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