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  * I don't have a proper solution for this problem right now.  I just copy one
  49  * level and don't nest.  It should handle limitted nesting but intelligently.
  50  *
  51  * The other thing is that you end up with a lot of garbage assignments where
  52  * we record "x could be anything. x->y could be anything. x->y->z->a->b->c
  53  * could *also* be anything!".  There should be a better way to filter this
  54  * useless information.
  55  *
  56  */
  57
  58 #include "scope.h"
  59 #include "smatch.h"
  60 #include "smatch_slist.h"
  61 #include "smatch_extra.h"
  62
  63 enum {
  64         COPY_NORMAL,
  65         COPY_MEMCPY,
  66         COPY_MEMSET,
  67 };
  68
  69 static struct symbol *get_struct_type(struct expression *expr)
  70 {
  71         struct symbol *type;
  72
  73         type = get_type(expr);
  74         if (!type)
  75                 return NULL;
  76         if (type->type == SYM_PTR) {
  77                 type = get_real_base_type(type);
  78                 if (!type)
  79                         return NULL;
  80         }
  81         if (type->type == SYM_STRUCT)
  82                 return type;
  83         if (type->type == SYM_UNION)
  84                 return type;
  85         return NULL;
  86 }
  87
  88 static struct expression *get_right_base_expr(struct symbol *left_type, struct expression *right)
  89 {
  90         struct symbol *struct_type;
  91
  92         if (!right)
  93                 return NULL;
  94
  95         struct_type = get_struct_type(right);
  96         if (!struct_type)
  97                 return NULL;
  98         if (struct_type != left_type)
  99                 return NULL;
 100
 101         if (right->type == EXPR_PREOP && right->op == '&')
 102                 right = strip_expr(right->unop);
 103
 104         if (right->type == EXPR_CALL)
 105                 return NULL;
 106
 107         if (is_pointer(right))
 108                 right = deref_expression(right);
 109
 110         return right;
 111 }
 112
 113 static struct expression *remove_addr(struct expression *expr)
 114 {
 115         struct symbol *type;
 116
 117         expr = strip_expr(expr);
 118         if (!expr)
 119                 return NULL;
 120
 121         if (expr->type == EXPR_PREOP && expr->op == '&')
 122                 return strip_expr(expr->unop);
 123         type = get_type(expr);
 124         if (!type)
 125                 return expr;
 126         if (type->type != SYM_PTR && type->type != SYM_ARRAY)
 127                 return expr;
 128
 129         return deref_expression(expr);
 130 }
 131
 132 static struct expression *faked_expression;
 133 struct expression *get_faked_expression(void)
 134 {
 135         if (!__in_fake_assign)
 136                 return NULL;
 137         return faked_expression;
 138 }
 139
 140 static void split_fake_expr(struct expression *expr)
 141 {
 142         __in_fake_assign++;
 143         __in_fake_struct_assign++;
 144         __split_expr(expr);
 145         __in_fake_struct_assign--;
 146         __in_fake_assign--;
 147 }
 148
 149 static void handle_non_struct_assignments(struct expression *left, struct expression *right)
 150 {
 151         struct symbol *type;
 152         struct expression *assign;
 153
 154         type = get_type(left);
 155         if (!type)
 156                 return;
 157         if (type->type == SYM_PTR) {
 158                 left = deref_expression(left);
 159                 if (right)
 160                         right = deref_expression(right);
 161                 else
 162                         right = unknown_value_expression(left);
 163                 assign = assign_expression(left, '=', right);
 164                 split_fake_expr(assign);
 165                 return;
 166         }
 167         if (type->type != SYM_BASETYPE)
 168                 return;
 169         right = strip_expr(right);
 170         if (!right)
 171                 right = unknown_value_expression(left);
 172         assign = assign_expression(left, '=', right);
 173         split_fake_expr(assign);
 174 }
 175
 176 static void set_inner_struct_members(int mode, struct expression *faked, struct expression *left, struct expression *right, struct symbol *member)
 177 {
 178         struct expression *left_member;
 179         struct expression *right_member = NULL;  /* silence GCC */
 180         struct expression *assign;
 181         struct symbol *base = get_real_base_type(member);
 182         struct symbol *tmp;
 183
 184         if (member->ident) {
 185                 left = member_expression(left, '.', member->ident);
 186                 if (mode != COPY_MEMSET && right)
 187                         right = member_expression(right, '.', member->ident);
 188         }
 189
 190         FOR_EACH_PTR(base->symbol_list, tmp) {
 191                 struct symbol *type;
 192
 193                 type = get_real_base_type(tmp);
 194                 if (!type)
 195                         continue;
 196
 197                 if (type->type == SYM_ARRAY)
 198                         continue;
 199                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 200                         set_inner_struct_members(mode, faked, left, right, tmp);
 201                         continue;
 202                 }
 203                 if (!tmp->ident)
 204                         continue;
 205
 206                 left_member = member_expression(left, '.', tmp->ident);
 207
 208                 switch (mode) {
 209                 case COPY_NORMAL:
 210                 case COPY_MEMCPY:
 211                         if (right)
 212                                 right_member = member_expression(right, '.', tmp->ident);
 213                         else
 214                                 right_member = unknown_value_expression(left_member);
 215                         break;
 216                 case COPY_MEMSET:
 217                         right_member = right;
 218                         break;
 219                 }
 220
 221                 assign = assign_expression(left_member, '=', right_member);
 222                 split_fake_expr(assign);
 223         } END_FOR_EACH_PTR(tmp);
 224 }
 225
 226 static void __struct_members_copy(int mode, struct expression *faked,
 227                                   struct expression *left,
 228                                   struct expression *right)
 229 {
 230         struct symbol *struct_type, *tmp, *type;
 231         struct expression *left_member;
 232         struct expression *right_member;
 233         struct expression *assign;
 234         int op = '.';
 235
 236
 237         if (__in_fake_assign)
 238                 return;
 239         faked_expression = faked;
 240
 241         left = strip_expr(left);
 242         right = strip_expr(right);
 243
 244         struct_type = get_struct_type(left);
 245         if (!struct_type) {
 246                 /*
 247                  * This is not a struct assignment obviously.  But this is where
 248                  * memcpy() is handled so it feels like a good place to add this
 249                  * code.
 250                  */
 251                 handle_non_struct_assignments(left, right);
 252                 goto done;
 253         }
 254
 255         if (is_pointer(left)) {
 256                 left = deref_expression(left);
 257                 op = '*';
 258         }
 259         if (mode != COPY_MEMSET)
 260                 right = get_right_base_expr(struct_type, right);
 261
 262         FOR_EACH_PTR(struct_type->symbol_list, tmp) {
 263                 type = get_real_base_type(tmp);
 264                 if (!type)
 265                         continue;
 266                 if (type->type == SYM_ARRAY)
 267                         continue;
 268
 269                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 270                         set_inner_struct_members(mode, faked, left, right, tmp);
 271                         continue;
 272                 }
 273
 274                 if (!tmp->ident)
 275                         continue;
 276
 277                 left_member = member_expression(left, op, tmp->ident);
 278                 right_member = NULL;
 279
 280                 switch (mode) {
 281                 case COPY_NORMAL:
 282                 case COPY_MEMCPY:
 283                         if (right)
 284                                 right_member = member_expression(right, op, tmp->ident);
 285                         else
 286                                 right_member = unknown_value_expression(left_member);
 287                         break;
 288                 case COPY_MEMSET:
 289                         right_member = right;
 290                         break;
 291                 }
 292                 if (!right_member) {
 293                         sm_perror("No right member");
 294                         continue;
 295                 }
 296                 assign = assign_expression(left_member, '=', right_member);
 297                 split_fake_expr(assign);
 298         } END_FOR_EACH_PTR(tmp);
 299
 300 done:
 301         faked_expression = NULL;
 302 }
 303
 304 static int returns_zeroed_mem(struct expression *expr)
 305 {
 306         char *fn;
 307
 308         if (expr->type != EXPR_CALL || expr->fn->type != EXPR_SYMBOL)
 309                 return 0;
 310         fn = expr_to_var(expr->fn);
 311         if (!fn)
 312                 return 0;
 313         if (strcmp(fn, "kcalloc") == 0)
 314                 return 1;
 315         if (option_project == PROJ_KERNEL && strstr(fn, "zalloc"))
 316                 return 1;
 317         return 0;
 318 }
 319
 320 static int copy_containter_states(struct expression *left, struct expression *right, int offset)
 321 {
 322         char *left_name = NULL, *right_name = NULL;
 323         struct symbol *left_sym, *right_sym;
 324         struct sm_state *sm, *new_sm;
 325         int ret = 0;
 326         int len;
 327         char buf[64];
 328         char new_name[128];
 329
 330         right_name = expr_to_var_sym(right, &right_sym);
 331         if (!right_name || !right_sym)
 332                 goto free;
 333         left_name = expr_to_var_sym(left, &left_sym);
 334         if (!left_name || !left_sym)
 335                 goto free;
 336
 337         len = snprintf(buf, sizeof(buf), "%s(-%d)", right_name, offset);
 338         if (len >= sizeof(buf))
 339                 goto free;
 340
 341         FOR_EACH_SM(__get_cur_stree(), sm) {
 342                 if (sm->sym != right_sym)
 343                         continue;
 344                 if (strncmp(sm->name, buf, len) != 0)
 345                         continue;
 346                 snprintf(new_name, sizeof(new_name), "%s%s", left_name, sm->name + len);
 347                 new_sm = clone_sm(sm);
 348                 new_sm->name = alloc_sname(new_name);
 349                 new_sm->sym = left_sym;
 350                 __set_sm(new_sm);
 351                 ret = 1;
 352         } END_FOR_EACH_SM(sm);
 353 free:
 354         free_string(left_name);
 355         free_string(right_name);
 356         return ret;
 357 }
 358
 359 static int handle_param_offsets(struct expression *expr)
 360 {
 361         struct expression *right;
 362         sval_t sval;
 363
 364         right = strip_expr(expr->right);
 365
 366         if (right->type != EXPR_BINOP || right->op != '-')
 367                 return 0;
 368
 369         if (!get_value(right->right, &sval))
 370                 return 0;
 371
 372         right = get_assigned_expr(right->left);
 373         if (!right)
 374                 return 0;
 375         return copy_containter_states(expr->left, right, sval.value);
 376 }
 377
 378 static void returns_container_of(struct expression *expr, int param, char *key, char *value)
 379 {
 380         struct expression *call, *arg;
 381         int offset;
 382
 383         if (expr->type != EXPR_ASSIGNMENT || expr->op != '=')
 384                 return;
 385         call = strip_expr(expr->right);
 386         if (call->type != EXPR_CALL)
 387                 return;
 388         if (param != -1)
 389                 return;
 390         param = atoi(key);
 391         offset = atoi(value);
 392
 393         arg = get_argument_from_call_expr(call->args, param);
 394         if (!arg)
 395                 return;
 396
 397         copy_containter_states(expr->left, arg, -offset);
 398 }
 399
 400 void __fake_struct_member_assignments(struct expression *expr)
 401 {
 402         struct symbol *left_type;
 403
 404         if (expr->op != '=')
 405                 return;
 406
 407         if (is_zero(expr->right))
 408                 return;
 409
 410         left_type = get_type(expr->left);
 411         if (!left_type ||
 412             (left_type->type != SYM_PTR &&
 413              left_type->type != SYM_STRUCT))
 414                 return;
 415
 416         if (handle_param_offsets(expr))
 417                 return;
 418
 419         if (returns_zeroed_mem(expr->right))
 420                 __struct_members_copy(COPY_MEMSET, expr, expr->left, zero_expr());
 421         else
 422                 __struct_members_copy(COPY_NORMAL, expr, expr->left, expr->right);
 423 }
 424
 425 static void match_memset(const char *fn, struct expression *expr, void *_size_arg)
 426 {
 427         struct expression *buf;
 428         struct expression *val;
 429
 430         buf = get_argument_from_call_expr(expr->args, 0);
 431         val = get_argument_from_call_expr(expr->args, 1);
 432
 433         buf = strip_expr(buf);
 434         __struct_members_copy(COPY_MEMSET, expr, remove_addr(buf), val);
 435 }
 436
 437 static void match_memcpy(const char *fn, struct expression *expr, void *_arg)
 438 {
 439         struct expression *dest;
 440         struct expression *src;
 441
 442         dest = get_argument_from_call_expr(expr->args, 0);
 443         src = get_argument_from_call_expr(expr->args, 1);
 444
 445         __struct_members_copy(COPY_MEMCPY, expr, remove_addr(dest), remove_addr(src));
 446 }
 447
 448 static void match_memcpy_unknown(const char *fn, struct expression *expr, void *_arg)
 449 {
 450         struct expression *dest;
 451
 452         dest = get_argument_from_call_expr(expr->args, 0);
 453         __struct_members_copy(COPY_MEMCPY, expr, remove_addr(dest), NULL);
 454 }
 455
 456 static void match_sscanf(const char *fn, struct expression *expr, void *unused)
 457 {
 458         struct expression *arg;
 459         int i;
 460
 461         i = -1;
 462         FOR_EACH_PTR(expr->args, arg) {
 463                 if (++i < 2)
 464                         continue;
 465                 __struct_members_copy(COPY_MEMCPY, expr, remove_addr(arg), NULL);
 466         } END_FOR_EACH_PTR(arg);
 467 }
 468
 469 static void unop_expr(struct expression *expr)
 470 {
 471         if (expr->op != SPECIAL_INCREMENT &&
 472             expr->op != SPECIAL_DECREMENT)
 473                 return;
 474
 475         if (!is_pointer(expr))
 476                 return;
 477         faked_expression = expr;
 478         __struct_members_copy(COPY_MEMCPY, expr, expr->unop, NULL);
 479         faked_expression = NULL;
 480 }
 481
 482 static void register_clears_param(void)
 483 {
 484         struct token *token;
 485         char name[256];
 486         const char *function;
 487         int param;
 488
 489         if (option_project == PROJ_NONE)
 490                 return;
 491
 492         snprintf(name, 256, "%s.clears_argument", option_project_str);
 493
 494         token = get_tokens_file(name);
 495         if (!token)
 496                 return;
 497         if (token_type(token) != TOKEN_STREAMBEGIN)
 498                 return;
 499         token = token->next;
 500         while (token_type(token) != TOKEN_STREAMEND) {
 501                 if (token_type(token) != TOKEN_IDENT)
 502                         return;
 503                 function = show_ident(token->ident);
 504                 token = token->next;
 505                 if (token_type(token) != TOKEN_NUMBER)
 506                         return;
 507                 param = atoi(token->number);
 508                 add_function_hook(function, &match_memcpy_unknown, INT_PTR(param));
 509                 token = token->next;
 510         }
 511         clear_token_alloc();
 512 }
 513
 514 static void db_param_cleared(struct expression *expr, int param, char *key, char *value)
 515 {
 516         struct expression *arg;
 517
 518         while (expr->type == EXPR_ASSIGNMENT)
 519                 expr = strip_expr(expr->right);
 520         if (expr->type != EXPR_CALL)
 521                 return;
 522
 523         /*
 524          * FIXME:  __struct_members_copy() requires an expression but
 525          * get_variable_from_key() returns a name/sym pair so that doesn't
 526          * work here.
 527          */
 528         if (strcmp(key, "$") != 0)
 529                 return;
 530
 531         arg = get_argument_from_call_expr(expr->args, param);
 532         if (!arg)
 533                 return;
 534
 535         if (strcmp(value, "0") == 0)
 536                 __struct_members_copy(COPY_MEMSET, expr, remove_addr(arg), zero_expr());
 537         else
 538                 __struct_members_copy(COPY_MEMCPY, expr, remove_addr(arg), NULL);
 539 }
 540
 541 void register_struct_assignment(int id)
 542 {
 543         add_function_hook("memset", &match_memset, NULL);
 544         add_function_hook("__memset", &match_memset, NULL);
 545
 546         add_function_hook("memcpy", &match_memcpy, INT_PTR(0));
 547         add_function_hook("memmove", &match_memcpy, INT_PTR(0));
 548         add_function_hook("__memcpy", &match_memcpy, INT_PTR(0));
 549         add_function_hook("__memmove", &match_memcpy, INT_PTR(0));
 550
 551         add_function_hook("sscanf", &match_sscanf, NULL);
 552
 553         add_hook(&unop_expr, OP_HOOK);
 554         register_clears_param();
 555         select_return_states_hook(PARAM_CLEARED, &db_param_cleared);
 556
 557         select_return_states_hook(CONTAINER, &returns_container_of);
 558 }