check_uninitialized.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 #include "smatch.h"
  19 #include "smatch_slist.h"
  20 #include "smatch_extra.h"
  21
  22 static int my_id;
  23
  24 STATE(uninitialized);
  25 STATE(initialized);
  26
  27 static bool uncertain_code_path(void)
  28 {
  29         if (implications_off || parse_error)
  30                 return true;
  31         if (is_impossible_path())
  32                 return true;
  33
  34         return false;
  35 }
  36
  37 static void pre_merge_hook(struct sm_state *cur, struct sm_state *other)
  38 {
  39         if (is_impossible_path())
  40                 set_state(my_id, cur->name, cur->sym, &initialized);
  41 }
  42
  43 static void mark_members_uninitialized(struct symbol *sym)
  44 {
  45         struct symbol *struct_type, *tmp, *base_type;
  46         char buf[256];
  47
  48         struct_type = get_real_base_type(sym);
  49         FOR_EACH_PTR(struct_type->symbol_list, tmp) {
  50                 if (!tmp->ident)
  51                         continue;
  52                 base_type = get_real_base_type(tmp);
  53                 if (!base_type ||
  54                     base_type->type == SYM_STRUCT ||
  55                     base_type->type == SYM_ARRAY ||
  56                     base_type->type == SYM_UNION)
  57                         continue;
  58                 snprintf(buf, sizeof(buf), "%s.%s", sym->ident->name, tmp->ident->name);
  59                 set_state(my_id, buf, sym, &uninitialized);
  60         } END_FOR_EACH_PTR(tmp);
  61 }
  62
  63 static void match_declarations(struct symbol *sym)
  64 {
  65         struct symbol *type;
  66
  67         if (!cur_func_sym)
  68                 return;
  69
  70         if (sym->initializer)
  71                 return;
  72
  73         type = get_real_base_type(sym);
  74         /* Smatch is crap at tracking arrays */
  75         if (type->type == SYM_ARRAY)
  76                 return;
  77         if (type->type == SYM_UNION)
  78                 return;
  79         if (sym->ctype.modifiers & MOD_STATIC)
  80                 return;
  81
  82         if (!sym->ident)
  83                 return;
  84
  85         if (type->type == SYM_STRUCT) {
  86                 mark_members_uninitialized(sym);
  87                 return;
  88         }
  89
  90         set_state(my_id, sym->ident->name, sym, &uninitialized);
  91 }
  92
  93 static int is_initialized(struct expression *expr)
  94 {
  95         struct sm_state *sm;
  96
  97         expr = strip_expr(expr);
  98         if (expr->type != EXPR_SYMBOL)
  99                 return 1;
 100         sm = get_sm_state_expr(my_id, expr);
 101         if (!sm)
 102                 return 1;
 103         if (!slist_has_state(sm->possible, &uninitialized))
 104                 return 1;
 105         return 0;
 106 }
 107
 108 static void warn_about_special_assign(struct expression *expr)
 109 {
 110         char *name;
 111
 112         if (!expr || expr->type != EXPR_ASSIGNMENT || expr->op == '=')
 113                 return;
 114
 115         if (uncertain_code_path())
 116                 return;
 117
 118         if (is_initialized(expr->left))
 119                 return;
 120
 121         name = expr_to_str(expr->left);
 122         sm_warning("uninitialized special assign '%s'", name);
 123         free_string(name);
 124 }
 125
 126 static void extra_mod_hook(const char *name, struct symbol *sym, struct expression *expr, struct smatch_state *state)
 127 {
 128         struct expression *parent = expr_get_parent_expr(expr);
 129
 130         if (!cur_func_sym)
 131                 return;
 132
 133         if (__in_fake_struct_assign && parent &&
 134             parent->type == EXPR_ASSIGNMENT &&
 135             is_fake_call(parent->right))
 136                 return;
 137         if (expr && expr->smatch_flags & Fake)
 138                 return;
 139         if (!sym || !sym->ident)
 140                 return;
 141         if (strcmp(name, sym->ident->name) != 0)
 142                 return;
 143         warn_about_special_assign(expr);
 144         set_state(my_id, name, sym, &initialized);
 145 }
 146
 147 static void match_assign(struct expression *expr)
 148 {
 149         struct expression *right;
 150
 151         if (is_fake_var_assign(expr))
 152                 return;
 153
 154         right = strip_expr(expr->right);
 155         if (right->type == EXPR_PREOP && right->op == '&')
 156                 set_state_expr(my_id, right->unop, &initialized);
 157 }
 158
 159 static void match_negative_comparison(struct expression *expr)
 160 {
 161         struct expression *success;
 162         struct sm_state *sm;
 163         sval_t max;
 164
 165         /*
 166          * In the kernel, people don't use "if (ret) {" and "if (ret < 0) {"
 167          * consistently.  Ideally Smatch would know the return but often it
 168          * doesn't.
 169          *
 170          */
 171
 172         if (option_project != PROJ_KERNEL)
 173                 return;
 174
 175         if (expr->type != EXPR_COMPARE || expr->op != '<')
 176                 return;
 177         if (!expr_is_zero(expr->right))
 178                 return;
 179         if (get_implied_max(expr->left, &max) && max.value == 0)
 180                 return;
 181
 182         success = compare_expression(expr->left, SPECIAL_EQUAL, expr->right);
 183         if (!assume(success))
 184                 return;
 185
 186         FOR_EACH_MY_SM(my_id, __get_cur_stree(), sm) {
 187                 if (sm->state == &initialized)
 188                         set_true_false_states(my_id, sm->name, sm->sym, NULL, &initialized);
 189         } END_FOR_EACH_SM(sm);
 190
 191         end_assume();
 192 }
 193
 194 static struct statement *clear_states;
 195 static void match_enum_switch(struct statement *stmt)
 196 {
 197         struct expression *expr;
 198         struct symbol *type;
 199
 200         if (stmt->type != STMT_COMPOUND)
 201                 return;
 202         stmt = stmt_get_parent_stmt(stmt);
 203         if (!stmt || stmt->type != STMT_SWITCH)
 204                 return;
 205
 206         /* This ended up way uglier than I imagined */
 207         if (__has_default_case())
 208                 return;
 209
 210         expr = strip_expr(stmt->switch_expression);
 211         type = expr->ctype;
 212         if (!type || type->type != SYM_ENUM)
 213                 return;
 214
 215         clear_states = stmt;
 216 }
 217
 218 static void match_enum_switch_after(struct statement *stmt)
 219 {
 220         struct sm_state *sm;
 221
 222         if (clear_states != stmt)
 223                 return;
 224
 225         FOR_EACH_MY_SM(my_id, __get_cur_stree(), sm) {
 226                 if (sm->state == &merged)
 227                         set_state(my_id, sm->name, sm->sym, &initialized);
 228         } END_FOR_EACH_SM(sm);
 229 }
 230
 231 static void match_dereferences(struct expression *expr)
 232 {
 233         char *name;
 234
 235         if (uncertain_code_path())
 236                 return;
 237
 238         if (expr->type != EXPR_PREOP)
 239                 return;
 240         if (is_initialized(expr->unop))
 241                 return;
 242
 243         name = expr_to_str(expr->unop);
 244         sm_error("potentially dereferencing uninitialized '%s'.", name);
 245         free_string(name);
 246
 247         set_state_expr(my_id, expr->unop, &initialized);
 248 }
 249
 250 static void match_condition(struct expression *expr)
 251 {
 252         char *name;
 253
 254         if (uncertain_code_path())
 255                 return;
 256
 257         if (is_initialized(expr))
 258                 return;
 259
 260         name = expr_to_str(expr);
 261         sm_error("potentially using uninitialized '%s'.", name);
 262         free_string(name);
 263
 264         set_state_expr(my_id, expr, &initialized);
 265 }
 266
 267 static void match_function_pointer_call(struct expression *expr)
 268 {
 269         struct expression *parent, *arg, *tmp;
 270
 271         /*
 272          * If you call a function pointer foo->read(&val) without checking
 273          * for errors then you knew what you were doing when you wrote that
 274          * code.  I'm not the police to try to prevent intentional bugs.
 275          *
 276          */
 277         parent = expr_get_parent_expr(expr);
 278         if (parent)
 279                 return;
 280         if (expr->fn->type == EXPR_SYMBOL)
 281                 return;
 282
 283         FOR_EACH_PTR(expr->args, arg) {
 284                 tmp = strip_expr(arg);
 285                 if (tmp->type != EXPR_PREOP || tmp->op != '&')
 286                         continue;
 287                 set_state_expr(my_id, tmp->unop, &initialized);
 288         } END_FOR_EACH_PTR(arg);
 289 }
 290
 291 static void match_call(struct expression *expr)
 292 {
 293         struct expression *arg;
 294         char *name;
 295
 296         if (uncertain_code_path())
 297                 return;
 298
 299         FOR_EACH_PTR(expr->args, arg) {
 300                 if (is_initialized(arg))
 301                         continue;
 302
 303                 name = expr_to_str(arg);
 304                 sm_warning("passing uninitialized '%s'", name);
 305                 free_string(name);
 306
 307                 set_state_expr(my_id, arg, &initialized);
 308         } END_FOR_EACH_PTR(arg);
 309 }
 310
 311 static int param_used_callback(void *found, int argc, char **argv, char **azColName)
 312 {
 313         *(int *)found = 1;
 314         return 0;
 315 }
 316
 317 static int member_is_used(struct expression *call, int param, char *printed_name)
 318 {
 319         int found;
 320
 321         /* for function pointers assume everything is used */
 322         if (call->fn->type != EXPR_SYMBOL)
 323                 return 0;
 324
 325         found = 0;
 326         run_sql(&param_used_callback, &found,
 327                 "select * from return_implies where %s and type = %d and parameter = %d and key = '%s';",
 328                 get_static_filter(call->fn->symbol), PARAM_USED, param, printed_name);
 329         return found;
 330 }
 331
 332 static void match_call_struct_members(struct expression *expr)
 333 {
 334         struct symbol *type, *sym;
 335         struct expression *arg;
 336         struct sm_state *sm;
 337         char *arg_name;
 338         char buf[256];
 339         int param;
 340
 341         return;
 342
 343         if (parse_error)
 344                 return;
 345
 346         param = -1;
 347         FOR_EACH_PTR(expr->args, arg) {
 348                 param++;
 349                 if (arg->type != EXPR_PREOP || arg->op != '&')
 350                         continue;
 351                 type = get_type(arg->unop);
 352                 if (!type || type->type != SYM_STRUCT)
 353                         continue;
 354                 arg_name = expr_to_var_sym(arg->unop, &sym);
 355                 if (!arg_name || !sym)
 356                         goto free;
 357                 FOR_EACH_MY_SM(my_id, __get_cur_stree(), sm) {
 358                         if (sm->sym != sym)
 359                                 continue;
 360                         if (!slist_has_state(sm->possible, &uninitialized))
 361                                 continue;
 362                         snprintf(buf, sizeof(buf), "$->%s", sm->name + strlen(arg_name) + 1);
 363                         if (!member_is_used(expr, param, buf))
 364                                 goto free;
 365                         sm_warning("struct member %s is uninitialized", sm->name);
 366                 } END_FOR_EACH_SM(sm);
 367
 368 free:
 369                 free_string(arg_name);
 370         } END_FOR_EACH_PTR(arg);
 371 }
 372
 373 static int is_being_modified(struct expression *expr)
 374 {
 375         struct expression *parent;
 376         struct statement *stmt;
 377
 378         parent = expr_get_parent_expr(expr);
 379         if (!parent)
 380                 return 0;
 381         while (parent->type == EXPR_PREOP && parent->op == '(') {
 382                 parent = expr_get_parent_expr(parent);
 383                 if (!parent)
 384                         return 0;
 385         }
 386         if (parent->type == EXPR_PREOP && parent->op == '&')
 387                 return 1;
 388         if (parent->type == EXPR_ASSIGNMENT && expr_equiv(parent->left, expr))
 389                 return 1;
 390
 391         stmt = last_ptr_list((struct ptr_list *)big_statement_stack);
 392         if (stmt && stmt->type == STMT_ASM)
 393                 return 1;
 394
 395         return 0;
 396 }
 397
 398 static bool is_just_silencing_used_variable(struct expression *expr)
 399 {
 400         struct symbol *type;
 401
 402         while ((expr = expr_get_parent_expr(expr))) {
 403                 if (expr->type == EXPR_PREOP && expr->op == '(')
 404                         continue;
 405                 if (expr->type == EXPR_CAST) {
 406                         type = expr->cast_type;
 407
 408                         if (!type)
 409                                 return false;
 410                         if (type->type == SYM_NODE)
 411                                 type = get_real_base_type(type);
 412                         if (type == &void_ctype)
 413                                 return true;
 414                 }
 415         }
 416
 417         return false;
 418 }
 419
 420 static void match_symbol(struct expression *expr)
 421 {
 422         char *name;
 423
 424         if (uncertain_code_path())
 425                 return;
 426
 427         if (is_initialized(expr))
 428                 return;
 429
 430         if (is_being_modified(expr))
 431                 return;
 432
 433         if (is_just_silencing_used_variable(expr))
 434                 return;
 435
 436         name = expr_to_str(expr);
 437         sm_error("uninitialized symbol '%s'.", name);
 438         free_string(name);
 439
 440         set_state_expr(my_id, expr, &initialized);
 441 }
 442
 443 static void match_untracked(struct expression *call, int param)
 444 {
 445         struct expression *arg;
 446
 447         arg = get_argument_from_call_expr(call->args, param);
 448         arg = strip_expr(arg);
 449         if (!arg || arg->type != EXPR_PREOP || arg->op != '&')
 450                 return;
 451         arg = strip_expr(arg->unop);
 452         set_state_expr(my_id, arg, &initialized);
 453 }
 454
 455 static void match_ignore_param(const char *fn, struct expression *expr, void *_arg_nr)
 456 {
 457         int arg_nr = PTR_INT(_arg_nr);
 458         struct expression *arg;
 459
 460         arg = get_argument_from_call_expr(expr->args, arg_nr);
 461         arg = strip_expr(arg);
 462         if (!arg)
 463                 return;
 464         if (arg->type != EXPR_PREOP || arg->op != '&')
 465                 return;
 466         arg = strip_expr(arg->unop);
 467         set_state_expr(my_id, arg, &initialized);
 468 }
 469
 470 static void register_ignored_params_from_file(void)
 471 {
 472         char name[256];
 473         struct token *token;
 474         const char *func;
 475         char prev_func[256];
 476         int param;
 477
 478         memset(prev_func, 0, sizeof(prev_func));
 479         snprintf(name, 256, "%s.ignore_uninitialized_param", option_project_str);
 480         name[255] = '\0';
 481         token = get_tokens_file(name);
 482         if (!token)
 483                 return;
 484         if (token_type(token) != TOKEN_STREAMBEGIN)
 485                 return;
 486         token = token->next;
 487         while (token_type(token) != TOKEN_STREAMEND) {
 488                 if (token_type(token) != TOKEN_IDENT)
 489                         return;
 490                 func = show_ident(token->ident);
 491
 492                 token = token->next;
 493                 if (token_type(token) != TOKEN_NUMBER)
 494                         return;
 495                 param = atoi(token->number);
 496
 497                 add_function_hook(func, &match_ignore_param, INT_PTR(param));
 498
 499                 token = token->next;
 500         }
 501         clear_token_alloc();
 502 }
 503
 504 void check_uninitialized(int id)
 505 {
 506         my_id = id;
 507
 508         add_hook(&match_declarations, DECLARATION_HOOK);
 509         add_extra_mod_hook(&extra_mod_hook);
 510         add_hook(&match_assign, ASSIGNMENT_HOOK);
 511         add_hook(&match_negative_comparison, CONDITION_HOOK);
 512         add_hook(&match_enum_switch, STMT_HOOK_AFTER);
 513         add_hook(&match_enum_switch_after, STMT_HOOK_AFTER);
 514         add_untracked_param_hook(&match_untracked);
 515         add_pre_merge_hook(my_id, &pre_merge_hook);
 516
 517         add_hook(&match_dereferences, DEREF_HOOK);
 518         add_hook(&match_condition, CONDITION_HOOK);
 519         add_hook(&match_call, FUNCTION_CALL_HOOK);
 520         add_hook(&match_function_pointer_call, FUNCTION_CALL_HOOK);
 521         add_hook(&match_call_struct_members, FUNCTION_CALL_HOOK);
 522         add_hook(&match_symbol, SYM_HOOK);
 523
 524         register_ignored_params_from_file();
 525 }