1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
//! Creating and computing forward interprocedural fixpoint problems.
//!
//! # General notes
//!
//! This module supports computation of fixpoint problems on the control flow graphs generated by the `graph` module.
//!
//! To compute a generalized fixpoint problem,
//! first construct a context object implementing the `Context`trait.
//! Use it to construct a `Computation` object.
//! The `Computation` object provides the necessary methods for the actual fixpoint computation.

use super::fixpoint::Context as GeneralFPContext;
use super::graph::*;
use super::interprocedural_fixpoint_generic::*;
use crate::intermediate_representation::*;
use petgraph::graph::EdgeIndex;
use petgraph::graph::NodeIndex;
use std::marker::PhantomData;

/// The context for an interprocedural fixpoint computation.
///
/// Basically, a `Context` object needs to contain a reference to the actual graph,
/// a method for merging node values,
/// and methods for computing the edge transitions for each different edge type.
///
/// All trait methods have access to the FixpointProblem structure, so that context informations are accessible through it.
///
/// All edge transition functions can return `None` to indicate that no information flows through the edge.
/// For example, this can be used to indicate edges that can never been taken.
pub trait Context<'a> {
    /// The type of the values that are assigned to nodes during the fixpoint computation.
    type Value: PartialEq + Eq + Clone;

    /// Get a reference to the graph that the fixpoint is computed on.
    fn get_graph(&self) -> &Graph<'a>;

    /// Merge two node values.
    fn merge(&self, value1: &Self::Value, value2: &Self::Value) -> Self::Value;

    /// Transition function for `Def` terms.
    /// The transition function for a basic block is computed
    /// by iteratively applying this function to the starting value for each `Def` term in the basic block.
    /// The iteration short-circuits and returns `None` if `update_def` returns `None` at any point.
    fn update_def(&self, value: &Self::Value, def: &Term<Def>) -> Option<Self::Value>;

    /// Transition function for (conditional and unconditional) `Jmp` terms.
    fn update_jump(
        &self,
        value: &Self::Value,
        jump: &Term<Jmp>,
        untaken_conditional: Option<&Term<Jmp>>,
        target: &Term<Blk>,
    ) -> Option<Self::Value>;

    /// Transition function for in-program calls.
    fn update_call(
        &self,
        value: &Self::Value,
        call: &Term<Jmp>,
        target: &Node,
        calling_convention: &Option<String>,
    ) -> Option<Self::Value>;

    /// Transition function for return instructions.
    /// Has access to the value at the callsite corresponding to the return edge.
    /// This way one can recover caller-specific information on return from a function.
    fn update_return(
        &self,
        value: Option<&Self::Value>,
        value_before_call: Option<&Self::Value>,
        call_term: &Term<Jmp>,
        return_term: &Term<Jmp>,
        calling_convention: &Option<String>,
    ) -> Option<Self::Value>;

    /// Transition function for calls to functions not contained in the binary.
    /// The corresponding edge goes from the callsite to the returned-to block.
    fn update_call_stub(&self, value: &Self::Value, call: &Term<Jmp>) -> Option<Self::Value>;

    /// This function is used to refine the value using the information on which branch was taken on a conditional jump.
    fn specialize_conditional(
        &self,
        value: &Self::Value,
        condition: &Expression,
        block_before_condition: &Term<Blk>,
        is_true: bool,
    ) -> Option<Self::Value>;
}

/// This struct is a wrapper to create a general fixpoint context out of an interprocedural fixpoint context.
pub struct GeneralizedContext<'a, T: Context<'a>> {
    context: T,
    _phantom_graph_reference: PhantomData<Graph<'a>>,
}

impl<'a, T: Context<'a>> GeneralizedContext<'a, T> {
    /// Create a new generalized context out of an interprocedural context object.
    pub fn new(context: T) -> Self {
        GeneralizedContext {
            context,
            _phantom_graph_reference: PhantomData,
        }
    }

    /// Get the inner context object.
    pub fn get_context(&self) -> &T {
        &self.context
    }
}

impl<'a, T: Context<'a>> GeneralFPContext for GeneralizedContext<'a, T> {
    type EdgeLabel = Edge<'a>;
    type NodeLabel = Node<'a>;
    type NodeValue = NodeValue<T::Value>;

    /// Get a reference to the underlying graph.
    fn get_graph(&self) -> &Graph<'a> {
        self.context.get_graph()
    }

    /// Merge two values using the merge function from the interprocedural context object.
    fn merge(&self, val1: &Self::NodeValue, val2: &Self::NodeValue) -> Self::NodeValue {
        use NodeValue::*;
        match (val1, val2) {
            (Value(value1), Value(value2)) => Value(self.context.merge(value1, value2)),
            (
                CallFlowCombinator {
                    call_stub: call1,
                    interprocedural_flow: return1,
                },
                CallFlowCombinator {
                    call_stub: call2,
                    interprocedural_flow: return2,
                },
            ) => CallFlowCombinator {
                call_stub: merge_option(call1, call2, |v1, v2| self.context.merge(v1, v2)),
                interprocedural_flow: merge_option(return1, return2, |v1, v2| {
                    self.context.merge(v1, v2)
                }),
            },
            _ => panic!("Malformed CFG in fixpoint computation"),
        }
    }

    /// Forward edge transition function.
    /// Applies the transition functions from the interprocedural context object
    /// corresponding to the type of the provided edge.
    fn update_edge(
        &self,
        node_value: &Self::NodeValue,
        edge: EdgeIndex,
    ) -> Option<Self::NodeValue> {
        let graph = self.context.get_graph();
        let (start_node, end_node) = graph.edge_endpoints(edge).unwrap();

        match graph.edge_weight(edge).unwrap() {
            Edge::Block => {
                let block_term = graph.node_weight(start_node).unwrap().get_block();
                let value = node_value.unwrap_value();
                let defs = &block_term.term.defs;
                let end_val = defs.iter().try_fold(value.clone(), |accum, def| {
                    self.context.update_def(&accum, def)
                });
                end_val.map(NodeValue::Value)
            }
            Edge::CallCombine(_) => Some(Self::NodeValue::Value(node_value.unwrap_value().clone())),
            Edge::Call(call) => self
                .context
                .update_call(
                    node_value.unwrap_value(),
                    call,
                    &graph[end_node],
                    &graph[end_node].get_sub().term.calling_convention,
                )
                .map(NodeValue::Value),
            Edge::CrCallStub => Some(NodeValue::CallFlowCombinator {
                call_stub: Some(node_value.unwrap_value().clone()),
                interprocedural_flow: None,
            }),
            Edge::CrReturnStub => Some(NodeValue::CallFlowCombinator {
                call_stub: None,
                interprocedural_flow: Some(node_value.unwrap_value().clone()),
            }),
            Edge::ReturnCombine(call_term) => match node_value {
                NodeValue::Value(_) => panic!("Unexpected interprocedural fixpoint graph state"),
                NodeValue::CallFlowCombinator {
                    call_stub,
                    interprocedural_flow,
                } => {
                    let (return_from_block, return_from_sub) = match graph.node_weight(start_node) {
                        Some(Node::CallReturn {
                            call: _,
                            return_: (return_from_block, return_from_sub),
                        }) => (return_from_block, return_from_sub),
                        _ => panic!("Malformed Control flow graph"),
                    };
                    let return_from_jmp = &return_from_block.term.jmps[0];
                    self.context
                        .update_return(
                            interprocedural_flow.as_ref(),
                            call_stub.as_ref(),
                            call_term,
                            return_from_jmp,
                            &return_from_sub.term.calling_convention,
                        )
                        .map(NodeValue::Value)
                }
            },
            Edge::ExternCallStub(call) => self
                .context
                .update_call_stub(node_value.unwrap_value(), call)
                .map(NodeValue::Value),
            Edge::Jump(jump, untaken_conditional) => {
                let value_after_condition = if let Jmp::CBranch {
                    target: _,
                    condition,
                } = &jump.term
                {
                    let block = graph[start_node].get_block();
                    self.context.specialize_conditional(
                        node_value.unwrap_value(),
                        condition,
                        block,
                        true,
                    )
                } else if let Some(untaken_conditional_jump) = untaken_conditional {
                    if let Jmp::CBranch {
                        target: _,
                        condition,
                    } = &untaken_conditional_jump.term
                    {
                        let block = graph[start_node].get_block();
                        self.context.specialize_conditional(
                            node_value.unwrap_value(),
                            condition,
                            block,
                            false,
                        )
                    } else {
                        panic!("Malformed control flow graph");
                    }
                } else {
                    Some(node_value.unwrap_value().clone())
                };
                if let Some(value) = value_after_condition {
                    self.context
                        .update_jump(
                            &value,
                            jump,
                            *untaken_conditional,
                            graph[end_node].get_block(),
                        )
                        .map(NodeValue::Value)
                } else {
                    None
                }
            }
        }
    }
}

/// Generate a new computation from the corresponding context and an optional default value for nodes.
pub fn create_computation<'a, T: Context<'a>>(
    problem: T,
    default_value: Option<T::Value>,
) -> super::fixpoint::Computation<GeneralizedContext<'a, T>> {
    let generalized_problem = GeneralizedContext::new(problem);
    super::fixpoint::Computation::new(generalized_problem, default_value.map(NodeValue::Value))
}

/// Returns a node ordering with callee nodes behind caller nodes.
pub fn create_bottom_up_worklist(graph: &Graph) -> Vec<NodeIndex> {
    let mut graph = graph.clone();
    graph.retain_edges(|frozen, edge| !matches!(frozen[edge], Edge::Call(..)));
    petgraph::algo::kosaraju_scc(&graph)
        .into_iter()
        .flatten()
        .collect()
}

/// Returns a node ordering with caller nodes behind callee nodes.
pub fn create_top_down_worklist(graph: &Graph) -> Vec<NodeIndex> {
    let mut graph = graph.clone();
    graph.retain_edges(|frozen, edge| !matches!(frozen[edge], Edge::CrReturnStub));
    petgraph::algo::kosaraju_scc(&graph)
        .into_iter()
        .flatten()
        .collect()
}

/// Generate a new computation from the corresponding context and an optional default value for nodes.
/// Uses a bottom up worklist order when computing the fixpoint.
///
/// The worklist order prefers callee nodes before caller nodes.
pub fn create_computation_with_bottom_up_worklist_order<'a, T: Context<'a>>(
    problem: T,
    default_value: Option<T::Value>,
) -> super::fixpoint::Computation<GeneralizedContext<'a, T>> {
    let priority_sorted_nodes: Vec<NodeIndex> = create_bottom_up_worklist(problem.get_graph());
    let generalized_problem = GeneralizedContext::new(problem);
    super::fixpoint::Computation::from_node_priority_list(
        generalized_problem,
        default_value.map(NodeValue::Value),
        priority_sorted_nodes,
    )
}

/// Generate a new computation from the corresponding context and an optional default value for nodes.
/// Uses a top down worklist order when computing the fixpoint.
///
/// The worklist order prefers caller nodes before callee nodes.
pub fn create_computation_with_top_down_worklist_order<'a, T: Context<'a>>(
    problem: T,
    default_value: Option<T::Value>,
) -> super::fixpoint::Computation<GeneralizedContext<'a, T>> {
    let priority_sorted_nodes: Vec<NodeIndex> = create_top_down_worklist(problem.get_graph());
    let generalized_problem = GeneralizedContext::new(problem);
    super::fixpoint::Computation::from_node_priority_list(
        generalized_problem,
        default_value.map(NodeValue::Value),
        priority_sorted_nodes,
    )
}

#[cfg(test)]
mod tests {

    use crate::{
        analysis::{
            expression_propagation::Context,
            forward_interprocedural_fixpoint::{
                create_computation_with_bottom_up_worklist_order,
                create_computation_with_top_down_worklist_order,
            },
        },
        expr,
        intermediate_representation::*,
    };
    use std::collections::{BTreeMap, HashMap};

    fn new_block(name: &str) -> Term<Blk> {
        Term {
            tid: Tid::new(name),
            term: Blk {
                defs: vec![],
                jmps: vec![],
                indirect_jmp_targets: Vec::new(),
            },
        }
    }

    /// Creates a project with one caller function of two blocks and one callee function of one block.
    fn mock_project() -> Project {
        let mut callee_block = new_block("callee block");
        callee_block.term.jmps.push(Term {
            tid: Tid::new("ret"),
            term: Jmp::Return(expr!("42:4")),
        });

        let called_function = Term {
            tid: Tid::new("called_function"),
            term: Sub {
                name: "called_function".to_string(),
                blocks: vec![callee_block],
                calling_convention: Some("_stdcall".to_string()),
            },
        };

        let mut caller_block_2 = new_block("caller_block_2");

        let mut caller_block_1 = new_block("caller_block_1");
        caller_block_1.term.jmps.push(Term {
            tid: Tid::new("call"),
            term: Jmp::Call {
                target: called_function.tid.clone(),
                return_: Some(caller_block_2.tid.clone()),
            },
        });

        caller_block_2.term.jmps.push(Term {
            tid: Tid::new("jmp"),
            term: Jmp::Branch(caller_block_1.tid.clone()),
        });

        let caller_function = Term {
            tid: Tid::new("caller_function"),
            term: Sub {
                name: "caller_function".to_string(),
                blocks: vec![caller_block_1, caller_block_2],
                calling_convention: Some("_stdcall".to_string()),
            },
        };
        let mut project = Project::mock_x64();
        project.program.term.subs = BTreeMap::from([
            (caller_function.tid.clone(), caller_function.clone()),
            (called_function.tid.clone(), called_function.clone()),
        ]);

        project
    }

    #[test]
    /// Checks if the nodes corresponding to the callee function are first in the worklist.
    fn check_bottom_up_worklist() {
        let project = mock_project();
        let graph = crate::analysis::graph::get_program_cfg(&project.program);
        let context = Context::new(&graph);
        let comp = create_computation_with_bottom_up_worklist_order(context, Some(HashMap::new()));
        // The last two nodes should belong to the callee
        for node in comp.get_worklist()[6..].iter() {
            match graph[*node] {
                crate::analysis::graph::Node::BlkStart(_, sub)
                | crate::analysis::graph::Node::BlkEnd(_, sub) => {
                    assert_eq!(sub.tid, Tid::new("called_function"))
                }
                _ => panic!(),
            }
        }
    }

    #[test]
    fn check_top_down_worklist() {
        let project = mock_project();
        let graph = crate::analysis::graph::get_program_cfg(&project.program);
        let context = Context::new(&graph);
        let comp = create_computation_with_top_down_worklist_order(context, Some(HashMap::new()));
        // The first two nodes should belong to the callee
        for node in comp.get_worklist()[..2].iter() {
            match graph[*node] {
                crate::analysis::graph::Node::BlkStart(_, sub)
                | crate::analysis::graph::Node::BlkEnd(_, sub) => {
                    assert_eq!(sub.tid, Tid::new("called_function"))
                }
                _ => panic!(),
            }
        }
    }
}