smatch_points_to_user_data.c

   1 /*
   2  * Copyright (C) 2020 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  * The problem here is that we can have:
  20  *
  21  * p = skb->data;
  22  *
  23  * In the olden days we would just set "*p = 0-255" which meant that it pointed
  24  * to user data.  But then if we say "if (*p == 11) {" that means that "*p" is
  25  * not user data any more, so then "*(p + 1)" is marked as not user data but it
  26  * is.
  27  *
  28  * So now we've separated out the stuff that points to a user_buf from the other
  29  * user data.
  30  *
  31  * There is a further complication because what if "p" points to a struct?  In
  32  * that case all the struct members are handled by smatch_kernel_user_data.c
  33  * but we still need to keep in mind that "*(p + 1)" is user data.  I'm not
  34  * totally 100% sure how this will work.
  35  *
  36  * Generally a user pointer should be a void pointer, or an array etc.  But if
  37  * it points to a struct that can only be used for pointer math.
  38  *
  39  */
  40
  41 #include "smatch.h"
  42 #include "smatch_slist.h"
  43 #include "smatch_extra.h"
  44
  45 static int my_id;
  46 STATE(user_data);
  47 STATE(user_data_set);
  48
  49 struct user_fn_info {
  50         const char *name;
  51         int type;
  52         int param;
  53         const char *key;
  54 };
  55
  56 // Old stuff that was here, but I no longer believe is user data
  57 // kmap_atomic()
  58 // skb_network_header
  59
  60 //      add_function_hook("memcpy_fromiovec", &match_user_copy, INT_PTR(0));
  61 //      add_function_hook("usb_control_msg", &match_user_copy, INT_PTR(6));
  62
  63 static struct user_fn_info func_table[] = {
  64         { "copy_from_user", USER_PTR_SET, 0, "$" },
  65         { "__copy_from_user", USER_PTR_SET, 0, "$" },
  66         { "kvm_read_guest_virt", USER_PTR_SET, 2, "$" },
  67         { "vpu_iface_receive_msg", USER_PTR_SET, 1, "$" },
  68         { "xdr_stream_decode_u32", USER_PTR_SET, 1, "$" },
  69
  70         { "(struct ksmbd_transport_ops)->read", USER_PTR_SET, 1, "$" },
  71         { "nlmsg_data", USER_PTR_SET, -1, "$" },
  72         { "nla_data", USER_PTR_SET, -1, "$" },
  73         { "memdup_user", USER_PTR_SET, -1, "$" },
  74         { "cfg80211_find_elem_match", USER_PTR_SET, -1, "$" },
  75         { "ieee80211_bss_get_elem", USER_PTR_SET, -1, "$" },
  76         { "cfg80211_find_elem", USER_PTR_SET, -1, "$" },
  77         { "ieee80211_bss_get_ie", USER_PTR_SET, -1, "$" },
  78
  79         { "brcmf_fweh_dequeue_event", USER_PTR_SET, -1, "&$->emsg" },
  80         { "wilc_wlan_rxq_remove", USER_PTR_SET, -1, "$->buffer" },
  81         { "cfg80211_find_vendor_ie", USER_PTR_SET, -1, "$" },
  82
  83         { "xdr_inline_decode", USER_PTR_SET, -1, "$" },
  84         { "ntfs_read_run_nb", USER_PTR_SET, 3, "$" },
  85         { "ntfs_read_bh", USER_PTR_SET, 3, "(0<~$0)" },
  86
  87         { "kstrtoull", USER_PTR_SET, 2, "$" },
  88         { "kstrtoll", USER_PTR_SET, 2, "$" },
  89         { "kstrtoul", USER_PTR_SET, 2, "$" },
  90         { "kstrtol", USER_PTR_SET, 2, "$" },
  91         { "kstrtoint", USER_PTR_SET, 2, "$" },
  92         { "kstrtou64", USER_PTR_SET, 2, "$" },
  93         { "kstrtos64", USER_PTR_SET, 2, "$" },
  94         { "kstrtou32", USER_PTR_SET, 2, "$" },
  95         { "kstrtos32", USER_PTR_SET, 2, "$" },
  96         { "kstrtou16", USER_PTR_SET, 2, "$" },
  97         { "kstrtos16", USER_PTR_SET, 2, "$" },
  98         { "kstrtou8", USER_PTR_SET, 2, "$" },
  99         { "kstrtos8", USER_PTR_SET, 2, "$" },
 100         { "kstrtoull_from_user", USER_PTR_SET, 2, "$" },
 101         { "kstrtoll_from_user", USER_PTR_SET, 2, "$" },
 102         { "kstrtoul_from_user", USER_PTR_SET, 2, "$" },
 103         { "kstrtol_from_user", USER_PTR_SET, 2, "$" },
 104         { "kstrtouint_from_user", USER_PTR_SET, 2, "$" },
 105         { "kstrtoint_from_user", USER_PTR_SET, 2, "$" },
 106         { "kstrtou16_from_user", USER_PTR_SET, 2, "$" },
 107         { "kstrtos16_from_user", USER_PTR_SET, 2, "$" },
 108         { "kstrtou8_from_user", USER_PTR_SET, 2, "$" },
 109         { "kstrtos8_from_user", USER_PTR_SET, 2, "$" },
 110         { "kstrtou64_from_user", USER_PTR_SET, 2, "$" },
 111         { "kstrtos64_from_user", USER_PTR_SET, 2, "$" },
 112         { "kstrtou32_from_user", USER_PTR_SET, 2, "$" },
 113         { "kstrtos32_from_user", USER_PTR_SET, 2, "$" },
 114 };
 115
 116 static struct user_fn_info call_table[] = {
 117         { "__handle_ksmbd_work", USER_DATA, 0, "$->request_buf" },
 118 };
 119
 120 bool is_skb_data(struct expression *expr)
 121 {
 122         struct symbol *sym;
 123
 124         expr = strip_expr(expr);
 125         if (!expr)
 126                 return false;
 127         if (expr->type != EXPR_DEREF)
 128                 return false;
 129
 130         if (!expr->member)
 131                 return false;
 132         if (strcmp(expr->member->name, "data") != 0)
 133                 return false;
 134
 135         sym = get_type(expr->deref);
 136         if (!sym)
 137                 return false;
 138         if (sym->type == SYM_PTR)
 139                 sym = get_real_base_type(sym);
 140         if (!sym || sym->type != SYM_STRUCT || !sym->ident)
 141                 return false;
 142         if (strcmp(sym->ident->name, "sk_buff") != 0)
 143                 return false;
 144
 145         return true;
 146 }
 147
 148 static bool is_array_of_user_data(struct expression *expr)
 149 {
 150         struct expression *deref;
 151         struct symbol *type;
 152
 153         if (expr->type == EXPR_PREOP && expr->op == '&') {
 154                 expr = strip_expr(expr->unop);
 155                 if (expr->type == EXPR_PREOP && expr->op == '*')
 156                         expr = strip_expr(expr->unop);
 157         }
 158
 159         /* This is for array elements &foo->data[4] */
 160         if (expr->type == EXPR_BINOP && expr->op == '+') {
 161                 if (points_to_user_data(expr->left))
 162                         return true;
 163                 if (points_to_user_data(expr->right))
 164                         return true;
 165         }
 166
 167         /* This is for if you have: foo = skb->data; frob(foo->array); */
 168         type = get_type(expr);
 169         if (!type || type->type != SYM_ARRAY)
 170                 return false;
 171
 172         if (expr->type != EXPR_DEREF)
 173                 return false;
 174         deref = strip_expr(expr->deref);
 175         if (deref->type != EXPR_PREOP || deref->op != '*')
 176                 return false;
 177         deref = strip_expr(deref->unop);
 178         return points_to_user_data(deref);
 179 }
 180
 181 static struct expression *remove_addr_stuff(struct expression *expr)
 182 {
 183         /* take "&foo->bar" and return "foo" */
 184         expr = strip_expr(expr);
 185         if (expr->type != EXPR_PREOP || expr->op != '&')
 186                 return expr;
 187         expr = strip_expr(expr->unop);
 188         while (expr && expr->type == EXPR_DEREF) {
 189                 expr = strip_expr(expr->deref);
 190                 if (expr->op == '.')
 191                         continue;
 192                 else
 193                         break;
 194         }
 195         if (expr->type == EXPR_PREOP && expr->op == '*')
 196                 expr = strip_expr(expr->unop);
 197         return expr;
 198 }
 199
 200 static bool math_points_to_user_data(struct expression *expr)
 201 {
 202         struct sm_state *sm;
 203
 204         // TODO: is_array_of_user_data() should probably be handled here
 205
 206         if (expr->type == EXPR_BINOP && expr->op == '+')
 207                 return math_points_to_user_data(expr->left);
 208
 209         expr = remove_addr_stuff(expr);
 210
 211         sm = get_sm_state_expr(my_id, expr);
 212         if (!sm)
 213                 return false;
 214         if (slist_has_state(sm->possible, &user_data) ||
 215             slist_has_state(sm->possible, &user_data_set))
 216                 return true;
 217         return false;
 218 }
 219
 220 bool points_to_user_data(struct expression *expr)
 221 {
 222         struct sm_state *sm;
 223
 224         expr = strip_expr(expr);
 225         if (!expr)
 226                 return false;
 227
 228         if (expr->type == EXPR_POSTOP)
 229                 expr = strip_expr(expr->unop);
 230
 231         if (is_fake_call(expr))
 232                 return false;
 233
 234         if (expr->type == EXPR_ASSIGNMENT)
 235                 return points_to_user_data(expr->left);
 236
 237         if (is_array_of_user_data(expr))
 238                 return true;
 239
 240         if (expr->type == EXPR_BINOP && expr->op == '+')
 241                 return math_points_to_user_data(expr);
 242
 243         if (is_skb_data(expr))
 244                 return true;
 245
 246         // FIXME if you have a struct pointer p then p->foo should be handled
 247         // by smatch_kernel_user_data.c but if you have (p + 1)->foo then this
 248         // should be handled here.
 249         sm = get_sm_state_expr(my_id, expr);
 250         if (!sm)
 251                 return false;
 252         if (slist_has_state(sm->possible, &user_data) ||
 253             slist_has_state(sm->possible, &user_data_set))
 254                 return true;
 255         return false;
 256 }
 257
 258 void set_array_user_ptr(struct expression *expr, bool is_new)
 259 {
 260         struct expression *tmp;
 261
 262         tmp = get_assigned_expr(expr);
 263         if (tmp)
 264                 set_state_expr(my_id, tmp, is_new ? &user_data_set : &user_data);
 265         set_state_expr(my_id, expr, is_new ? &user_data_set : &user_data);
 266 }
 267
 268 static void match_assign(struct expression *expr)
 269 {
 270         if (is_fake_call(expr->right))
 271                 return;
 272
 273         if (!is_ptr_type(get_type(expr->left)))
 274                 return;
 275
 276         if (points_to_user_data(expr->right)) {
 277                 // FIXME: if the types are different then mark the stuff on
 278                 // the left as user data.
 279                 set_state_expr(my_id, expr->left, &user_data);
 280                 return;
 281         }
 282
 283         // FIXME: just use a modification hook
 284         if (get_state_expr(my_id, expr->left))
 285                 set_state_expr(my_id, expr->left, &undefined);
 286 }
 287
 288 static void match_memcpy(const char *fn, struct expression *expr, void *_unused)
 289 {
 290         struct expression *dest, *src;
 291
 292         dest = get_argument_from_call_expr(expr->args, 0);
 293         src = get_argument_from_call_expr(expr->args, 1);
 294
 295         if (points_to_user_data(src)) {
 296                 set_state_expr(my_id, expr->left, &user_data_set);
 297                 return;
 298         }
 299
 300         if (get_state_expr(my_id, dest))
 301                 set_state_expr(my_id, dest, &undefined);
 302 }
 303
 304 static void fake_assign_helper(struct expression *expr, void *data)
 305 {
 306         struct expression *left = expr->left;
 307         struct symbol *type;
 308         bool set = data;
 309
 310         type = get_type(left);
 311         if (!type)
 312                 return;
 313         if (type->type == SYM_BASETYPE)
 314                 mark_as_user_data(left, set);
 315         else if (type->type == SYM_ARRAY)
 316                 set_array_user_ptr(expr, set);
 317 }
 318
 319 static void returns_user_ptr_helper(struct expression *expr, const char *name, struct symbol *sym, bool set)
 320 {
 321         struct expression *call, *arg;
 322
 323         call = expr;
 324         while (call && call->type == EXPR_ASSIGNMENT)
 325                 call = strip_expr(expr->right);
 326         if (!call || call->type != EXPR_CALL)
 327                 return;
 328
 329         if (!set && !we_pass_user_data(call))
 330                 return;
 331
 332         arg = gen_expression_from_name_sym(name, sym);
 333         if (!arg)
 334                 return;
 335
 336         create_recursive_fake_assignments(deref_expression(arg), &fake_assign_helper, INT_PTR(set));
 337
 338         if (arg->type == EXPR_PREOP && arg->op == '&') {
 339                 struct symbol *type;
 340
 341                 type = get_type(arg->unop);
 342                 if (!type || type->type != SYM_ARRAY)
 343                         return;
 344         }
 345
 346         set_state_expr(my_id, arg, set ? &user_data_set : &user_data);
 347 }
 348
 349 static void returns_user_ptr(struct expression *expr, const char *name, struct symbol *sym, void *data)
 350 {
 351         returns_user_ptr_helper(expr, name, sym, false);
 352 }
 353
 354 static void returns_user_ptr_set(struct expression *expr, const char *name, struct symbol *sym, void *data)
 355 {
 356         returns_user_ptr_helper(expr, name, sym, true);
 357 }
 358
 359 static void set_param_user_ptr(const char *name, struct symbol *sym, char *value)
 360 {
 361         set_state(my_id, name, sym, &user_data);
 362 }
 363
 364 static void set_caller_param_key_user_ptr(struct expression *expr, const char *name,
 365                                     struct symbol *sym, void *data)
 366 {
 367         set_state(my_id, name, sym, &user_data);
 368 }
 369
 370 static void caller_info_callback(struct expression *call, int param, char *printed_name, struct sm_state *sm)
 371 {
 372         if (is_socket_stuff(sm->sym))
 373                 return;
 374
 375         if (!slist_has_state(sm->possible, &user_data) &&
 376             !slist_has_state(sm->possible, &user_data_set))
 377                 return;
 378
 379         sql_insert_caller_info(call, USER_PTR, param, printed_name, "");
 380 }
 381
 382 static void return_info_callback(int return_id, char *return_ranges,
 383                                  struct expression *returned_expr,
 384                                  int param,
 385                                  const char *printed_name,
 386                                  struct sm_state *sm)
 387 {
 388         int type;
 389
 390         /* is this even possible? */
 391         if (strcmp(printed_name, "&$") == 0)
 392                 return;
 393
 394         if (is_socket_stuff(sm->sym))
 395                 return;
 396
 397         if (param >= 0) {
 398                 if (!slist_has_state(sm->possible, &user_data_set))
 399                         return;
 400                 type = USER_PTR_SET;
 401         } else {
 402                 if (slist_has_state(sm->possible, &user_data_set))
 403                         type = USER_PTR_SET;
 404                 else if (slist_has_state(sm->possible, &user_data))
 405                         type = USER_PTR;
 406                 else
 407                         return;
 408         }
 409         if (parent_is_gone_var_sym(sm->name, sm->sym))
 410                 return;
 411
 412         sql_insert_return_states(return_id, return_ranges, type,
 413                                  param, printed_name, "");
 414 }
 415
 416 void register_points_to_user_data(int id)
 417 {
 418         struct user_fn_info *info;
 419         int i;
 420
 421         my_id = id;
 422
 423         if (option_project != PROJ_KERNEL)
 424                 return;
 425
 426         add_hook(&match_assign, ASSIGNMENT_HOOK);
 427
 428         add_function_hook("memcpy", &match_memcpy, NULL);
 429         add_function_hook("__memcpy", &match_memcpy, NULL);
 430
 431         add_caller_info_callback(my_id, caller_info_callback);
 432         add_return_info_callback(my_id, return_info_callback);
 433
 434         select_caller_name_sym(set_param_user_ptr, USER_PTR);
 435         for (i = 0; i < ARRAY_SIZE(call_table); i++) {
 436                 info = &call_table[i];
 437                 add_function_param_key_hook_early(info->name,
 438                                                   &set_caller_param_key_user_ptr,
 439                                                   info->param, info->key, info);
 440         }
 441
 442         select_return_param_key(USER_PTR, &returns_user_ptr);
 443         select_return_param_key(USER_PTR_SET, &returns_user_ptr_set);
 444         for (i = 0; i < ARRAY_SIZE(func_table); i++) {
 445                 info = &func_table[i];
 446                 add_function_param_key_hook_late(info->name,
 447                                                  &returns_user_ptr_set,
 448                                                  info->param, info->key, info);
 449         }
 450
 451 }