Merge pull request #11311 from MathiasVP/repair-mustflow

C++: Repair `MustFlow` library for use-use flow

Author: MathiasVP

cpp/ql/lib/change-notes/2022-11-16-must-flow.md

@@ -0,0 +1,4 @@
+---
+category: breaking
+---
+The predicates in the `MustFlow::Configuration` class used by the `MustFlow` library (`semmle.code.cpp.ir.dataflow.MustFlow`) have changed to be defined directly in terms of the C++ IR instead of IR dataflow nodes.

cpp/ql/lib/semmle/code/cpp/ir/dataflow/MustFlow.qll

@@ -5,7 +5,6 @@
  */
 
 private import cpp
-import semmle.code.cpp.ir.dataflow.DataFlow
 private import semmle.code.cpp.ir.IR
 
 /**
@@ -25,18 +24,18 @@ abstract class MustFlowConfiguration extends string {
   /**
    * Holds if `source` is a relevant data flow source.
    */
-  abstract predicate isSource(DataFlow::Node source);
+  abstract predicate isSource(Instruction source);
 
   /**
    * Holds if `sink` is a relevant data flow sink.
    */
-  abstract predicate isSink(DataFlow::Node sink);
+  abstract predicate isSink(Operand sink);
 
   /**
    * Holds if the additional flow step from `node1` to `node2` must be taken
    * into account in the analysis.
    */
-  predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
+  predicate isAdditionalFlowStep(Operand node1, Instruction node2) { none() }
 
   /** Holds if this configuration allows flow from arguments to parameters. */
   predicate allowInterproceduralFlow() { any() }
@@ -48,30 +47,30 @@ abstract class MustFlowConfiguration extends string {
    * included in the module `PathGraph`.
    */
   final predicate hasFlowPath(MustFlowPathNode source, MustFlowPathSink sink) {
-    this.isSource(source.getNode()) and
+    this.isSource(source.getInstruction()) and
     source.getASuccessor+() = sink
   }
 }
 
 /** Holds if `node` flows from a source. */
 pragma[nomagic]
-private predicate flowsFromSource(DataFlow::Node node, MustFlowConfiguration config) {
+private predicate flowsFromSource(Instruction node, MustFlowConfiguration config) {
   config.isSource(node)
   or
-  exists(DataFlow::Node mid |
+  exists(Instruction mid |
     step(mid, node, config) and
     flowsFromSource(mid, pragma[only_bind_into](config))
   )
 }
 
 /** Holds if `node` flows to a sink. */
 pragma[nomagic]
-private predicate flowsToSink(DataFlow::Node node, MustFlowConfiguration config) {
+private predicate flowsToSink(Instruction node, MustFlowConfiguration config) {
   flowsFromSource(node, pragma[only_bind_into](config)) and
   (
-    config.isSink(node)
+    config.isSink(node.getAUse())
     or
-    exists(DataFlow::Node mid |
+    exists(Instruction mid |
       step(node, mid, config) and
       flowsToSink(mid, pragma[only_bind_into](config))
     )
@@ -198,12 +197,13 @@ private module Cached {
   }
 
   cached
-  predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) {
-    instructionToOperandStep(nodeFrom.asInstruction(), nodeTo.asOperand())
-    or
-    flowThroughCallable(nodeFrom.asInstruction(), nodeTo.asInstruction())
+  predicate step(Instruction nodeFrom, Instruction nodeTo) {
+    exists(Operand mid |
+      instructionToOperandStep(nodeFrom, mid) and
+      operandToInstructionStep(mid, nodeTo)
+    )
     or
-    operandToInstructionStep(nodeFrom.asOperand(), nodeTo.asInstruction())
+    flowThroughCallable(nodeFrom, nodeTo)
   }
 }
 
@@ -213,12 +213,12 @@ private module Cached {
  * way around.
  */
 pragma[inline]
-private Declaration getEnclosingCallable(DataFlow::Node n) {
-  pragma[only_bind_into](result) = pragma[only_bind_out](n).getEnclosingCallable()
+private IRFunction getEnclosingCallable(Instruction n) {
+  pragma[only_bind_into](result) = pragma[only_bind_out](n).getEnclosingIRFunction()
 }
 
 /** Holds if `nodeFrom` flows to `nodeTo`. */
-private predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo, MustFlowConfiguration config) {
+private predicate step(Instruction nodeFrom, Instruction nodeTo, MustFlowConfiguration config) {
   exists(config) and
   Cached::step(pragma[only_bind_into](nodeFrom), pragma[only_bind_into](nodeTo)) and
   (
@@ -227,45 +227,45 @@ private predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo, MustFlowC
     getEnclosingCallable(nodeFrom) = getEnclosingCallable(nodeTo)
   )
   or
-  config.isAdditionalFlowStep(nodeFrom, nodeTo)
+  config.isAdditionalFlowStep(nodeFrom.getAUse(), nodeTo)
 }
 
 private newtype TLocalPathNode =
-  MkLocalPathNode(DataFlow::Node n, MustFlowConfiguration config) {
+  MkLocalPathNode(Instruction n, MustFlowConfiguration config) {
     flowsToSink(n, config) and
     (
       config.isSource(n)
       or
-      exists(MustFlowPathNode mid | step(mid.getNode(), n, config))
+      exists(MustFlowPathNode mid | step(mid.getInstruction(), n, config))
     )
   }
 
 /** A `Node` that is in a path from a source to a sink. */
 class MustFlowPathNode extends TLocalPathNode {
-  DataFlow::Node n;
+  Instruction n;
 
   MustFlowPathNode() { this = MkLocalPathNode(n, _) }
 
   /** Gets the underlying node. */
-  DataFlow::Node getNode() { result = n }
+  Instruction getInstruction() { result = n }
 
   /** Gets a textual representation of this node. */
-  string toString() { result = n.toString() }
+  string toString() { result = n.getAst().toString() }
 
   /** Gets the location of this element. */
   Location getLocation() { result = n.getLocation() }
 
   /** Gets a successor node, if any. */
   MustFlowPathNode getASuccessor() {
-    step(this.getNode(), result.getNode(), this.getConfiguration())
+    step(this.getInstruction(), result.getInstruction(), this.getConfiguration())
   }
 
   /** Gets the associated configuration. */
   MustFlowConfiguration getConfiguration() { this = MkLocalPathNode(_, result) }
 }
 
 private class MustFlowPathSink extends MustFlowPathNode {
-  MustFlowPathSink() { this.getConfiguration().isSink(this.getNode()) }
+  MustFlowPathSink() { this.getConfiguration().isSink(this.getInstruction().getAUse()) }
 }
 
 /**

cpp/ql/src/Likely Bugs/Memory Management/ReturnStackAllocatedMemory.ql

@@ -26,11 +26,11 @@ predicate intentionallyReturnsStackPointer(Function f) {
 class ReturnStackAllocatedMemoryConfig extends MustFlowConfiguration {
   ReturnStackAllocatedMemoryConfig() { this = "ReturnStackAllocatedMemoryConfig" }
 
-  override predicate isSource(DataFlow::Node source) {
+  override predicate isSource(Instruction source) {
     // Holds if `source` is a node that represents the use of a stack variable
     exists(VariableAddressInstruction var, Function func |
-      var = source.asInstruction() and
-      func = var.getEnclosingFunction() and
+      var = source and
+      func = source.getEnclosingFunction() and
       var.getAstVariable() instanceof StackVariable and
       // Pointer-to-member types aren't properly handled in the dbscheme.
       not var.getResultType() instanceof PointerToMemberType and
@@ -40,15 +40,15 @@ class ReturnStackAllocatedMemoryConfig extends MustFlowConfiguration {
     )
   }
 
-  override predicate isSink(DataFlow::Node sink) {
+  override predicate isSink(Operand sink) {
     // Holds if `sink` is a node that represents the `StoreInstruction` that is subsequently used in
     // a `ReturnValueInstruction`.
     // We use the `StoreInstruction` instead of the instruction that defines the
     // `ReturnValueInstruction`'s source value operand because the former has better location information.
     exists(StoreInstruction store |
       store.getDestinationAddress().(VariableAddressInstruction).getIRVariable() instanceof
         IRReturnVariable and
-      sink.asOperand() = store.getSourceValueOperand()
+      sink = store.getSourceValueOperand()
     )
   }
 
@@ -77,10 +77,10 @@ class ReturnStackAllocatedMemoryConfig extends MustFlowConfiguration {
    * }
    * ```
    */
-  override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    node2.asInstruction().(FieldAddressInstruction).getObjectAddressOperand() = node1.asOperand()
+  override predicate isAdditionalFlowStep(Operand node1, Instruction node2) {
+    node2.(FieldAddressInstruction).getObjectAddressOperand() = node1
     or
-    node2.asInstruction().(PointerOffsetInstruction).getLeftOperand() = node1.asOperand()
+    node2.(PointerOffsetInstruction).getLeftOperand() = node1
   }
 }
 
@@ -89,6 +89,6 @@ from
   ReturnStackAllocatedMemoryConfig conf
 where
   conf.hasFlowPath(pragma[only_bind_into](source), pragma[only_bind_into](sink)) and
-  source.getNode().asInstruction() = var
-select sink.getNode(), source, sink, "May return stack-allocated memory from $@.", var.getAst(),
-  var.getAst().toString()
+  source.getInstruction() = var
+select sink.getInstruction(), source, sink, "May return stack-allocated memory from $@.",
+  var.getAst(), var.getAst().toString()

cpp/ql/src/Likely Bugs/OO/UnsafeUseOfThis.ql

@@ -22,37 +22,40 @@ import PathGraph
 class UnsafeUseOfThisConfig extends MustFlowConfiguration {
   UnsafeUseOfThisConfig() { this = "UnsafeUseOfThisConfig" }
 
-  override predicate isSource(DataFlow::Node source) { isSource(source, _, _) }
+  override predicate isSource(Instruction source) { isSource(source, _, _) }
 
-  override predicate isSink(DataFlow::Node sink) { isSink(sink, _) }
+  override predicate isSink(Operand sink) { isSink(sink, _) }
 }
 
-/** Holds if `instr` is a `this` pointer used by the call instruction `call`. */
-predicate isSink(DataFlow::Node sink, CallInstruction call) {
+/** Holds if `sink` is a `this` pointer used by the call instruction `call`. */
+predicate isSink(Operand sink, CallInstruction call) {
   exists(PureVirtualFunction func |
     call.getStaticCallTarget() = func and
-    call.getThisArgument() = sink.asInstruction() and
+    call.getThisArgumentOperand() = sink and
     // Weed out implicit calls to destructors of a base class
     not func instanceof Destructor
   )
 }
 
-/** Holds if `init` initializes the `this` pointer in class `c`. */
-predicate isSource(DataFlow::Node source, string msg, Class c) {
-  exists(InitializeParameterInstruction init | init = source.asInstruction() |
-    (
-      exists(Constructor func |
-        not func instanceof CopyConstructor and
-        not func instanceof MoveConstructor and
-        func = init.getEnclosingFunction() and
-        msg = "construction"
-      )
-      or
-      init.getEnclosingFunction() instanceof Destructor and msg = "destruction"
-    ) and
-    init.getIRVariable() instanceof IRThisVariable and
-    init.getEnclosingFunction().getDeclaringType() = c
-  )
+/**
+ * Holds if `source` initializes the `this` pointer in class `c`.
+ *
+ * The string `msg` describes whether the enclosing function is a
+ * constructor or destructor.
+ */
+predicate isSource(InitializeParameterInstruction source, string msg, Class c) {
+  (
+    exists(Constructor func |
+      not func instanceof CopyConstructor and
+      not func instanceof MoveConstructor and
+      func = source.getEnclosingFunction() and
+      msg = "construction"
+    )
+    or
+    source.getEnclosingFunction() instanceof Destructor and msg = "destruction"
+  ) and
+  source.getIRVariable() instanceof IRThisVariable and
+  source.getEnclosingFunction().getDeclaringType() = c
 }
 
 /**
@@ -68,8 +71,8 @@ predicate flows(
 ) {
   exists(UnsafeUseOfThisConfig conf |
     conf.hasFlowPath(source, sink) and
-    isSource(source.getNode(), msg, sourceClass) and
-    isSink(sink.getNode(), call)
+    isSource(source.getInstruction(), msg, sourceClass) and
+    isSink(sink.getInstruction().getAUse(), call)
   )
 }