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