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