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         expr = strip_expr(expr);
 111
 112         if (expr->type == EXPR_PREOP && expr->op == '&')
 113                 return strip_expr(expr->unop);
 114         return expr;
 115 }
 116
 117 static struct expression *faked_expression;
 118 struct expression *get_faked_expression(void)
 119 {
 120         if (!__in_fake_assign)
 121                 return NULL;
 122         return faked_expression;
 123 }
 124
 125 static void set_inner_struct_members(int mode, struct expression *faked, struct expression *left, struct expression *right, struct symbol *member)
 126 {
 127         struct expression *left_member;
 128         struct expression *right_member = NULL;  /* silence GCC */
 129         struct expression *assign;
 130         struct symbol *base = get_real_base_type(member);
 131         struct symbol *tmp;
 132
 133         if (member->ident) {
 134                 left = member_expression(left, '.', member->ident);
 135                 if (mode != COPY_MEMSET && right)
 136                         right = member_expression(right, '.', member->ident);
 137         }
 138
 139         FOR_EACH_PTR(base->symbol_list, tmp) {
 140                 struct symbol *type;
 141
 142                 type = get_real_base_type(tmp);
 143                 if (!type)
 144                         continue;
 145
 146                 if (type->type == SYM_ARRAY)
 147                         continue;
 148                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 149                         set_inner_struct_members(mode, faked, left, right, tmp);
 150                         continue;
 151                 }
 152                 if (!tmp->ident)
 153                         continue;
 154
 155                 left_member = member_expression(left, '.', tmp->ident);
 156
 157                 switch (mode) {
 158                 case COPY_NORMAL:
 159                 case COPY_MEMCPY:
 160                         if (right)
 161                                 right_member = member_expression(right, '.', tmp->ident);
 162                         else
 163                                 right_member = unknown_value_expression(left_member);
 164                         break;
 165                 case COPY_MEMSET:
 166                         right_member = right;
 167                         break;
 168                 }
 169
 170                 assign = assign_expression(left_member, right_member);
 171                 __in_fake_assign++;
 172                 __split_expr(assign);
 173                 __in_fake_assign--;
 174         } END_FOR_EACH_PTR(tmp);
 175
 176
 177 }
 178
 179 static void __struct_members_copy(int mode, struct expression *faked,
 180                                   struct expression *left,
 181                                   struct expression *right)
 182 {
 183         struct symbol *struct_type, *tmp, *type;
 184         struct expression *left_member;
 185         struct expression *right_member;
 186         struct expression *assign;
 187         int op = '.';
 188
 189
 190         if (__in_fake_assign)
 191                 return;
 192         faked_expression = faked;
 193
 194         left = strip_expr(left);
 195         right = strip_expr(right);
 196
 197         struct_type = get_struct_type(left);
 198         if (!struct_type) {
 199                 /*
 200                  * This is not a struct assignment obviously.  But this is where
 201                  * memcpy() is handled so it feels like a good place to add this
 202                  * code.
 203                  */
 204
 205                 type = get_type(left);
 206                 if (!type || type->type != SYM_BASETYPE)
 207                         goto done;
 208
 209                 right = strip_expr(right);
 210                 if (right && right->type == EXPR_PREOP && right->op == '&')
 211                         right = remove_addr(right);
 212                 else
 213                         right = unknown_value_expression(left);
 214                 assign = assign_expression(left, right);
 215                 __in_fake_assign++;
 216                 __split_expr(assign);
 217                 __in_fake_assign--;
 218                 goto done;
 219         }
 220
 221         if (is_pointer(left)) {
 222                 left = deref_expression(left);
 223                 op = '*';
 224         }
 225         if (mode != COPY_MEMSET)
 226                 right = get_right_base_expr(struct_type, right);
 227
 228         FOR_EACH_PTR(struct_type->symbol_list, tmp) {
 229                 type = get_real_base_type(tmp);
 230                 if (!type)
 231                         continue;
 232                 if (type->type == SYM_ARRAY)
 233                         continue;
 234
 235                 if (type->type == SYM_UNION || type->type == SYM_STRUCT) {
 236                         set_inner_struct_members(mode, faked, left, right, tmp);
 237                         continue;
 238                 }
 239
 240                 if (!tmp->ident)
 241                         continue;
 242
 243                 left_member = member_expression(left, op, tmp->ident);
 244                 right_member = NULL;
 245
 246                 switch (mode) {
 247                 case COPY_NORMAL:
 248                 case COPY_MEMCPY:
 249                         if (right)
 250                                 right_member = member_expression(right, op, tmp->ident);
 251                         else
 252                                 right_member = unknown_value_expression(left_member);
 253                         break;
 254                 case COPY_MEMSET:
 255                         right_member = right;
 256                         break;
 257                 }
 258                 if (!right_member) {
 259                         sm_msg("internal.  No right member");
 260                         continue;
 261                 }
 262                 assign = assign_expression(left_member, right_member);
 263                 __in_fake_assign++;
 264                 __split_expr(assign);
 265                 __in_fake_assign--;
 266         } END_FOR_EACH_PTR(tmp);
 267
 268 done:
 269         faked_expression = NULL;
 270 }
 271
 272 static int returns_zeroed_mem(struct expression *expr)
 273 {
 274         char *fn;
 275
 276         if (expr->type != EXPR_CALL || expr->fn->type != EXPR_SYMBOL)
 277                 return 0;
 278         fn = expr_to_var(expr->fn);
 279         if (!fn)
 280                 return 0;
 281         if (strcmp(fn, "kcalloc") == 0)
 282                 return 1;
 283         if (option_project == PROJ_KERNEL && strstr(fn, "zalloc"))
 284                 return 1;
 285         return 0;
 286 }
 287
 288 void __fake_struct_member_assignments(struct expression *expr)
 289 {
 290         struct symbol *struct_type;
 291
 292         if (is_zero(expr->right))
 293                 return;
 294
 295         struct_type = get_struct_type(expr->left);
 296         if (!struct_type)
 297                 return;
 298
 299         if (returns_zeroed_mem(expr->right))
 300                 __struct_members_copy(COPY_MEMSET, expr, expr->left, zero_expr());
 301         else
 302                 __struct_members_copy(COPY_NORMAL, expr, expr->left, expr->right);
 303 }
 304
 305 static void match_memset(const char *fn, struct expression *expr, void *_size_arg)
 306 {
 307         struct expression *buf;
 308         struct expression *val;
 309
 310         buf = get_argument_from_call_expr(expr->args, 0);
 311         val = get_argument_from_call_expr(expr->args, 1);
 312
 313         buf = strip_expr(buf);
 314         __struct_members_copy(COPY_MEMSET, expr, remove_addr(buf), val);
 315 }
 316
 317 static void match_memcpy(const char *fn, struct expression *expr, void *_arg)
 318 {
 319         struct expression *dest;
 320         struct expression *src;
 321
 322         dest = get_argument_from_call_expr(expr->args, 0);
 323         src = get_argument_from_call_expr(expr->args, 1);
 324
 325         __struct_members_copy(COPY_MEMCPY, expr, remove_addr(dest), remove_addr(src));
 326 }
 327
 328 static void match_memcpy_unknown(const char *fn, struct expression *expr, void *_arg)
 329 {
 330         struct expression *dest;
 331
 332         dest = get_argument_from_call_expr(expr->args, 0);
 333         __struct_members_copy(COPY_MEMCPY, expr, remove_addr(dest), NULL);
 334 }
 335
 336 static void match_sscanf(const char *fn, struct expression *expr, void *unused)
 337 {
 338         struct expression *arg;
 339         int i;
 340
 341         i = -1;
 342         FOR_EACH_PTR(expr->args, arg) {
 343                 if (++i < 2)
 344                         continue;
 345                 __struct_members_copy(COPY_MEMCPY, expr, remove_addr(arg), NULL);
 346         } END_FOR_EACH_PTR(arg);
 347 }
 348
 349 static void unop_expr(struct expression *expr)
 350 {
 351         if (expr->op != SPECIAL_INCREMENT &&
 352             expr->op != SPECIAL_DECREMENT)
 353                 return;
 354
 355         if (!is_pointer(expr))
 356                 return;
 357         __struct_members_copy(COPY_MEMCPY, expr, expr->unop, NULL);
 358 }
 359
 360 static void register_clears_param(void)
 361 {
 362         struct token *token;
 363         char name[256];
 364         const char *function;
 365         int param;
 366
 367         if (option_project == PROJ_NONE)
 368                 return;
 369
 370         snprintf(name, 256, "%s.clears_argument", option_project_str);
 371
 372         token = get_tokens_file(name);
 373         if (!token)
 374                 return;
 375         if (token_type(token) != TOKEN_STREAMBEGIN)
 376                 return;
 377         token = token->next;
 378         while (token_type(token) != TOKEN_STREAMEND) {
 379                 if (token_type(token) != TOKEN_IDENT)
 380                         return;
 381                 function = show_ident(token->ident);
 382                 token = token->next;
 383                 if (token_type(token) != TOKEN_NUMBER)
 384                         return;
 385                 param = atoi(token->number);
 386                 add_function_hook(function, &match_memcpy_unknown, INT_PTR(param));
 387                 token = token->next;
 388         }
 389         clear_token_alloc();
 390 }
 391
 392 static void db_param_cleared(struct expression *expr, int param, char *key, char *value)
 393 {
 394         struct expression *arg;
 395
 396         while (expr->type == EXPR_ASSIGNMENT)
 397                 expr = strip_expr(expr->right);
 398         if (expr->type != EXPR_CALL)
 399                 return;
 400
 401         /*
 402          * FIXME:  __struct_members_copy() requires an expression but
 403          * get_variable_from_key() returns a name/sym pair so that doesn't
 404          * work here.
 405          */
 406         if (strcmp(key, "$") != 0)
 407                 return;
 408
 409         arg = get_argument_from_call_expr(expr->args, param);
 410         if (!arg)
 411                 return;
 412
 413         if (strcmp(value, "0") == 0)
 414                 __struct_members_copy(COPY_MEMSET, expr, remove_addr(arg), zero_expr());
 415         else
 416                 __struct_members_copy(COPY_MEMCPY, expr, remove_addr(arg), NULL);
 417 }
 418
 419 void register_struct_assignment(int id)
 420 {
 421         add_function_hook("memset", &match_memset, NULL);
 422         add_function_hook("__memset", &match_memset, NULL);
 423
 424         add_function_hook("memcpy", &match_memcpy, INT_PTR(0));
 425         add_function_hook("memmove", &match_memcpy, INT_PTR(0));
 426         add_function_hook("__memcpy", &match_memcpy, INT_PTR(0));
 427         add_function_hook("__memmove", &match_memcpy, INT_PTR(0));
 428
 429         add_function_hook("sscanf", &match_sscanf, NULL);
 430
 431         add_hook(&unop_expr, OP_HOOK);
 432         register_clears_param();
 433         select_return_states_hook(PARAM_CLEARED, &db_param_cleared);
 434 }