angr 9.2.192__cp311-cp311-macosx_10_12_x86_64.whl

angr/analyses/decompiler/structuring/dream.py

@@ -0,0 +1,1217 @@
+# pylint:disable=multiple-statements,line-too-long,consider-using-enumerate
+from __future__ import annotations
+from typing import Any, TYPE_CHECKING
+import logging
+from collections import defaultdict, OrderedDict
+
+import networkx
+
+import claripy
+import angr.ailment as ailment
+
+from angr.utils.graph import GraphUtils
+from angr.knowledge_plugins.cfg import IndirectJumpType
+from angr.analyses.decompiler.graph_region import GraphRegion
+from angr.analyses.decompiler.empty_node_remover import EmptyNodeRemover
+from angr.analyses.decompiler.jumptable_entry_condition_rewriter import JumpTableEntryConditionRewriter
+from angr.analyses.decompiler.condition_processor import ConditionProcessor
+from angr.analyses.decompiler.region_simplifiers.cascading_cond_transformer import CascadingConditionTransformer
+from angr.analyses.decompiler.utils import (
+    extract_jump_targets,
+    get_ast_subexprs,
+    switch_extract_cmp_bounds,
+    remove_last_statement,
+    first_nonlabel_nonphi_node,
+)
+from .structurer_nodes import (
+    SequenceNode,
+    CodeNode,
+    ConditionNode,
+    ConditionalBreakNode,
+    LoopNode,
+    SwitchCaseNode,
+    BreakNode,
+    ContinueNode,
+    MultiNode,
+    CascadingConditionNode,
+    BaseNode,
+    EmptyBlockNotice,
+)
+from .structurer_base import StructurerBase
+
+
+if TYPE_CHECKING:
+    from angr.knowledge_plugins.functions import Function
+
+l = logging.getLogger(name=__name__)
+
+
+#
+# The main analysis
+#
+
+
+class DreamStructurer(StructurerBase):
+    """
+    Structure a region using a structuring algorithm that is similar to the one in Dream decompiler (described in the
+    "no more gotos" paper). Note that this implementation has quite a few improvements over the original described
+    version and *should not* be used to evaluate the performance of the original algorithm described in that paper.
+
+    The current function graph is provided so that we can detect certain edge cases, for example, jump table entries no
+    longer exist due to empty node removal during structuring or prior steps.
+    """
+
+    NAME = "dream"
+
+    def __init__(
+        self,
+        region,
+        parent_map=None,
+        condition_processor=None,
+        func: Function | None = None,
+        case_entry_to_switch_head: dict[int, int] | None = None,
+        parent_region=None,
+        **kwargs,
+    ):
+        super().__init__(
+            region,
+            parent_map=parent_map,
+            condition_processor=condition_processor,
+            func=func,
+            case_entry_to_switch_head=case_entry_to_switch_head,
+            parent_region=parent_region,
+            **kwargs,
+        )
+
+        self._analyze()
+
+    def _analyze(self):
+        has_cycle = self._has_cycle()
+        # sanity checks
+        if self._region.cyclic:
+            if not has_cycle:
+                l.critical(
+                    "Region %r is supposed to be a cyclic region but there is no cycle inside. This is usually "
+                    "due to the existence of loop headers with more than one in-edges, which angr decompiler "
+                    "does not support yet. The decompilation result will be wrong.",
+                    self._region,
+                )
+            self._analyze_cyclic()
+        else:
+            if has_cycle:
+                l.critical(
+                    "Region %r is supposed to be an acyclic region but there are cycles inside. This is usually "
+                    "due to the existence of loop headers with more than one in-edges, which angr decompiler "
+                    "does not support yet. The decompilation result will be wrong.",
+                    self._region,
+                )
+            self._analyze_acyclic()
+
+    def _analyze_cyclic(self):
+        loop_head = self._region.head
+
+        loop_subgraph = self._region.graph
+        successors = self._region.successors
+        assert successors is not None
+
+        assert len(successors) <= 1
+
+        loop_node = self._make_endless_loop(loop_head, loop_subgraph, successors)
+
+        loop_node = self._refine_loop(loop_node)
+
+        seq = SequenceNode(
+            loop_head.addr, nodes=[loop_node] + [succ for succ in successors if succ in self._region.graph]
+        )
+
+        self.result = seq
+
+    def _analyze_acyclic(self):
+        # let's generate conditions first
+        self.cond_proc.recover_reaching_conditions(
+            self._region, with_successors=True, case_entry_to_switch_head=self._case_entry_to_switch_head
+        )
+
+        # make the sequence node and pack reaching conditions into CodeNode instances
+        seq = self._make_sequence()
+
+        self._new_sequences.append(seq)
+
+        while self._new_sequences:
+            seq_ = self._new_sequences.pop(0)
+            if len(seq_.nodes) <= 1:
+                continue
+            self._structure_sequence(seq_)
+
+        seq = EmptyNodeRemover(seq).result
+
+        # unpack nodes and remove CodeNode wrappers
+        seq = self._unpack_sequence(seq)
+
+        self.result = seq
+
+    def _find_loop_nodes_and_successors(self):
+        graph = self._region.graph
+        head = self._region.head
+
+        # find initial loop nodes
+        loop_nodes = None
+        components = networkx.strongly_connected_components(graph)
+        for component in components:
+            if head in component:
+                loop_nodes = component
+                break
+        if loop_nodes is None:
+            # this should never happen - loop head always forms a cycle
+            raise TypeError("A bug (impossible case) in the algorithm is triggered.")
+
+        # extend loop nodes
+        while True:
+            loop_nodes_updated = False
+            for loop_node in loop_nodes:
+                for succ in graph.successors(loop_node):
+                    if succ not in loop_nodes:
+                        # determine if this successor's all predecessors are in the loop
+                        predecessors = graph.predecessors(succ)
+                        if all(pred in loop_nodes for pred in predecessors):
+                            # yes!
+                            loop_nodes.add(succ)
+                            loop_nodes_updated = True
+                            break
+                if loop_nodes_updated:
+                    break
+            if not loop_nodes_updated:
+                break
+
+        # find loop nodes and successors
+        loop_subgraph = networkx.subgraph(graph, loop_nodes)
+        loop_node_addrs = {node.addr for node in loop_subgraph}
+
+        # Case A: The loop successor is inside the current region (does it happen at all?)
+        loop_successors = set()
+
+        for node, successors in networkx.bfs_successors(graph, head):
+            if node.addr in loop_node_addrs:
+                for suc in successors:
+                    if suc not in loop_subgraph:
+                        loop_successors.add(suc)
+
+        # Case B: The loop successor is the successor to this region in the parent graph
+        if not loop_successors and self._parent_map is not None:
+            current_region = self._region
+            parent_region = self._parent_map.get(current_region, None)
+            while parent_region and not loop_successors:
+                parent_graph = parent_region.graph
+                for node, successors in networkx.bfs_successors(parent_graph, current_region):
+                    if node.addr == current_region.addr:
+                        for suc in successors:
+                            if suc not in loop_subgraph:
+                                loop_successors.add(suc)
+                current_region = parent_region
+                parent_region = self._parent_map.get(current_region, None)
+
+        return loop_subgraph, loop_successors
+
+    def _make_endless_loop(self, loop_head, loop_subgraph, loop_successors):
+        loop_body = self._to_loop_body_sequence(loop_head, loop_subgraph, loop_successors)
+
+        # create a while(true) loop with sequence node being the loop body
+        return LoopNode("while", None, loop_body, addr=loop_head.addr)
+
+    def _refine_loop(self, loop_node):
+        while True:
+            # while
+            r, loop_node = self._refine_loop_while(loop_node)
+            if r:
+                continue
+
+            # do-while
+            r, loop_node = self._refine_loop_dowhile(loop_node)
+            if r:
+                continue
+
+            # no more changes
+            break
+
+        return loop_node
+
+    @staticmethod
+    def _refine_loop_while(loop_node):
+        if loop_node.sort == "while" and loop_node.condition is None and loop_node.sequence_node.nodes:
+            # it's an endless loop
+            first_node = first_nonlabel_nonphi_node(loop_node.sequence_node)
+            inner_first_node = first_node.node if type(first_node) is CodeNode else first_node
+            if type(inner_first_node) is ConditionalBreakNode:
+                while_cond = ConditionProcessor.simplify_condition(claripy.Not(inner_first_node.condition))
+                new_seq = loop_node.sequence_node.copy()
+                new_seq.nodes = [nn for nn in new_seq.nodes if nn is not first_node]
+                new_loop_node = LoopNode("while", while_cond, new_seq, addr=loop_node.addr)
+
+                return True, new_loop_node
+
+        return False, loop_node
+
+    @staticmethod
+    def _refine_loop_dowhile(loop_node):
+        if loop_node.sort == "while" and loop_node.condition is None and loop_node.sequence_node.nodes:
+            # it's an endless loop
+            last_node = loop_node.sequence_node.nodes[-1]
+            if type(last_node) is ConditionalBreakNode:
+                while_cond = ConditionProcessor.simplify_condition(claripy.Not(last_node.condition))
+                new_seq = loop_node.sequence_node.copy()
+                new_seq.nodes = new_seq.nodes[:-1]
+                new_loop_node = LoopNode("do-while", while_cond, new_seq)
+
+                return True, new_loop_node
+
+        return False, loop_node
+
+    def _to_loop_body_sequence(self, loop_head, loop_subgraph, loop_successors):
+        graph = self._region.graph_with_successors
+        assert graph is not None
+        loop_region_graph = networkx.DiGraph()
+
+        # TODO: Make sure the loop body has been structured
+
+        queue = [loop_head]
+        traversed = set()
+        loop_successor_addrs = {succ.addr for succ in loop_successors}
+        replaced_nodes = {}
+        outedges = []
+
+        while queue:
+            node = queue[0]
+            queue = queue[1:]
+
+            loop_region_graph.add_node(node)
+            traversed.add(node)
+
+            successors_and_data = list(graph.out_edges(node, data=True))  # successors are all inside the current region
+
+            for _, dst, edge_data in successors_and_data:
+                # sanity check
+                if dst.addr in loop_successor_addrs:
+                    outedges.append((node, dst, edge_data))
+                    continue
+                if dst not in loop_subgraph and dst.addr not in loop_successor_addrs:
+                    # what's this node?
+                    l.error("Found a node that belongs to neither loop body nor loop successors. Something is wrong.")
+                    # raise Exception()
+
+                if replaced_nodes.get(dst, dst) is not loop_head:
+                    loop_region_graph.add_edge(node, replaced_nodes.get(dst, dst), **edge_data)
+                if dst in traversed or dst in queue:
+                    continue
+                queue.append(dst)
+
+        # Create a graph region and structure it
+        loop_region_graph_with_successors = networkx.DiGraph(loop_region_graph)
+        loop_successors = set()  # update loop_successors with nodes in outedges
+        for src, dst, edge_data in outedges:
+            loop_region_graph_with_successors.add_edge(src, dst, **edge_data)
+            loop_successors.add(dst)
+        region = GraphRegion(
+            loop_head, loop_region_graph, successors=None, graph_with_successors=None, cyclic=False, full_graph=None
+        )
+        structurer = self.project.analyses[DreamStructurer].prep()(
+            region, condition_processor=self.cond_proc, func=self.function, jump_tables=self.jump_tables
+        )
+        seq = structurer.result
+
+        # traverse this node and rewrite all conditional jumps that go outside the loop to breaks
+        self._rewrite_conditional_jumps_to_breaks(seq, loop_successor_addrs)
+        # traverse this node and rewrite all jumps that go to the beginning of the loop to continue
+        self._rewrite_jumps_to_continues(seq)
+
+        seq = self._remove_redundant_jumps(seq)
+        seq = self._remove_conditional_jumps(seq)
+        seq = EmptyNodeRemover(seq).result
+
+        while True:
+            r, seq = self._merge_conditional_breaks(seq)
+            if r:
+                continue
+            r, seq = self._merge_nesting_conditionals(seq)
+            if r:
+                continue
+            break
+
+        return EmptyNodeRemover(seq).result
+
+    def _make_sequence(self):
+        seq = SequenceNode(None)
+
+        for node in GraphUtils.quasi_topological_sort_nodes(self._region.graph):
+            seq.add_node(CodeNode(node, self.cond_proc.reaching_conditions.get(node, None)))
+
+        if seq.nodes:
+            seq.addr = seq.nodes[0].addr
+
+        return seq
+
+    @staticmethod
+    def _unpack_sequence(seq):
+        def _handle_Code(node, **kwargs):  # pylint:disable=unused-argument
+            node = node.node
+            return walker._handle(node)
+
+        def _handle_Sequence(node, **kwargs):  # pylint:disable=unused-argument
+            for i in range(len(node.nodes)):  # pylint:disable=consider-using-enumerate
+                node.nodes[i] = walker._handle(node.nodes[i])
+            return node
+
+        def _handle_ConditionNode(node, **kwargs):  # pylint:disable=unused-argument
+            if node.true_node is not None:
+                node.true_node = walker._handle(node.true_node)
+            if node.false_node is not None:
+                node.false_node = walker._handle(node.false_node)
+            return node
+
+        def _handle_CascadingConditionNode(node: CascadingConditionNode, **kwargs):  # pylint:disable=unused-argument
+            new_cond_and_nodes = []
+            for cond, child_node in node.condition_and_nodes:
+                new_cond_and_nodes.append((cond, walker._handle(child_node)))
+            node.condition_and_nodes = new_cond_and_nodes
+
+            if node.else_node is not None:
+                node.else_node = walker._handle(node.else_node)  # type: ignore
+            return node
+
+        def _handle_SwitchCaseNode(node, **kwargs):  # pylint:disable=unused-argument
+            for i in list(node.cases.keys()):
+                node.cases[i] = walker._handle(node.cases[i])
+            if node.default_node is not None:
+                node.default_node = walker._handle(node.default_node)
+            return node
+
+        def _handle_Default(node, **kwargs):  # pylint:disable=unused-argument
+            return node
+
+        handlers = {
+            CodeNode: _handle_Code,
+            SequenceNode: _handle_Sequence,
+            ConditionNode: _handle_ConditionNode,
+            CascadingConditionNode: _handle_CascadingConditionNode,
+            SwitchCaseNode: _handle_SwitchCaseNode,
+            # don't do anything
+            LoopNode: _handle_Default,
+            ContinueNode: _handle_Default,
+            ConditionalBreakNode: _handle_Default,
+            BreakNode: _handle_Default,
+            MultiNode: _handle_Default,
+            ailment.Block: _handle_Default,
+        }
+
+        walker = SequenceWalker(handlers=handlers)
+        walker.walk(seq)
+
+        return seq
+
+    def _structure_sequence(self, seq):
+        self._make_switch_cases(seq)
+
+        # this is hackish...
+        # seq.nodes = new_seq.nodes
+
+        self._merge_same_conditioned_nodes(seq)
+        self._structure_common_subexpression_conditions(seq)
+        self._make_ites(seq)
+        self._remove_redundant_jumps(seq)
+
+        empty_node_remover = EmptyNodeRemover(seq)
+        new_seq = empty_node_remover.result
+        # update self._new_sequences
+        self._update_new_sequences(set(empty_node_remover.removed_sequences), empty_node_remover.replaced_sequences)
+
+        # we need to do it in-place
+        seq.nodes = new_seq.nodes
+
+        self._replace_complex_reaching_conditions(seq)
+        self._make_condition_nodes(seq)
+        self._make_cascading_condition_nodes(seq)
+
+        while True:
+            r, seq = self._merge_conditional_breaks(seq)
+            if r:
+                continue
+            r, seq = self._merge_nesting_conditionals(seq)
+            if r:
+                continue
+            break
+
+    def _merge_same_conditioned_nodes(self, seq):
+        # search for nodes with the same reaching condition and then merge them into one sequence node
+        i = 0
+        while i < len(seq.nodes) - 1:
+            node_0 = seq.nodes[i]
+            if type(node_0) is not CodeNode:
+                i += 1
+                continue
+            rcond_0 = node_0.reaching_condition
+            if rcond_0 is None:
+                i += 1
+                continue
+            node_1 = seq.nodes[i + 1]
+            if type(node_1) is not CodeNode:
+                i += 1
+                continue
+            rcond_1 = node_1.reaching_condition
+            if rcond_1 is None:
+                i += 1
+                continue
+            r = claripy.simplify(rcond_0 == rcond_1)
+            if claripy.is_true(r):
+                # node_0 and node_1 should be put into the same sequence node
+                new_node = CodeNode(
+                    self._merge_nodes(node_0.node, node_1.node),
+                    node_0.reaching_condition,
+                )
+                seq.nodes = [*seq.nodes[:i], new_node, *seq.nodes[i + 2 :]]
+                continue
+            i += 1
+
+    #
+    # Dealing with switch-case structures
+    #
+
+    def _make_switch_cases(self, seq):
+        """
+        Search for nodes that look like switch-cases and convert them to switch cases.
+
+        :param seq:     The Sequence node.
+        :return:        None
+        """
+
+        addr2nodes: dict[int, set[CodeNode]] = defaultdict(set)
+        for node in seq.nodes:
+            addr2nodes[node.addr].add(node)
+
+        while True:
+            for i in range(len(seq.nodes)):
+                node = seq.nodes[i]
+
+                # Jumptable_AddressLoadedFromMemory
+                r = self._make_switch_cases_address_loaded_from_memory(seq, i, node, addr2nodes)
+                if r:
+                    # we found a node that looks like a switch-case. seq.nodes are changed. resume to find the next such
+                    # case
+                    break
+
+                # Jumptable_AddressComputed
+                r = self._make_switch_cases_address_computed(seq, i, node, addr2nodes)
+                if r:
+                    break
+
+            else:
+                # we did not find any node that looks like a switch-case. exit.
+                break
+
+    def _make_switch_cases_address_loaded_from_memory(self, seq, i, node, addr2nodes: dict[int, set[CodeNode]]) -> bool:
+        """
+        A typical jump table involves multiple nodes, which look like the following:
+
+        Head:  s_50 = Conv(32->64, (Load(addr=stack_base-28, size=4, endness=Iend_LE) - 0x3f<32>))<8>
+               if (((Load(addr=stack_base-28, size=4, endness=Iend_LE) - 0x3f<32>) <= 0x36<32>))
+                    { Goto A<64> } else { Goto B<64> }
+
+        A:     (with an indirect jump)
+               Goto((
+                Conv(32->64, Load(addr=(0x40964c<64> + (Load(addr=stack_base-80, size=8, endness=Iend_LE) Mul 0x4<8>)),
+                size=4, endness=Iend_LE)) + 0x40964c<64>))
+
+        B:     (the default case)
+        """
+
+        try:
+            last_stmt = self.cond_proc.get_last_statement(node)
+            if not isinstance(last_stmt, ailment.Stmt.ConditionalJump):
+                return False
+        except EmptyBlockNotice:
+            return False
+        successor_addrs = extract_jump_targets(last_stmt)
+        if len(successor_addrs) != 2:
+            return False
+
+        for t in successor_addrs:
+            if t in addr2nodes and t in self.jump_tables:
+                # this is a candidate!
+                target = t
+                break
+        else:
+            return False
+
+        jump_table = self.jump_tables[target]
+        if jump_table.type != IndirectJumpType.Jumptable_AddressLoadedFromMemory:
+            return False
+
+        # extract the comparison expression, lower-, and upper-bounds from the last statement
+        cmp = switch_extract_cmp_bounds(last_stmt)
+        if not cmp:
+            return False
+        cmp_expr, cmp_lb, _cmp_ub = cmp  # pylint:disable=unused-variable
+
+        # the real indirect jump
+        if len(addr2nodes[target]) != 1:
+            return False
+        node_a = next(iter(addr2nodes[target]))
+        # the default case
+        node_b_addr = next(iter(t for t in successor_addrs if t != target))
+
+        # Node A might have been structured. Un-structure it if that is the case.
+        r, node_a = self._switch_unpack_sequence_node(seq, node_a, node_b_addr, jump_table, addr2nodes)
+        if not r:
+            return False
+
+        # build switch-cases
+        assert jump_table.jumptable_entries is not None
+        cases, node_default, to_remove = self._switch_build_cases(
+            seq, cmp_lb, jump_table.jumptable_entries, i, node_b_addr, addr2nodes
+        )
+        # if we don't know what the end address of this switch-case structure is, let's figure it out
+        switch_end_addr = (
+            node_b_addr
+            if node_default is None
+            else self._switch_find_switch_end_addr(cases, node_default, {nn.addr for nn in self._region.graph})
+        )
+        if switch_end_addr is not None:
+            self._switch_handle_gotos(cases, node_default, switch_end_addr)
+
+        assert last_stmt.tags["ins_addr"] is not None
+        self._make_switch_cases_core(
+            seq,
+            i,
+            node,
+            cmp_expr,
+            cases,
+            node_default,
+            last_stmt.tags["ins_addr"],
+            addr2nodes,
+            to_remove,
+            node_a=node_a,
+            jumptable_addr=jump_table.addr,
+        )
+
+        return True
+
+    def _make_switch_cases_address_computed(self, seq, i, node, addr2nodes: dict[int, set[CodeNode]]) -> bool:
+        if node.addr not in self.jump_tables:
+            return False
+        jump_table = self.jump_tables[node.addr]
+        if jump_table.type != IndirectJumpType.Jumptable_AddressComputed:
+            return False
+
+        try:
+            last_stmts = self.cond_proc.get_last_statements(node)
+        except EmptyBlockNotice:
+            return False
+        if len(last_stmts) != 1:
+            return False
+        last_stmt = last_stmts[0]
+
+        if not isinstance(last_stmt, ailment.Stmt.ConditionalJump):
+            return False
+
+        # Typical look:
+        #   t2 = (r5<4> - 0x22<32>)
+        #   if ((t2 <= 0x1c<32>)) { Goto (0x41d10c<32> + (t2 << 0x2<8>)) } else { Goto 0x41d108<32> }
+        #
+        # extract the comparison expression, lower-, and upper-bounds from the last statement
+        cmp = switch_extract_cmp_bounds(last_stmt)
+        if not cmp:
+            return False
+        cmp_expr, cmp_lb, _cmp_ub = cmp  # pylint:disable=unused-variable
+
+        jumptable_entries = jump_table.jumptable_entries
+        assert jumptable_entries is not None
+
+        if isinstance(last_stmt.false_target, ailment.Expr.Const):
+            default_addr = last_stmt.false_target.value
+            assert isinstance(default_addr, int)
+        else:
+            return False
+
+        cases, node_default, to_remove = self._switch_build_cases(
+            seq, cmp_lb, jumptable_entries, i, default_addr, addr2nodes
+        )
+        if node_default is None:
+            # there must be a default case
+            return False
+
+        self._make_switch_cases_core(
+            seq,
+            i,
+            node,
+            cmp_expr,
+            cases,
+            node_default,
+            node.addr,
+            addr2nodes,
+            to_remove,
+            jumptable_addr=jump_table.addr,
+        )
+
+        return True
+
+    def _make_switch_cases_core(
+        self,
+        seq,
+        i,
+        node,
+        cmp_expr,
+        cases: OrderedDict,
+        node_default,
+        addr,
+        addr2nodes,
+        to_remove,
+        *,
+        jumptable_addr: int,
+        node_a=None,
+    ):
+        scnode = SwitchCaseNode(cmp_expr, cases, node_default, addr=addr)
+        scnode = CodeNode(scnode, node.reaching_condition)
+
+        # insert the switch-case node
+        seq.insert_node(i + 1, scnode)
+        # remove all those entry nodes
+        if node_default is not None:
+            to_remove.add(node_default)
+        for node_ in to_remove:
+            seq.remove_node(node_)
+            addr2nodes[node_.addr].discard(node_)
+            if not addr2nodes[node_.addr]:
+                del addr2nodes[node_.addr]
+        # remove the last statement in node
+        remove_last_statement(node)
+        if BaseNode.test_empty_node(node):
+            seq.remove_node(node)
+        if node_a is not None:
+            # remove the last statement in node_a
+            remove_last_statement(node_a)
+            if BaseNode.test_empty_node(node_a):
+                seq.remove_node(node_a)
+
+        # rewrite conditions in the entire SequenceNode to remove jump table entry conditions
+        rewriter = JumpTableEntryConditionRewriter(self.cond_proc.jump_table_conds[jumptable_addr])
+        rewriter.walk(seq)  # update SequenceNodes in-place
+
+    def _switch_unpack_sequence_node(
+        self, seq: SequenceNode, node_a, node_b_addr: int, jumptable, addr2nodes: dict[int, set[CodeNode]]
+    ) -> tuple[bool, CodeNode | None]:
+        """
+        We might have already structured the actual body of the switch-case structure into a single Sequence node (node
+        A). If that is the case, we un-structure the sequence node in this method.
+
+        :param seq:                 The original Sequence node.
+        :param node_a:              Node A.
+        :param node_b_addr:         Address of node B.
+        :param jumptable:           The corresponding jump table instance.
+        :param addr2nodes:          A dict of addresses to their corresponding nodes in `seq`.
+        :return:                    A boolean value indicating the result and an updated node_a. The boolean value is
+                                    True if unpacking is not necessary or we successfully unpacked the sequence node,
+                                    False otherwise.
+        """
+
+        jumptable_entries = jumptable.jumptable_entries
+
+        node_a_block_addrs = {n.addr for n in node_a.node.nodes} if isinstance(node_a.node, SequenceNode) else set()
+        #
+        # if that is the case, we un-structure it here
+        if all(entry_addr in addr2nodes for entry_addr in jumptable_entries):
+            return True, node_a
+        if self._switch_check_existence_of_jumptable_entries(
+            jumptable_entries, node_a_block_addrs, set(addr2nodes.keys()), node_a.addr, node_b_addr
+        ):
+            # unpacking is needed
+            for n in node_a.node.nodes:
+                assert n.addr is not None
+                if isinstance(n, ConditionNode):
+                    unpacked = self._switch_unpack_condition_node(n, jumptable)
+                    if unpacked is None:
+                        # unsupported. bail
+                        return False, None
+                    if n.addr in addr2nodes:
+                        del addr2nodes[n.addr]
+                    addr2nodes[n.addr].add(unpacked)
+                    seq.add_node(unpacked)
+                else:
+                    the_node = CodeNode(n, None)
+                    if n.addr in addr2nodes:
+                        del addr2nodes[n.addr]
+                    addr2nodes[n.addr].add(the_node)
+                    seq.add_node(the_node)
+            if node_a != addr2nodes[node_a.addr]:
+                # update node_a
+                seq.remove_node(node_a)
+                node_a = next(iter(addr2nodes[node_a.addr]))
+            return True, node_a
+
+        # a jumptable entry is missing. it's very likely marked as the successor of the entire switch-case region. we
+        # should have been handling it when dealing with multi-exit regions. ignore it here.
+        return True, node_a
+
+    def _switch_unpack_condition_node(self, cond_node: ConditionNode, jumptable) -> CodeNode | None:
+        """
+        Unpack condition nodes by only removing one condition in the form of
+        <Bool jump_table_402020 == 0x402ac4>.
+
+        :param cond_node:   The condition node to unpack.
+        :return:            The new unpacked node.
+        """
+
+        # FIXME: With the new jump table entry condition, this function is probably never used. Remove sequence node
+        # FIXME: unpacking logic if that is the case.
+
+        cond = cond_node.condition
+
+        # look for a condition in the form of xxx == jump_target
+        eq_condition = None
+        remaining_cond = None
+        true_node = None
+        false_node = None
+
+        jumptable_var = self.cond_proc.create_jump_target_var(jumptable.addr)
+
+        if cond.op == "And":
+            for arg in cond.args:
+                if (
+                    arg.op == "__eq__"
+                    and arg.args[0] is jumptable_var
+                    and isinstance(arg.args[1], claripy.ast.Bits)
+                    and arg.args[1].concrete
+                ):
+                    # found it
+                    eq_condition = arg
+                    remaining_cond = claripy.And(*(arg_ for arg_ in cond.args if arg_ is not arg))
+                    true_node = cond_node.true_node
+                    false_node = cond_node.false_node
+                    break
+            else:
+                # unsupported
+                return None
+        elif cond.op == "__eq__":
+            if cond.args[0] is jumptable_var and isinstance(cond.args[1], claripy.ast.Bits) and cond.args[1].concrete:
+                # found it
+                eq_condition = cond
+                true_node = cond_node.true_node
+                false_node = cond_node.false_node
+                remaining_cond = None
+            else:
+                # unsupported
+                return None
+        else:
+            # unsupported
+            return None
+
+        if remaining_cond is None:
+            if true_node is not None and false_node is None:
+                return CodeNode(true_node, eq_condition)
+            # unsupported
+            return None
+
+        return CodeNode(
+            ConditionNode(cond_node.addr, claripy.true(), remaining_cond, true_node, false_node=false_node),
+            eq_condition,
+        )
+
+    def _switch_check_existence_of_jumptable_entries(
+        self,
+        jumptable_entries,
+        node_a_block_addrs: set[int],
+        known_node_addrs: set[int],
+        node_a_addr: int,
+        node_b_addr: int,
+    ) -> bool:
+        """
+        Check if all entries in the given jump table exist in the given set of nodes of a SequenceNode.
+
+        :param jumptable_entries:   Addresses of jump table entries.
+        :param node_a_block_addrs:  A set of addresses for nodes that belong to Node A.
+        :return:                    True if the check passes, False otherwise.
+        """
+
+        all_node_addrs = node_a_block_addrs | known_node_addrs | {node_b_addr}
+        expected_node_a_addrs = set()
+        for entry_addr in jumptable_entries:
+            if entry_addr in all_node_addrs:
+                expected_node_a_addrs.add(entry_addr)
+                continue
+            # the entry may go missing if the entire node has been folded into its successor node.
+            # in this case, we check if (a) this entry node has only one successor, and (b) this successor exists in
+            # seq_node_addrs.
+            if self.function is not None:
+                entry_node = self.function.get_node(entry_addr)
+                if entry_node is not None:
+                    successors = []
+                    for _, dst, data in self.function.graph.out_edges(entry_node, data=True):
+                        if data.get("type", "transition") != "call":
+                            successors.append(dst)
+                    if len(successors) == 1 and successors[0].addr in all_node_addrs:
+                        # found the single successor
+                        expected_node_a_addrs.add(successors[0].addr)
+                        continue
+            # it's also possible that this is just a jump that breaks out of the switch-case. we simply ignore it.
+            continue
+
+        # finally, make sure all expected nodes exist
+        # not sure what is going on... if it's false
+        return node_a_block_addrs.issuperset((expected_node_a_addrs | {node_a_addr}) - {node_b_addr})
+
+    def _switch_find_jumptable_entry_node(self, entry_addr: int, addr2nodes: dict[int, set[CodeNode]]) -> Any | None:
+        """
+        Find the correct node for a given jump table entry address in addr2nodes.
+
+        This method is needed because prior optimization steps may remove some blocks (e.g., empty blocks or blocks that
+        only have branch instructions). If the given jump table entry address corresponds to a removed block, it will
+        not be found inside addr2nodes dict. In such cases, we need to follow graph edges in the CFG and find the first
+        block whose address is inside addr2nodes dict.
+
+        :param entry_addr:  Address of the jump table entry.
+        :return:            The correct node if we can find it, or None if we fail to find one.
+        """
+
+        if entry_addr in addr2nodes and len(addr2nodes[entry_addr]) == 1:
+            return next(iter(addr2nodes[entry_addr]))
+        # magic
+        if self.function is None:
+            return None
+
+        addr = entry_addr
+        node = self.function.get_node(addr)
+        for _ in range(5):  # we try at most five steps
+            if node is None:
+                return None
+            successors = []
+            for _, dst, data in self.function.graph.out_edges(node, data=True):
+                if data.get("type", "transition") != "call":
+                    successors.append(dst)
+            if len(successors) != 1:
+                return None
+            successor = successors[0]
+            if successor.addr in addr2nodes:
+                # found it!
+                return next(iter(addr2nodes[successor.addr]))
+            # keep looking
+            node = successor
+        return None
+
+    def _switch_build_cases(
+        self,
+        seq: SequenceNode,
+        cmp_lb: int,
+        jumptable_entries: list[int],
+        head_node_idx: int,
+        node_b_addr: int,
+        addr2nodes: dict[int, set[CodeNode]],
+    ) -> tuple[OrderedDict, Any, Any]:
+        """
+        Discover all cases for the switch-case structure and build the switch-cases dict.
+
+        :param seq:                 The original Sequence node.
+        :param cmp_lb:              The lower bound of the jump table comparison.
+        :param jumptable_entries:   Addresses of indirect jump targets in the jump table.
+        :param head_node_addr:      The index of the head block of this jump table in `seq`.
+        :param node_b_addr:         Address of node B. Potentially, node B is the default node.
+        :param addr2nodes:          A dict of addresses to their corresponding nodes in `seq`.
+        :return:                    A tuple of (dict of cases, the default node if exists, nodes to remove).
+        """
+
+        cases: OrderedDict[int | tuple[int, ...], SequenceNode] = OrderedDict()
+        to_remove = set()
+        node_default = addr2nodes.get(node_b_addr)
+        if node_default is not None:
+            node_default = next(iter(node_default))
+
+        entry_addrs_set = set(jumptable_entries)
+        converted_nodes: dict[int, Any] = {}
+        entry_addr_to_ids = defaultdict(set)
+
+        for j, entry_addr in enumerate(jumptable_entries):
+            cases_idx = cmp_lb + j
+            if entry_addr == node_b_addr:
+                # jump to default or end of the switch-case structure - ignore this case
+                continue
+
+            entry_addr_to_ids[entry_addr].add(cases_idx)
+
+            if entry_addr in converted_nodes:
+                continue
+
+            entry_node = self._switch_find_jumptable_entry_node(entry_addr, addr2nodes)
+            if entry_node is None:
+                # Missing entries. They are probably *after* the entire switch-case construct. Replace it with an empty
+                # Goto node.
+                case_inner_node = ailment.Block(
+                    0,
+                    0,
+                    statements=[
+                        ailment.Stmt.Jump(
+                            None,
+                            ailment.Expr.Const(None, None, entry_addr, self.project.arch.bits),
+                            ins_addr=0,
+                            stmt_idx=0,
+                        )
+                    ],
+                )
+                case_node = SequenceNode(0, nodes=[CodeNode(case_inner_node, claripy.true())])
+                converted_nodes[entry_addr] = case_node
+                continue
+
+            case_node = SequenceNode(entry_node.addr, nodes=[CodeNode(entry_node.node, claripy.true())])
+            to_remove.add(entry_node)
+            entry_node_idx = seq.nodes.index(entry_node)
+
+            if entry_node_idx <= head_node_idx:
+                # it's jumping to a block that dominates the head. it's likely to be an optimized continue; statement
+                # in a switch-case wrapped inside a while loop.
+                # replace it with an empty Goto node
+                case_inner_node = ailment.Block(
+                    0,
+                    0,
+                    statements=[
+                        ailment.Stmt.Jump(
+                            None,
+                            ailment.Expr.Const(None, None, entry_addr, self.project.arch.bits),
+                            ins_addr=0,
+                            stmt_idx=0,
+                        )
+                    ],
+                )
+                case_node = SequenceNode(0, nodes=[CodeNode(case_inner_node, claripy.true())])
+                converted_nodes[entry_addr] = case_node
+                continue
+
+            # find nodes that this entry node dominates
+            cond_subexprs = list(get_ast_subexprs(entry_node.reaching_condition))
+            guarded_nodes = None
+            for subexpr in cond_subexprs:
+                guarded_node_candidates = self._nodes_guarded_by_common_subexpr(seq, subexpr, entry_node_idx + 1)
+                if guarded_nodes is None:
+                    guarded_nodes = {node_ for _, node_, _ in guarded_node_candidates}
+                else:
+                    guarded_nodes = guarded_nodes.intersection({node_ for _, node_, _ in guarded_node_candidates})
+
+            if guarded_nodes is not None:
+                # keep the topological order of nodes in Sequence node
+                sorted_guarded_nodes = [node_ for node_ in seq.nodes[entry_node_idx + 1 :] if node_ in guarded_nodes]
+                for node_ in sorted_guarded_nodes:
+                    if node_ is not entry_node and node_.addr not in entry_addrs_set:
+                        # fix reaching condition
+                        reaching_condition_subexprs = set(get_ast_subexprs(node_.reaching_condition)).difference(
+                            set(cond_subexprs)
+                        )
+                        new_reaching_condition = claripy.And(*reaching_condition_subexprs)
+                        new_node = CodeNode(node_.node, new_reaching_condition)
+                        case_node.add_node(new_node)
+                        to_remove.add(node_)
+
+            # do we have a default node?
+            case_last_stmt = self.cond_proc.get_last_statement(case_node)
+            if isinstance(case_last_stmt, ailment.Stmt.Jump):
+                targets = extract_jump_targets(case_last_stmt)
+                if len(targets) == 1 and targets[0] == node_b_addr:
+                    # jump to the default case is rare - it's more likely that there is no default for this
+                    # switch-case struct
+                    node_default = None
+
+            converted_nodes[entry_addr] = case_node
+
+        for entry_addr, converted_node in converted_nodes.items():
+            cases_ids = entry_addr_to_ids[entry_addr]
+            if len(cases_ids) == 1:
+                cases[next(iter(cases_ids))] = converted_node
+            else:
+                cases[tuple(sorted(cases_ids))] = converted_node
+
+            self._new_sequences.append(converted_node)
+
+        # reorganize cases to handle fallthroughs
+        cases = self._reorganize_switch_cases(cases)
+
+        return cases, node_default, to_remove
+
+    #
+    # Dealing with If-Then-Else structures
+    #
+
+    def _make_ites(self, seq):
+        # search for a == ^a pairs
+
+        while True:
+            break_hard = False
+            for i in range(len(seq.nodes)):
+                node_0 = seq.nodes[i]
+                if type(node_0) is not CodeNode:
+                    continue
+                rcond_0 = node_0.reaching_condition
+                if rcond_0 is None:
+                    continue
+                if claripy.is_true(rcond_0) or claripy.is_false(rcond_0):
+                    continue
+                for j in range(i + 1, len(seq.nodes)):
+                    node_1 = seq.nodes[j]
+                    if type(node_1) is not CodeNode:
+                        continue
+                    if node_0 is node_1:
+                        continue
+                    rcond_1 = node_1.reaching_condition
+                    if rcond_1 is None:
+                        continue
+                    cond_ = claripy.simplify(claripy.Not(rcond_0) == rcond_1)
+                    if claripy.is_true(cond_):
+                        # node_0 and node_1 should be structured using an if-then-else
+                        self._make_ite(seq, node_0, node_1)
+                        break_hard = True
+                        break
+                if break_hard:
+                    break
+            else:
+                break
+
+    def _structure_common_subexpression_conditions(self, seq):
+        # use common subexpressions to structure nodes and create more if-then-else instances
+
+        i = 0
+        while i < len(seq.nodes) - 1:
+            structured = False
+            node_0 = seq.nodes[i]
+            if not isinstance(node_0, CodeNode):
+                i += 1
+                continue
+            rcond_0 = node_0.reaching_condition
+            if rcond_0 is None:
+                i += 1
+                continue
+            subexprs_0 = list(get_ast_subexprs(rcond_0))
+
+            for common_subexpr in subexprs_0:
+                if claripy.is_true(common_subexpr):
+                    continue
+                candidates = self._nodes_guarded_by_common_subexpr(seq, common_subexpr, i + 1)
+                if candidates:
+                    candidates.insert(0, (i, node_0, subexprs_0))
+                    new_node = self._create_seq_node_guarded_by_common_subexpr(common_subexpr, candidates)
+                    self._new_sequences.append(new_node)
+
+                    # remove all old nodes and replace them with the new node
+                    for idx, _, _ in candidates:
+                        seq.nodes[idx] = None
+                    seq.nodes[i] = CodeNode(new_node, common_subexpr)
+                    seq.nodes = [n for n in seq.nodes if n is not None]
+                    structured = True
+                    break
+
+            if not structured:
+                i += 1
+
+    @staticmethod
+    def _nodes_guarded_by_common_subexpr(seq, common_subexpr, starting_idx):
+        candidates = []
+
+        if common_subexpr is claripy.true():
+            return []
+        for j, node_1 in enumerate(seq.nodes[starting_idx:]):
+            rcond_1 = getattr(node_1, "reaching_condition", None)
+            if rcond_1 is None:
+                continue
+            subexprs_1 = list(get_ast_subexprs(rcond_1))
+            if any(subexpr_1 is common_subexpr for subexpr_1 in subexprs_1):
+                # we found one!
+                candidates.append((starting_idx + j, node_1, subexprs_1))
+
+        return candidates
+
+    @staticmethod
+    def _create_seq_node_guarded_by_common_subexpr(common_subexpr, candidates):
+        new_nodes = []
+
+        for _, node, subexprs in candidates:
+            # :)
+            new_subexprs = [ex for ex in subexprs if ex is not common_subexpr]
+            new_node = CodeNode(
+                node.node,
+                claripy.And(*new_subexprs),
+            )
+            new_nodes.append(new_node)
+
+        return SequenceNode(None if not new_nodes else new_nodes[0].addr, nodes=new_nodes)
+
+    def _replace_complex_reaching_conditions(self, seq: SequenceNode):
+        for i in range(len(seq.nodes)):
+            node = seq.nodes[i]
+
+            if (
+                isinstance(node, CodeNode)
+                and node.reaching_condition is not None
+                and node.reaching_condition.op == "Or"
+                and node.node in self.cond_proc.guarding_conditions
+            ):
+                guarding_condition = self.cond_proc.guarding_conditions[node.node]
+                # the op of guarding condition is always "Or"
+                if (
+                    len(guarding_condition.args) < len(node.reaching_condition.args)
+                    and guarding_condition.depth < node.reaching_condition.depth
+                ):
+                    node.reaching_condition = guarding_condition
+
+    def _make_condition_nodes(self, seq):
+        # make all conditionally-reachable nodes ConditionNodes
+        for i in range(len(seq.nodes)):
+            node = seq.nodes[i]
+
+            if isinstance(node, CodeNode):
+                if isinstance(node.node, SequenceNode) and node.node not in self._new_sequences:
+                    self._make_condition_nodes(node.node)
+
+                if node.reaching_condition is not None and not claripy.is_true(node.reaching_condition):
+                    if isinstance(node.node, ConditionalBreakNode):
+                        # Put conditions together and simplify them
+                        cond = claripy.And(node.reaching_condition, node.node.condition)
+                        new_node = CodeNode(ConditionalBreakNode(node.node.addr, cond, node.node.target), None)
+                    else:
+                        new_node = ConditionNode(node.addr, None, node.reaching_condition, node, None)
+                    seq.nodes[i] = new_node
+
+    @staticmethod
+    def _make_cascading_condition_nodes(seq: SequenceNode):
+        """
+        Convert nested condition nodes into a CascadingConditionNode.
+        """
+        CascadingConditionTransformer(seq)
+
+    def _make_ite(self, seq, node_0, node_1):
+        # ensure order
+        if node_0.addr > node_1.addr:
+            node_0, node_1 = node_1, node_0
+
+        node_0_pos = seq.node_position(node_0)
+        node_1_pos = seq.node_position(node_1)
+        pos = max(node_0_pos, node_1_pos)
+
+        node_0_, node_1_ = node_0.copy(), node_1.copy()
+        # clear their reaching conditions
+        node_0_.reaching_condition = None
+        node_1_.reaching_condition = None
+
+        node_0_kids = self._nodes_guarded_by_common_subexpr(seq, node_0.reaching_condition, node_0_pos + 1)
+        node_0_kids.insert(0, (node_0_pos, node_0_, [node_0.reaching_condition]))
+        node_1_kids = self._nodes_guarded_by_common_subexpr(seq, node_1.reaching_condition, node_1_pos + 1)
+        node_1_kids.insert(0, (node_1_pos, node_1_, [node_1.reaching_condition]))
+
+        new_node_0 = self._create_seq_node_guarded_by_common_subexpr(node_0.reaching_condition, node_0_kids)
+        new_node_1 = self._create_seq_node_guarded_by_common_subexpr(node_1.reaching_condition, node_1_kids)
+
+        self._new_sequences.append(new_node_0)
+        self._new_sequences.append(new_node_1)
+
+        seq_addr = seq.addr
+
+        # erase all nodes in the candidates
+        for idx, _, _ in node_0_kids + node_1_kids:
+            seq.nodes[idx] = None
+
+        seq.insert_node(pos, ConditionNode(seq_addr, None, node_0.reaching_condition, new_node_0, new_node_1))
+        seq.nodes = [n for n in seq.nodes if n is not None]
+
+
+# delayed import
+from angr.analyses.decompiler.sequence_walker import SequenceWalker  # pylint:disable=wrong-import-position