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