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