/* UndefinedBehaviorSanitizer, undefined behavior detector. Copyright (C) 2013-2014 Free Software Foundation, Inc. Contributed by Marek Polacek This file is part of GCC. GCC is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3, or (at your option) any later version. GCC is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with GCC; see the file COPYING3. If not see . */ #include "config.h" #include "system.h" #include "coretypes.h" #include "tree.h" #include "stor-layout.h" #include "stringpool.h" #include "cgraph.h" #include "tree-pass.h" #include "tree-ssa-alias.h" #include "tree-pretty-print.h" #include "internal-fn.h" #include "gimple-expr.h" #include "gimple.h" #include "gimple-iterator.h" #include "gimple-ssa.h" #include "gimple-walk.h" #include "hashtab.h" #include "output.h" #include "tm_p.h" #include "toplev.h" #include "cfgloop.h" #include "ubsan.h" #include "c-family/c-common.h" #include "rtl.h" #include "expr.h" #include "tree-ssanames.h" #include "asan.h" #include "gimplify-me.h" #include "intl.h" #include "realmpfr.h" #include "dfp.h" /* Map from a tree to a VAR_DECL tree. */ struct GTY(()) tree_type_map { struct tree_map_base type; tree decl; }; #define tree_type_map_eq tree_map_base_eq #define tree_type_map_marked_p tree_map_base_marked_p /* Hash from a tree in a tree_type_map. */ unsigned int tree_type_map_hash (const void *item) { return TYPE_UID (((const struct tree_type_map *)item)->type.from); } static GTY ((if_marked ("tree_type_map_marked_p"), param_is (struct tree_type_map))) htab_t decl_tree_for_type; /* Lookup a VAR_DECL for TYPE, and return it if we find one. */ static tree decl_for_type_lookup (tree type) { /* If the hash table is not initialized yet, create it now. */ if (decl_tree_for_type == NULL) { decl_tree_for_type = htab_create_ggc (10, tree_type_map_hash, tree_type_map_eq, 0); /* That also means we don't have to bother with the lookup. */ return NULL_TREE; } struct tree_type_map *h, in; in.type.from = type; h = (struct tree_type_map *) htab_find_with_hash (decl_tree_for_type, &in, TYPE_UID (type)); return h ? h->decl : NULL_TREE; } /* Insert a mapping TYPE->DECL in the VAR_DECL for type hashtable. */ static void decl_for_type_insert (tree type, tree decl) { struct tree_type_map *h; void **slot; h = ggc_alloc (); h->type.from = type; h->decl = decl; slot = htab_find_slot_with_hash (decl_tree_for_type, h, TYPE_UID (type), INSERT); *(struct tree_type_map **) slot = h; } /* Helper routine, which encodes a value in the pointer_sized_int_node. Arguments with precision <= POINTER_SIZE are passed directly, the rest is passed by reference. T is a value we are to encode. IN_EXPAND_P is true if this function is called during expansion. */ tree ubsan_encode_value (tree t, bool in_expand_p) { tree type = TREE_TYPE (t); const unsigned int bitsize = GET_MODE_BITSIZE (TYPE_MODE (type)); if (bitsize <= POINTER_SIZE) switch (TREE_CODE (type)) { case BOOLEAN_TYPE: case ENUMERAL_TYPE: case INTEGER_TYPE: return fold_build1 (NOP_EXPR, pointer_sized_int_node, t); case REAL_TYPE: { tree itype = build_nonstandard_integer_type (bitsize, true); t = fold_build1 (VIEW_CONVERT_EXPR, itype, t); return fold_convert (pointer_sized_int_node, t); } default: gcc_unreachable (); } else { if (!DECL_P (t) || !TREE_ADDRESSABLE (t)) { /* The reason for this is that we don't want to pessimize code by making vars unnecessarily addressable. */ tree var = create_tmp_var (type, NULL); tree tem = build2 (MODIFY_EXPR, void_type_node, var, t); if (in_expand_p) { rtx mem = assign_stack_temp_for_type (TYPE_MODE (type), GET_MODE_SIZE (TYPE_MODE (type)), type); SET_DECL_RTL (var, mem); expand_assignment (var, t, false); return build_fold_addr_expr (var); } t = build_fold_addr_expr (var); return build2 (COMPOUND_EXPR, TREE_TYPE (t), tem, t); } else return build_fold_addr_expr (t); } } /* Build struct __ubsan_type_descriptor { unsigned short __typekind; unsigned short __typeinfo; char __typename[]; } type. */ static tree ubsan_type_descriptor_type (void) { static const char *field_names[3] = { "__typekind", "__typeinfo", "__typename" }; tree fields[3], ret; tree itype = build_range_type (sizetype, size_zero_node, NULL_TREE); tree flex_arr_type = build_array_type (char_type_node, itype); ret = make_node (RECORD_TYPE); for (int i = 0; i < 3; i++) { fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, get_identifier (field_names[i]), (i == 2) ? flex_arr_type : short_unsigned_type_node); DECL_CONTEXT (fields[i]) = ret; if (i) DECL_CHAIN (fields[i - 1]) = fields[i]; } TYPE_FIELDS (ret) = fields[0]; TYPE_NAME (ret) = get_identifier ("__ubsan_type_descriptor"); layout_type (ret); return ret; } /* Build struct __ubsan_source_location { const char *__filename; unsigned int __line; unsigned int __column; } type. */ static tree ubsan_source_location_type (void) { static const char *field_names[3] = { "__filename", "__line", "__column" }; tree fields[3], ret; tree const_char_type = build_qualified_type (char_type_node, TYPE_QUAL_CONST); ret = make_node (RECORD_TYPE); for (int i = 0; i < 3; i++) { fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, get_identifier (field_names[i]), (i == 0) ? build_pointer_type (const_char_type) : unsigned_type_node); DECL_CONTEXT (fields[i]) = ret; if (i) DECL_CHAIN (fields[i - 1]) = fields[i]; } TYPE_FIELDS (ret) = fields[0]; TYPE_NAME (ret) = get_identifier ("__ubsan_source_location"); layout_type (ret); return ret; } /* Helper routine that returns a CONSTRUCTOR of __ubsan_source_location type with its fields filled from a location_t LOC. */ static tree ubsan_source_location (location_t loc) { expanded_location xloc; tree type = ubsan_source_location_type (); xloc = expand_location (loc); if (xloc.file == NULL) xloc.file = ""; /* Fill in the values from LOC. */ size_t len = strlen (xloc.file); tree str = build_string (len + 1, xloc.file); TREE_TYPE (str) = build_array_type (char_type_node, build_index_type (size_int (len))); TREE_READONLY (str) = 1; TREE_STATIC (str) = 1; str = build_fold_addr_expr (str); tree ctor = build_constructor_va (type, 3, NULL_TREE, str, NULL_TREE, build_int_cst (unsigned_type_node, xloc.line), NULL_TREE, build_int_cst (unsigned_type_node, xloc.column)); TREE_CONSTANT (ctor) = 1; TREE_STATIC (ctor) = 1; return ctor; } /* This routine returns a magic number for TYPE. */ static unsigned short get_ubsan_type_info_for_type (tree type) { gcc_assert (TYPE_SIZE (type) && tree_fits_uhwi_p (TYPE_SIZE (type))); if (TREE_CODE (type) == REAL_TYPE) return tree_to_uhwi (TYPE_SIZE (type)); else if (INTEGRAL_TYPE_P (type)) { int prec = exact_log2 (tree_to_uhwi (TYPE_SIZE (type))); gcc_assert (prec != -1); return (prec << 1) | !TYPE_UNSIGNED (type); } else return 0; } /* Helper routine that returns ADDR_EXPR of a VAR_DECL of a type descriptor. It first looks into the hash table; if not found, create the VAR_DECL, put it into the hash table and return the ADDR_EXPR of it. TYPE describes a particular type. PSTYLE is an enum controlling how we want to print the type. */ tree ubsan_type_descriptor (tree type, enum ubsan_print_style pstyle) { /* See through any typedefs. */ type = TYPE_MAIN_VARIANT (type); tree decl = decl_for_type_lookup (type); /* It is possible that some of the earlier created DECLs were found unused, in that case they weren't emitted and varpool_get_node returns NULL node on them. But now we really need them. Thus, renew them here. */ if (decl != NULL_TREE && varpool_get_node (decl)) return build_fold_addr_expr (decl); tree dtype = ubsan_type_descriptor_type (); tree type2 = type; const char *tname = NULL; char *pretty_name; unsigned char deref_depth = 0; unsigned short tkind, tinfo; /* Get the name of the type, or the name of the pointer type. */ if (pstyle == UBSAN_PRINT_POINTER) { gcc_assert (POINTER_TYPE_P (type)); type2 = TREE_TYPE (type); /* Remove any '*' operators from TYPE. */ while (POINTER_TYPE_P (type2)) deref_depth++, type2 = TREE_TYPE (type2); if (TREE_CODE (type2) == METHOD_TYPE) type2 = TYPE_METHOD_BASETYPE (type2); } /* If an array, get its type. */ type2 = strip_array_types (type2); if (pstyle == UBSAN_PRINT_ARRAY) { while (POINTER_TYPE_P (type2)) deref_depth++, type2 = TREE_TYPE (type2); } if (TYPE_NAME (type2) != NULL) { if (TREE_CODE (TYPE_NAME (type2)) == IDENTIFIER_NODE) tname = IDENTIFIER_POINTER (TYPE_NAME (type2)); else if (DECL_NAME (TYPE_NAME (type2)) != NULL) tname = IDENTIFIER_POINTER (DECL_NAME (TYPE_NAME (type2))); } if (tname == NULL) /* We weren't able to determine the type name. */ tname = ""; /* Decorate the type name with '', '*', "struct", or "union". */ pretty_name = (char *) alloca (strlen (tname) + 16 + deref_depth); if (pstyle == UBSAN_PRINT_POINTER) { int pos = sprintf (pretty_name, "'%s%s%s%s%s%s%s", TYPE_VOLATILE (type2) ? "volatile " : "", TYPE_READONLY (type2) ? "const " : "", TYPE_RESTRICT (type2) ? "restrict " : "", TYPE_ATOMIC (type2) ? "_Atomic " : "", TREE_CODE (type2) == RECORD_TYPE ? "struct " : TREE_CODE (type2) == UNION_TYPE ? "union " : "", tname, deref_depth == 0 ? "" : " "); while (deref_depth-- > 0) pretty_name[pos++] = '*'; pretty_name[pos++] = '\''; pretty_name[pos] = '\0'; } else if (pstyle == UBSAN_PRINT_ARRAY) { /* Pretty print the array dimensions. */ gcc_assert (TREE_CODE (type) == ARRAY_TYPE); tree t = type; int pos = sprintf (pretty_name, "'%s ", tname); while (deref_depth-- > 0) pretty_name[pos++] = '*'; while (TREE_CODE (t) == ARRAY_TYPE) { pretty_name[pos++] = '['; tree dom = TYPE_DOMAIN (t); if (dom && TREE_CODE (TYPE_MAX_VALUE (dom)) == INTEGER_CST) pos += sprintf (&pretty_name[pos], HOST_WIDE_INT_PRINT_DEC, tree_to_shwi (TYPE_MAX_VALUE (dom)) + 1); else /* ??? We can't determine the variable name; print VLA unspec. */ pretty_name[pos++] = '*'; pretty_name[pos++] = ']'; t = TREE_TYPE (t); } pretty_name[pos++] = '\''; pretty_name[pos] = '\0'; /* Save the tree with stripped types. */ type = t; } else sprintf (pretty_name, "'%s'", tname); switch (TREE_CODE (type)) { case BOOLEAN_TYPE: case ENUMERAL_TYPE: case INTEGER_TYPE: tkind = 0x0000; break; case REAL_TYPE: /* FIXME: libubsan right now only supports float, double and long double type formats. */ if (TYPE_MODE (type) == TYPE_MODE (float_type_node) || TYPE_MODE (type) == TYPE_MODE (double_type_node) || TYPE_MODE (type) == TYPE_MODE (long_double_type_node)) tkind = 0x0001; else tkind = 0xffff; break; default: tkind = 0xffff; break; } tinfo = get_ubsan_type_info_for_type (type); /* Create a new VAR_DECL of type descriptor. */ char tmp_name[32]; static unsigned int type_var_id_num; ASM_GENERATE_INTERNAL_LABEL (tmp_name, "Lubsan_type", type_var_id_num++); decl = build_decl (UNKNOWN_LOCATION, VAR_DECL, get_identifier (tmp_name), dtype); TREE_STATIC (decl) = 1; TREE_PUBLIC (decl) = 0; DECL_ARTIFICIAL (decl) = 1; DECL_IGNORED_P (decl) = 1; DECL_EXTERNAL (decl) = 0; size_t len = strlen (pretty_name); tree str = build_string (len + 1, pretty_name); TREE_TYPE (str) = build_array_type (char_type_node, build_index_type (size_int (len))); TREE_READONLY (str) = 1; TREE_STATIC (str) = 1; tree ctor = build_constructor_va (dtype, 3, NULL_TREE, build_int_cst (short_unsigned_type_node, tkind), NULL_TREE, build_int_cst (short_unsigned_type_node, tinfo), NULL_TREE, str); TREE_CONSTANT (ctor) = 1; TREE_STATIC (ctor) = 1; DECL_INITIAL (decl) = ctor; varpool_finalize_decl (decl); /* Save the VAR_DECL into the hash table. */ decl_for_type_insert (type, decl); return build_fold_addr_expr (decl); } /* Create a structure for the ubsan library. NAME is a name of the new structure. The arguments in ... are of __ubsan_type_descriptor type and there are at most two of them. MISMATCH are data used by ubsan pointer checking. */ tree ubsan_create_data (const char *name, const location_t *ploc, const struct ubsan_mismatch_data *mismatch, ...) { va_list args; tree ret, t; tree fields[5]; vec *saved_args = NULL; size_t i = 0; location_t loc = UNKNOWN_LOCATION; /* Firstly, create a pointer to type descriptor type. */ tree td_type = ubsan_type_descriptor_type (); TYPE_READONLY (td_type) = 1; td_type = build_pointer_type (td_type); /* Create the structure type. */ ret = make_node (RECORD_TYPE); if (ploc != NULL) { loc = LOCATION_LOCUS (*ploc); fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, NULL_TREE, ubsan_source_location_type ()); DECL_CONTEXT (fields[i]) = ret; i++; } va_start (args, mismatch); for (t = va_arg (args, tree); t != NULL_TREE; i++, t = va_arg (args, tree)) { gcc_checking_assert (i < 3); /* Save the tree arguments for later use. */ vec_safe_push (saved_args, t); fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, NULL_TREE, td_type); DECL_CONTEXT (fields[i]) = ret; if (i) DECL_CHAIN (fields[i - 1]) = fields[i]; } va_end (args); if (mismatch != NULL) { /* We have to add two more decls. */ fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, NULL_TREE, pointer_sized_int_node); DECL_CONTEXT (fields[i]) = ret; DECL_CHAIN (fields[i - 1]) = fields[i]; i++; fields[i] = build_decl (UNKNOWN_LOCATION, FIELD_DECL, NULL_TREE, unsigned_char_type_node); DECL_CONTEXT (fields[i]) = ret; DECL_CHAIN (fields[i - 1]) = fields[i]; i++; } TYPE_FIELDS (ret) = fields[0]; TYPE_NAME (ret) = get_identifier (name); layout_type (ret); /* Now, fill in the type. */ char tmp_name[32]; static unsigned int ubsan_var_id_num; ASM_GENERATE_INTERNAL_LABEL (tmp_name, "Lubsan_data", ubsan_var_id_num++); tree var = build_decl (UNKNOWN_LOCATION, VAR_DECL, get_identifier (tmp_name), ret); TREE_STATIC (var) = 1; TREE_PUBLIC (var) = 0; DECL_ARTIFICIAL (var) = 1; DECL_IGNORED_P (var) = 1; DECL_EXTERNAL (var) = 0; vec *v; vec_alloc (v, i); tree ctor = build_constructor (ret, v); /* If desirable, set the __ubsan_source_location element. */ if (ploc != NULL) CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, ubsan_source_location (loc)); size_t nelts = vec_safe_length (saved_args); for (i = 0; i < nelts; i++) { t = (*saved_args)[i]; CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, t); } if (mismatch != NULL) { /* Append the pointer data. */ CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, mismatch->align); CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, mismatch->ckind); } TREE_CONSTANT (ctor) = 1; TREE_STATIC (ctor) = 1; DECL_INITIAL (var) = ctor; varpool_finalize_decl (var); return var; } /* Instrument the __builtin_unreachable call. We just call the libubsan routine instead. */ tree ubsan_instrument_unreachable (location_t loc) { if (flag_sanitize_undefined_trap_on_error) return build_call_expr_loc (loc, builtin_decl_explicit (BUILT_IN_TRAP), 0); initialize_sanitizer_builtins (); tree data = ubsan_create_data ("__ubsan_unreachable_data", &loc, NULL, NULL_TREE); tree t = builtin_decl_explicit (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE); return build_call_expr_loc (loc, t, 1, build_fold_addr_expr_loc (loc, data)); } /* Return true if T is a call to a libubsan routine. */ bool is_ubsan_builtin_p (tree t) { return TREE_CODE (t) == FUNCTION_DECL && strncmp (IDENTIFIER_POINTER (DECL_NAME (t)), "__builtin___ubsan_", 18) == 0; } /* Expand the UBSAN_BOUNDS special builtin function. */ void ubsan_expand_bounds_ifn (gimple_stmt_iterator *gsi) { gimple stmt = gsi_stmt (*gsi); location_t loc = gimple_location (stmt); gcc_assert (gimple_call_num_args (stmt) == 3); /* Pick up the arguments of the UBSAN_BOUNDS call. */ tree type = TREE_TYPE (TREE_TYPE (gimple_call_arg (stmt, 0))); tree index = gimple_call_arg (stmt, 1); tree orig_index_type = TREE_TYPE (index); tree bound = gimple_call_arg (stmt, 2); gimple_stmt_iterator gsi_orig = *gsi; /* Create condition "if (index > bound)". */ basic_block then_bb, fallthru_bb; gimple_stmt_iterator cond_insert_point = create_cond_insert_point (gsi, 0/*before_p*/, false, true, &then_bb, &fallthru_bb); index = fold_convert (TREE_TYPE (bound), index); index = force_gimple_operand_gsi (&cond_insert_point, index, true/*simple_p*/, NULL_TREE, false/*before*/, GSI_NEW_STMT); gimple g = gimple_build_cond (GT_EXPR, index, bound, NULL_TREE, NULL_TREE); gimple_set_location (g, loc); gsi_insert_after (&cond_insert_point, g, GSI_NEW_STMT); /* Generate __ubsan_handle_out_of_bounds call. */ *gsi = gsi_after_labels (then_bb); if (flag_sanitize_undefined_trap_on_error) g = gimple_build_call (builtin_decl_explicit (BUILT_IN_TRAP), 0); else { tree data = ubsan_create_data ("__ubsan_out_of_bounds_data", &loc, NULL, ubsan_type_descriptor (type, UBSAN_PRINT_ARRAY), ubsan_type_descriptor (orig_index_type), NULL_TREE); data = build_fold_addr_expr_loc (loc, data); enum built_in_function bcode = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_OUT_OF_BOUNDS : BUILT_IN_UBSAN_HANDLE_OUT_OF_BOUNDS_ABORT; tree fn = builtin_decl_explicit (bcode); tree val = force_gimple_operand_gsi (gsi, ubsan_encode_value (index), true, NULL_TREE, true, GSI_SAME_STMT); g = gimple_build_call (fn, 2, data, val); } gimple_set_location (g, loc); gsi_insert_before (gsi, g, GSI_SAME_STMT); /* Get rid of the UBSAN_BOUNDS call from the IR. */ unlink_stmt_vdef (stmt); gsi_remove (&gsi_orig, true); /* Point GSI to next logical statement. */ *gsi = gsi_start_bb (fallthru_bb); } /* Expand UBSAN_NULL internal call. */ void ubsan_expand_null_ifn (gimple_stmt_iterator gsi) { gimple stmt = gsi_stmt (gsi); location_t loc = gimple_location (stmt); gcc_assert (gimple_call_num_args (stmt) == 2); tree ptr = gimple_call_arg (stmt, 0); tree ckind = gimple_call_arg (stmt, 1); basic_block cur_bb = gsi_bb (gsi); /* Split the original block holding the pointer dereference. */ edge e = split_block (cur_bb, stmt); /* Get a hold on the 'condition block', the 'then block' and the 'else block'. */ basic_block cond_bb = e->src; basic_block fallthru_bb = e->dest; basic_block then_bb = create_empty_bb (cond_bb); add_bb_to_loop (then_bb, cond_bb->loop_father); loops_state_set (LOOPS_NEED_FIXUP); /* Make an edge coming from the 'cond block' into the 'then block'; this edge is unlikely taken, so set up the probability accordingly. */ e = make_edge (cond_bb, then_bb, EDGE_TRUE_VALUE); e->probability = PROB_VERY_UNLIKELY; /* Connect 'then block' with the 'else block'. This is needed as the ubsan routines we call in the 'then block' are not noreturn. The 'then block' only has one outcoming edge. */ make_single_succ_edge (then_bb, fallthru_bb, EDGE_FALLTHRU); /* Set up the fallthrough basic block. */ e = find_edge (cond_bb, fallthru_bb); e->flags = EDGE_FALSE_VALUE; e->count = cond_bb->count; e->probability = REG_BR_PROB_BASE - PROB_VERY_UNLIKELY; /* Update dominance info for the newly created then_bb; note that fallthru_bb's dominance info has already been updated by split_bock. */ if (dom_info_available_p (CDI_DOMINATORS)) set_immediate_dominator (CDI_DOMINATORS, then_bb, cond_bb); /* Put the ubsan builtin call into the newly created BB. */ gimple g; if (flag_sanitize_undefined_trap_on_error) g = gimple_build_call (builtin_decl_implicit (BUILT_IN_TRAP), 0); else { enum built_in_function bcode = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH : BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT; tree fn = builtin_decl_implicit (bcode); const struct ubsan_mismatch_data m = { build_zero_cst (pointer_sized_int_node), ckind }; tree data = ubsan_create_data ("__ubsan_null_data", &loc, &m, ubsan_type_descriptor (TREE_TYPE (ptr), UBSAN_PRINT_POINTER), NULL_TREE); data = build_fold_addr_expr_loc (loc, data); g = gimple_build_call (fn, 2, data, build_zero_cst (pointer_sized_int_node)); } gimple_set_location (g, loc); gimple_stmt_iterator gsi2 = gsi_start_bb (then_bb); gsi_insert_after (&gsi2, g, GSI_NEW_STMT); /* Unlink the UBSAN_NULLs vops before replacing it. */ unlink_stmt_vdef (stmt); g = gimple_build_cond (EQ_EXPR, ptr, build_int_cst (TREE_TYPE (ptr), 0), NULL_TREE, NULL_TREE); gimple_set_location (g, loc); /* Replace the UBSAN_NULL with a GIMPLE_COND stmt. */ gsi_replace (&gsi, g, false); } /* Instrument a member call. We check whether 'this' is NULL. */ static void instrument_member_call (gimple_stmt_iterator *iter) { tree this_parm = gimple_call_arg (gsi_stmt (*iter), 0); tree kind = build_int_cst (unsigned_char_type_node, UBSAN_MEMBER_CALL); gimple g = gimple_build_call_internal (IFN_UBSAN_NULL, 2, this_parm, kind); gimple_set_location (g, gimple_location (gsi_stmt (*iter))); gsi_insert_before (iter, g, GSI_SAME_STMT); } /* Instrument a memory reference. T is the pointer, IS_LHS says whether the pointer is on the left hand side of the assignment. */ static void instrument_mem_ref (tree t, gimple_stmt_iterator *iter, bool is_lhs) { enum ubsan_null_ckind ikind = is_lhs ? UBSAN_STORE_OF : UBSAN_LOAD_OF; if (RECORD_OR_UNION_TYPE_P (TREE_TYPE (TREE_TYPE (t)))) ikind = UBSAN_MEMBER_ACCESS; tree kind = build_int_cst (unsigned_char_type_node, ikind); gimple g = gimple_build_call_internal (IFN_UBSAN_NULL, 2, t, kind); gimple_set_location (g, gimple_location (gsi_stmt (*iter))); gsi_insert_before (iter, g, GSI_SAME_STMT); } /* Perform the pointer instrumentation. */ static void instrument_null (gimple_stmt_iterator gsi, bool is_lhs) { gimple stmt = gsi_stmt (gsi); tree t = is_lhs ? gimple_get_lhs (stmt) : gimple_assign_rhs1 (stmt); t = get_base_address (t); const enum tree_code code = TREE_CODE (t); if (code == MEM_REF && TREE_CODE (TREE_OPERAND (t, 0)) == SSA_NAME) instrument_mem_ref (TREE_OPERAND (t, 0), &gsi, is_lhs); else if (code == ADDR_EXPR && POINTER_TYPE_P (TREE_TYPE (t)) && TREE_CODE (TREE_TYPE (TREE_TYPE (t))) == METHOD_TYPE) instrument_member_call (&gsi); } /* Build an ubsan builtin call for the signed-integer-overflow sanitization. CODE says what kind of builtin are we building, LOC is a location, LHSTYPE is the type of LHS, OP0 and OP1 are operands of the binary operation. */ tree ubsan_build_overflow_builtin (tree_code code, location_t loc, tree lhstype, tree op0, tree op1) { if (flag_sanitize_undefined_trap_on_error) return build_call_expr_loc (loc, builtin_decl_explicit (BUILT_IN_TRAP), 0); tree data = ubsan_create_data ("__ubsan_overflow_data", &loc, NULL, ubsan_type_descriptor (lhstype), NULL_TREE); enum built_in_function fn_code; switch (code) { case PLUS_EXPR: fn_code = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW : BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT; break; case MINUS_EXPR: fn_code = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW : BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT; break; case MULT_EXPR: fn_code = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW : BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT; break; case NEGATE_EXPR: fn_code = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW : BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT; break; default: gcc_unreachable (); } tree fn = builtin_decl_explicit (fn_code); return build_call_expr_loc (loc, fn, 2 + (code != NEGATE_EXPR), build_fold_addr_expr_loc (loc, data), ubsan_encode_value (op0, true), op1 ? ubsan_encode_value (op1, true) : NULL_TREE); } /* Perform the signed integer instrumentation. GSI is the iterator pointing at statement we are trying to instrument. */ static void instrument_si_overflow (gimple_stmt_iterator gsi) { gimple stmt = gsi_stmt (gsi); tree_code code = gimple_assign_rhs_code (stmt); tree lhs = gimple_assign_lhs (stmt); tree lhstype = TREE_TYPE (lhs); tree a, b; gimple g; /* If this is not a signed operation, don't instrument anything here. Also punt on bit-fields. */ if (!INTEGRAL_TYPE_P (lhstype) || TYPE_OVERFLOW_WRAPS (lhstype) || GET_MODE_BITSIZE (TYPE_MODE (lhstype)) != TYPE_PRECISION (lhstype)) return; switch (code) { case MINUS_EXPR: case PLUS_EXPR: case MULT_EXPR: /* Transform i = u {+,-,*} 5; into i = UBSAN_CHECK_{ADD,SUB,MUL} (u, 5); */ a = gimple_assign_rhs1 (stmt); b = gimple_assign_rhs2 (stmt); g = gimple_build_call_internal (code == PLUS_EXPR ? IFN_UBSAN_CHECK_ADD : code == MINUS_EXPR ? IFN_UBSAN_CHECK_SUB : IFN_UBSAN_CHECK_MUL, 2, a, b); gimple_call_set_lhs (g, lhs); gsi_replace (&gsi, g, false); break; case NEGATE_EXPR: /* Represent i = -u; as i = UBSAN_CHECK_SUB (0, u); */ a = build_int_cst (lhstype, 0); b = gimple_assign_rhs1 (stmt); g = gimple_build_call_internal (IFN_UBSAN_CHECK_SUB, 2, a, b); gimple_call_set_lhs (g, lhs); gsi_replace (&gsi, g, false); break; case ABS_EXPR: /* Transform i = ABS_EXPR; into _N = UBSAN_CHECK_SUB (0, u); i = ABS_EXPR<_N>; */ a = build_int_cst (lhstype, 0); b = gimple_assign_rhs1 (stmt); g = gimple_build_call_internal (IFN_UBSAN_CHECK_SUB, 2, a, b); a = make_ssa_name (lhstype, NULL); gimple_call_set_lhs (g, a); gimple_set_location (g, gimple_location (stmt)); gsi_insert_before (&gsi, g, GSI_SAME_STMT); gimple_assign_set_rhs1 (stmt, a); update_stmt (stmt); break; default: break; } } /* Instrument loads from (non-bitfield) bool and C++ enum values to check if the memory value is outside of the range of the valid type values. */ static void instrument_bool_enum_load (gimple_stmt_iterator *gsi) { gimple stmt = gsi_stmt (*gsi); tree rhs = gimple_assign_rhs1 (stmt); tree type = TREE_TYPE (rhs); tree minv = NULL_TREE, maxv = NULL_TREE; if (TREE_CODE (type) == BOOLEAN_TYPE && (flag_sanitize & SANITIZE_BOOL)) { minv = boolean_false_node; maxv = boolean_true_node; } else if (TREE_CODE (type) == ENUMERAL_TYPE && (flag_sanitize & SANITIZE_ENUM) && TREE_TYPE (type) != NULL_TREE && TREE_CODE (TREE_TYPE (type)) == INTEGER_TYPE && (TYPE_PRECISION (TREE_TYPE (type)) < GET_MODE_PRECISION (TYPE_MODE (type)))) { minv = TYPE_MIN_VALUE (TREE_TYPE (type)); maxv = TYPE_MAX_VALUE (TREE_TYPE (type)); } else return; int modebitsize = GET_MODE_BITSIZE (TYPE_MODE (type)); HOST_WIDE_INT bitsize, bitpos; tree offset; enum machine_mode mode; int volatilep = 0, unsignedp = 0; tree base = get_inner_reference (rhs, &bitsize, &bitpos, &offset, &mode, &unsignedp, &volatilep, false); tree utype = build_nonstandard_integer_type (modebitsize, 1); if ((TREE_CODE (base) == VAR_DECL && DECL_HARD_REGISTER (base)) || (bitpos % modebitsize) != 0 || bitsize != modebitsize || GET_MODE_BITSIZE (TYPE_MODE (utype)) != modebitsize || TREE_CODE (gimple_assign_lhs (stmt)) != SSA_NAME) return; location_t loc = gimple_location (stmt); tree ptype = build_pointer_type (TREE_TYPE (rhs)); tree atype = reference_alias_ptr_type (rhs); gimple g = gimple_build_assign (make_ssa_name (ptype, NULL), build_fold_addr_expr (rhs)); gimple_set_location (g, loc); gsi_insert_before (gsi, g, GSI_SAME_STMT); tree mem = build2 (MEM_REF, utype, gimple_assign_lhs (g), build_int_cst (atype, 0)); tree urhs = make_ssa_name (utype, NULL); g = gimple_build_assign (urhs, mem); gimple_set_location (g, loc); gsi_insert_before (gsi, g, GSI_SAME_STMT); minv = fold_convert (utype, minv); maxv = fold_convert (utype, maxv); if (!integer_zerop (minv)) { g = gimple_build_assign_with_ops (MINUS_EXPR, make_ssa_name (utype, NULL), urhs, minv); gimple_set_location (g, loc); gsi_insert_before (gsi, g, GSI_SAME_STMT); } gimple_stmt_iterator gsi2 = *gsi; basic_block then_bb, fallthru_bb; *gsi = create_cond_insert_point (gsi, true, false, true, &then_bb, &fallthru_bb); g = gimple_build_cond (GT_EXPR, gimple_assign_lhs (g), int_const_binop (MINUS_EXPR, maxv, minv), NULL_TREE, NULL_TREE); gimple_set_location (g, loc); gsi_insert_after (gsi, g, GSI_NEW_STMT); gimple_assign_set_rhs_with_ops (&gsi2, NOP_EXPR, urhs, NULL_TREE); update_stmt (stmt); gsi2 = gsi_after_labels (then_bb); if (flag_sanitize_undefined_trap_on_error) g = gimple_build_call (builtin_decl_explicit (BUILT_IN_TRAP), 0); else { tree data = ubsan_create_data ("__ubsan_invalid_value_data", &loc, NULL, ubsan_type_descriptor (type), NULL_TREE); data = build_fold_addr_expr_loc (loc, data); enum built_in_function bcode = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE : BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT; tree fn = builtin_decl_explicit (bcode); tree val = force_gimple_operand_gsi (&gsi2, ubsan_encode_value (urhs), true, NULL_TREE, true, GSI_SAME_STMT); g = gimple_build_call (fn, 2, data, val); } gimple_set_location (g, loc); gsi_insert_before (&gsi2, g, GSI_SAME_STMT); } /* Instrument float point-to-integer conversion. TYPE is an integer type of destination, EXPR is floating-point expression. */ tree ubsan_instrument_float_cast (location_t loc, tree type, tree expr) { tree expr_type = TREE_TYPE (expr); tree t, tt, fn, min, max; enum machine_mode mode = TYPE_MODE (expr_type); int prec = TYPE_PRECISION (type); bool uns_p = TYPE_UNSIGNED (type); /* Float to integer conversion first truncates toward zero, so even signed char c = 127.875f; is not problematic. Therefore, we should complain only if EXPR is unordered or smaller or equal than TYPE_MIN_VALUE - 1.0 or greater or equal than TYPE_MAX_VALUE + 1.0. */ if (REAL_MODE_FORMAT (mode)->b == 2) { /* For maximum, TYPE_MAX_VALUE might not be representable in EXPR_TYPE, e.g. if TYPE is 64-bit long long and EXPR_TYPE is IEEE single float, but TYPE_MAX_VALUE + 1.0 is either representable or infinity. */ REAL_VALUE_TYPE maxval = dconst1; SET_REAL_EXP (&maxval, REAL_EXP (&maxval) + prec - !uns_p); real_convert (&maxval, mode, &maxval); max = build_real (expr_type, maxval); /* For unsigned, assume -1.0 is always representable. */ if (uns_p) min = build_minus_one_cst (expr_type); else { /* TYPE_MIN_VALUE is generally representable (or -inf), but TYPE_MIN_VALUE - 1.0 might not be. */ REAL_VALUE_TYPE minval = dconstm1, minval2; SET_REAL_EXP (&minval, REAL_EXP (&minval) + prec - 1); real_convert (&minval, mode, &minval); real_arithmetic (&minval2, MINUS_EXPR, &minval, &dconst1); real_convert (&minval2, mode, &minval2); if (real_compare (EQ_EXPR, &minval, &minval2) && !real_isinf (&minval)) { /* If TYPE_MIN_VALUE - 1.0 is not representable and rounds to TYPE_MIN_VALUE, we need to subtract more. As REAL_MODE_FORMAT (mode)->p is the number of base digits, we want to subtract a number that will be 1 << (REAL_MODE_FORMAT (mode)->p - 1) times smaller than minval. */ minval2 = dconst1; gcc_assert (prec > REAL_MODE_FORMAT (mode)->p); SET_REAL_EXP (&minval2, REAL_EXP (&minval2) + prec - 1 - REAL_MODE_FORMAT (mode)->p + 1); real_arithmetic (&minval2, MINUS_EXPR, &minval, &minval2); real_convert (&minval2, mode, &minval2); } min = build_real (expr_type, minval2); } } else if (REAL_MODE_FORMAT (mode)->b == 10) { /* For _Decimal128 up to 34 decimal digits, - sign, dot, e, exponent. */ char buf[64]; mpfr_t m; int p = REAL_MODE_FORMAT (mode)->p; REAL_VALUE_TYPE maxval, minval; /* Use mpfr_snprintf rounding to compute the smallest representable decimal number greater or equal than 1 << (prec - !uns_p). */ mpfr_init2 (m, prec + 2); mpfr_set_ui_2exp (m, 1, prec - !uns_p, GMP_RNDN); mpfr_snprintf (buf, sizeof buf, "%.*RUe", p - 1, m); decimal_real_from_string (&maxval, buf); max = build_real (expr_type, maxval); /* For unsigned, assume -1.0 is always representable. */ if (uns_p) min = build_minus_one_cst (expr_type); else { /* Use mpfr_snprintf rounding to compute the largest representable decimal number less or equal than (-1 << (prec - 1)) - 1. */ mpfr_set_si_2exp (m, -1, prec - 1, GMP_RNDN); mpfr_sub_ui (m, m, 1, GMP_RNDN); mpfr_snprintf (buf, sizeof buf, "%.*RDe", p - 1, m); decimal_real_from_string (&minval, buf); min = build_real (expr_type, minval); } mpfr_clear (m); } else return NULL_TREE; if (flag_sanitize_undefined_trap_on_error) fn = build_call_expr_loc (loc, builtin_decl_explicit (BUILT_IN_TRAP), 0); else { /* Create the __ubsan_handle_float_cast_overflow fn call. */ tree data = ubsan_create_data ("__ubsan_float_cast_overflow_data", NULL, NULL, ubsan_type_descriptor (expr_type), ubsan_type_descriptor (type), NULL_TREE); enum built_in_function bcode = flag_sanitize_recover ? BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW : BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT; fn = builtin_decl_explicit (bcode); fn = build_call_expr_loc (loc, fn, 2, build_fold_addr_expr_loc (loc, data), ubsan_encode_value (expr, false)); } t = fold_build2 (UNLE_EXPR, boolean_type_node, expr, min); tt = fold_build2 (UNGE_EXPR, boolean_type_node, expr, max); return fold_build3 (COND_EXPR, void_type_node, fold_build2 (TRUTH_OR_EXPR, boolean_type_node, t, tt), fn, integer_zero_node); } namespace { const pass_data pass_data_ubsan = { GIMPLE_PASS, /* type */ "ubsan", /* name */ OPTGROUP_NONE, /* optinfo_flags */ true, /* has_execute */ TV_TREE_UBSAN, /* tv_id */ ( PROP_cfg | PROP_ssa ), /* properties_required */ 0, /* properties_provided */ 0, /* properties_destroyed */ 0, /* todo_flags_start */ TODO_update_ssa, /* todo_flags_finish */ }; class pass_ubsan : public gimple_opt_pass { public: pass_ubsan (gcc::context *ctxt) : gimple_opt_pass (pass_data_ubsan, ctxt) {} /* opt_pass methods: */ virtual bool gate (function *) { return flag_sanitize & (SANITIZE_NULL | SANITIZE_SI_OVERFLOW | SANITIZE_BOOL | SANITIZE_ENUM); } virtual unsigned int execute (function *); }; // class pass_ubsan unsigned int pass_ubsan::execute (function *fun) { basic_block bb; gimple_stmt_iterator gsi; initialize_sanitizer_builtins (); FOR_EACH_BB_FN (bb, fun) { for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi);) { gimple stmt = gsi_stmt (gsi); if (is_gimple_debug (stmt) || gimple_clobber_p (stmt)) { gsi_next (&gsi); continue; } if ((flag_sanitize & SANITIZE_SI_OVERFLOW) && is_gimple_assign (stmt)) instrument_si_overflow (gsi); if (flag_sanitize & SANITIZE_NULL) { if (gimple_store_p (stmt)) instrument_null (gsi, true); if (gimple_assign_load_p (stmt)) instrument_null (gsi, false); } if (flag_sanitize & (SANITIZE_BOOL | SANITIZE_ENUM) && gimple_assign_load_p (stmt)) instrument_bool_enum_load (&gsi); gsi_next (&gsi); } } return 0; } } // anon namespace gimple_opt_pass * make_pass_ubsan (gcc::context *ctxt) { return new pass_ubsan (ctxt); } #include "gt-ubsan.h"