smatch_points_to_host_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 /* The below code is an adjustment of the smatch_points_to_user_data.c to
  19  * track (untrusted and potentially malicious) data pointers received by a
  20  * guest kernel via various low-level port IO, MMIO, MSR, CPUID and PCI
  21  * config space access functions (see func_table for a full list) from an
  22  * untrusted host/VMM under confidential computing threat model
  23  * (where a guest VM does not trust the host/VMM).
  24  * We call this data as 'host' data here. Being able to track host data
  25  * allows creating new smatch patterns that can perform various checks
  26  * on such data, i.e. verify that that there are no spectre v1 gadgets
  27  * on this attack surface or even simply report locations where host data
  28  * is being processed in the kernel source code (useful for a targeted code
  29  * audit).
  30  * Similar as smatch_points_to_user_data.c it works with
  31  * smatch_kernel_host_data.c code.
  32  *
  33  *  Copyright (C) 2022 Elena Reshetova Intel Corporation.
  34  */
  35
  36 #include "smatch.h"
  37 #include "smatch_slist.h"
  38 #include "smatch_extra.h"
  39
  40 static int my_id;
  41 STATE(host_data);
  42 STATE(host_data_set);
  43
  44 static const char *returns_pointer_to_host_data[] = {
  45         "virtqueue_get_buf"
  46 };
  47
  48 struct host_fn_info {
  49         const char *name;
  50         int type;
  51         int param;
  52         const char *key;
  53         const sval_t *implies_start, *implies_end;
  54         func_hook *call_back;
  55 };
  56
  57 struct host_fn_info func_pointer_table[] = {
  58             /* Functions that return host data as argument 1 */
  59         { "memcpy_fromio", HOST_DATA, 0, "*$" },
  60         { "acpi_os_read_iomem", HOST_DATA, 1, "*$" },
  61         { "mmio_insb", HOST_DATA, 1, "*$" },
  62         { "mmio_insw", HOST_DATA, 1, "*$" },
  63         { "mmio_insl", HOST_DATA, 1, "*$" },
  64         { "acpi_read_bit_register", HOST_DATA, 1, "*$" },
  65           /* Functions that return host data as argument 3 */
  66         { "__virtio_cread_many", HOST_DATA, 2, "*$" },
  67         { "pci_user_read_config_word", HOST_DATA, 2, "*$" },
  68         { "pci_user_read_config_dword", HOST_DATA, 2, "*$" },
  69         { "pci_user_read_config_byte", HOST_DATA, 2, "*$" },
  70           /* Functions that return host data as argument 4 */
  71         { "pci_bus_read_config_byte", HOST_DATA, 3, "*$" },
  72         { "pci_bus_read_config_word", HOST_DATA, 3, "*$" },
  73         { "pci_bus_read_config_dword", HOST_DATA, 3, "*$" },
  74         /* Functions that return host data as argument 6 */
  75         { "raw_pci_read", HOST_DATA, 5, "*$" },
  76         /* Functions that return host data as arguments 2-5 */
  77         { "cpuid", HOST_DATA, 1, "*$" },
  78         { "cpuid", HOST_DATA, 2, "*$" },
  79         { "cpuid", HOST_DATA, 3, "*$" },
  80         { "cpuid", HOST_DATA, 4, "*$" },
  81         /* Functions that return host data as arguments 3-6 */
  82         { "cpuid_count", HOST_DATA, 2, "*$" },
  83         { "cpuid_count", HOST_DATA, 3, "*$" },
  84         { "cpuid_count", HOST_DATA, 4, "*$" },
  85         { "cpuid_count", HOST_DATA, 5, "*$" },
  86 };
  87
  88 bool is_host_data_fn(struct symbol *fn)
  89 {
  90         int i;
  91
  92         if (!fn || !fn->ident)
  93                 return false;
  94
  95         for (i = 0; i < ARRAY_SIZE(returns_pointer_to_host_data); i++) {
  96                 if (strcmp(fn->ident->name, returns_pointer_to_host_data[i]) == 0) {
  97 //                      func_gets_host_data = true;
  98                         return true;
  99                 }
 100         }
 101         return false;
 102 }
 103
 104 static bool is_points_to_host_data_fn(struct expression *expr)
 105 {
 106         expr = strip_expr(expr);
 107         if (!expr || expr->type != EXPR_CALL || expr->fn->type != EXPR_SYMBOL ||
 108             !expr->fn->symbol)
 109                 return false;
 110         return is_host_data_fn(expr->fn->symbol);
 111 }
 112
 113 static bool is_array_of_host_data(struct expression *expr)
 114 {
 115         struct expression *deref;
 116         struct symbol *type;
 117
 118         if (expr->type == EXPR_PREOP && expr->op == '&') {
 119                 expr = strip_expr(expr->unop);
 120                 if (expr->type == EXPR_PREOP && expr->op == '*')
 121                         expr = strip_expr(expr->unop);
 122         }
 123
 124         /* This is for array elements &foo->data[4] */
 125         if (expr->type == EXPR_BINOP && expr->op == '+') {
 126                 if (points_to_host_data(expr->left))
 127                         return true;
 128                 if (points_to_host_data(expr->right))
 129                         return true;
 130         }
 131
 132         /* This is for if you have: foo = skb->data; frob(foo->array); */
 133         type = get_type(expr);
 134         if (!type || type->type != SYM_ARRAY)
 135                 return false;
 136
 137         if (expr->type != EXPR_DEREF)
 138                 return false;
 139         deref = strip_expr(expr->deref);
 140         if (deref->type != EXPR_PREOP || deref->op != '*')
 141                 return false;
 142         deref = strip_expr(deref->unop);
 143         return points_to_host_data(deref);
 144 }
 145
 146 bool points_to_host_data(struct expression *expr)
 147 {
 148         struct sm_state *sm;
 149
 150         if (!expr)
 151                 return false;
 152
 153         expr = strip_expr(expr);
 154         if (!expr)
 155                 return false;
 156
 157         if (is_fake_call(expr))
 158                 return false;
 159
 160         if (expr->type == EXPR_ASSIGNMENT)
 161                 return points_to_host_data(expr->left);
 162
 163         if (is_array_of_host_data(expr))
 164                 return true;
 165
 166         if (expr->type == EXPR_BINOP && expr->op == '+')
 167                 expr = strip_expr(expr->left);
 168
 169         if (is_points_to_host_data_fn(expr))
 170                 return true;
 171
 172         sm = get_sm_state_expr(my_id, expr);
 173         if (!sm)
 174                 return false;
 175         if (slist_has_state(sm->possible, &host_data) ||
 176             slist_has_state(sm->possible, &host_data_set))
 177                 return true;
 178         return false;
 179 }
 180
 181 void set_points_to_host_data(struct expression *expr, bool is_new)
 182 {
 183         struct expression *tmp;
 184
 185         tmp = get_assigned_expr(expr);
 186         if (tmp)
 187                 set_state_expr(my_id, tmp, is_new ? &host_data_set : &host_data);
 188         set_state_expr(my_id, expr, is_new ? &host_data_set : &host_data);
 189 }
 190
 191 static void match_assign_host(struct expression *expr)
 192 {
 193
 194         if (is_fake_call(expr->right))
 195                 return;
 196
 197         if (!is_ptr_type(get_type(expr->left))){
 198                 return;
 199         }
 200
 201         if (points_to_host_data(expr->right)) {
 202                 set_points_to_host_data(expr->left, false);
 203                 return;
 204         }
 205
 206         if (get_state_expr(my_id, expr->left)){
 207                 set_state_expr(my_id, expr->left, &undefined);
 208         }
 209 }
 210
 211 static void match_memcpy_host(const char *fn, struct expression *expr, void *_unused)
 212 {
 213         struct expression *dest, *src;
 214
 215         dest = get_argument_from_call_expr(expr->args, 0);
 216         src = get_argument_from_call_expr(expr->args, 1);
 217
 218         if (points_to_host_data(src)) {
 219                 set_points_to_host_data(dest, false);
 220                 return;
 221         }
 222
 223         if (get_state_expr(my_id, dest))
 224                 set_state_expr(my_id, dest, &undefined);
 225 }
 226
 227 static void match_host_function(const char *fn, struct expression *expr, void *_unused)
 228 {
 229         struct expression *dest;
 230
 231         for (int i = 0; i < ARRAY_SIZE(func_pointer_table); i++)
 232                 if (sym_name_is(func_pointer_table[i].name, expr->fn)){
 233                         dest = get_argument_from_call_expr(expr->args, func_pointer_table[i].param);
 234                         dest = strip_expr(dest);
 235                         if (!dest)
 236                                 return;
 237                         set_state_expr(my_id, dest, &host_data);
 238                 }
 239 }
 240
 241 static void return_info_callback_host(int return_id, char *return_ranges,
 242                                  struct expression *returned_expr,
 243                                  int param,
 244                                  const char *printed_name,
 245                                  struct sm_state *sm)
 246 {
 247         int type;
 248
 249         if (strncmp(printed_name, "&$", 2) == 0)
 250                 return;
 251
 252         if (!slist_has_state(sm->possible, &host_data))
 253                 return;
 254
 255         if (param >= 0) {
 256                 if (!slist_has_state(sm->possible, &host_data_set))
 257                         return;
 258                 type = HOST_PTR_SET;
 259         } else {
 260                 if (slist_has_state(sm->possible, &host_data_set))
 261                         type = HOST_PTR_SET;
 262                 else if (slist_has_state(sm->possible, &host_data))
 263                         type = HOST_PTR;
 264                 else
 265                         return;
 266         }
 267         if (parent_is_gone_var_sym(sm->name, sm->sym))
 268                 return;
 269
 270         sql_insert_return_states(return_id, return_ranges, type,
 271                                  param, printed_name, "");
 272 }
 273
 274 static void returns_host_ptr_helper(struct expression *expr, int param, char *key, char *value, bool set)
 275 {
 276         struct expression *arg;
 277         struct expression *call;
 278         char *name;
 279         struct symbol *sym;
 280
 281         call = expr;
 282         while (call->type == EXPR_ASSIGNMENT)
 283                 call = strip_expr(call->right);
 284         if (call->type != EXPR_CALL)
 285                 return;
 286
 287         if (!set && !we_pass_host_data(call))
 288                 return;
 289
 290         if (param == -1) {
 291                 if (expr->type != EXPR_ASSIGNMENT) {
 292                         /* Nothing to do.  Fake assignments should handle it */
 293                         return;
 294                 }
 295                 arg = expr->left;
 296                 goto set_host;
 297         }
 298
 299         arg = get_argument_from_call_expr(call->args, param);
 300         if (!arg)
 301                 return;
 302 set_host:
 303         name = get_variable_from_key(arg, key, &sym);
 304         if (!name || !sym)
 305                 goto free;
 306         if (set)
 307                 set_state(my_id, name, sym, &host_data_set);
 308         else
 309                 set_state(my_id, name, sym, &host_data);
 310 free:
 311         free_string(name);
 312 }
 313
 314 static void returns_host_ptr(struct expression *expr, int param, char *key, char *value)
 315 {
 316         returns_host_ptr_helper(expr, param, key, value, false);
 317 }
 318
 319 static void returns_host_ptr_set(struct expression *expr, int param, char *key, char *value)
 320 {
 321         returns_host_ptr_helper(expr, param, key, value, true);
 322 }
 323
 324 static void set_param_host_ptr(const char *name, struct symbol *sym, char *key, char *value)
 325 {
 326         struct expression *expr;
 327         char *fullname;
 328
 329         expr = symbol_expression(sym);
 330         fullname = get_variable_from_key(expr, key, NULL);
 331         if (!fullname)
 332                 return;
 333         set_state(my_id, fullname, sym, &host_data);
 334 }
 335
 336 static void caller_info_callback_host(struct expression *call, int param, char *printed_name, struct sm_state *sm)
 337 {
 338         if (!slist_has_state(sm->possible, &host_data) &&
 339             !slist_has_state(sm->possible, &host_data_set))
 340                 return;
 341         sql_insert_caller_info(call, HOST_PTR, param, printed_name, "");
 342 }
 343
 344 void register_points_to_host_data(int id)
 345 {
 346         my_id = id;
 347
 348         if (option_project != PROJ_KERNEL)
 349                 return;
 350
 351         add_hook(&match_assign_host, ASSIGNMENT_HOOK);
 352
 353         for (int i = 0; i < ARRAY_SIZE(func_pointer_table); i++)
 354                 add_function_hook(func_pointer_table[i].name, &match_host_function, NULL);
 355
 356         add_function_hook("memcpy", &match_memcpy_host, NULL);
 357         add_function_hook("__memcpy", &match_memcpy_host, NULL);
 358
 359         add_caller_info_callback(my_id, caller_info_callback_host);
 360         add_return_info_callback(my_id, return_info_callback_host);
 361
 362         select_caller_info_hook(set_param_host_ptr, HOST_PTR);
 363         select_return_states_hook(HOST_PTR, &returns_host_ptr);
 364         select_return_states_hook(HOST_PTR_SET, &returns_host_ptr_set);
 365 }