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