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_copy_to_scratch", USER_PTR_SET, -1, "$" },
  84         { "xdr_inline_decode", USER_PTR_SET, -1, "$" },
  85         { "ntfs_read_run_nb", USER_PTR_SET, 3, "$" },
  86         { "ntfs_read_bh", USER_PTR_SET, 3, "(0<~$0)" },
  87
  88         { "kstrtoull", USER_PTR_SET, 2, "$" },
  89         { "kstrtoll", USER_PTR_SET, 2, "$" },
  90         { "kstrtoul", USER_PTR_SET, 2, "$" },
  91         { "kstrtol", USER_PTR_SET, 2, "$" },
  92         { "kstrtoint", USER_PTR_SET, 2, "$" },
  93         { "kstrtou64", USER_PTR_SET, 2, "$" },
  94         { "kstrtos64", USER_PTR_SET, 2, "$" },
  95         { "kstrtou32", USER_PTR_SET, 2, "$" },
  96         { "kstrtos32", USER_PTR_SET, 2, "$" },
  97         { "kstrtou16", USER_PTR_SET, 2, "$" },
  98         { "kstrtos16", USER_PTR_SET, 2, "$" },
  99         { "kstrtou8", USER_PTR_SET, 2, "$" },
 100         { "kstrtos8", USER_PTR_SET, 2, "$" },
 101         { "kstrtoull_from_user", USER_PTR_SET, 2, "$" },
 102         { "kstrtoll_from_user", USER_PTR_SET, 2, "$" },
 103         { "kstrtoul_from_user", USER_PTR_SET, 2, "$" },
 104         { "kstrtol_from_user", USER_PTR_SET, 2, "$" },
 105         { "kstrtouint_from_user", USER_PTR_SET, 2, "$" },
 106         { "kstrtoint_from_user", USER_PTR_SET, 2, "$" },
 107         { "kstrtou16_from_user", USER_PTR_SET, 2, "$" },
 108         { "kstrtos16_from_user", USER_PTR_SET, 2, "$" },
 109         { "kstrtou8_from_user", USER_PTR_SET, 2, "$" },
 110         { "kstrtos8_from_user", USER_PTR_SET, 2, "$" },
 111         { "kstrtou64_from_user", USER_PTR_SET, 2, "$" },
 112         { "kstrtos64_from_user", USER_PTR_SET, 2, "$" },
 113         { "kstrtou32_from_user", USER_PTR_SET, 2, "$" },
 114         { "kstrtos32_from_user", USER_PTR_SET, 2, "$" },
 115 };
 116
 117 static struct user_fn_info call_table[] = {
 118         { "__handle_ksmbd_work", USER_DATA, 0, "$->request_buf" },
 119 };
 120
 121 bool is_skb_data(struct expression *expr)
 122 {
 123         struct symbol *sym;
 124
 125         expr = strip_expr(expr);
 126         if (!expr)
 127                 return false;
 128         if (expr->type != EXPR_DEREF)
 129                 return false;
 130
 131         if (!expr->member)
 132                 return false;
 133         if (strcmp(expr->member->name, "data") != 0)
 134                 return false;
 135
 136         sym = get_type(expr->deref);
 137         if (!sym)
 138                 return false;
 139         if (sym->type == SYM_PTR)
 140                 sym = get_real_base_type(sym);
 141         if (!sym || sym->type != SYM_STRUCT || !sym->ident)
 142                 return false;
 143         if (strcmp(sym->ident->name, "sk_buff") != 0)
 144                 return false;
 145
 146         return true;
 147 }
 148
 149 static bool is_array_of_user_data(struct expression *expr)
 150 {
 151         struct expression *deref;
 152         struct symbol *type;
 153
 154         if (expr->type == EXPR_PREOP && expr->op == '&') {
 155                 expr = strip_expr(expr->unop);
 156                 if (expr->type == EXPR_PREOP && expr->op == '*')
 157                         expr = strip_expr(expr->unop);
 158         }
 159
 160         /* This is for array elements &foo->data[4] */
 161         if (expr->type == EXPR_BINOP && expr->op == '+') {
 162                 if (points_to_user_data(expr->left))
 163                         return true;
 164                 if (points_to_user_data(expr->right))
 165                         return true;
 166         }
 167
 168         /* This is for if you have: foo = skb->data; frob(foo->array); */
 169         type = get_type(expr);
 170         if (!type || type->type != SYM_ARRAY)
 171                 return false;
 172
 173         if (expr->type != EXPR_DEREF)
 174                 return false;
 175         deref = strip_expr(expr->deref);
 176         if (deref->type != EXPR_PREOP || deref->op != '*')
 177                 return false;
 178         deref = strip_expr(deref->unop);
 179         return points_to_user_data(deref);
 180 }
 181
 182 static struct expression *remove_addr_stuff(struct expression *expr)
 183 {
 184         /* take "&foo->bar" and return "foo" */
 185         expr = strip_expr(expr);
 186         if (expr->type != EXPR_PREOP || expr->op != '&')
 187                 return expr;
 188         expr = strip_expr(expr->unop);
 189         while (expr && expr->type == EXPR_DEREF) {
 190                 expr = strip_expr(expr->deref);
 191                 if (expr->op == '.')
 192                         continue;
 193                 else
 194                         break;
 195         }
 196         if (expr->type == EXPR_PREOP && expr->op == '*')
 197                 expr = strip_expr(expr->unop);
 198         return expr;
 199 }
 200
 201 static bool math_points_to_user_data(struct expression *expr)
 202 {
 203         struct sm_state *sm;
 204
 205         // TODO: is_array_of_user_data() should probably be handled here
 206
 207         if (expr->type == EXPR_BINOP && expr->op == '+')
 208                 return math_points_to_user_data(expr->left);
 209
 210         expr = remove_addr_stuff(expr);
 211
 212         sm = get_sm_state_expr(my_id, expr);
 213         if (!sm)
 214                 return false;
 215         if (slist_has_state(sm->possible, &user_data) ||
 216             slist_has_state(sm->possible, &user_data_set))
 217                 return true;
 218         return false;
 219 }
 220
 221 bool points_to_user_data(struct expression *expr)
 222 {
 223         struct sm_state *sm;
 224
 225         expr = strip_expr(expr);
 226         if (!expr)
 227                 return false;
 228
 229         if (expr->type == EXPR_POSTOP)
 230                 expr = strip_expr(expr->unop);
 231
 232         if (is_fake_call(expr))
 233                 return false;
 234
 235         if (expr->type == EXPR_ASSIGNMENT)
 236                 return points_to_user_data(expr->left);
 237
 238         if (is_array_of_user_data(expr))
 239                 return true;
 240
 241         if (expr->type == EXPR_BINOP && expr->op == '+')
 242                 return math_points_to_user_data(expr);
 243
 244         if (is_skb_data(expr))
 245                 return true;
 246
 247         // FIXME if you have a struct pointer p then p->foo should be handled
 248         // by smatch_kernel_user_data.c but if you have (p + 1)->foo then this
 249         // should be handled here.
 250         sm = get_sm_state_expr(my_id, expr);
 251         if (!sm)
 252                 return false;
 253         if (slist_has_state(sm->possible, &user_data) ||
 254             slist_has_state(sm->possible, &user_data_set))
 255                 return true;
 256         return false;
 257 }
 258
 259 void set_array_user_ptr(struct expression *expr, bool is_new)
 260 {
 261         struct expression *tmp;
 262
 263         tmp = get_assigned_expr(expr);
 264         if (tmp)
 265                 set_state_expr(my_id, tmp, is_new ? &user_data_set : &user_data);
 266         set_state_expr(my_id, expr, is_new ? &user_data_set : &user_data);
 267 }
 268
 269 static void match_assign(struct expression *expr)
 270 {
 271         if (is_fake_call(expr->right))
 272                 return;
 273
 274         if (!is_ptr_type(get_type(expr->left)))
 275                 return;
 276
 277         if (points_to_user_data(expr->right)) {
 278                 // FIXME: if the types are different then mark the stuff on
 279                 // the left as user data.
 280                 set_state_expr(my_id, expr->left, &user_data);
 281                 return;
 282         }
 283
 284         // FIXME: just use a modification hook
 285         if (get_state_expr(my_id, expr->left))
 286                 set_state_expr(my_id, expr->left, &undefined);
 287 }
 288
 289 static void match_memcpy(const char *fn, struct expression *expr, void *_unused)
 290 {
 291         struct expression *dest, *src;
 292
 293         dest = get_argument_from_call_expr(expr->args, 0);
 294         src = get_argument_from_call_expr(expr->args, 1);
 295
 296         if (points_to_user_data(src)) {
 297                 set_state_expr(my_id, expr->left, &user_data_set);
 298                 return;
 299         }
 300
 301         if (get_state_expr(my_id, dest))
 302                 set_state_expr(my_id, dest, &undefined);
 303 }
 304
 305 static void fake_assign_helper(struct expression *expr, void *data)
 306 {
 307         struct expression *left = expr->left;
 308         struct symbol *type;
 309         bool set = data;
 310
 311         type = get_type(left);
 312         if (!type)
 313                 return;
 314         if (type->type == SYM_BASETYPE)
 315                 mark_as_user_data(left, set);
 316         else if (type->type == SYM_ARRAY)
 317                 set_array_user_ptr(left, set);
 318 }
 319
 320 static void returns_user_ptr_helper(struct expression *expr, const char *name, struct symbol *sym, bool set)
 321 {
 322         struct expression *call, *arg;
 323
 324         call = expr;
 325         while (call && call->type == EXPR_ASSIGNMENT)
 326                 call = strip_expr(expr->right);
 327         if (!call || call->type != EXPR_CALL)
 328                 return;
 329
 330         if (!set && !we_pass_user_data(call))
 331                 return;
 332
 333         arg = gen_expression_from_name_sym(name, sym);
 334         if (!arg)
 335                 return;
 336
 337         create_recursive_fake_assignments(deref_expression(arg), &fake_assign_helper, INT_PTR(set));
 338
 339         if (arg->type == EXPR_PREOP && arg->op == '&') {
 340                 struct symbol *type;
 341
 342                 type = get_type(arg->unop);
 343                 if (!type || type->type != SYM_ARRAY)
 344                         return;
 345         }
 346
 347         set_state_expr(my_id, arg, set ? &user_data_set : &user_data);
 348 }
 349
 350 static void returns_user_ptr(struct expression *expr, const char *name, struct symbol *sym, void *data)
 351 {
 352         returns_user_ptr_helper(expr, name, sym, false);
 353 }
 354
 355 static void returns_user_ptr_set(struct expression *expr, const char *name, struct symbol *sym, void *data)
 356 {
 357         returns_user_ptr_helper(expr, name, sym, true);
 358 }
 359
 360 static void set_param_user_ptr(const char *name, struct symbol *sym, char *value)
 361 {
 362         set_state(my_id, name, sym, &user_data);
 363 }
 364
 365 static void set_caller_param_key_user_ptr(struct expression *expr, const char *name,
 366                                     struct symbol *sym, void *data)
 367 {
 368         set_state(my_id, name, sym, &user_data);
 369 }
 370
 371 static void caller_info_callback(struct expression *call, int param, char *printed_name, struct sm_state *sm)
 372 {
 373         if (is_socket_stuff(sm->sym))
 374                 return;
 375
 376         if (!slist_has_state(sm->possible, &user_data) &&
 377             !slist_has_state(sm->possible, &user_data_set))
 378                 return;
 379
 380         sql_insert_caller_info(call, USER_PTR, param, printed_name, "");
 381 }
 382
 383 static void return_info_callback(int return_id, char *return_ranges,
 384                                  struct expression *returned_expr,
 385                                  int param,
 386                                  const char *printed_name,
 387                                  struct sm_state *sm)
 388 {
 389         int type;
 390
 391         /* is this even possible? */
 392         if (strcmp(printed_name, "&$") == 0)
 393                 return;
 394
 395         if (is_socket_stuff(sm->sym))
 396                 return;
 397
 398         if (param >= 0) {
 399                 if (!slist_has_state(sm->possible, &user_data_set))
 400                         return;
 401                 type = USER_PTR_SET;
 402         } else {
 403                 if (slist_has_state(sm->possible, &user_data_set))
 404                         type = USER_PTR_SET;
 405                 else if (slist_has_state(sm->possible, &user_data))
 406                         type = USER_PTR;
 407                 else
 408                         return;
 409         }
 410         if (parent_is_gone_var_sym(sm->name, sm->sym))
 411                 return;
 412
 413         sql_insert_return_states(return_id, return_ranges, type,
 414                                  param, printed_name, "");
 415 }
 416
 417 void register_points_to_user_data(int id)
 418 {
 419         struct user_fn_info *info;
 420         int i;
 421
 422         my_id = id;
 423
 424         if (option_project != PROJ_KERNEL)
 425                 return;
 426
 427         add_hook(&match_assign, ASSIGNMENT_HOOK);
 428
 429         add_function_hook("memcpy", &match_memcpy, NULL);
 430         add_function_hook("__memcpy", &match_memcpy, NULL);
 431
 432         add_caller_info_callback(my_id, caller_info_callback);
 433         add_return_info_callback(my_id, return_info_callback);
 434
 435         select_caller_name_sym(set_param_user_ptr, USER_PTR);
 436         for (i = 0; i < ARRAY_SIZE(call_table); i++) {
 437                 info = &call_table[i];
 438                 add_function_param_key_hook_early(info->name,
 439                                                   &set_caller_param_key_user_ptr,
 440                                                   info->param, info->key, info);
 441         }
 442
 443         select_return_param_key(USER_PTR, &returns_user_ptr);
 444         select_return_param_key(USER_PTR_SET, &returns_user_ptr_set);
 445         for (i = 0; i < ARRAY_SIZE(func_table); i++) {
 446                 info = &func_table[i];
 447                 add_function_param_key_hook_late(info->name,
 448                                                  &returns_user_ptr_set,
 449                                                  info->param, info->key, info);
 450         }
 451
 452 }