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