hohn/codeql
1
0
Fork 0
mirror of https://github.com/github/codeql.git synced 2025-12-21 03:06:31 +01:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #13863 from aschackmull/dataflow/pack4

Browse Source 
Dataflow: Move the shared library to a properly shared qlpack.
This commit is contained in:
Anders Schack-Mulligen
2023-08-02 14:19:49 +02:00
committed by GitHub
parent 7bea18773c 1ad51e754e
commit 7bc8bf616f
72 changed files with 7177 additions and 53563 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
cpp/ql/lib/qlpack.yml
View File

@@ -6,6 +6,7 @@ extractor: cpp
library: true
upgrades: upgrades
dependencies:
codeql/dataflow: ${workspace}
codeql/ssa: ${workspace}
codeql/tutorial: ${workspace}
codeql/util: ${workspace}

4
cpp/ql/lib/semmle/code/cpp/dataflow/DataFlow.qll
View File

@@ -26,6 +26,8 @@ import cpp
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow {
import semmle.code.cpp.dataflow.internal.DataFlow
private import semmle.code.cpp.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<CppOldDataFlow>
import semmle.code.cpp.dataflow.internal.DataFlowImpl1
}

450
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlow.qll
View File

@@ -1,450 +0,0 @@
/**
* Provides an implementation of global (interprocedural) data flow. This file
* re-exports the local (intraprocedural) data flow analysis from
* `DataFlowImplSpecific::Public` and adds a global analysis, mainly exposed
* through the `Global` and `GlobalWithState` modules.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
import DataFlowImplCommonPublic
private import DataFlowImpl
/** An input configuration for data flow. */
signature module ConfigSig {
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source);
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink);
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
default predicate isBarrier(Node node) { none() }
/** Holds if data flow into `node` is prohibited. */
default predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
default predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
default predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Holds if `node` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
default predicate neverSkip(Node node) {
isAdditionalFlowStep(node, _) or isAdditionalFlowStep(_, node)
}
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
default int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
default FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `flowPath`. */
default predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `flowPath`. */
default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (as it is in a `path-problem` query).
*/
default predicate includeHiddenNodes() { none() }
}
/** An input configuration for data flow using flow state. */
signature module StateConfigSig {
bindingset[this]
class FlowState;
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state);
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state);
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
default predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
default predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
default predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
default predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
default predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
default predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Holds if `node` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
default predicate neverSkip(Node node) {
isAdditionalFlowStep(node, _) or
isAdditionalFlowStep(_, node) or
isAdditionalFlowStep(node, _, _, _) or
isAdditionalFlowStep(_, _, node, _)
}
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
default int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
default FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `flowPath`. */
default predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `flowPath`. */
default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (as it is in a `path-problem` query).
*/
default predicate includeHiddenNodes() { none() }
}
/**
* Gets the exploration limit for `partialFlow` and `partialFlowRev`
* measured in approximate number of interprocedural steps.
*/
signature int explorationLimitSig();
/**
* The output of a global data flow computation.
*/
signature module GlobalFlowSig {
/**
* A `Node` augmented with a call context (except for sinks) and an access path.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
class PathNode;
/**
* Holds if data can flow from `source` to `sink`.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate flowPath(PathNode source, PathNode sink);
/**
* Holds if data can flow from `source` to `sink`.
*/
predicate flow(Node source, Node sink);
/**
* Holds if data can flow from some source to `sink`.
*/
predicate flowTo(Node sink);
/**
* Holds if data can flow from some source to `sink`.
*/
predicate flowToExpr(DataFlowExpr sink);
}
/**
* Constructs a global data flow computation.
*/
module Global<ConfigSig Config> implements GlobalFlowSig {
private module C implements FullStateConfigSig {
import DefaultState<Config>
import Config
}
import Impl<C>
}
/** DEPRECATED: Use `Global` instead. */
deprecated module Make<ConfigSig Config> implements GlobalFlowSig {
import Global<Config>
}
/**
* Constructs a global data flow computation using flow state.
*/
module GlobalWithState<StateConfigSig Config> implements GlobalFlowSig {
private module C implements FullStateConfigSig {
import Config
}
import Impl<C>
}
/** DEPRECATED: Use `GlobalWithState` instead. */
deprecated module MakeWithState<StateConfigSig Config> implements GlobalFlowSig {
import GlobalWithState<Config>
}
signature class PathNodeSig {
/** Gets a textual representation of this element. */
string toString();
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
);
/** Gets the underlying `Node`. */
Node getNode();
}
signature module PathGraphSig<PathNodeSig PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
predicate edges(PathNode a, PathNode b);
/** Holds if `n` is a node in the graph of data flow path explanations. */
predicate nodes(PathNode n, string key, string val);
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out);
}
/**
* Constructs a `PathGraph` from two `PathGraph`s by disjoint union.
*/
module MergePathGraph<
PathNodeSig PathNode1, PathNodeSig PathNode2, PathGraphSig<PathNode1> Graph1,
PathGraphSig<PathNode2> Graph2>
{
private newtype TPathNode =
TPathNode1(PathNode1 p) or
TPathNode2(PathNode2 p)
/** A node in a graph of path explanations that is formed by disjoint union of the two given graphs. */
class PathNode extends TPathNode {
/** Gets this as a projection on the first given `PathGraph`. */
PathNode1 asPathNode1() { this = TPathNode1(result) }
/** Gets this as a projection on the second given `PathGraph`. */
PathNode2 asPathNode2() { this = TPathNode2(result) }
/** Gets a textual representation of this element. */
string toString() {
result = this.asPathNode1().toString() or
result = this.asPathNode2().toString()
}
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
this.asPathNode1().hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn) or
this.asPathNode2().hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
Node getNode() {
result = this.asPathNode1().getNode() or
result = this.asPathNode2().getNode()
}
}
/**
* Provides the query predicates needed to include a graph in a path-problem query.
*/
module PathGraph implements PathGraphSig<PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
query predicate edges(PathNode a, PathNode b) {
Graph1::edges(a.asPathNode1(), b.asPathNode1()) or
Graph2::edges(a.asPathNode2(), b.asPathNode2())
}
/** Holds if `n` is a node in the graph of data flow path explanations. */
query predicate nodes(PathNode n, string key, string val) {
Graph1::nodes(n.asPathNode1(), key, val) or
Graph2::nodes(n.asPathNode2(), key, val)
}
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
query predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out) {
Graph1::subpaths(arg.asPathNode1(), par.asPathNode1(), ret.asPathNode1(), out.asPathNode1()) or
Graph2::subpaths(arg.asPathNode2(), par.asPathNode2(), ret.asPathNode2(), out.asPathNode2())
}
}
}
/**
* Constructs a `PathGraph` from three `PathGraph`s by disjoint union.
*/
module MergePathGraph3<
PathNodeSig PathNode1, PathNodeSig PathNode2, PathNodeSig PathNode3,
PathGraphSig<PathNode1> Graph1, PathGraphSig<PathNode2> Graph2, PathGraphSig<PathNode3> Graph3>
{
private module MergedInner = MergePathGraph<PathNode1, PathNode2, Graph1, Graph2>;
private module Merged =
MergePathGraph<MergedInner::PathNode, PathNode3, MergedInner::PathGraph, Graph3>;
/** A node in a graph of path explanations that is formed by disjoint union of the three given graphs. */
class PathNode instanceof Merged::PathNode {
/** Gets this as a projection on the first given `PathGraph`. */
PathNode1 asPathNode1() { result = super.asPathNode1().asPathNode1() }
/** Gets this as a projection on the second given `PathGraph`. */
PathNode2 asPathNode2() { result = super.asPathNode1().asPathNode2() }
/** Gets this as a projection on the third given `PathGraph`. */
PathNode3 asPathNode3() { result = super.asPathNode2() }
/** Gets a textual representation of this element. */
string toString() { result = super.toString() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
Node getNode() { result = super.getNode() }
}
/**
* Provides the query predicates needed to include a graph in a path-problem query.
*/
module PathGraph implements PathGraphSig<PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
query predicate edges(PathNode a, PathNode b) { Merged::PathGraph::edges(a, b) }
/** Holds if `n` is a node in the graph of data flow path explanations. */
query predicate nodes(PathNode n, string key, string val) {
Merged::PathGraph::nodes(n, key, val)
}
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
query predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out) {
Merged::PathGraph::subpaths(arg, par, ret, out)
}
}
}

8
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowDispatch.qll
View File

@@ -5,8 +5,8 @@ private import DataFlowUtil
/**
* Gets a function that might be called by `call`.
*/
Function viableCallable(Call call) {
result = call.getTarget()
Function viableCallable(DataFlowCall call) {
result = call.(Call).getTarget()
or
// If the target of the call does not have a body in the snapshot, it might
// be because the target is just a header declaration, and the real target
@@ -58,13 +58,13 @@ private predicate functionSignature(Function f, string qualifiedName, int nparam
* Holds if the set of viable implementations that can be called by `call`
* might be improved by knowing the call context.
*/
predicate mayBenefitFromCallContext(Call call, Function f) { none() }
predicate mayBenefitFromCallContext(DataFlowCall call, Function f) { none() }
/**
* Gets a viable dispatch target of `call` in the context `ctx`. This is
* restricted to those `call`s for which a context might make a difference.
*/
Function viableImplInCallContext(Call call, Call ctx) { none() }
Function viableImplInCallContext(DataFlowCall call, DataFlowCall ctx) { none() }
/** A parameter position represented by an integer. */
class ParameterPosition extends int {

4747
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl.qll
View File

File diff suppressed because it is too large Load Diff

1484
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplCommon.qll
View File

File diff suppressed because it is too large Load Diff

10
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplSpecific.qll
View File

@@ -1,6 +1,9 @@
/**
* Provides C++-specific definitions for use in the data flow library.
*/
private import codeql.dataflow.DataFlowParameter
module Private {
import DataFlowPrivate
import DataFlowDispatch
@@ -9,3 +12,10 @@ module Private {
module Public {
import DataFlowUtil
}
module CppOldDataFlow implements DataFlowParameter {
import Private
import Public
Node exprNode(DataFlowExpr e) { result = Public::exprNode(e) }
}

22
cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowPrivate.qll
View File

@@ -153,10 +153,11 @@ predicate jumpStep(Node n1, Node n2) { none() }
* Thus, `node2` references an object with a field `f` that contains the
* value of `node1`.
*/
predicate storeStep(Node node1, Content f, PostUpdateNode node2) {
predicate storeStep(Node node1, ContentSet f, Node node2) {
exists(ClassAggregateLiteral aggr, Field field |
// The following line requires `node2` to be both an `ExprNode` and a
// The following lines requires `node2` to be both an `ExprNode` and a
// `PostUpdateNode`, which means it must be an `ObjectInitializerNode`.
node2 instanceof PostUpdateNode and
node2.asExpr() = aggr and
f.(FieldContent).getField() = field and
aggr.getAFieldExpr(field) = node1.asExpr()
@@ -167,12 +168,13 @@ predicate storeStep(Node node1, Content f, PostUpdateNode node2) {
node1.asExpr() = a and
a.getLValue() = fa
) and
node2.getPreUpdateNode().asExpr() = fa.getQualifier() and
node2.(PostUpdateNode).getPreUpdateNode().asExpr() = fa.getQualifier() and
f.(FieldContent).getField() = fa.getTarget()
)
or
exists(ConstructorFieldInit cfi |
node2.getPreUpdateNode().(PreConstructorInitThis).getConstructorFieldInit() = cfi and
node2.(PostUpdateNode).getPreUpdateNode().(PreConstructorInitThis).getConstructorFieldInit() =
cfi and
f.(FieldContent).getField() = cfi.getTarget() and
node1.asExpr() = cfi.getExpr()
)
@@ -183,7 +185,7 @@ predicate storeStep(Node node1, Content f, PostUpdateNode node2) {
* Thus, `node1` references an object with a field `f` whose value ends up in
* `node2`.
*/
predicate readStep(Node node1, Content f, Node node2) {
predicate readStep(Node node1, ContentSet f, Node node2) {
exists(FieldAccess fr |
node1.asExpr() = fr.getQualifier() and
fr.getTarget() = f.(FieldContent).getField() and
@@ -195,7 +197,7 @@ predicate readStep(Node node1, Content f, Node node2) {
/**
* Holds if values stored inside content `c` are cleared at node `n`.
*/
predicate clearsContent(Node n, Content c) {
predicate clearsContent(Node n, ContentSet c) {
none() // stub implementation
}
@@ -235,12 +237,6 @@ class CastNode extends Node {
CastNode() { none() } // stub implementation
}
/**
* Holds if `n` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
predicate neverSkipInPathGraph(Node n) { none() }
class DataFlowCallable = Function;
class DataFlowExpr = Expr;
@@ -265,8 +261,6 @@ class DataFlowCall extends Expr instanceof Call {
predicate isUnreachableInCall(Node n, DataFlowCall call) { none() } // stub implementation
int accessPathLimit() { result = 5 }
/**
* Holds if access paths with `c` at their head always should be tracked at high
* precision. This disables adaptive access path precision for such access paths.

4
cpp/ql/lib/semmle/code/cpp/dataflow/new/DataFlow.qll
View File

@@ -26,6 +26,8 @@ import cpp
* global (inter-procedural) data flow analyses.
*/
module DataFlow {
import semmle.code.cpp.ir.dataflow.internal.DataFlow
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<CppDataFlow>
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

4
cpp/ql/lib/semmle/code/cpp/ir/dataflow/DataFlow.qll
View File

@@ -22,6 +22,8 @@
import cpp
module DataFlow {
import semmle.code.cpp.ir.dataflow.internal.DataFlow
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<CppDataFlow>
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

450
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlow.qll
View File

@@ -1,450 +0,0 @@
/**
* Provides an implementation of global (interprocedural) data flow. This file
* re-exports the local (intraprocedural) data flow analysis from
* `DataFlowImplSpecific::Public` and adds a global analysis, mainly exposed
* through the `Global` and `GlobalWithState` modules.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
import DataFlowImplCommonPublic
private import DataFlowImpl
/** An input configuration for data flow. */
signature module ConfigSig {
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source);
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink);
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
default predicate isBarrier(Node node) { none() }
/** Holds if data flow into `node` is prohibited. */
default predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
default predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
default predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Holds if `node` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
default predicate neverSkip(Node node) {
isAdditionalFlowStep(node, _) or isAdditionalFlowStep(_, node)
}
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
default int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
default FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `flowPath`. */
default predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `flowPath`. */
default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (as it is in a `path-problem` query).
*/
default predicate includeHiddenNodes() { none() }
}
/** An input configuration for data flow using flow state. */
signature module StateConfigSig {
bindingset[this]
class FlowState;
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state);
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state);
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
default predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
default predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
default predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
default predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
default predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
default predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Holds if `node` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
default predicate neverSkip(Node node) {
isAdditionalFlowStep(node, _) or
isAdditionalFlowStep(_, node) or
isAdditionalFlowStep(node, _, _, _) or
isAdditionalFlowStep(_, _, node, _)
}
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
default int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
default FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `flowPath`. */
default predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `flowPath`. */
default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (as it is in a `path-problem` query).
*/
default predicate includeHiddenNodes() { none() }
}
/**
* Gets the exploration limit for `partialFlow` and `partialFlowRev`
* measured in approximate number of interprocedural steps.
*/
signature int explorationLimitSig();
/**
* The output of a global data flow computation.
*/
signature module GlobalFlowSig {
/**
* A `Node` augmented with a call context (except for sinks) and an access path.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
class PathNode;
/**
* Holds if data can flow from `source` to `sink`.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate flowPath(PathNode source, PathNode sink);
/**
* Holds if data can flow from `source` to `sink`.
*/
predicate flow(Node source, Node sink);
/**
* Holds if data can flow from some source to `sink`.
*/
predicate flowTo(Node sink);
/**
* Holds if data can flow from some source to `sink`.
*/
predicate flowToExpr(DataFlowExpr sink);
}
/**
* Constructs a global data flow computation.
*/
module Global<ConfigSig Config> implements GlobalFlowSig {
private module C implements FullStateConfigSig {
import DefaultState<Config>
import Config
}
import Impl<C>
}
/** DEPRECATED: Use `Global` instead. */
deprecated module Make<ConfigSig Config> implements GlobalFlowSig {
import Global<Config>
}
/**
* Constructs a global data flow computation using flow state.
*/
module GlobalWithState<StateConfigSig Config> implements GlobalFlowSig {
private module C implements FullStateConfigSig {
import Config
}
import Impl<C>
}
/** DEPRECATED: Use `GlobalWithState` instead. */
deprecated module MakeWithState<StateConfigSig Config> implements GlobalFlowSig {
import GlobalWithState<Config>
}
signature class PathNodeSig {
/** Gets a textual representation of this element. */
string toString();
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
);
/** Gets the underlying `Node`. */
Node getNode();
}
signature module PathGraphSig<PathNodeSig PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
predicate edges(PathNode a, PathNode b);
/** Holds if `n` is a node in the graph of data flow path explanations. */
predicate nodes(PathNode n, string key, string val);
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out);
}
/**
* Constructs a `PathGraph` from two `PathGraph`s by disjoint union.
*/
module MergePathGraph<
PathNodeSig PathNode1, PathNodeSig PathNode2, PathGraphSig<PathNode1> Graph1,
PathGraphSig<PathNode2> Graph2>
{
private newtype TPathNode =
TPathNode1(PathNode1 p) or
TPathNode2(PathNode2 p)
/** A node in a graph of path explanations that is formed by disjoint union of the two given graphs. */
class PathNode extends TPathNode {
/** Gets this as a projection on the first given `PathGraph`. */
PathNode1 asPathNode1() { this = TPathNode1(result) }
/** Gets this as a projection on the second given `PathGraph`. */
PathNode2 asPathNode2() { this = TPathNode2(result) }
/** Gets a textual representation of this element. */
string toString() {
result = this.asPathNode1().toString() or
result = this.asPathNode2().toString()
}
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
this.asPathNode1().hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn) or
this.asPathNode2().hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
Node getNode() {
result = this.asPathNode1().getNode() or
result = this.asPathNode2().getNode()
}
}
/**
* Provides the query predicates needed to include a graph in a path-problem query.
*/
module PathGraph implements PathGraphSig<PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
query predicate edges(PathNode a, PathNode b) {
Graph1::edges(a.asPathNode1(), b.asPathNode1()) or
Graph2::edges(a.asPathNode2(), b.asPathNode2())
}
/** Holds if `n` is a node in the graph of data flow path explanations. */
query predicate nodes(PathNode n, string key, string val) {
Graph1::nodes(n.asPathNode1(), key, val) or
Graph2::nodes(n.asPathNode2(), key, val)
}
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
query predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out) {
Graph1::subpaths(arg.asPathNode1(), par.asPathNode1(), ret.asPathNode1(), out.asPathNode1()) or
Graph2::subpaths(arg.asPathNode2(), par.asPathNode2(), ret.asPathNode2(), out.asPathNode2())
}
}
}
/**
* Constructs a `PathGraph` from three `PathGraph`s by disjoint union.
*/
module MergePathGraph3<
PathNodeSig PathNode1, PathNodeSig PathNode2, PathNodeSig PathNode3,
PathGraphSig<PathNode1> Graph1, PathGraphSig<PathNode2> Graph2, PathGraphSig<PathNode3> Graph3>
{
private module MergedInner = MergePathGraph<PathNode1, PathNode2, Graph1, Graph2>;
private module Merged =
MergePathGraph<MergedInner::PathNode, PathNode3, MergedInner::PathGraph, Graph3>;
/** A node in a graph of path explanations that is formed by disjoint union of the three given graphs. */
class PathNode instanceof Merged::PathNode {
/** Gets this as a projection on the first given `PathGraph`. */
PathNode1 asPathNode1() { result = super.asPathNode1().asPathNode1() }
/** Gets this as a projection on the second given `PathGraph`. */
PathNode2 asPathNode2() { result = super.asPathNode1().asPathNode2() }
/** Gets this as a projection on the third given `PathGraph`. */
PathNode3 asPathNode3() { result = super.asPathNode2() }
/** Gets a textual representation of this element. */
string toString() { result = super.toString() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
Node getNode() { result = super.getNode() }
}
/**
* Provides the query predicates needed to include a graph in a path-problem query.
*/
module PathGraph implements PathGraphSig<PathNode> {
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
query predicate edges(PathNode a, PathNode b) { Merged::PathGraph::edges(a, b) }
/** Holds if `n` is a node in the graph of data flow path explanations. */
query predicate nodes(PathNode n, string key, string val) {
Merged::PathGraph::nodes(n, key, val)
}
/**
* Holds if `(arg, par, ret, out)` forms a subpath-tuple, that is, flow through
* a subpath between `par` and `ret` with the connecting edges `arg -> par` and
* `ret -> out` is summarized as the edge `arg -> out`.
*/
query predicate subpaths(PathNode arg, PathNode par, PathNode ret, PathNode out) {
Merged::PathGraph::subpaths(arg, par, ret, out)
}
}
}

6
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowDispatch.qll
View File

@@ -9,7 +9,7 @@ private import DataFlowImplCommon as DataFlowImplCommon
* Gets a function that might be called by `call`.
*/
cached
Function viableCallable(CallInstruction call) {
DataFlowCallable viableCallable(DataFlowCall call) {
DataFlowImplCommon::forceCachingInSameStage() and
result = call.getStaticCallTarget()
or
@@ -235,7 +235,7 @@ private predicate functionSignature(Function f, string qualifiedName, int nparam
* Holds if the set of viable implementations that can be called by `call`
* might be improved by knowing the call context.
*/
predicate mayBenefitFromCallContext(CallInstruction call, Function f) {
predicate mayBenefitFromCallContext(DataFlowCall call, DataFlowCallable f) {
mayBenefitFromCallContext(call, f, _)
}
@@ -259,7 +259,7 @@ private predicate mayBenefitFromCallContext(
* Gets a viable dispatch target of `call` in the context `ctx`. This is
* restricted to those `call`s for which a context might make a difference.
*/
Function viableImplInCallContext(CallInstruction call, CallInstruction ctx) {
DataFlowCallable viableImplInCallContext(DataFlowCall call, DataFlowCall ctx) {
result = viableCallable(call) and
exists(int i, Function f |
mayBenefitFromCallContext(pragma[only_bind_into](call), f, i) and

4747
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll
View File

File diff suppressed because it is too large Load Diff

1484
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImplCommon.qll
View File

File diff suppressed because it is too large Load Diff

10
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImplSpecific.qll
View File

@@ -1,6 +1,9 @@
/**
* Provides IR-specific definitions for use in the data flow library.
*/
private import codeql.dataflow.DataFlowParameter
module Private {
import DataFlowPrivate
import DataFlowDispatch
@@ -9,3 +12,10 @@ module Private {
module Public {
import DataFlowUtil
}
module CppDataFlow implements DataFlowParameter {
import Private
import Public
Node exprNode(DataFlowExpr e) { result = Public::exprNode(e) }
}

20
cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowPrivate.qll
View File

@@ -681,9 +681,7 @@ predicate storeStepImpl(Node node1, Content c, PostFieldUpdateNode node2, boolea
* Thus, `node2` references an object with a field `f` that contains the
* value of `node1`.
*/
predicate storeStep(Node node1, Content c, PostFieldUpdateNode node2) {
storeStepImpl(node1, c, node2, _)
}
predicate storeStep(Node node1, ContentSet c, Node node2) { storeStepImpl(node1, c, node2, _) }
/**
* Holds if `operandFrom` flows to `operandTo` using a sequence of conversion-like
@@ -744,7 +742,7 @@ predicate nodeHasInstruction(Node node, Instruction instr, int indirectionIndex)
* Thus, `node1` references an object with a field `f` whose value ends up in
* `node2`.
*/
predicate readStep(Node node1, Content c, Node node2) {
predicate readStep(Node node1, ContentSet c, Node node2) {
exists(FieldAddress fa1, Operand operand, int numberOfLoads, int indirectionIndex2 |
nodeHasOperand(node2, operand, indirectionIndex2) and
// The `1` here matches the `node2.getIndirectionIndex() = 1` conjunct
@@ -767,7 +765,7 @@ predicate readStep(Node node1, Content c, Node node2) {
/**
* Holds if values stored inside content `c` are cleared at node `n`.
*/
predicate clearsContent(Node n, Content c) {
predicate clearsContent(Node n, ContentSet c) {
n =
any(PostUpdateNode pun, Content d | d.impliesClearOf(c) and storeStepImpl(_, d, pun, true) | pun)
.getPreUpdateNode() and
@@ -792,7 +790,7 @@ predicate clearsContent(Node n, Content c) {
storeStepImpl(_, d, pun, true) and
pun.getPreUpdateNode() = n
|
c.getIndirectionIndex() = d.getIndirectionIndex()
c.(Content).getIndirectionIndex() = d.getIndirectionIndex()
)
)
}
@@ -833,12 +831,6 @@ class CastNode extends Node {
CastNode() { none() } // stub implementation
}
/**
* Holds if `n` should never be skipped over in the `PathGraph` and in path
* explanations.
*/
predicate neverSkipInPathGraph(Node n) { none() }
/**
* A function that may contain code or a variable that may contain itself. When
* flow crosses from one _enclosing callable_ to another, the interprocedural
@@ -853,7 +845,7 @@ class DataFlowType = Type;
/** A function call relevant for data flow. */
class DataFlowCall extends CallInstruction {
Function getEnclosingCallable() { result = this.getEnclosingFunction() }
DataFlowCallable getEnclosingCallable() { result = this.getEnclosingFunction() }
}
module IsUnreachableInCall {
@@ -924,8 +916,6 @@ module IsUnreachableInCall {
import IsUnreachableInCall
int accessPathLimit() { result = 5 }
/**
* Holds if access paths with `c` at their head always should be tracked at high
* precision. This disables adaptive access path precision for such access paths.