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