbol *type;
  75
  76         type = get_type(expr);
  77         if (!type)
  78                 return NULL;
  79         if (type->type == SYM_PTR) {
  80                 type = get_real_base_type(type);
  81                 if (!type)
  82                         return NULL;
  83         }
  84         if (type->type == SYM_STRUCT)
  85                 return type;
  86         if (type->type == SYM_UNION)
  87                 return type;
  88         return NULL;
  89 }
  90
  91 static struct expression *get_right_base_expr(struct symbol *left_type, struct expression *right)
  92 {
  93         struct symbol *struct_type;
  94
  95         if (!right)
  96                 return NULL;
  97
  98         struct_type = get_struct_type(right);
  99         if (!struct_type)
 100                 return NULL;
 101         if (struct_type != left_type)
 102                 return NULL;
 103
 104         if (right->type == EXPR_PREOP && right->op == '&')
 105                 right = strip_expr(right->unop);
 106
 107         if (right->type == EXPR_CALL)
 108                 return NULL;
 109
 110         return right;
 111 }
 112
 113 static struct expression *add_dereference(struct expression *expr)
 114 {
 115         struct symbol *type;
 116
 117         /*
 118          * We want to re-write "memcpy(foo, bar, sizeof(*foo));" as
 119          * "*foo = *bar;".
 120          */
 121
 122         expr = strip_expr(expr);
 123         if (!expr)
 124                 return NULL;
 125
 126         if (expr->type == EXPR_PREOP && expr->op == '&')
 127                 return strip_expr(expr->unop);
 128         type = get_type(expr);
 129         if (!type)
 130                 return NULL;
 131         if (type->type != SYM_PTR && type->type != SYM_ARRAY)
 132                 return NULL;
 133
 134         return deref_expression(expr);
 135 }
 136
 137 static struct expression *faked_expression;
 138 struct expression *get_faked_expression(void)
 139 {
 140         if (!__in_fake_assign)
 141                 return NULL;
 142         return faked_expression;
 143 }
 144
 145 static void split_fake_expr(struct expression *expr, void *unused)
 146 {
 147         __in_fake_assign++;
 148         __in_fake_struct_assign++;
 149         __split_expr(expr);
 150         __in_fake_struct_assign--;
 151         __in_fake_assign--;
 152 }
 153
 154 static void handle_non_struct_assignments(struct expression *left, struct expression *right,
 155                         void (*assign_handler)(struct expression *expr, void *data),
 156                         void *data)
 157 {
 158         struct symbol *left_type, *right_type;
 159         struct expression *assign;
 160
 161         while (right && right->type == EXPR_ASSIGNMENT)
 162                 right = strip_parens(right->left);
 163
 164         if (!right)
 165                 right = unknown_value_expression(left);
 166
 167         left_type = get_type(left);
 168         right_type = get_type(right);
 169         if (!left_type || !right_type)
 170                 return;
 171
 172         if (right_type->type == SYM_ARRAY)
 173                 right = unknown_value_expression(left);
 174
 175         if (left_type->type == SYM_PTR) {
 176                 left = deref_expression(left);
 177                 assign = assign_expression(left, '=', right);
 178                 assign_handler(assign, data);
 179                 return;
 180         }
 181         if (left_type->type != SYM_BASETYPE)
 182                 return;
 183         assign = assign_expression(left, '=', right);
 184         assign_handler(assign, data);
 185 }
 186
 187 static void set_inner_struct_members(int mode, struct expression *faked,
 188                 struct expression *left, struct expression *right, struct symbol *member,
 189                 void (*assign_handler)(struct expression *expr, void *data),
 190                 void *data)
 191 {
 192         struct expression *left_member;
 193         struct expression *right_expr = NULL;  /* silence GCC */
 194         struct expression *assign;
 195         struct symbol *base = get_real_base_type(member);
 196         struct symbol *tmp;
 197
 198         if (member->ident) {
 199                 left = member_expression(left, '.', member->ident);
 200                 if (mode == COPY_NORMAL && right)
 201                         right = member_expression(right, '.', member->ident);
 202         }
 203
 204         FOR_EACH_PTR(base->symbol_list, tmp) {
 205                 struct symbol *type;
 206
 207                 type = get_real_base_type(tmp);
 208                 if (!type)
 209                         continue;
 210
 211                 if (type->type == SYM_ARRAY)
 212                         continue;
 213                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 214                         set_inner_struct_members(mode, faked, left, right, tmp, assign_handler, data);
 215                         continue;
 216                 }
 217                 if (!tmp->ident)
 218                         continue;
 219
 220                 left_member = member_expression(left, '.', tmp->ident);
 221
 222                 right_expr = NULL;
 223                 if (mode == COPY_NORMAL && right)
 224                         right_expr = member_expression(right, '.', tmp->ident);
 225                 if (mode == COPY_ZERO)
 226                         right_expr = zero_expr();
 227                 if (!right_expr)
 228                         right_expr = unknown_value_expression(left_member);
 229
 230                 assign = assign_expression(left_member, '=', right_expr);
 231                 assign_handler(assign, data);
 232         } END_FOR_EACH_PTR(tmp);
 233 }
 234
 235 static void __struct_members_copy(int mode, struct expression *faked,
 236                                   struct expression *left,
 237                                   struct expression *right,
 238                                   void (*assign_handler)(struct expression *expr, void *data),
 239                                   void *data)
 240 {
 241         struct symbol *struct_type, *tmp, *type;
 242         struct expression *left_member;
 243         struct expression *right_expr;
 244         struct expression *assign;
 245
 246         if (__in_fake_assign || !left)
 247                 return;
 248         faked_expression = faked;
 249
 250         left = strip_expr(left);
 251         right = strip_expr(right);
 252
 253         if (left->type == EXPR_PREOP && left->op == '*' && is_pointer(left))
 254                 left = preop_expression(left, '(');
 255
 256         struct_type = get_struct_type(left);
 257         if (!struct_type) {
 258                 /*
 259                  * This is not a struct assignment obviously.  But this is where
 260                  * memcpy() is handled so it feels like a good place to add this
 261                  * code.
 262                  */
 263                 handle_non_struct_assignments(left, right, assign_handler, data);
 264                 goto done;
 265         }
 266
 267         if (mode == COPY_NORMAL)
 268                 right = get_right_base_expr(struct_type, right);
 269
 270         FOR_EACH_PTR(struct_type->symbol_list, tmp) {
 271                 type = get_real_base_type(tmp);
 272                 if (!type)
 273                         continue;
 274                 if (type->type == SYM_ARRAY)
 275                         continue;
 276
 277                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 278                         set_inner_struct_members(mode, faked, left, right, tmp, assign_handler, data);
 279                         continue;
 280                 }
 281
 282                 if (!tmp->ident)
 283                         continue;
 284
 285                 left_member = member_expression(left, '.', tmp->ident);
 286                 right_expr = NULL;
 287
 288                 if (mode == COPY_NORMAL && right)
 289                         right_expr = member_expression(right, '.', tmp->ident);
 290                 if (mode == COPY_ZERO)
 291                         right_expr = zero_expr();
 292                 if (!right_expr)
 293                         right_expr = unknown_value_expression(left_member);
 294
 295                 assign = assign_expression(left_member, '=', right_expr);
 296                 assign_handler(assign, data);
 297         } END_FOR_EACH_PTR(tmp);
 298
 299 done:
 300         faked_expression = NULL;
 301 }
 302
 303 static int returns_zeroed_mem(struct expression *expr)
 304 {
 305         struct expression *tmp;
 306         char *fn;
 307
 308         if (expr->type != EXPR_CALL || expr->fn->type != EXPR_SYMBOL)
 309                 return 0;
 310
 311         if (is_fake_call(expr)) {
 312                 tmp = get_faked_expression();
 313                 if (!tmp || tmp->type != EXPR_ASSIGNMENT || tmp->op != '=')
 314                         return 0;
 315                 expr = tmp->right;
 316                 if (expr->type != EXPR_CALL || expr->fn->type != EXPR_SYMBOL)
 317                         return 0;
 318         }
 319
 320         fn = expr_to_var(expr->fn);
 321         if (!fn)
 322                 return 0;
 323         if (strcmp(fn, "kcalloc") == 0)
 324                 return 1;
 325         if (option_project == PROJ_KERNEL && strstr(fn, "zalloc"))
 326                 return 1;
 327         return 0;
 328 }
 329
 330 static int copy_containter_states(struct expression *left, struct expression *right, int offset)
 331 {
 332         char *left_name = NULL, *right_name = NULL;
 333         struct symbol *left_sym, *right_sym;
 334         struct sm_state *sm, *new_sm;
 335         int ret = 0;
 336         int len;
 337         char buf[64];
 338         char new_name[128];
 339
 340         right_name = expr_to_var_sym(right, &right_sym);
 341         if (!right_name || !right_sym)
 342                 goto free;
 343         left_name = expr_to_var_sym(left, &left_sym);
 344         if (!left_name || !left_sym)
 345                 goto free;
 346
 347         len = snprintf(buf, sizeof(buf), "%s(-%d)", right_name, offset);
 348         if (len >= sizeof(buf))
 349                 goto free;
 350
 351         FOR_EACH_SM_SAFE(__get_cur_stree(), sm) {
 352                 if (sm->sym != right_sym)
 353                         continue;
 354                 if (strncmp(sm->name, buf, len) != 0)
 355                         continue;
 356                 snprintf(new_name, sizeof(new_name), "%s%s", left_name, sm->name + len);
 357                 new_sm = clone_sm(sm);
 358                 new_sm->name = alloc_sname(new_name);
 359                 new_sm->sym = left_sym;
 360                 __set_sm(new_sm);
 361                 ret = 1;
 362         } END_FOR_EACH_SM_SAFE(sm);
 363 free:
 364         free_string(left_name);
 365         free_string(right_name);
 366         return ret;
 367 }
 368
 369 static int handle_param_offsets(struct expression *expr)
 370 {
 371         struct expression *right;
 372         sval_t sval;
 373
 374         right = strip_expr(expr->right);
 375
 376         if (right->type != EXPR_BINOP || right->op != '-')
 377                 return 0;
 378
 379         if (!get_value(right->right, &sval))
 380                 return 0;
 381
 382         right = get_assigned_expr(right->left);
 383         if (!right)
 384                 return 0;
 385         return copy_containter_states(expr->left, right, sval.value);
 386 }
 387
 388 static void returns_container_of(struct expression *expr, int param, char *key, char *value)
 389 {
 390         struct expression *call, *arg;
 391         int offset;
 392
 393         if (expr->type != EXPR_ASSIGNMENT || expr->op != '=')
 394                 return;
 395         call = strip_expr(expr->right);
 396         if (call->type != EXPR_CALL)
 397                 return;
 398         if (param != -1)
 399                 return;
 400         param = atoi(key);
 401         offset = atoi(value);
 402
 403         arg = get_argument_from_call_expr(call->args, param);
 404         if (!arg)
 405                 return;
 406
 407         copy_containter_states(expr->left, arg, -offset);
 408 }
 409
 410 void __fake_struct_member_assignments(struct expression *expr)
 411 {
 412         struct expression *left, *right;
 413         struct symbol *type;
 414         int mode;
 415
 416         if (expr->op != '=')
 417                 return;
 418
 419         if (is_noderef_ptr(expr->right))
 420                 return;
 421
 422         type = get_type(expr->left);
 423         if (!type)
 424                 return;
 425         if (type->type != SYM_PTR && type->type != SYM_STRUCT)
 426                 return;
 427
 428         if (handle_param_offsets(expr))
 429                 return;
 430
 431         left = expr->left;
 432         right = expr->right;
 433
 434         if (type->type == SYM_PTR) {
 435                 /* Convert "p = q;" to "*p = *q;" */
 436                 left = add_dereference(left);
 437                 right = add_dereference(right);
 438         }
 439
 440         if (returns_zeroed_mem(expr->right))
 441                 mode = COPY_ZERO;
 442         else if (types_equiv(get_type(left), get_type(right)))
 443                 mode = COPY_NORMAL;
 444         else
 445                 mode = COPY_UNKNOWN;
 446
 447         __struct_members_copy(mode, expr, left, right, split_fake_expr, NULL);
 448 }
 449
 450 static void match_memset(const char *fn, struct expression *expr, void *_size_arg)
 451 {
 452         struct expression *buf;
 453         struct expression *val;
 454         int mode;
 455
 456         buf = get_argument_from_call_expr(expr->args, 0);
 457         val = get_argument_from_call_expr(expr->args, 1);
 458
 459         buf = strip_expr(buf);
 460         if (expr_is_zero(val))
 461                 mode = COPY_ZERO;
 462         else
 463                 mode = COPY_UNKNOWN;
 464         __struct_members_copy(mode, expr, add_dereference(buf), NULL, split_fake_expr, NULL);
 465 }
 466
 467 static void match_memcpy(const char *fn, struct expression *expr, void *_arg)
 468 {
 469         struct expression *dest;
 470         struct expression *src;
 471         int mode;
 472
 473         dest = get_argument_from_call_expr(expr->args, 0);
 474         src = get_argument_from_call_expr(expr->args, 1);
 475
 476         if (types_equiv(get_type(src), get_type(dest)))
 477                 mode = COPY_NORMAL;
 478         else
 479                 mode = COPY_UNKNOWN;
 480
 481         __struct_members_copy(mode, expr, add_dereference(dest), add_dereference(src), split_fake_expr, NULL);
 482 }
 483
 484 static void match_memdup(const char *fn, struct expression *call_expr,
 485                         struct expression *expr, void *_unused)
 486 {
 487         struct expression *left, *right, *arg;
 488         int mode;
 489
 490         if (!expr || expr->type != EXPR_ASSIGNMENT)
 491                 return;
 492
 493         left = strip_expr(expr->left);
 494         right = strip_expr(expr->right);
 495
 496         if (right->type != EXPR_CALL)
 497                 return;
 498         arg = get_argument_from_call_expr(right->args, 0);
 499
 500         if (types_equiv(get_type(left), get_type(right)))
 501                 mode = COPY_NORMAL;
 502         else
 503                 mode = COPY_UNKNOWN;
 504
 505         __struct_members_copy(mode, expr, add_dereference(left), add_dereference(arg), split_fake_expr, NULL);
 506 }
 507
 508 static void match_memcpy_unknown(const char *fn, struct expression *expr, void *_arg)
 509 {
 510         struct expression *dest;
 511
 512         dest = get_argument_from_call_expr(expr->args, 0);
 513         __struct_members_copy(COPY_UNKNOWN, expr, add_dereference(dest), NULL, split_fake_expr, NULL);
 514 }
 515
 516 static void match_sscanf(const char *fn, struct expression *expr, void *unused)
 517 {
 518         struct expression *arg;
 519         int i;
 520
 521         i = -1;
 522         FOR_EACH_PTR(expr->args, arg) {
 523                 if (++i < 2)
 524                         continue;
 525                 __struct_members_copy(COPY_UNKNOWN, expr, add_dereference(arg), NULL, split_fake_expr, NULL);
 526         } END_FOR_EACH_PTR(arg);
 527 }
 528
 529 static void unop_expr(struct expression *expr)
 530 {
 531         if (expr->op != SPECIAL_INCREMENT &&
 532             expr->op != SPECIAL_DECREMENT)
 533                 return;
 534
 535         if (!is_pointer(expr))
 536                 return;
 537         faked_expression = expr;
 538         __struct_members_copy(COPY_UNKNOWN, expr, expr->unop, NULL, split_fake_expr, NULL);
 539         faked_expression = NULL;
 540 }
 541
 542 static void register_clears_param(void)
 543 {
 544         struct token *token;
 545         char name[256];
 546         const char *function;
 547         int param;
 548
 549         if (option_project == PROJ_NONE)
 550                 return;
 551
 552         snprintf(name, 256, "%s.clears_argument", option_project_str);
 553
 554         token = get_tokens_file(name);
 555         if (!token)
 556                 return;
 557         if (token_type(token) != TOKEN_STREAMBEGIN)
 558                 return;
 559         token = token->next;
 560         while (token_type(token) != TOKEN_STREAMEND) {
 561                 if (token_type(token) != TOKEN_IDENT)
 562                         return;
 563                 function = show_ident(token->ident);
 564                 token = token->next;
 565                 if (token_type(token) != TOKEN_NUMBER)
 566                         return;
 567                 param = atoi(token->number);
 568                 add_function_hook(function, &match_memcpy_unknown, INT_PTR(param));
 569                 token = token->next;
 570         }
 571         clear_token_alloc();
 572 }
 573
 574 void create_recursive_fake_assignments(struct expression *expr,
 575                 void (*assign_handler)(struct expression *expr, void *data),
 576                 void *data)
 577 {
 578         __struct_members_copy(COPY_UNKNOWN, expr, expr, NULL, assign_handler, data);
 579 }
 580
 581 static void db_buf_cleared(struct expression *expr, int param, char *key, char *value)
 582 {
 583         struct expression *arg;
 584
 585         arg = gen_expr_from_param_key(expr, param, key);
 586         if (!arg)
 587                 return;
 588
 589         __in_buf_clear++;
 590         if (strcmp(value, "0") == 0)
 591                 __struct_members_copy(COPY_ZERO, expr, arg, NULL, split_fake_expr, NULL);
 592         else
 593                 __struct_members_copy(COPY_UNKNOWN, expr, arg, NULL, split_fake_expr, NULL);
 594         __in_buf_clear--;
 595 }
 596
 597 static void db_param_add_set(struct expression *expr, int param, char *key, char *value)
 598 {
 599         struct expression *arg;
 600         struct symbol *type;
 601
 602         arg = gen_expr_from_param_key(expr, param, key);
 603         if (!arg)
 604                 return;
 605         type = get_type(arg);
 606         if (!type)
 607                 return;
 608         if (type->type != SYM_STRUCT && type->type != SYM_UNION)
 609                 return;
 610
 611         __struct_members_copy(COPY_UNKNOWN, expr, arg, NULL, split_fake_expr, NULL);
 612 }
 613
 614 void register_struct_assignment(int id)
 615 {
 616         add_function_data((unsigned long *)&faked_expression);
 617
 618         add_function_hook("memset", &match_memset, NULL);
 619         add_function_hook("__memset", &match_memset, NULL);
 620         add_function_hook("__builtin_memset", &match_memset, NULL);
 621
 622         add_function_hook("memcpy", &match_memcpy, INT_PTR(0));
 623         add_function_hook("memmove", &match_memcpy, INT_PTR(0));
 624         add_function_hook("__memcpy", &match_memcpy, INT_PTR(0));
 625         add_function_hook("__memmove", &match_memcpy, INT_PTR(0));
 626         add_function_hook("__builtin_memcpy", &match_memcpy, INT_PTR(0));
 627         add_function_hook("__builtin_memmove", &match_memcpy, INT_PTR(0));
 628
 629         if (option_project == PROJ_KERNEL) {
 630                 return_implies_state_sval("kmemdup", valid_ptr_min_sval, valid_ptr_max_sval, &match_memdup, NULL);
 631                 add_function_hook("copy_from_user", &match_memcpy, INT_PTR(0));
 632                 add_function_hook("_copy_from_user", &match_memcpy, INT_PTR(0));
 633         }
 634
 635         add_function_hook("sscanf", &match_sscanf, NULL);
 636
 637         add_hook(&unop_expr, OP_HOOK);
 638         register_clears_param();
 639         select_return_states_hook(BUF_CLEARED, &db_buf_cleared);
 640         select_return_states_hook(PARAM_ADD, &db_param_add_set);
 641         select_return_states_hook(PARAM_SET, &db_param_add_set);
 642
 643         select_return_states_hook(CONTAINER, &returns_container_of);
 644 }