binary/elf/symbolization

ELF-specific rules and relations for symbolization

reg_has_got(EA:address, Reg:register)

Value of register ‘Reg’ at address ‘EA’ is a GOT-relative base reference.

• Uses: got_section, loaded_section

• Used by: symbol_minus_symbol, symbolic_expr, symbolic_operand_attribute

• Recursive: cmp_defines, first_block_in_byte_interval, overlapping_instruction, split_load_for_symbolization, relocation_adjustment, call_tls_get_addr, simple_data_access_pattern, reg_def_use.def_used, base_relative_operand, split_load_total_points, block, litpool_confidence, reg_has_got, value_reg_limit, stack_def_use.live_var_used, split_load_point, split_load_operand, flags_and_jump_pair, wis_schedule_iter, unresolved_interval, code_in_block, compare_and_jump_indirect_op_valid, __agg_subclause3, inferred_main_in_reg, reg_used_for, is_padding, resolved_reaches, compare_and_jump_indirect, __agg_subclause2, padding_block_candidate, reg_reg_arithmetic_operation_defs, def_used_for_address, impossible_block, plt_block, arm_jump_table_block_instruction, composite_data_access, possible_target_from, jump_table_candidate_refined, base_relative_operation, split_load, __agg_single3, reg_has_base_image, no_value_reg_limit, relative_address_start, wis_schedule, plt_entry, reg_def_use.used_in_block, adjusts_stack_in_block, possible_target, initialized_data_segment, reg_def_use.live_var_at_block_end, cmp_reg_to_reg, indefinite_litpool_ref, negative_block_heuristic, contains_implausible_instr_seq, indexed_pc_relative_load, next_block_in_byte_interval, code_in_block_candidate, symbolic_expr_from_relocation, unlikely_have_symbolic_immediate, incomplete_block, arm_jump_table_data_block_limit, inferred_main_dispatch, arch.simple_data_load, data_block_limit, block_total_points, invalid_jump_table_candidate, stack_def_use.live_var_def, reg_def_use.live_var_def, arm_jump_table_block_start, block_overlap, stack_base_reg_move, may_fallthrough, __agg_single2, block_heuristic, __agg_subclause9, block_implies_block, segment_target_range, stack_def_use.live_var_at_prior_used, cinf_ldr_add_pc, no_return_call_propagated, unresolved_block_overlap, jump_table_max, common_tail, unresolved_interval_order, tls_get_addr, block_points_proportional, likely_fallthrough, block_boundaries, jump_table_element_access, block_points, invalid, jump_table_target, reg_def_use.used, arm_jump_table_candidate, reg_def_use.live_var_at_prior_used, no_return_call, arm_jump_table_skip_first_entry, stack_def_use.defined_in_block, split_load_conflict, arm_jump_table_candidate_start, stack_def_use.live_var_at_block_end, reg_def_use.last_def_in_block, arm_jump_table_cmp_limit, data_in_code, candidate_block_is_not_padding, inter_procedural_edge, known_block, instruction_memory_access_size, overlap_with_litpool, block_limit, function_inference.function_entry_initial, init_symbol_minus_symbol_candidate_arm, const_value_reg_used, self_contained_segment, stack_def_use.last_def_in_block, must_fallthrough, straight_line_last_def, padding_block_limit, jump_table_candidate, no_return_call_refined, block_candidate_dependency_edge, split_load_candidate, stack_def_use.block_last_def, last_value_reg_limit, relative_address, no_return_block, litpool_ref, symbol_minus_symbol_litpool_access_pattern, block_last_instruction, gp_relative_operand, data_in_code_propagate, block_candidate_boundaries, basic_target, compare_and_jump_immediate, wis_memo, __agg_subclause6, nop_in_padding_candidate, reg_def_use.defined_in_block, __agg_single6, relocation_adjustment_total, candidate_block_is_padding, wis_prior, reg_def_use.live_var_used, __agg_subclause7, wis_has_prior, compare_and_jump_register, block_instruction_next, arm_jump_table_data_block, value_reg, contains_plausible_instr_seq, relative_jump_table_entry_candidate, adrp_used, stack_def_use.def_used, got_relative_operand, __agg_subclause10, value_reg_edge, arch.reg_relative_load, reg_def_use.flow_def, hi_load_prop, jump_table_start, branch_to_calculated_pc_rel_addr, tls_desc_call, discarded_block, stack_def_use.live_var_used_in_block, data_segment, value_reg_unsupported, block_next, discarded_split_load, indexed_pc_relative_load_relative, litpool_symbolic_operand, data_block_candidate, reg_def_use.return_block_end, code_in_block_candidate_refined, after_end, jump_table_signed, reg_def_use.block_last_def, base_relative_jump, unresolved_interval_best_block, straight_line_def_used, unresolved_block, data_access, correlated_live_reg, jump_table_prelude, litpool_boundaries, init_ldr_add_pc, arch.extend_load, stack_def_use.used_in_block, stack_def_use.ref_in_block, reg_def_use.ref_in_block, reg_def_use.return_val_used

got_reference_pointer(EA:address)

Address ‘EA’ is used as a GOT-relative base reference. ‘EA’ is the beginning of either the .got or .got.plt section.

• Uses: loaded_section

• Used by: got_relative_operand, inferred_special_symbol, jump_table_start, plt_entry, symbol_minus_symbol, symbolic_expr, symbolic_operand_attribute

got_relative_operand(EA:address, Index:operand_index, Dest:address)

Operand instruction at ‘EA’ at index ‘Index’ is a GOT-relative reference to an address ‘Dest’.

• Uses: arch.reg_reg_arithmetic_operation, got_reference_pointer, instruction_displacement_offset, instruction_get_op, op_immediate_and_reg, op_indirect_mapped, relocation, symbol

• Used by: moved_label_candidate, moved_label_class, symbolic_operand_attribute, symbolic_operand_candidate, tls_global_dynamic

• Recursive: cmp_defines, first_block_in_byte_interval, overlapping_instruction, split_load_for_symbolization, relocation_adjustment, call_tls_get_addr, simple_data_access_pattern, reg_def_use.def_used, base_relative_operand, split_load_total_points, block, litpool_confidence, reg_has_got, value_reg_limit, stack_def_use.live_var_used, split_load_point, split_load_operand, flags_and_jump_pair, wis_schedule_iter, unresolved_interval, code_in_block, compare_and_jump_indirect_op_valid, __agg_subclause3, inferred_main_in_reg, reg_used_for, is_padding, resolved_reaches, compare_and_jump_indirect, __agg_subclause2, padding_block_candidate, reg_reg_arithmetic_operation_defs, def_used_for_address, impossible_block, plt_block, arm_jump_table_block_instruction, composite_data_access, possible_target_from, jump_table_candidate_refined, base_relative_operation, split_load, __agg_single3, reg_has_base_image, no_value_reg_limit, relative_address_start, wis_schedule, plt_entry, reg_def_use.used_in_block, adjusts_stack_in_block, possible_target, initialized_data_segment, reg_def_use.live_var_at_block_end, cmp_reg_to_reg, indefinite_litpool_ref, negative_block_heuristic, contains_implausible_instr_seq, indexed_pc_relative_load, next_block_in_byte_interval, code_in_block_candidate, symbolic_expr_from_relocation, unlikely_have_symbolic_immediate, incomplete_block, arm_jump_table_data_block_limit, inferred_main_dispatch, arch.simple_data_load, data_block_limit, block_total_points, invalid_jump_table_candidate, stack_def_use.live_var_def, reg_def_use.live_var_def, arm_jump_table_block_start, block_overlap, stack_base_reg_move, may_fallthrough, __agg_single2, block_heuristic, __agg_subclause9, block_implies_block, segment_target_range, stack_def_use.live_var_at_prior_used, cinf_ldr_add_pc, no_return_call_propagated, unresolved_block_overlap, jump_table_max, common_tail, unresolved_interval_order, tls_get_addr, block_points_proportional, likely_fallthrough, block_boundaries, jump_table_element_access, block_points, invalid, jump_table_target, reg_def_use.used, arm_jump_table_candidate, reg_def_use.live_var_at_prior_used, no_return_call, arm_jump_table_skip_first_entry, stack_def_use.defined_in_block, split_load_conflict, arm_jump_table_candidate_start, stack_def_use.live_var_at_block_end, reg_def_use.last_def_in_block, arm_jump_table_cmp_limit, data_in_code, candidate_block_is_not_padding, inter_procedural_edge, known_block, instruction_memory_access_size, overlap_with_litpool, block_limit, function_inference.function_entry_initial, init_symbol_minus_symbol_candidate_arm, const_value_reg_used, self_contained_segment, stack_def_use.last_def_in_block, must_fallthrough, straight_line_last_def, padding_block_limit, jump_table_candidate, no_return_call_refined, block_candidate_dependency_edge, split_load_candidate, stack_def_use.block_last_def, last_value_reg_limit, relative_address, no_return_block, litpool_ref, symbol_minus_symbol_litpool_access_pattern, block_last_instruction, gp_relative_operand, data_in_code_propagate, block_candidate_boundaries, basic_target, compare_and_jump_immediate, wis_memo, __agg_subclause6, nop_in_padding_candidate, reg_def_use.defined_in_block, __agg_single6, relocation_adjustment_total, candidate_block_is_padding, wis_prior, reg_def_use.live_var_used, __agg_subclause7, wis_has_prior, compare_and_jump_register, block_instruction_next, arm_jump_table_data_block, value_reg, contains_plausible_instr_seq, relative_jump_table_entry_candidate, adrp_used, stack_def_use.def_used, got_relative_operand, __agg_subclause10, value_reg_edge, arch.reg_relative_load, reg_def_use.flow_def, hi_load_prop, jump_table_start, branch_to_calculated_pc_rel_addr, tls_desc_call, discarded_block, stack_def_use.live_var_used_in_block, data_segment, value_reg_unsupported, block_next, discarded_split_load, indexed_pc_relative_load_relative, litpool_symbolic_operand, data_block_candidate, reg_def_use.return_block_end, code_in_block_candidate_refined, after_end, jump_table_signed, reg_def_use.block_last_def, base_relative_jump, unresolved_interval_best_block, straight_line_def_used, unresolved_block, data_access, correlated_live_reg, jump_table_prelude, litpool_boundaries, init_ldr_add_pc, arch.extend_load, stack_def_use.used_in_block, stack_def_use.ref_in_block, reg_def_use.ref_in_block, reg_def_use.return_val_used

start_function(ea:address)

• Uses: binary_format, entry_point, function_symbol

• Used by: __agg_subclause2, __agg_subclause3, function_inference.function_entry_initial, inferred_special_symbol

do_infer_main_function()

Decide whether the main function should be inferred.

WARNING: Predicate not present in compiled Datalog program (Dead Code)

inferred_main_dispatch(EA:address)

Locate where main() is dispatched (i.e., where __libc_start_main is called.)

• Uses: binary_format, binary_type, function_symbol

• Recursive: cmp_defines, first_block_in_byte_interval, overlapping_instruction, split_load_for_symbolization, relocation_adjustment, call_tls_get_addr, simple_data_access_pattern, reg_def_use.def_used, base_relative_operand, split_load_total_points, block, litpool_confidence, reg_has_got, value_reg_limit, stack_def_use.live_var_used, split_load_point, split_load_operand, flags_and_jump_pair, wis_schedule_iter, unresolved_interval, code_in_block, compare_and_jump_indirect_op_valid, __agg_subclause3, inferred_main_in_reg, reg_used_for, is_padding, resolved_reaches, compare_and_jump_indirect, __agg_subclause2, padding_block_candidate, reg_reg_arithmetic_operation_defs, def_used_for_address, impossible_block, plt_block, arm_jump_table_block_instruction, composite_data_access, possible_target_from, jump_table_candidate_refined, base_relative_operation, split_load, __agg_single3, reg_has_base_image, no_value_reg_limit, relative_address_start, wis_schedule, plt_entry, reg_def_use.used_in_block, adjusts_stack_in_block, possible_target, initialized_data_segment, reg_def_use.live_var_at_block_end, cmp_reg_to_reg, indefinite_litpool_ref, negative_block_heuristic, contains_implausible_instr_seq, indexed_pc_relative_load, next_block_in_byte_interval, code_in_block_candidate, symbolic_expr_from_relocation, unlikely_have_symbolic_immediate, incomplete_block, arm_jump_table_data_block_limit, inferred_main_dispatch, arch.simple_data_load, data_block_limit, block_total_points, invalid_jump_table_candidate, stack_def_use.live_var_def, reg_def_use.live_var_def, arm_jump_table_block_start, block_overlap, stack_base_reg_move, may_fallthrough, __agg_single2, block_heuristic, __agg_subclause9, block_implies_block, segment_target_range, stack_def_use.live_var_at_prior_used, cinf_ldr_add_pc, no_return_call_propagated, unresolved_block_overlap, jump_table_max, common_tail, unresolved_interval_order, tls_get_addr, block_points_proportional, likely_fallthrough, block_boundaries, jump_table_element_access, block_points, invalid, jump_table_target, reg_def_use.used, arm_jump_table_candidate, reg_def_use.live_var_at_prior_used, no_return_call, arm_jump_table_skip_first_entry, stack_def_use.defined_in_block, split_load_conflict, arm_jump_table_candidate_start, stack_def_use.live_var_at_block_end, reg_def_use.last_def_in_block, arm_jump_table_cmp_limit, data_in_code, candidate_block_is_not_padding, inter_procedural_edge, known_block, instruction_memory_access_size, overlap_with_litpool, block_limit, function_inference.function_entry_initial, init_symbol_minus_symbol_candidate_arm, const_value_reg_used, self_contained_segment, stack_def_use.last_def_in_block, must_fallthrough, straight_line_last_def, padding_block_limit, jump_table_candidate, no_return_call_refined, block_candidate_dependency_edge, split_load_candidate, stack_def_use.block_last_def, last_value_reg_limit, relative_address, no_return_block, litpool_ref, symbol_minus_symbol_litpool_access_pattern, block_last_instruction, gp_relative_operand, data_in_code_propagate, block_candidate_boundaries, basic_target, compare_and_jump_immediate, wis_memo, __agg_subclause6, nop_in_padding_candidate, reg_def_use.defined_in_block, __agg_single6, relocation_adjustment_total, candidate_block_is_padding, wis_prior, reg_def_use.live_var_used, __agg_subclause7, wis_has_prior, compare_and_jump_register, block_instruction_next, arm_jump_table_data_block, value_reg, contains_plausible_instr_seq, relative_jump_table_entry_candidate, adrp_used, stack_def_use.def_used, got_relative_operand, __agg_subclause10, value_reg_edge, arch.reg_relative_load, reg_def_use.flow_def, hi_load_prop, jump_table_start, branch_to_calculated_pc_rel_addr, tls_desc_call, discarded_block, stack_def_use.live_var_used_in_block, data_segment, value_reg_unsupported, block_next, discarded_split_load, indexed_pc_relative_load_relative, litpool_symbolic_operand, data_block_candidate, reg_def_use.return_block_end, code_in_block_candidate_refined, after_end, jump_table_signed, reg_def_use.block_last_def, base_relative_jump, unresolved_interval_best_block, straight_line_def_used, unresolved_block, data_access, correlated_live_reg, jump_table_prelude, litpool_boundaries, init_ldr_add_pc, arch.extend_load, stack_def_use.used_in_block, stack_def_use.ref_in_block, reg_def_use.ref_in_block, reg_def_use.return_val_used

inferred_main_in_reg(EA:address, Reg:register)

The location and register where the address of main is loaded.

• Uses: arch.integer_reg_param, instruction, instruction_get_src_op, next, op_indirect_contains_reg, op_regdirect_contains_reg

• Used by: inferred_main_function

• Recursive: cmp_defines, first_block_in_byte_interval, overlapping_instruction, split_load_for_symbolization, relocation_adjustment, call_tls_get_addr, simple_data_access_pattern, reg_def_use.def_used, base_relative_operand, split_load_total_points, block, litpool_confidence, reg_has_got, value_reg_limit, stack_def_use.live_var_used, split_load_point, split_load_operand, flags_and_jump_pair, wis_schedule_iter, unresolved_interval, code_in_block, compare_and_jump_indirect_op_valid, __agg_subclause3, inferred_main_in_reg, reg_used_for, is_padding, resolved_reaches, compare_and_jump_indirect, __agg_subclause2, padding_block_candidate, reg_reg_arithmetic_operation_defs, def_used_for_address, impossible_block, plt_block, arm_jump_table_block_instruction, composite_data_access, possible_target_from, jump_table_candidate_refined, base_relative_operation, split_load, __agg_single3, reg_has_base_image, no_value_reg_limit, relative_address_start, wis_schedule, plt_entry, reg_def_use.used_in_block, adjusts_stack_in_block, possible_target, initialized_data_segment, reg_def_use.live_var_at_block_end, cmp_reg_to_reg, indefinite_litpool_ref, negative_block_heuristic, contains_implausible_instr_seq, indexed_pc_relative_load, next_block_in_byte_interval, code_in_block_candidate, symbolic_expr_from_relocation, unlikely_have_symbolic_immediate, incomplete_block, arm_jump_table_data_block_limit, inferred_main_dispatch, arch.simple_data_load, data_block_limit, block_total_points, invalid_jump_table_candidate, stack_def_use.live_var_def, reg_def_use.live_var_def, arm_jump_table_block_start, block_overlap, stack_base_reg_move, may_fallthrough, __agg_single2, block_heuristic, __agg_subclause9, block_implies_block, segment_target_range, stack_def_use.live_var_at_prior_used, cinf_ldr_add_pc, no_return_call_propagated, unresolved_block_overlap, jump_table_max, common_tail, unresolved_interval_order, tls_get_addr, block_points_proportional, likely_fallthrough, block_boundaries, jump_table_element_access, block_points, invalid, jump_table_target, reg_def_use.used, arm_jump_table_candidate, reg_def_use.live_var_at_prior_used, no_return_call, arm_jump_table_skip_first_entry, stack_def_use.defined_in_block, split_load_conflict, arm_jump_table_candidate_start, stack_def_use.live_var_at_block_end, reg_def_use.last_def_in_block, arm_jump_table_cmp_limit, data_in_code, candidate_block_is_not_padding, inter_procedural_edge, known_block, instruction_memory_access_size, overlap_with_litpool, block_limit, function_inference.function_entry_initial, init_symbol_minus_symbol_candidate_arm, const_value_reg_used, self_contained_segment, stack_def_use.last_def_in_block, must_fallthrough, straight_line_last_def, padding_block_limit, jump_table_candidate, no_return_call_refined, block_candidate_dependency_edge, split_load_candidate, stack_def_use.block_last_def, last_value_reg_limit, relative_address, no_return_block, litpool_ref, symbol_minus_symbol_litpool_access_pattern, block_last_instruction, gp_relative_operand, data_in_code_propagate, block_candidate_boundaries, basic_target, compare_and_jump_immediate, wis_memo, __agg_subclause6, nop_in_padding_candidate, reg_def_use.defined_in_block, __agg_single6, relocation_adjustment_total, candidate_block_is_padding, wis_prior, reg_def_use.live_var_used, __agg_subclause7, wis_has_prior, compare_and_jump_register, block_instruction_next, arm_jump_table_data_block, value_reg, contains_plausible_instr_seq, relative_jump_table_entry_candidate, adrp_used, stack_def_use.def_used, got_relative_operand, __agg_subclause10, value_reg_edge, arch.reg_relative_load, reg_def_use.flow_def, hi_load_prop, jump_table_start, branch_to_calculated_pc_rel_addr, tls_desc_call, discarded_block, stack_def_use.live_var_used_in_block, data_segment, value_reg_unsupported, block_next, discarded_split_load, indexed_pc_relative_load_relative, litpool_symbolic_operand, data_block_candidate, reg_def_use.return_block_end, code_in_block_candidate_refined, after_end, jump_table_signed, reg_def_use.block_last_def, base_relative_jump, unresolved_interval_best_block, straight_line_def_used, unresolved_block, data_access, correlated_live_reg, jump_table_prelude, litpool_boundaries, init_ldr_add_pc, arch.extend_load, stack_def_use.used_in_block, stack_def_use.ref_in_block, reg_def_use.ref_in_block, reg_def_use.return_val_used

inferred_main_function(Main_location:address)

Infer the location of the main function.

• Uses: address_in_data, code_in_block, inferred_main_in_reg, instruction_get_op, op_indirect_contains_reg, reg_def_use.def_used, value_reg

• Recursive: symbolic_expr_symbol_minus_symbol, symbolic_expr, +disconnected3, data_object_conflict, labeled_data_candidate, code_pointer_in_data, +disconnected2, labeled_ea, code_in_refined_block, moved_label, block_needs_merging, best_symexpr_symbol, discarded_jump_table_entry, best_func_symbol, data_limit, moved_displacement_candidate, symbol_minus_symbol, refined_block, +disconnected1, moved_data_label, next_address_in_data, symbolic_expr_attribute, relative_jump_table_entry, inferred_main_function, symbol_minus_symbol_candidate, next_data_limit, main_function, symbolic_data, symbolic_operand_point, resolved_transfer, code_in_split_block, base_relative_symbolic_operand, address_array_aux, data_access_limit, symbol_minus_symbol_from_relocation, string_candidate_refined, split_block, +disconnected6, data_object_total_points, string_candidate, data_object, address_array, discarded_data_object, block_needs_splitting_at, inferred_special_symbol, symbol_score, after_address_in_data, symbolic_operand_attribute, data_object_point, moved_pc_relative_candidate, symbolic_operand, function_inference.function_entry, jump_table, data_limit_after_access, boundary_sym_expr, data_object_candidate, label_conflict, got_reference, moved_label_candidate, preferred_data_access

reloc_type_relpc(Type:symbol)

Type of relocation relevant to pc-relative attribute

• Used by: relocation_adjustment

reloc_type_with_name(Type:symbol)

Type of relocation possibly with non-empty Symbol name

• Used by: symbolic_expr_from_relocation

reloc_type_nameless(Type:symbol)

Type of relocation possibly with empty Symbol name

• Used by: symbolic_expr_from_relocation

copy_relocated_symbol(EA:address, Name:symbol)

Collect all the symbols whose address, size and type are the same as the symbol with COPY relocation.

• Uses: relocation, symbol

• Recursive: copy_relocated_symbol

best_ifunc_symbol(EA:address, SymbolName:symbol)

Best GNU_IFUNC symbol for plt entries. Uses ifunc_symbol_score to deterministically select a single ifunc symbol for each EA.

• Uses: ifunc_symbol_score, relocation, symbol

• Used by: plt_entry

ifunc_symbol_score(EA:address, SymbolName:symbol, Score:unsigned)

Provide a score to GNU_IFUNC symbols based on scope, visibility, and their table index.

• Uses: defined_symbol, ifunc_scope_score, symbol_visibility_score

• Used by: best_ifunc_symbol

ifunc_scope_score(Scope:symbol, Score:unsigned)

Auxiliary predicate to define ifunc_symbol_score Prefer symbols with LOCAL over GLOBAL, and with GLOBAl over anything else (WEAK).

• Uses: symbol

• Used by: ifunc_symbol_score