Compare commits

..

1 Commits

Author SHA1 Message Date
Alexandre Boulgakov
3c695e3555 Swift: Remove workaround for LWG/issue3657. 2023-06-13 22:56:15 +01:00
3417 changed files with 114691 additions and 171296 deletions

View File

@@ -1,9 +1,3 @@
common --enable_platform_specific_config
build --repo_env=CC=clang --repo_env=CXX=clang++
build:linux --cxxopt=-std=c++20
build:macos --cxxopt=-std=c++20 --cpu=darwin_x86_64
build:windows --cxxopt=/std:c++20 --cxxopt=/Zc:preprocessor
build --repo_env=CC=clang --repo_env=CXX=clang++ --cxxopt="-std=c++20"
try-import %workspace%/local.bazelrc

View File

@@ -1 +1 @@
6.3.1
6.1.2

View File

@@ -1,29 +0,0 @@
name: "Check implicit this warnings"
on:
workflow_dispatch:
pull_request:
paths:
- "**qlpack.yml"
branches:
- main
- "rc/*"
jobs:
check:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Check that implicit this warnings is enabled for all packs
shell: bash
run: |
EXIT_CODE=0
packs="$(find . -iname 'qlpack.yml')"
for pack_file in ${packs}; do
option="$(yq '.warnOnImplicitThis' ${pack_file})"
if [ "${option}" != "true" ]; then
echo "::error file=${pack_file}::warnOnImplicitThis property must be set to 'true' for pack ${pack_file}"
EXIT_CODE=1
fi
done
exit "${EXIT_CODE}"

View File

@@ -7,17 +7,15 @@ on:
- .github/workflows/go-tests-other-os.yml
- .github/actions/**
- codeql-workspace.yml
env:
GO_VERSION: '~1.21.0'
jobs:
test-mac:
name: Test MacOS
runs-on: macos-latest
steps:
- name: Set up Go ${{ env.GO_VERSION }}
- name: Set up Go 1.20
uses: actions/setup-go@v4
with:
go-version: ${{ env.GO_VERSION }}
go-version: '1.20'
id: go
- name: Check out code
@@ -49,10 +47,10 @@ jobs:
name: Test Windows
runs-on: windows-latest-xl
steps:
- name: Set up Go ${{ env.GO_VERSION }}
- name: Set up Go 1.20
uses: actions/setup-go@v4
with:
go-version: ${{ env.GO_VERSION }}
go-version: '1.20'
id: go
- name: Check out code

View File

@@ -15,17 +15,15 @@ on:
- .github/workflows/go-tests.yml
- .github/actions/**
- codeql-workspace.yml
env:
GO_VERSION: '~1.21.0'
jobs:
test-linux:
name: Test Linux (Ubuntu)
runs-on: ubuntu-latest-xl
steps:
- name: Set up Go ${{ env.GO_VERSION }}
- name: Set up Go 1.20
uses: actions/setup-go@v4
with:
go-version: ${{ env.GO_VERSION }}
go-version: '1.20'
id: go
- name: Check out code

View File

@@ -5,9 +5,9 @@ repos:
rev: v3.2.0
hooks:
- id: trailing-whitespace
exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)|.*\.patch
exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)
- id: end-of-file-fixer
exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)|.*\.patch
exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)
- repo: https://github.com/pre-commit/mirrors-clang-format
rev: v13.0.1
@@ -21,11 +21,6 @@ repos:
- id: autopep8
files: ^misc/codegen/.*\.py
- repo: https://github.com/warchant/pre-commit-buildifier
rev: 0.0.2
hooks:
- id: buildifier
- repo: local
hooks:
- id: codeql-format

View File

@@ -14,16 +14,14 @@ If you have an idea for a query that you would like to share with other CodeQL u
1. **Directory structure**
There are eight language-specific query directories in this repository:
There are six language-specific query directories in this repository:
* C/C++: `cpp/ql/src`
* C#: `csharp/ql/src`
* Go: `go/ql/src`
* Java/Kotlin: `java/ql/src`
* Java: `java/ql/src`
* JavaScript: `javascript/ql/src`
* Python: `python/ql/src`
* Ruby: `ruby/ql/src`
* Swift: `swift/ql/src`
Each language-specific directory contains further subdirectories that group queries based on their `@tags` or purpose.
- Experimental queries and libraries are stored in the `experimental` subdirectory within each language-specific directory in the [CodeQL repository](https://github.com/github/codeql). For example, experimental Java queries and libraries are stored in `java/ql/src/experimental` and any corresponding tests in `java/ql/test/experimental`.

View File

@@ -1,4 +1,24 @@
{
"DataFlow Java/C++/C#/Go/Python/Ruby/Swift": [
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlow.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlow.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlow.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlow.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlow.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlow.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlow.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/DataFlow.qll"
],
"DataFlowImpl Java/C++/C#/Go/Python/Ruby/Swift": [
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl.qll"
],
"DataFlow Java/C++/C#/Go/Python/Ruby/Swift Legacy Configuration": [
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl1.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl2.qll",
@@ -22,6 +42,7 @@
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl5.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl1.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl2.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImplForStringsNewReplacer.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl1.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl2.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl3.qll",
@@ -32,6 +53,26 @@
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplForPathname.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl1.qll"
],
"DataFlow Java/C++/C#/Go/Python/Ruby/Swift Common": [
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImplCommon.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplCommon.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImplCommon.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImplCommon.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImplCommon.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImplCommon.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplCommon.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImplCommon.qll"
],
"TaintTracking Java/C++/C#/Go/Python/Ruby/Swift": [
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTracking.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTracking.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking1/TaintTracking.qll",
"go/ql/lib/semmle/go/dataflow/internal/tainttracking1/TaintTracking.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/tainttracking1/TaintTracking.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/tainttracking1/TaintTracking.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/tainttracking1/TaintTracking.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/tainttracking1/TaintTracking.qll"
],
"TaintTracking Legacy Configuration Java/C++/C#/Go/Python/Ruby/Swift": [
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
@@ -473,14 +514,15 @@
"ruby/ql/lib/codeql/ruby/security/internal/SensitiveDataHeuristics.qll",
"swift/ql/lib/codeql/swift/security/internal/SensitiveDataHeuristics.qll"
],
"CFG": [
"csharp/ql/lib/semmle/code/csharp/controlflow/internal/ControlFlowGraphImplShared.qll",
"ruby/ql/lib/codeql/ruby/controlflow/internal/ControlFlowGraphImplShared.qll",
"swift/ql/lib/codeql/swift/controlflow/internal/ControlFlowGraphImplShared.qll"
],
"TypeTracker": [
"python/ql/lib/semmle/python/dataflow/new/internal/TypeTracker.qll",
"ruby/ql/lib/codeql/ruby/typetracking/TypeTracker.qll"
],
"SummaryTypeTracker": [
"python/ql/lib/semmle/python/dataflow/new/internal/SummaryTypeTracker.qll",
"ruby/ql/lib/codeql/ruby/typetracking/internal/SummaryTypeTracker.qll"
],
"AccessPathSyntax": [
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/AccessPathSyntax.qll",
"go/ql/lib/semmle/go/dataflow/internal/AccessPathSyntax.qll",
@@ -556,9 +598,5 @@
"EncryptionKeySizes Python/Java": [
"python/ql/lib/semmle/python/security/internal/EncryptionKeySizes.qll",
"java/ql/lib/semmle/code/java/security/internal/EncryptionKeySizes.qll"
],
"Python model summaries test extension": [
"python/ql/test/experimental/dataflow/model-summaries/InlineTaintTest.ext.yml",
"python/ql/test/experimental/dataflow/model-summaries/NormalDataflowTest.ext.yml"
]
}

View File

@@ -327,7 +327,7 @@ namespace Semmle.Autobuild.Cpp.Tests
{
Actions.RunProcess[@"cmd.exe /C nuget restore C:\Project\test.sln -DisableParallelProcessing"] = 1;
Actions.RunProcess[@"cmd.exe /C C:\Project\.nuget\nuget.exe restore C:\Project\test.sln -DisableParallelProcessing"] = 0;
Actions.RunProcess[@"cmd.exe /C CALL ^""C:\Program^ Files^ ^(x86^)\Microsoft^ Visual^ Studio^ 14.0\VC\vcvarsall.bat^"" && set Platform=&& type NUL && msbuild C:\Project\test.sln /t:rebuild /p:Platform=""x86"" /p:Configuration=""Release"""] = 0;
Actions.RunProcess[@"cmd.exe /C CALL ^""C:\Program Files ^(x86^)\Microsoft Visual Studio 14.0\VC\vcvarsall.bat^"" && set Platform=&& type NUL && msbuild C:\Project\test.sln /t:rebuild /p:Platform=""x86"" /p:Configuration=""Release"""] = 0;
Actions.RunProcessOut[@"C:\Program Files (x86)\Microsoft Visual Studio\Installer\vswhere.exe -prerelease -legacy -property installationPath"] = "";
Actions.RunProcess[@"C:\Program Files (x86)\Microsoft Visual Studio\Installer\vswhere.exe -prerelease -legacy -property installationPath"] = 1;
Actions.RunProcess[@"C:\Program Files (x86)\Microsoft Visual Studio\Installer\vswhere.exe -prerelease -legacy -property installationVersion"] = 0;

View File

@@ -1,2 +0,0 @@
description: Remove _Float128 type
compatibility: full

View File

@@ -2,4 +2,3 @@ name: codeql/cpp-downgrades
groups: cpp
downgrades: .
library: true
warnOnImplicitThis: true

View File

@@ -4,4 +4,3 @@ groups:
- examples
dependencies:
codeql/cpp-all: ${workspace}
warnOnImplicitThis: true

View File

@@ -1,70 +1,3 @@
## 0.9.2
### Minor Analysis Improvements
* Only the 2 level indirection of `argv` (corresponding to `**argv`) is consided for `FlowSource`.
## 0.9.1
No user-facing changes.
## 0.9.0
### Breaking Changes
* The `shouldPrintFunction` predicate from `PrintAstConfiguration` has been replaced by `shouldPrintDeclaration`. Users should now override `shouldPrintDeclaration` if they want to limit the declarations that should be printed.
* The `shouldPrintFunction` predicate from `PrintIRConfiguration` has been replaced by `shouldPrintDeclaration`. Users should now override `shouldPrintDeclaration` if they want to limit the declarations that should be printed.
### Major Analysis Improvements
* The `PrintAST` library now also prints global and namespace variables and their initializers.
### Minor Analysis Improvements
* The `_Float128x` type is no longer exposed as a builtin type. As this type could not occur any code base, this should only affect queries that explicitly looked at the builtin types.
## 0.8.1
### Deprecated APIs
* The library `semmle.code.cpp.dataflow.DataFlow` has been deprecated. Please use `semmle.code.cpp.dataflow.new.DataFlow` instead.
### New Features
* The `DataFlow::StateConfigSig` signature module has gained default implementations for `isBarrier/2` and `isAdditionalFlowStep/4`.
Hence it is no longer needed to provide `none()` implementations of these predicates if they are not needed.
### Minor Analysis Improvements
* Data flow configurations can now include a predicate `neverSkip(Node node)`
in order to ensure inclusion of certain nodes in the path explanations. The
predicate defaults to the end-points of the additional flow steps provided in
the configuration, which means that such steps now always are visible by
default in path explanations.
* The `IRGuards` library has improved handling of pointer addition and subtraction operations.
## 0.8.0
### New Features
* The `ProductFlow::StateConfigSig` signature now includes default predicates for `isBarrier1`, `isBarrier2`, `isAdditionalFlowStep1`, and `isAdditionalFlowStep1`. Hence, it is no longer needed to provide `none()` implementations of these predicates if they are not needed.
### Minor Analysis Improvements
* Deleted the deprecated `getURL` predicate from the `Container`, `Folder`, and `File` classes. Use the `getLocation` predicate instead.
## 0.7.4
No user-facing changes.
## 0.7.3
### Minor Analysis Improvements
* Deleted the deprecated `hasCopyConstructor` predicate from the `Class` class in `Class.qll`.
* Deleted many deprecated predicates and classes with uppercase `AST`, `SSA`, `CFG`, `API`, etc. in their names. Use the PascalCased versions instead.
* Deleted the deprecated `CodeDuplication.qll` file.
## 0.7.2
### New Features

View File

@@ -1,7 +1,6 @@
## 0.7.3
### Minor Analysis Improvements
---
category: minorAnalysis
---
* Deleted the deprecated `hasCopyConstructor` predicate from the `Class` class in `Class.qll`.
* Deleted many deprecated predicates and classes with uppercase `AST`, `SSA`, `CFG`, `API`, etc. in their names. Use the PascalCased versions instead.
* Deleted the deprecated `CodeDuplication.qll` file.
* Deleted the deprecated `CodeDuplication.qll` file.

View File

@@ -1,3 +0,0 @@
## 0.7.4
No user-facing changes.

View File

@@ -1,9 +0,0 @@
## 0.8.0
### New Features
* The `ProductFlow::StateConfigSig` signature now includes default predicates for `isBarrier1`, `isBarrier2`, `isAdditionalFlowStep1`, and `isAdditionalFlowStep1`. Hence, it is no longer needed to provide `none()` implementations of these predicates if they are not needed.
### Minor Analysis Improvements
* Deleted the deprecated `getURL` predicate from the `Container`, `Folder`, and `File` classes. Use the `getLocation` predicate instead.

View File

@@ -1,19 +0,0 @@
## 0.8.1
### Deprecated APIs
* The library `semmle.code.cpp.dataflow.DataFlow` has been deprecated. Please use `semmle.code.cpp.dataflow.new.DataFlow` instead.
### New Features
* The `DataFlow::StateConfigSig` signature module has gained default implementations for `isBarrier/2` and `isAdditionalFlowStep/4`.
Hence it is no longer needed to provide `none()` implementations of these predicates if they are not needed.
### Minor Analysis Improvements
* Data flow configurations can now include a predicate `neverSkip(Node node)`
in order to ensure inclusion of certain nodes in the path explanations. The
predicate defaults to the end-points of the additional flow steps provided in
the configuration, which means that such steps now always are visible by
default in path explanations.
* The `IRGuards` library has improved handling of pointer addition and subtraction operations.

View File

@@ -1,14 +0,0 @@
## 0.9.0
### Breaking Changes
* The `shouldPrintFunction` predicate from `PrintAstConfiguration` has been replaced by `shouldPrintDeclaration`. Users should now override `shouldPrintDeclaration` if they want to limit the declarations that should be printed.
* The `shouldPrintFunction` predicate from `PrintIRConfiguration` has been replaced by `shouldPrintDeclaration`. Users should now override `shouldPrintDeclaration` if they want to limit the declarations that should be printed.
### Major Analysis Improvements
* The `PrintAST` library now also prints global and namespace variables and their initializers.
### Minor Analysis Improvements
* The `_Float128x` type is no longer exposed as a builtin type. As this type could not occur any code base, this should only affect queries that explicitly looked at the builtin types.

View File

@@ -1,3 +0,0 @@
## 0.9.1
No user-facing changes.

View File

@@ -1,5 +0,0 @@
## 0.9.2
### Minor Analysis Improvements
* Only the 2 level indirection of `argv` (corresponding to `**argv`) is consided for `FlowSource`.

View File

@@ -1,2 +1,2 @@
---
lastReleaseVersion: 0.9.2
lastReleaseVersion: 0.7.2

View File

@@ -18,10 +18,10 @@ external string selectedSourceFile();
class Cfg extends PrintAstConfiguration {
/**
* Holds if the AST for `decl` should be printed.
* Print All declarations from the selected file.
* Holds if the AST for `func` should be printed.
* Print All functions from the selected file.
*/
override predicate shouldPrintDeclaration(Declaration decl) {
decl.getFile() = getFileBySourceArchiveName(selectedSourceFile())
override predicate shouldPrintFunction(Function func) {
func.getFile() = getFileBySourceArchiveName(selectedSourceFile())
}
}

View File

@@ -1,12 +1,11 @@
name: codeql/cpp-all
version: 0.9.2
version: 0.7.3-dev
groups: cpp
dbscheme: semmlecode.cpp.dbscheme
extractor: cpp
library: true
upgrades: upgrades
dependencies:
codeql/dataflow: ${workspace}
codeql/ssa: ${workspace}
codeql/tutorial: ${workspace}
codeql/util: ${workspace}

View File

@@ -34,6 +34,14 @@ class Container extends Locatable, @container {
*/
string getAbsolutePath() { none() } // overridden by subclasses
/**
* DEPRECATED: Use `getLocation` instead.
* Gets a URL representing the location of this container.
*
* For more information see [Providing URLs](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/#providing-urls).
*/
deprecated string getURL() { none() } // overridden by subclasses
/**
* Gets the relative path of this file or folder from the root folder of the
* analyzed source location. The relative path of the root folder itself is
@@ -175,6 +183,12 @@ class Folder extends Container, @folder {
}
override string getAPrimaryQlClass() { result = "Folder" }
/**
* DEPRECATED: Use `getLocation` instead.
* Gets the URL of this folder.
*/
deprecated override string getURL() { result = "file://" + this.getAbsolutePath() + ":0:0:0:0" }
}
/**
@@ -199,6 +213,12 @@ class File extends Container, @file {
result.hasLocationInfo(_, 0, 0, 0, 0)
}
/**
* DEPRECATED: Use `getLocation` instead.
* Gets the URL of this file.
*/
deprecated override string getURL() { result = "file://" + this.getAbsolutePath() + ":0:0:0:0" }
/** Holds if this file was compiled as C (at any point). */
predicate compiledAsC() { fileannotations(underlyingElement(this), 1, "compiled as c", "1") }

View File

@@ -6,9 +6,11 @@ private import PrintAST
* that requests that function, or no `PrintASTConfiguration` exists.
*/
private predicate shouldPrintDeclaration(Declaration decl) {
not (decl instanceof Function or decl instanceof GlobalOrNamespaceVariable)
not decl instanceof Function
or
exists(PrintAstConfiguration config | config.shouldPrintDeclaration(decl))
not exists(PrintAstConfiguration c)
or
exists(PrintAstConfiguration config | config.shouldPrintFunction(decl))
}
/**

View File

@@ -9,13 +9,13 @@ import cpp
import PrintAST
/**
* Temporarily tweak this class or make a copy to control which declarations are
* Temporarily tweak this class or make a copy to control which functions are
* printed.
*/
class Cfg extends PrintAstConfiguration {
/**
* TWEAK THIS PREDICATE AS NEEDED.
* Holds if the AST for `decl` should be printed.
* Holds if the AST for `func` should be printed.
*/
override predicate shouldPrintDeclaration(Declaration decl) { any() }
override predicate shouldPrintFunction(Function func) { any() }
}

View File

@@ -1,9 +1,9 @@
/**
* Provides queries to pretty-print a C++ AST as a graph.
*
* By default, this will print the AST for all functions and global and namespace variables in
* the database. To change this behavior, extend `PrintASTConfiguration` and override
* `shouldPrintDeclaration` to hold for only the declarations you wish to view the AST for.
* By default, this will print the AST for all functions in the database. To change this behavior,
* extend `PrintASTConfiguration` and override `shouldPrintFunction` to hold for only the functions
* you wish to view the AST for.
*/
import cpp
@@ -12,7 +12,7 @@ private import semmle.code.cpp.Print
private newtype TPrintAstConfiguration = MkPrintAstConfiguration()
/**
* The query can extend this class to control which declarations are printed.
* The query can extend this class to control which functions are printed.
*/
class PrintAstConfiguration extends TPrintAstConfiguration {
/**
@@ -21,16 +21,14 @@ class PrintAstConfiguration extends TPrintAstConfiguration {
string toString() { result = "PrintASTConfiguration" }
/**
* Holds if the AST for `decl` should be printed. By default, holds for all
* functions and global and namespace variables. Currently, does not support any
* other declaration types.
* Holds if the AST for `func` should be printed. By default, holds for all
* functions.
*/
predicate shouldPrintDeclaration(Declaration decl) { any() }
predicate shouldPrintFunction(Function func) { any() }
}
private predicate shouldPrintDeclaration(Declaration decl) {
exists(PrintAstConfiguration config | config.shouldPrintDeclaration(decl)) and
(decl instanceof Function or decl instanceof GlobalOrNamespaceVariable)
private predicate shouldPrintFunction(Function func) {
exists(PrintAstConfiguration config | config.shouldPrintFunction(func))
}
bindingset[s]
@@ -71,7 +69,7 @@ private predicate locationSortKeys(Locatable ast, string file, int line, int col
)
}
private Declaration getAnEnclosingDeclaration(Locatable ast) {
private Function getEnclosingFunction(Locatable ast) {
result = ast.(Expr).getEnclosingFunction()
or
result = ast.(Stmt).getEnclosingFunction()
@@ -80,10 +78,6 @@ private Declaration getAnEnclosingDeclaration(Locatable ast) {
or
result = ast.(Parameter).getFunction()
or
result = ast.(Expr).getEnclosingDeclaration()
or
result = ast.(Initializer).getDeclaration()
or
result = ast
}
@@ -92,21 +86,21 @@ private Declaration getAnEnclosingDeclaration(Locatable ast) {
* nodes for things like parameter lists and constructor init lists.
*/
private newtype TPrintAstNode =
TAstNode(Locatable ast) { shouldPrintDeclaration(getAnEnclosingDeclaration(ast)) } or
TAstNode(Locatable ast) { shouldPrintFunction(getEnclosingFunction(ast)) } or
TDeclarationEntryNode(DeclStmt stmt, DeclarationEntry entry) {
// We create a unique node for each pair of (stmt, entry), to avoid having one node with
// multiple parents due to extractor bug CPP-413.
stmt.getADeclarationEntry() = entry and
shouldPrintDeclaration(stmt.getEnclosingFunction())
shouldPrintFunction(stmt.getEnclosingFunction())
} or
TParametersNode(Function func) { shouldPrintDeclaration(func) } or
TParametersNode(Function func) { shouldPrintFunction(func) } or
TConstructorInitializersNode(Constructor ctor) {
ctor.hasEntryPoint() and
shouldPrintDeclaration(ctor)
shouldPrintFunction(ctor)
} or
TDestructorDestructionsNode(Destructor dtor) {
dtor.hasEntryPoint() and
shouldPrintDeclaration(dtor)
shouldPrintFunction(dtor)
}
/**
@@ -164,10 +158,10 @@ class PrintAstNode extends TPrintAstNode {
/**
* Holds if this node should be printed in the output. By default, all nodes
* within functions and global and namespace variables are printed, but the query
* can override `PrintASTConfiguration.shouldPrintDeclaration` to filter the output.
* within a function are printed, but the query can override
* `PrintASTConfiguration.shouldPrintFunction` to filter the output.
*/
final predicate shouldPrint() { shouldPrintDeclaration(this.getEnclosingDeclaration()) }
final predicate shouldPrint() { shouldPrintFunction(this.getEnclosingFunction()) }
/**
* Gets the children of this node.
@@ -235,15 +229,10 @@ class PrintAstNode extends TPrintAstNode {
abstract string getChildAccessorPredicateInternal(int childIndex);
/**
* Gets the `Declaration` that contains this node.
* Gets the `Function` that contains this node.
*/
private Declaration getEnclosingDeclaration() { result = this.getParent*().getDeclaration() }
/**
* Gets the `Declaration` this node represents.
*/
private Declaration getDeclaration() {
result = this.(AstNode).getAst() and shouldPrintDeclaration(result)
private Function getEnclosingFunction() {
result = this.getParent*().(FunctionNode).getFunction()
}
}
@@ -582,53 +571,16 @@ class DestructorDestructionsNode extends PrintAstNode, TDestructorDestructionsNo
final Destructor getDestructor() { result = dtor }
}
abstract private class FunctionOrGlobalOrNamespaceVariableNode extends AstNode {
override string toString() { result = qlClass(ast) + getIdentityString(ast) }
private int getOrder() {
this =
rank[result](FunctionOrGlobalOrNamespaceVariableNode node, Declaration decl, string file,
int line, int column |
node.getAst() = decl and
locationSortKeys(decl, file, line, column)
|
node order by file, line, column, getIdentityString(decl)
)
}
override string getProperty(string key) {
result = super.getProperty(key)
or
key = "semmle.order" and result = this.getOrder().toString()
}
}
/**
* A node representing a `GlobalOrNamespaceVariable`.
*/
class GlobalOrNamespaceVariableNode extends FunctionOrGlobalOrNamespaceVariableNode {
GlobalOrNamespaceVariable var;
GlobalOrNamespaceVariableNode() { var = ast }
override PrintAstNode getChildInternal(int childIndex) {
childIndex = 0 and
result.(AstNode).getAst() = var.getInitializer()
}
override string getChildAccessorPredicateInternal(int childIndex) {
childIndex = 0 and result = "getInitializer()"
}
}
/**
* A node representing a `Function`.
*/
class FunctionNode extends FunctionOrGlobalOrNamespaceVariableNode {
class FunctionNode extends AstNode {
Function func;
FunctionNode() { func = ast }
override string toString() { result = qlClass(func) + getIdentityString(func) }
override PrintAstNode getChildInternal(int childIndex) {
childIndex = 0 and
result.(ParametersNode).getFunction() = func
@@ -652,10 +604,31 @@ class FunctionNode extends FunctionOrGlobalOrNamespaceVariableNode {
or
childIndex = 3 and result = "<destructions>"
}
private int getOrder() {
this =
rank[result](FunctionNode node, Function function, string file, int line, int column |
node.getAst() = function and
locationSortKeys(function, file, line, column)
|
node order by file, line, column, getIdentityString(function)
)
}
override string getProperty(string key) {
result = super.getProperty(key)
or
key = "semmle.order" and result = this.getOrder().toString()
}
/**
* Gets the `Function` this node represents.
*/
final Function getFunction() { result = func }
}
private string getChildAccessorWithoutConversions(Locatable parent, Element child) {
shouldPrintDeclaration(getAnEnclosingDeclaration(parent)) and
shouldPrintFunction(getEnclosingFunction(parent)) and
(
exists(Stmt s | s = parent |
namedStmtChildPredicates(s, child, result)
@@ -674,7 +647,7 @@ private string getChildAccessorWithoutConversions(Locatable parent, Element chil
}
private predicate namedStmtChildPredicates(Locatable s, Element e, string pred) {
shouldPrintDeclaration(getAnEnclosingDeclaration(s)) and
shouldPrintFunction(getEnclosingFunction(s)) and
(
exists(int n | s.(BlockStmt).getStmt(n) = e and pred = "getStmt(" + n + ")")
or
@@ -762,14 +735,12 @@ private predicate namedStmtChildPredicates(Locatable s, Element e, string pred)
}
private predicate namedExprChildPredicates(Expr expr, Element ele, string pred) {
shouldPrintDeclaration(expr.getEnclosingDeclaration()) and
shouldPrintFunction(expr.getEnclosingFunction()) and
(
expr.(Access).getTarget() = ele and pred = "getTarget()"
or
expr.(VariableAccess).getQualifier() = ele and pred = "getQualifier()"
or
expr.(FunctionAccess).getQualifier() = ele and pred = "getQualifier()"
or
exists(Field f |
expr.(ClassAggregateLiteral).getAFieldExpr(f) = ele and
pred = "getAFieldExpr(" + f.toString() + ")"

View File

@@ -814,6 +814,9 @@ private predicate floatingPointTypeMapping(
// _Float128
kind = 49 and base = 2 and domain = TRealDomain() and realKind = 49 and extended = false
or
// _Float128x
kind = 50 and base = 2 and domain = TRealDomain() and realKind = 50 and extended = true
or
// _Float16
kind = 52 and base = 2 and domain = TRealDomain() and realKind = 52 and extended = false
or

View File

@@ -627,20 +627,6 @@ private predicate sub_lt(
x = int_value(rhs.getRight()) and
k = c - x
)
or
exists(PointerSubInstruction lhs, int c, int x |
compares_lt(cmp, lhs.getAUse(), right, c, isLt, testIsTrue) and
left = lhs.getLeftOperand() and
x = int_value(lhs.getRight()) and
k = c + x
)
or
exists(PointerSubInstruction rhs, int c, int x |
compares_lt(cmp, left, rhs.getAUse(), c, isLt, testIsTrue) and
right = rhs.getLeftOperand() and
x = int_value(rhs.getRight()) and
k = c - x
)
}
// left + x < right + c => left < right + (c-x)
@@ -667,26 +653,6 @@ private predicate add_lt(
) and
k = c + x
)
or
exists(PointerAddInstruction lhs, int c, int x |
compares_lt(cmp, lhs.getAUse(), right, c, isLt, testIsTrue) and
(
left = lhs.getLeftOperand() and x = int_value(lhs.getRight())
or
left = lhs.getRightOperand() and x = int_value(lhs.getLeft())
) and
k = c - x
)
or
exists(PointerAddInstruction rhs, int c, int x |
compares_lt(cmp, left, rhs.getAUse(), c, isLt, testIsTrue) and
(
right = rhs.getLeftOperand() and x = int_value(rhs.getRight())
or
right = rhs.getRightOperand() and x = int_value(rhs.getLeft())
) and
k = c + x
)
}
// left - x == right + c => left == right + (c+x)
@@ -707,20 +673,6 @@ private predicate sub_eq(
x = int_value(rhs.getRight()) and
k = c - x
)
or
exists(PointerSubInstruction lhs, int c, int x |
compares_eq(cmp, lhs.getAUse(), right, c, areEqual, testIsTrue) and
left = lhs.getLeftOperand() and
x = int_value(lhs.getRight()) and
k = c + x
)
or
exists(PointerSubInstruction rhs, int c, int x |
compares_eq(cmp, left, rhs.getAUse(), c, areEqual, testIsTrue) and
right = rhs.getLeftOperand() and
x = int_value(rhs.getRight()) and
k = c - x
)
}
// left + x == right + c => left == right + (c-x)
@@ -747,26 +699,6 @@ private predicate add_eq(
) and
k = c + x
)
or
exists(PointerAddInstruction lhs, int c, int x |
compares_eq(cmp, lhs.getAUse(), right, c, areEqual, testIsTrue) and
(
left = lhs.getLeftOperand() and x = int_value(lhs.getRight())
or
left = lhs.getRightOperand() and x = int_value(lhs.getLeft())
) and
k = c - x
)
or
exists(PointerAddInstruction rhs, int c, int x |
compares_eq(cmp, left, rhs.getAUse(), c, areEqual, testIsTrue) and
(
right = rhs.getLeftOperand() and x = int_value(rhs.getRight())
or
right = rhs.getRightOperand() and x = int_value(rhs.getLeft())
) and
k = c + x
)
}
/** The int value of integer constant expression. */

View File

@@ -20,14 +20,10 @@
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow {
private import semmle.code.cpp.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<CppOldDataFlow>
module DataFlow {
import semmle.code.cpp.dataflow.internal.DataFlow
import semmle.code.cpp.dataflow.internal.DataFlowImpl1
}

View File

@@ -12,11 +12,9 @@
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow2` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow2 {
module DataFlow2 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl2
}

View File

@@ -12,11 +12,9 @@
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow3` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow3 {
module DataFlow3 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl3
}

View File

@@ -12,11 +12,9 @@
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow4` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow4 {
module DataFlow4 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl4
}

View File

@@ -19,16 +19,10 @@ import semmle.code.cpp.dataflow.DataFlow
import semmle.code.cpp.dataflow.DataFlow2
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.TaintTracking` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
deprecated module TaintTracking {
import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
import TaintFlowMake<CppOldDataFlow, CppOldTaintTracking>
module TaintTracking {
import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTracking
import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -12,11 +12,9 @@
*/
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.TaintTracking2` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
deprecated module TaintTracking2 {
module TaintTracking2 {
import semmle.code.cpp.dataflow.internal.tainttracking2.TaintTrackingImpl
}

View File

@@ -0,0 +1,412 @@
/**
* 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() }
/**
* 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`.
*/
predicate isBarrier(Node node, FlowState state);
/** 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`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2);
/**
* 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() }
/**
* 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 = Merged::PathGraph;
}

View File

@@ -5,8 +5,8 @@ private import DataFlowUtil
/**
* Gets a function that might be called by `call`.
*/
Function viableCallable(DataFlowCall call) {
result = call.(Call).getTarget()
Function viableCallable(Call call) {
result = 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(DataFlowCall call, Function f) { none() }
predicate mayBenefitFromCallContext(Call 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(DataFlowCall call, DataFlowCall ctx) { none() }
Function viableImplInCallContext(Call call, Call ctx) { none() }
/** A parameter position represented by an integer. */
class ParameterPosition extends int {

File diff suppressed because it is too large Load Diff

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

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

View File

@@ -153,11 +153,10 @@ 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, ContentSet f, Node node2) {
predicate storeStep(Node node1, Content f, PostUpdateNode node2) {
exists(ClassAggregateLiteral aggr, Field field |
// The following lines requires `node2` to be both an `ExprNode` and a
// The following line 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()
@@ -168,13 +167,12 @@ predicate storeStep(Node node1, ContentSet f, Node node2) {
node1.asExpr() = a and
a.getLValue() = fa
) and
node2.(PostUpdateNode).getPreUpdateNode().asExpr() = fa.getQualifier() and
node2.getPreUpdateNode().asExpr() = fa.getQualifier() and
f.(FieldContent).getField() = fa.getTarget()
)
or
exists(ConstructorFieldInit cfi |
node2.(PostUpdateNode).getPreUpdateNode().(PreConstructorInitThis).getConstructorFieldInit() =
cfi and
node2.getPreUpdateNode().(PreConstructorInitThis).getConstructorFieldInit() = cfi and
f.(FieldContent).getField() = cfi.getTarget() and
node1.asExpr() = cfi.getExpr()
)
@@ -185,7 +183,7 @@ predicate storeStep(Node node1, ContentSet f, Node node2) {
* Thus, `node1` references an object with a field `f` whose value ends up in
* `node2`.
*/
predicate readStep(Node node1, ContentSet f, Node node2) {
predicate readStep(Node node1, Content f, Node node2) {
exists(FieldAccess fr |
node1.asExpr() = fr.getQualifier() and
fr.getTarget() = f.(FieldContent).getField() and
@@ -197,7 +195,7 @@ predicate readStep(Node node1, ContentSet f, Node node2) {
/**
* Holds if values stored inside content `c` are cleared at node `n`.
*/
predicate clearsContent(Node n, ContentSet c) {
predicate clearsContent(Node n, Content c) {
none() // stub implementation
}
@@ -261,6 +259,8 @@ 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.

View File

@@ -1,10 +0,0 @@
/**
* Provides C++-specific definitions for use in the taint tracking library.
*/
private import codeql.dataflow.TaintTracking
private import DataFlowImplSpecific
module CppOldTaintTracking implements InputSig<CppOldDataFlow> {
import TaintTrackingUtil
}

View File

@@ -39,7 +39,7 @@ predicate defaultAdditionalTaintStep(DataFlow::Node src, DataFlow::Node sink) {
* of `c` at sinks and inputs to additional taint steps.
*/
bindingset[node]
predicate defaultImplicitTaintRead(DataFlow::Node node, DataFlow::ContentSet c) { none() }
predicate defaultImplicitTaintRead(DataFlow::Node node, DataFlow::Content c) { none() }
/**
* Holds if `node` should be a sanitizer in all global taint flow configurations

View File

@@ -0,0 +1,74 @@
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
private module AddTaintDefaults<DataFlowInternal::FullStateConfigSig Config> implements
DataFlowInternal::FullStateConfigSig
{
import Config
predicate isBarrier(DataFlow::Node node) {
Config::isBarrier(node) or defaultTaintSanitizer(node)
}
predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
Config::isAdditionalFlowStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2)
}
predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
Config::allowImplicitRead(node, c)
or
(
Config::isSink(node, _) or
Config::isAdditionalFlowStep(node, _) or
Config::isAdditionalFlowStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
}
/**
* Constructs a global taint tracking computation.
*/
module Global<DataFlow::ConfigSig Config> implements DataFlow::GlobalFlowSig {
private module Config0 implements DataFlowInternal::FullStateConfigSig {
import DataFlowInternal::DefaultState<Config>
import Config
}
private module C implements DataFlowInternal::FullStateConfigSig {
import AddTaintDefaults<Config0>
}
import DataFlowInternal::Impl<C>
}
/** DEPRECATED: Use `Global` instead. */
deprecated module Make<DataFlow::ConfigSig Config> implements DataFlow::GlobalFlowSig {
import Global<Config>
}
/**
* Constructs a global taint tracking computation using flow state.
*/
module GlobalWithState<DataFlow::StateConfigSig Config> implements DataFlow::GlobalFlowSig {
private module Config0 implements DataFlowInternal::FullStateConfigSig {
import Config
}
private module C implements DataFlowInternal::FullStateConfigSig {
import AddTaintDefaults<Config0>
}
import DataFlowInternal::Impl<C>
}
/** DEPRECATED: Use `GlobalWithState` instead. */
deprecated module MakeWithState<DataFlow::StateConfigSig Config> implements DataFlow::GlobalFlowSig {
import GlobalWithState<Config>
}

View File

@@ -26,8 +26,6 @@ import cpp
* global (inter-procedural) data flow analyses.
*/
module 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.DataFlow
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

View File

@@ -23,10 +23,6 @@ import semmle.code.cpp.dataflow.new.DataFlow2
* global (inter-procedural) taint-tracking analyses.
*/
module TaintTracking {
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
import TaintFlowMake<CppDataFlow, CppTaintTracking>
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTracking
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -368,11 +368,6 @@ class FunctionAccess extends Access, @routineexpr {
/** Gets the accessed function. */
override Function getTarget() { funbind(underlyingElement(this), unresolveElement(result)) }
/**
* Gets the expression generating the function being accessed.
*/
Expr getQualifier() { this.getChild(-1) = result }
/** Gets a textual representation of this function access. */
override string toString() {
if exists(this.getTarget())

View File

@@ -152,19 +152,7 @@ class Expr extends StmtParent, @expr {
else result = this.getValue()
}
/**
* Holds if this expression has a value that can be determined at compile time.
*
* An expression has a value that can be determined at compile time when:
* - it is a compile-time constant, e.g., a literal value or the result of a constexpr
* compile-time constant;
* - it is an address of a (member) function, an address of a constexpr variable
* initialized to a constant address, or an address of an lvalue, or any of the
* previous with a constant value added to or subtracted from the address;
* - it is a reference to a (member) function, a reference to a constexpr variable
* initialized to a constant address, or a reference to an lvalue;
* - it is a non-template parameter of a uninstantiated template.
*/
/** Holds if this expression has a value that can be determined at compile time. */
cached
predicate isConstant() {
valuebind(_, underlyingElement(this))

View File

@@ -4,8 +4,8 @@
* This file contains the actual implementation of `PrintIR.ql`. For test cases and very small
* databases, `PrintIR.ql` can be run directly to dump the IR for the entire database. For most
* uses, however, it is better to write a query that imports `PrintIR.qll`, extends
* `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
* to dump.
* `PrintIRConfiguration`, and overrides `shouldPrintFunction()` to select a subset of functions to
* dump.
*/
import implementation.aliased_ssa.PrintIR

View File

@@ -22,8 +22,6 @@
import cpp
module 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.DataFlow
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

View File

@@ -19,10 +19,6 @@ import semmle.code.cpp.ir.dataflow.DataFlow
import semmle.code.cpp.ir.dataflow.DataFlow2
module TaintTracking {
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
import TaintFlowMake<CppDataFlow, CppTaintTracking>
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTracking
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -0,0 +1,412 @@
/**
* 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() }
/**
* 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`.
*/
predicate isBarrier(Node node, FlowState state);
/** 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`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2);
/**
* 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() }
/**
* 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 = Merged::PathGraph;
}

View File

@@ -9,7 +9,7 @@ private import DataFlowImplCommon as DataFlowImplCommon
* Gets a function that might be called by `call`.
*/
cached
DataFlowCallable viableCallable(DataFlowCall call) {
Function viableCallable(CallInstruction 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(DataFlowCall call, DataFlowCallable f) {
predicate mayBenefitFromCallContext(CallInstruction call, Function 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.
*/
DataFlowCallable viableImplInCallContext(DataFlowCall call, DataFlowCall ctx) {
Function viableImplInCallContext(CallInstruction call, CallInstruction ctx) {
result = viableCallable(call) and
exists(int i, Function f |
mayBenefitFromCallContext(pragma[only_bind_into](call), f, i) and

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

@@ -276,8 +276,6 @@ private module Config implements FullStateConfigSig {
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
@@ -315,8 +313,6 @@ private module Config implements FullStateConfigSig {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }

View File

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

View File

@@ -193,89 +193,86 @@ private class SingleUseOperandNode0 extends OperandNode0, TSingleUseOperandNode0
SingleUseOperandNode0() { this = TSingleUseOperandNode0(op) }
}
private module IndirectOperands {
/**
* INTERNAL: Do not use.
*
* A node that represents the indirect value of an operand in the IR
* after `index` number of loads.
*
* Note: Unlike `RawIndirectOperand`, a value of type `IndirectOperand` may
* be an `OperandNode`.
*/
class IndirectOperand extends Node {
Operand operand;
int indirectionIndex;
IndirectOperand() {
this.(RawIndirectOperand).getOperand() = operand and
this.(RawIndirectOperand).getIndirectionIndex() = indirectionIndex
or
nodeHasOperand(this, Ssa::getIRRepresentationOfIndirectOperand(operand, indirectionIndex),
indirectionIndex - 1)
}
/** Gets the underlying operand. */
Operand getOperand() { result = operand }
/** Gets the underlying indirection index. */
int getIndirectionIndex() { result = indirectionIndex }
/**
* INTERNAL: Do not use.
*
* A node that represents the indirect value of an operand in the IR
* after `index` number of loads.
*
* Note: Unlike `RawIndirectOperand`, a value of type `IndirectOperand` may
* be an `OperandNode`.
* Holds if this `IndirectOperand` is represented directly in the IR instead of
* a `RawIndirectionOperand` with operand `op` and indirection index `index`.
*/
abstract class IndirectOperand extends Node {
/** Gets the underlying operand and the underlying indirection index. */
abstract predicate hasOperandAndIndirectionIndex(Operand operand, int indirectionIndex);
}
private class IndirectOperandFromRaw extends IndirectOperand instanceof RawIndirectOperand {
override predicate hasOperandAndIndirectionIndex(Operand operand, int indirectionIndex) {
operand = RawIndirectOperand.super.getOperand() and
indirectionIndex = RawIndirectOperand.super.getIndirectionIndex()
}
}
private class IndirectOperandFromIRRepr extends IndirectOperand {
Operand operand;
int indirectionIndex;
IndirectOperandFromIRRepr() {
exists(Operand repr |
repr = Ssa::getIRRepresentationOfIndirectOperand(operand, indirectionIndex) and
nodeHasOperand(this, repr, indirectionIndex - 1)
)
}
override predicate hasOperandAndIndirectionIndex(Operand op, int index) {
op = operand and index = indirectionIndex
}
predicate isIRRepresentationOf(Operand op, int index) {
this instanceof OperandNode and
(
op = operand and
index = indirectionIndex
)
}
}
import IndirectOperands
/**
* INTERNAL: Do not use.
*
* A node that represents the indirect value of an instruction in the IR
* after `index` number of loads.
*
* Note: Unlike `RawIndirectInstruction`, a value of type `IndirectInstruction` may
* be an `InstructionNode`.
*/
class IndirectInstruction extends Node {
Instruction instr;
int indirectionIndex;
IndirectInstruction() {
this.(RawIndirectInstruction).getInstruction() = instr and
this.(RawIndirectInstruction).getIndirectionIndex() = indirectionIndex
or
nodeHasInstruction(this, Ssa::getIRRepresentationOfIndirectInstruction(instr, indirectionIndex),
indirectionIndex - 1)
}
/** Gets the underlying instruction. */
Instruction getInstruction() { result = instr }
/** Gets the underlying indirection index. */
int getIndirectionIndex() { result = indirectionIndex }
private module IndirectInstructions {
/**
* INTERNAL: Do not use.
*
* A node that represents the indirect value of an instruction in the IR
* after `index` number of loads.
*
* Note: Unlike `RawIndirectInstruction`, a value of type `IndirectInstruction` may
* be an `InstructionNode`.
* Holds if this `IndirectInstruction` is represented directly in the IR instead of
* a `RawIndirectionInstruction` with instruction `i` and indirection index `index`.
*/
abstract class IndirectInstruction extends Node {
/** Gets the underlying operand and the underlying indirection index. */
abstract predicate hasInstructionAndIndirectionIndex(Instruction instr, int index);
}
private class IndirectInstructionFromRaw extends IndirectInstruction instanceof RawIndirectInstruction
{
override predicate hasInstructionAndIndirectionIndex(Instruction instr, int index) {
instr = RawIndirectInstruction.super.getInstruction() and
index = RawIndirectInstruction.super.getIndirectionIndex()
}
}
private class IndirectInstructionFromIRRepr extends IndirectInstruction {
Instruction instr;
int indirectionIndex;
IndirectInstructionFromIRRepr() {
exists(Instruction repr |
repr = Ssa::getIRRepresentationOfIndirectInstruction(instr, indirectionIndex) and
nodeHasInstruction(this, repr, indirectionIndex - 1)
)
}
override predicate hasInstructionAndIndirectionIndex(Instruction i, int index) {
i = instr and index = indirectionIndex
}
predicate isIRRepresentationOf(Instruction i, int index) {
this instanceof InstructionNode and
(
i = instr and
index = indirectionIndex
)
}
}
import IndirectInstructions
/** Gets the callable in which this node occurs. */
DataFlowCallable nodeGetEnclosingCallable(Node n) { result = n.getEnclosingCallable() }
@@ -321,11 +318,9 @@ private class PrimaryArgumentNode extends ArgumentNode, OperandNode {
private class SideEffectArgumentNode extends ArgumentNode, SideEffectOperandNode {
override predicate argumentOf(DataFlowCall dfCall, ArgumentPosition pos) {
exists(int indirectionIndex |
pos = TIndirectionPosition(argumentIndex, pragma[only_bind_into](indirectionIndex)) and
this.getCallInstruction() = dfCall and
super.hasAddressOperandAndIndirectionIndex(_, pragma[only_bind_into](indirectionIndex))
)
this.getCallInstruction() = dfCall and
pos.(IndirectionPosition).getArgumentIndex() = this.getArgumentIndex() and
pos.(IndirectionPosition).getIndirectionIndex() = super.getIndirectionIndex()
}
}
@@ -653,16 +648,13 @@ predicate jumpStep(Node n1, Node n2) {
* Holds if data can flow from `node1` to `node2` via an assignment to `f`.
* Thus, `node2` references an object with a field `f` that contains the
* value of `node1`.
*
* The boolean `certain` is true if the destination address does not involve
* any pointer arithmetic, and false otherwise.
*/
predicate storeStepImpl(Node node1, Content c, PostFieldUpdateNode node2, boolean certain) {
predicate storeStep(Node node1, Content c, PostFieldUpdateNode node2) {
exists(int indirectionIndex1, int numberOfLoads, StoreInstruction store |
nodeHasInstruction(node1, store, pragma[only_bind_into](indirectionIndex1)) and
node2.getIndirectionIndex() = 1 and
numberOfLoadsFromOperand(node2.getFieldAddress(), store.getDestinationAddressOperand(),
numberOfLoads, certain)
numberOfLoads)
|
exists(FieldContent fc | fc = c |
fc.getField() = node2.getUpdatedField() and
@@ -676,32 +668,21 @@ predicate storeStepImpl(Node node1, Content c, PostFieldUpdateNode node2, boolea
)
}
/**
* Holds if data can flow from `node1` to `node2` via an assignment to `f`.
* Thus, `node2` references an object with a field `f` that contains the
* value of `node1`.
*/
predicate storeStep(Node node1, ContentSet c, Node node2) { storeStepImpl(node1, c, node2, _) }
/**
* Holds if `operandFrom` flows to `operandTo` using a sequence of conversion-like
* operations and exactly `n` `LoadInstruction` operations.
*/
private predicate numberOfLoadsFromOperandRec(
Operand operandFrom, Operand operandTo, int ind, boolean certain
) {
private predicate numberOfLoadsFromOperandRec(Operand operandFrom, Operand operandTo, int ind) {
exists(Instruction load | Ssa::isDereference(load, operandFrom) |
operandTo = operandFrom and ind = 0 and certain = true
operandTo = operandFrom and ind = 0
or
numberOfLoadsFromOperand(load.getAUse(), operandTo, ind - 1, certain)
numberOfLoadsFromOperand(load.getAUse(), operandTo, ind - 1)
)
or
exists(Operand op, Instruction instr, boolean isPointerArith, boolean certain0 |
exists(Operand op, Instruction instr |
instr = op.getDef() and
conversionFlow(operandFrom, instr, isPointerArith, _) and
numberOfLoadsFromOperand(op, operandTo, ind, certain0)
|
if isPointerArith = true then certain = false else certain = certain0
conversionFlow(operandFrom, instr, _, _) and
numberOfLoadsFromOperand(op, operandTo, ind)
)
}
@@ -709,16 +690,13 @@ private predicate numberOfLoadsFromOperandRec(
* Holds if `operandFrom` flows to `operandTo` using a sequence of conversion-like
* operations and exactly `n` `LoadInstruction` operations.
*/
private predicate numberOfLoadsFromOperand(
Operand operandFrom, Operand operandTo, int n, boolean certain
) {
numberOfLoadsFromOperandRec(operandFrom, operandTo, n, certain)
private predicate numberOfLoadsFromOperand(Operand operandFrom, Operand operandTo, int n) {
numberOfLoadsFromOperandRec(operandFrom, operandTo, n)
or
not Ssa::isDereference(_, operandFrom) and
not conversionFlow(operandFrom, _, _, _) and
operandFrom = operandTo and
n = 0 and
certain = true
n = 0
}
// Needed to join on both an operand and an index at the same time.
@@ -742,13 +720,13 @@ 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, ContentSet c, Node node2) {
predicate readStep(Node node1, Content 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
// in `storeStep`.
nodeHasOperand(node1, fa1.getObjectAddressOperand(), 1) and
numberOfLoadsFromOperand(fa1, operand, numberOfLoads, _)
numberOfLoadsFromOperand(fa1, operand, numberOfLoads)
|
exists(FieldContent fc | fc = c |
fc.getField() = fa1.getField() and
@@ -765,34 +743,8 @@ predicate readStep(Node node1, ContentSet c, Node node2) {
/**
* Holds if values stored inside content `c` are cleared at node `n`.
*/
predicate clearsContent(Node n, ContentSet c) {
n =
any(PostUpdateNode pun, Content d | d.impliesClearOf(c) and storeStepImpl(_, d, pun, true) | pun)
.getPreUpdateNode() and
(
// The crement operations and pointer addition and subtraction self-assign. We do not
// want to clear the contents if it is indirectly pointed at by any of these operations,
// as part of the contents might still be accessible afterwards. If there is no such
// indirection clearing the contents is safe.
not exists(Operand op, Cpp::Operation p |
n.(IndirectOperand).hasOperandAndIndirectionIndex(op, _) and
(
p instanceof Cpp::AssignPointerAddExpr or
p instanceof Cpp::AssignPointerSubExpr or
p instanceof Cpp::CrementOperation
)
|
p.getAnOperand() = op.getUse().getAst()
)
or
forex(PostUpdateNode pun, Content d |
pragma[only_bind_into](d).impliesClearOf(pragma[only_bind_into](c)) and
storeStepImpl(_, d, pun, true) and
pun.getPreUpdateNode() = n
|
c.(Content).getIndirectionIndex() = d.getIndirectionIndex()
)
)
predicate clearsContent(Node n, Content c) {
none() // stub implementation
}
/**
@@ -845,76 +797,12 @@ class DataFlowType = Type;
/** A function call relevant for data flow. */
class DataFlowCall extends CallInstruction {
DataFlowCallable getEnclosingCallable() { result = this.getEnclosingFunction() }
Function getEnclosingCallable() { result = this.getEnclosingFunction() }
}
module IsUnreachableInCall {
private import semmle.code.cpp.ir.ValueNumbering
private import semmle.code.cpp.controlflow.IRGuards as G
predicate isUnreachableInCall(Node n, DataFlowCall call) { none() } // stub implementation
private class ConstantIntegralTypeArgumentNode extends PrimaryArgumentNode {
int value;
ConstantIntegralTypeArgumentNode() {
value = op.getDef().(IntegerConstantInstruction).getValue().toInt()
}
int getValue() { result = value }
}
pragma[nomagic]
private predicate ensuresEq(Operand left, Operand right, int k, IRBlock block, boolean areEqual) {
any(G::IRGuardCondition guard).ensuresEq(left, right, k, block, areEqual)
}
pragma[nomagic]
private predicate ensuresLt(Operand left, Operand right, int k, IRBlock block, boolean areEqual) {
any(G::IRGuardCondition guard).ensuresLt(left, right, k, block, areEqual)
}
predicate isUnreachableInCall(Node n, DataFlowCall call) {
exists(
DirectParameterNode paramNode, ConstantIntegralTypeArgumentNode arg,
IntegerConstantInstruction constant, int k, Operand left, Operand right, IRBlock block
|
// arg flows into `paramNode`
DataFlowImplCommon::viableParamArg(call, paramNode, arg) and
left = constant.getAUse() and
right = valueNumber(paramNode.getInstruction()).getAUse() and
block = n.getBasicBlock()
|
// and there's a guard condition which ensures that the result of `left == right + k` is `areEqual`
exists(boolean areEqual |
ensuresEq(pragma[only_bind_into](left), pragma[only_bind_into](right),
pragma[only_bind_into](k), pragma[only_bind_into](block), areEqual)
|
// this block ensures that left = right + k, but it holds that `left != right + k`
areEqual = true and
constant.getValue().toInt() != arg.getValue() + k
or
// this block ensures that or `left != right + k`, but it holds that `left = right + k`
areEqual = false and
constant.getValue().toInt() = arg.getValue() + k
)
or
// or there's a guard condition which ensures that the result of `left < right + k` is `isLessThan`
exists(boolean isLessThan |
ensuresLt(pragma[only_bind_into](left), pragma[only_bind_into](right),
pragma[only_bind_into](k), pragma[only_bind_into](block), isLessThan)
|
isLessThan = true and
// this block ensures that `left < right + k`, but it holds that `left >= right + k`
constant.getValue().toInt() >= arg.getValue() + k
or
// this block ensures that `left >= right + k`, but it holds that `left < right + k`
isLessThan = false and
constant.getValue().toInt() < arg.getValue() + k
)
)
}
}
import IsUnreachableInCall
int accessPathLimit() { result = 5 }
/**
* Holds if access paths with `c` at their head always should be tracked at high
@@ -951,7 +839,7 @@ predicate additionalLambdaFlowStep(Node nodeFrom, Node nodeTo, boolean preserves
* One example would be to allow flow like `p.foo = p.bar;`, which is disallowed
* by default as a heuristic.
*/
predicate allowParameterReturnInSelf(ParameterNode p) { p instanceof IndirectParameterNode }
predicate allowParameterReturnInSelf(ParameterNode p) { none() }
private predicate fieldHasApproxName(Field f, string s) {
s = f.getName().charAt(0) and
@@ -1078,7 +966,7 @@ private IRVariable getIRVariableForParameterNode(ParameterNode p) {
/** Holds if `v` is the source variable corresponding to the parameter represented by `p`. */
pragma[nomagic]
private predicate parameterNodeHasSourceVariable(ParameterNode p, Ssa::SourceVariable v) {
private predicate parameterNodeHasSourceVariable(ParameterNode p, Ssa::SourceIRVariable v) {
v.getIRVariable() = getIRVariableForParameterNode(p) and
exists(Position pos | p.isParameterOf(_, pos) |
pos instanceof DirectPosition and

View File

@@ -274,7 +274,7 @@ class Node extends TIRDataFlowNode {
* represents the value of `**x` going into `f`.
*/
Expr asIndirectArgument(int index) {
this.(SideEffectOperandNode).hasAddressOperandAndIndirectionIndex(_, index) and
this.(SideEffectOperandNode).getIndirectionIndex() = index and
result = this.(SideEffectOperandNode).getArgument()
}
@@ -317,7 +317,7 @@ class Node extends TIRDataFlowNode {
index = 0 and
result = this.(ExplicitParameterNode).getParameter()
or
this.(IndirectParameterNode).hasInstructionAndIndirectionIndex(_, index) and
this.(IndirectParameterNode).getIndirectionIndex() = index and
result = this.(IndirectParameterNode).getParameter()
}
@@ -577,20 +577,15 @@ class SsaPhiNode extends Node, TSsaPhiNode {
*
* A node representing a value after leaving a function.
*/
class SideEffectOperandNode extends Node instanceof IndirectOperand {
class SideEffectOperandNode extends Node, IndirectOperand {
CallInstruction call;
int argumentIndex;
SideEffectOperandNode() {
IndirectOperand.super.hasOperandAndIndirectionIndex(call.getArgumentOperand(argumentIndex), _)
}
SideEffectOperandNode() { operand = call.getArgumentOperand(argumentIndex) }
CallInstruction getCallInstruction() { result = call }
/** Gets the underlying operand and the underlying indirection index. */
predicate hasAddressOperandAndIndirectionIndex(Operand operand, int indirectionIndex) {
IndirectOperand.super.hasOperandAndIndirectionIndex(operand, indirectionIndex)
}
Operand getAddressOperand() { result = operand }
int getArgumentIndex() { result = argumentIndex }
@@ -670,10 +665,10 @@ class InitialGlobalValue extends Node, TInitialGlobalValue {
*
* A node representing an indirection of a parameter.
*/
class IndirectParameterNode extends Node instanceof IndirectInstruction {
class IndirectParameterNode extends Node, IndirectInstruction {
InitializeParameterInstruction init;
IndirectParameterNode() { IndirectInstruction.super.hasInstructionAndIndirectionIndex(init, _) }
IndirectParameterNode() { this.getInstruction() = init }
int getArgumentIndex() { init.hasIndex(result) }
@@ -682,12 +677,7 @@ class IndirectParameterNode extends Node instanceof IndirectInstruction {
override Declaration getEnclosingCallable() { result = this.getFunction() }
override Declaration getFunction() { result = init.getEnclosingFunction() }
/** Gets the underlying operand and the underlying indirection index. */
predicate hasInstructionAndIndirectionIndex(Instruction instr, int index) {
IndirectInstruction.super.hasInstructionAndIndirectionIndex(instr, index)
}
override Declaration getFunction() { result = this.getInstruction().getEnclosingFunction() }
override Location getLocationImpl() { result = this.getParameter().getLocation() }
@@ -709,8 +699,7 @@ class IndirectReturnNode extends Node {
IndirectReturnNode() {
this instanceof FinalParameterNode
or
this.(IndirectOperand)
.hasOperandAndIndirectionIndex(any(ReturnValueInstruction ret).getReturnAddressOperand(), _)
this.(IndirectOperand).getOperand() = any(ReturnValueInstruction ret).getReturnAddressOperand()
}
override Declaration getEnclosingCallable() { result = this.getFunction() }
@@ -733,7 +722,7 @@ class IndirectReturnNode extends Node {
int getIndirectionIndex() {
result = this.(FinalParameterNode).getIndirectionIndex()
or
this.(IndirectOperand).hasOperandAndIndirectionIndex(_, result)
result = this.(IndirectOperand).getIndirectionIndex()
}
}
@@ -781,12 +770,26 @@ class IndirectArgumentOutNode extends Node, TIndirectArgumentOutNode, PartialDef
override Expr getDefinedExpr() { result = operand.getDef().getUnconvertedResultExpression() }
}
pragma[nomagic]
predicate indirectReturnOutNodeOperand0(CallInstruction call, Operand operand, int indirectionIndex) {
Ssa::hasRawIndirectInstruction(call, indirectionIndex) and
operandForFullyConvertedCall(operand, call)
}
pragma[nomagic]
predicate indirectReturnOutNodeInstruction0(
CallInstruction call, Instruction instr, int indirectionIndex
) {
Ssa::hasRawIndirectInstruction(call, indirectionIndex) and
instructionForFullyConvertedCall(instr, call)
}
/**
* Holds if `node` is an indirect operand with columns `(operand, indirectionIndex)`, and
* `operand` represents a use of the fully converted value of `call`.
*/
private predicate hasOperand(Node node, CallInstruction call, int indirectionIndex, Operand operand) {
operandForFullyConvertedCall(operand, call) and
indirectReturnOutNodeOperand0(call, operand, indirectionIndex) and
hasOperandAndIndex(node, operand, indirectionIndex)
}
@@ -799,7 +802,7 @@ private predicate hasOperand(Node node, CallInstruction call, int indirectionInd
private predicate hasInstruction(
Node node, CallInstruction call, int indirectionIndex, Instruction instr
) {
instructionForFullyConvertedCall(instr, call) and
indirectReturnOutNodeInstruction0(call, instr, indirectionIndex) and
hasInstructionAndIndex(node, instr, indirectionIndex)
}
@@ -1103,8 +1106,7 @@ predicate exprNodeShouldBeInstruction(Node node, Expr e) {
/** Holds if `node` should be an `IndirectInstruction` that maps `node.asIndirectExpr()` to `e`. */
predicate indirectExprNodeShouldBeIndirectInstruction(IndirectInstruction node, Expr e) {
exists(Instruction instr |
node.hasInstructionAndIndirectionIndex(instr, _) and
not indirectExprNodeShouldBeIndirectOperand(_, e)
instr = node.getInstruction() and not indirectExprNodeShouldBeIndirectOperand(_, e)
|
e = instr.(VariableAddressInstruction).getAst().(Expr).getFullyConverted()
or
@@ -1305,8 +1307,8 @@ pragma[noinline]
private predicate indirectParameterNodeHasArgumentIndexAndIndex(
IndirectParameterNode node, int argumentIndex, int indirectionIndex
) {
node.hasInstructionAndIndirectionIndex(_, indirectionIndex) and
node.getArgumentIndex() = argumentIndex
node.getArgumentIndex() = argumentIndex and
node.getIndirectionIndex() = indirectionIndex
}
/** A synthetic parameter to model the pointed-to object of a pointer parameter. */
@@ -1477,14 +1479,18 @@ VariableNode variableNode(Variable v) {
*/
Node uninitializedNode(LocalVariable v) { none() }
pragma[noinline]
predicate hasOperandAndIndex(IndirectOperand indirectOperand, Operand operand, int indirectionIndex) {
indirectOperand.hasOperandAndIndirectionIndex(operand, indirectionIndex)
indirectOperand.getOperand() = operand and
indirectOperand.getIndirectionIndex() = indirectionIndex
}
pragma[noinline]
predicate hasInstructionAndIndex(
IndirectInstruction indirectInstr, Instruction instr, int indirectionIndex
) {
indirectInstr.hasInstructionAndIndirectionIndex(instr, indirectionIndex)
indirectInstr.getInstruction() = instr and
indirectInstr.getIndirectionIndex() = indirectionIndex
}
cached
@@ -1520,25 +1526,6 @@ private module Cached {
)
}
/**
* Holds if `operand.getDef() = instr`, but there exists a `StoreInstruction` that
* writes to an address that is equivalent to the value computed by `instr` in
* between `instr` and `operand`, and therefore there should not be flow from `*instr`
* to `*operand`.
*/
pragma[nomagic]
private predicate isStoredToBetween(Instruction instr, Operand operand) {
simpleOperandLocalFlowStep(pragma[only_bind_into](instr), pragma[only_bind_into](operand)) and
exists(StoreInstruction store, IRBlock block, int storeIndex, int instrIndex, int operandIndex |
store.getDestinationAddress() = instr and
block.getInstruction(storeIndex) = store and
block.getInstruction(instrIndex) = instr and
block.getInstruction(operandIndex) = operand.getUse() and
instrIndex < storeIndex and
storeIndex < operandIndex
)
}
private predicate indirectionInstructionFlow(
RawIndirectInstruction nodeFrom, IndirectOperand nodeTo
) {
@@ -1548,8 +1535,7 @@ private module Cached {
simpleOperandLocalFlowStep(pragma[only_bind_into](instr), pragma[only_bind_into](operand))
|
hasOperandAndIndex(nodeTo, operand, pragma[only_bind_into](indirectionIndex)) and
hasInstructionAndIndex(nodeFrom, instr, pragma[only_bind_into](indirectionIndex)) and
not isStoredToBetween(instr, operand)
hasInstructionAndIndex(nodeFrom, instr, pragma[only_bind_into](indirectionIndex))
)
}
@@ -1670,7 +1656,8 @@ module ExprFlowCached {
private predicate isIndirectBaseOfArrayAccess(IndirectOperand n, Expr e) {
exists(LoadInstruction load, PointerArithmeticInstruction pai |
pai = load.getSourceAddress() and
n.hasOperandAndIndirectionIndex(pai.getLeftOperand(), 1) and
pai.getLeftOperand() = n.getOperand() and
n.getIndirectionIndex() = 1 and
e = load.getConvertedResultExpression()
)
}
@@ -1838,20 +1825,6 @@ class Content extends TContent {
predicate hasLocationInfo(string path, int sl, int sc, int el, int ec) {
path = "" and sl = 0 and sc = 0 and el = 0 and ec = 0
}
/** Gets the indirection index of this `Content`. */
abstract int getIndirectionIndex();
/**
* INTERNAL: Do not use.
*
* Holds if a write to this `Content` implies that `c` is
* also cleared.
*
* For example, a write to a field `f` implies that any content of
* the form `*f` is also cleared.
*/
abstract predicate impliesClearOf(Content c);
}
/** A reference through a non-union instance field. */
@@ -1869,21 +1842,10 @@ class FieldContent extends Content, TFieldContent {
Field getField() { result = f }
/** Gets the indirection index of this `FieldContent`. */
pragma[inline]
override int getIndirectionIndex() {
int getIndirectionIndex() {
pragma[only_bind_into](result) = pragma[only_bind_out](indirectionIndex)
}
override predicate impliesClearOf(Content c) {
exists(FieldContent fc |
fc = c and
fc.getField() = f and
// If `this` is `f` then `c` is cleared if it's of the
// form `*f`, `**f`, etc.
fc.getIndirectionIndex() >= indirectionIndex
)
}
}
/** A reference through an instance field of a union. */
@@ -1908,21 +1870,9 @@ class UnionContent extends Content, TUnionContent {
/** Gets the indirection index of this `UnionContent`. */
pragma[inline]
override int getIndirectionIndex() {
int getIndirectionIndex() {
pragma[only_bind_into](result) = pragma[only_bind_out](indirectionIndex)
}
override predicate impliesClearOf(Content c) {
exists(UnionContent uc |
uc = c and
uc.getUnion() = u and
// If `this` is `u` then `c` is cleared if it's of the
// form `*u`, `**u`, etc. (and we ignore `bytes` because
// we know the entire union is overwritten because it's a
// union).
uc.getIndirectionIndex() >= indirectionIndex
)
}
}
/**

View File

@@ -448,8 +448,6 @@ module TaintedWithPath {
}
predicate isBarrierIn(DataFlow::Node node) { nodeIsBarrierIn(node) }
predicate neverSkip(Node node) { none() }
}
private module AdjustedFlow = TaintTracking::Global<AdjustedConfig>;

View File

@@ -13,7 +13,7 @@ class FieldFlowPropertyProvider extends IRPropertyProvider {
override string getOperandProperty(Operand operand, string key) {
exists(PostFieldUpdateNode pfun, Content content |
key = "store " + content.toString() and
pfun.getPreUpdateNode().(IndirectOperand).hasOperandAndIndirectionIndex(operand, _) and
operand = pfun.getPreUpdateNode().(IndirectOperand).getOperand() and
result =
strictconcat(string element, Node node |
storeStep(node, content, pfun) and
@@ -25,7 +25,7 @@ class FieldFlowPropertyProvider extends IRPropertyProvider {
or
exists(Node node2, Content content |
key = "read " + content.toString() and
node2.(IndirectOperand).hasOperandAndIndirectionIndex(operand, _) and
operand = node2.(IndirectOperand).getOperand() and
result =
strictconcat(string element, Node node1 |
readStep(node1, content, node2) and

View File

@@ -18,12 +18,9 @@ private string stars(int k) {
}
string starsForNode(Node node) {
exists(int indirectionIndex |
node.(IndirectInstruction).hasInstructionAndIndirectionIndex(_, indirectionIndex) or
node.(IndirectOperand).hasOperandAndIndirectionIndex(_, indirectionIndex)
|
result = stars(indirectionIndex)
)
result = stars(node.(IndirectInstruction).getIndirectionIndex())
or
result = stars(node.(IndirectOperand).getIndirectionIndex())
or
not node instanceof IndirectInstruction and
not node instanceof IndirectOperand and

View File

@@ -192,13 +192,13 @@ module ProductFlow {
* Holds if data flow through `node` is prohibited through the first projection of the product
* dataflow graph when the flow state is `state`.
*/
default predicate isBarrier1(DataFlow::Node node, FlowState1 state) { none() }
predicate isBarrier1(DataFlow::Node node, FlowState1 state);
/**
* Holds if data flow through `node` is prohibited through the second projection of the product
* dataflow graph when the flow state is `state`.
*/
default predicate isBarrier2(DataFlow::Node node, FlowState2 state) { none() }
predicate isBarrier2(DataFlow::Node node, FlowState2 state);
/**
* Holds if data flow through `node` is prohibited through the first projection of the product
@@ -237,11 +237,9 @@ module ProductFlow {
*
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
default predicate isAdditionalFlowStep1(
predicate isAdditionalFlowStep1(
DataFlow::Node node1, FlowState1 state1, DataFlow::Node node2, FlowState1 state2
) {
none()
}
);
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps in
@@ -255,11 +253,9 @@ module ProductFlow {
*
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
default predicate isAdditionalFlowStep2(
predicate isAdditionalFlowStep2(
DataFlow::Node node1, FlowState2 state1, DataFlow::Node node2, FlowState2 state2
) {
none()
}
);
/**
* Holds if data flow into `node` is prohibited in the first projection of the product
@@ -297,22 +293,6 @@ module ProductFlow {
reachable(source1, source2, sink1, sink2)
}
/** Holds if data can flow from `(source1, source2)` to `(sink1, sink2)`. */
predicate flow(
DataFlow::Node source1, DataFlow::Node source2, DataFlow::Node sink1, DataFlow::Node sink2
) {
exists(
Flow1::PathNode pSource1, Flow2::PathNode pSource2, Flow1::PathNode pSink1,
Flow2::PathNode pSink2
|
pSource1.getNode() = source1 and
pSource2.getNode() = source2 and
pSink1.getNode() = sink1 and
pSink2.getNode() = sink2 and
flowPath(pSource1, pSource2, pSink1, pSink2)
)
}
private module Config1 implements DataFlow::StateConfigSig {
class FlowState = FlowState1;
@@ -379,6 +359,7 @@ module ProductFlow {
Config::isSinkPair(node1.getNode(), node1.getState(), node2.getNode(), node2.getState())
}
pragma[assume_small_delta]
pragma[nomagic]
private predicate fwdReachableInterprocEntry(Flow1::PathNode node1, Flow2::PathNode node2) {
isSourcePair(node1, node2)
@@ -415,6 +396,7 @@ module ProductFlow {
fwdIsSuccessorExit(pragma[only_bind_into](mid1), pragma[only_bind_into](mid2), succ1, succ2)
}
pragma[assume_small_delta]
private predicate fwdIsSuccessor(
Flow1::PathNode pred1, Flow2::PathNode pred2, Flow1::PathNode succ1, Flow2::PathNode succ2
) {
@@ -424,6 +406,7 @@ module ProductFlow {
)
}
pragma[assume_small_delta]
pragma[nomagic]
private predicate revReachableInterprocEntry(Flow1::PathNode node1, Flow2::PathNode node2) {
fwdReachableInterprocEntry(node1, node2) and

View File

@@ -10,35 +10,32 @@ private import ssa0.SsaInternals as SsaInternals0
import SsaInternalsCommon
private module SourceVariables {
int getMaxIndirectionForIRVariable(IRVariable var) {
exists(Type type, boolean isGLValue |
var.getLanguageType().hasType(type, isGLValue) and
if isGLValue = true
then result = 1 + getMaxIndirectionsForType(type)
else result = getMaxIndirectionsForType(type)
)
}
cached
private newtype TSourceVariable =
TMkSourceVariable(SsaInternals0::SourceVariable base, int ind) {
ind = [0 .. countIndirectionsForCppType(base.getLanguageType()) + 1]
TSourceIRVariable(BaseIRVariable baseVar, int ind) {
ind = [0 .. getMaxIndirectionForIRVariable(baseVar.getIRVariable())]
} or
TCallVariable(AllocationInstruction call, int ind) {
ind = [0 .. countIndirectionsForCppType(getResultLanguageType(call))]
}
class SourceVariable extends TSourceVariable {
SsaInternals0::SourceVariable base;
abstract class SourceVariable extends TSourceVariable {
int ind;
SourceVariable() { this = TMkSourceVariable(base, ind) }
/** Gets the IR variable associated with this `SourceVariable`, if any. */
IRVariable getIRVariable() { result = base.(BaseIRVariable).getIRVariable() }
/**
* Gets the base source variable (i.e., the variable without any
* indirections) of this source variable.
*/
SsaInternals0::SourceVariable getBaseVariable() { result = base }
bindingset[ind]
SourceVariable() { any() }
/** Gets a textual representation of this element. */
string toString() {
ind = 0 and
result = this.getBaseVariable().toString()
or
ind > 0 and
result = this.getBaseVariable().toString() + " indirection"
}
abstract string toString();
/**
* Gets the number of loads performed on the base source variable
@@ -46,19 +43,65 @@ private module SourceVariables {
*/
int getIndirection() { result = ind }
/**
* Gets the base source variable (i.e., the variable without any
* indirections) of this source variable.
*/
abstract BaseSourceVariable getBaseVariable();
/** Holds if this variable is a glvalue. */
predicate isGLValue() { ind = 0 }
predicate isGLValue() { none() }
/**
* Gets the type of this source variable. If `isGLValue()` holds, then
* the type of this source variable should be thought of as "pointer
* to `getType()`".
*/
DataFlowType getType() {
if this.isGLValue()
then result = base.getType()
else result = getTypeImpl(base.getType(), ind - 1)
abstract DataFlowType getType();
}
class SourceIRVariable extends SourceVariable, TSourceIRVariable {
BaseIRVariable var;
SourceIRVariable() { this = TSourceIRVariable(var, ind) }
IRVariable getIRVariable() { result = var.getIRVariable() }
override BaseIRVariable getBaseVariable() { result.getIRVariable() = this.getIRVariable() }
override string toString() {
ind = 0 and
result = this.getIRVariable().toString()
or
ind > 0 and
result = this.getIRVariable().toString() + " indirection"
}
override predicate isGLValue() { ind = 0 }
override DataFlowType getType() {
if ind = 0 then result = var.getType() else result = getTypeImpl(var.getType(), ind - 1)
}
}
class CallVariable extends SourceVariable, TCallVariable {
AllocationInstruction call;
CallVariable() { this = TCallVariable(call, ind) }
AllocationInstruction getCall() { result = call }
override BaseCallVariable getBaseVariable() { result.getCallInstruction() = call }
override string toString() {
ind = 0 and
result = "Call"
or
ind > 0 and
result = "Call indirection"
}
override DataFlowType getType() { result = getTypeImpl(call.getResultType(), ind) }
}
}
@@ -94,9 +137,8 @@ predicate hasRawIndirectInstruction(Instruction instr, int indirectionIndex) {
cached
private newtype TDefOrUseImpl =
TDefImpl(BaseSourceVariableInstruction base, Operand address, int indirectionIndex) {
isDef(_, _, address, base, _, indirectionIndex) and
(
TDefImpl(Operand address, int indirectionIndex) {
exists(Instruction base | isDef(_, _, address, base, _, indirectionIndex) |
// We only include the definition if the SSA pruning stage
// concluded that the definition is live after the write.
any(SsaInternals0::Def def).getAddressOperand() = address
@@ -106,8 +148,8 @@ private newtype TDefOrUseImpl =
base.(VariableAddressInstruction).getAstVariable() instanceof GlobalLikeVariable
)
} or
TUseImpl(BaseSourceVariableInstruction base, Operand operand, int indirectionIndex) {
isUse(_, operand, base, _, indirectionIndex) and
TUseImpl(Operand operand, int indirectionIndex) {
isUse(_, operand, _, _, indirectionIndex) and
not isDef(_, _, operand, _, _, _)
} or
TGlobalUse(GlobalLikeVariable v, IRFunction f, int indirectionIndex) {
@@ -194,7 +236,7 @@ abstract private class DefOrUseImpl extends TDefOrUseImpl {
/**
* Gets the instruction that computes the base of this definition or use.
* This is always a `VariableAddressInstruction` or an `CallInstruction`.
* This is always a `VariableAddressInstruction` or an `AllocationInstruction`.
*/
abstract BaseSourceVariableInstruction getBase();
@@ -266,17 +308,15 @@ abstract class DefImpl extends DefOrUseImpl {
}
private class DirectDef extends DefImpl, TDefImpl {
BaseSourceVariableInstruction base;
DirectDef() { this = TDefImpl(address, ind) }
DirectDef() { this = TDefImpl(base, address, ind) }
override BaseSourceVariableInstruction getBase() { isDef(_, _, address, result, _, _) }
override BaseSourceVariableInstruction getBase() { result = base }
override int getIndirection() { isDef(_, _, address, _, result, ind) }
override int getIndirection() { isDef(_, _, address, base, result, ind) }
override Node0Impl getValue() { isDef(_, result, address, _, _, _) }
override Node0Impl getValue() { isDef(_, result, address, base, _, _) }
override predicate isCertain() { isDef(true, _, address, base, _, ind) }
override predicate isCertain() { isDef(true, _, address, _, _, ind) }
}
private class IteratorDef extends DefImpl, TIteratorDef {
@@ -319,7 +359,6 @@ abstract class UseImpl extends DefOrUseImpl {
abstract private class OperandBasedUse extends UseImpl {
Operand operand;
BaseSourceVariableInstruction base;
bindingset[ind]
OperandBasedUse() { any() }
@@ -327,44 +366,50 @@ abstract private class OperandBasedUse extends UseImpl {
final override predicate hasIndexInBlock(IRBlock block, int index) {
// See the comment in `ssa0`'s `OperandBasedUse` for an explanation of this
// predicate's implementation.
if base.getAst() = any(Cpp::PostfixCrementOperation c).getOperand()
then
exists(Operand op, int indirectionIndex, int indirection |
indirectionIndex = this.getIndirectionIndex() and
indirection = this.getIndirection() and
op =
min(Operand cand, int i |
isUse(_, cand, base, indirection, indirectionIndex) and
block.getInstruction(i) = cand.getUse()
|
cand order by i
) and
block.getInstruction(index) = op.getUse()
)
else operand.getUse() = block.getInstruction(index)
exists(BaseSourceVariableInstruction base | base = this.getBase() |
if base.getAst() = any(Cpp::PostfixCrementOperation c).getOperand()
then
exists(Operand op, int indirectionIndex, int indirection |
indirectionIndex = this.getIndirectionIndex() and
indirection = this.getIndirection() and
op =
min(Operand cand, int i |
isUse(_, cand, base, indirection, indirectionIndex) and
block.getInstruction(i) = cand.getUse()
|
cand order by i
) and
block.getInstruction(index) = op.getUse()
)
else operand.getUse() = block.getInstruction(index)
)
}
final override BaseSourceVariableInstruction getBase() { result = base }
final Operand getOperand() { result = operand }
final override Cpp::Location getLocation() { result = operand.getLocation() }
}
private class DirectUse extends OperandBasedUse, TUseImpl {
DirectUse() { this = TUseImpl(base, operand, ind) }
DirectUse() { this = TUseImpl(operand, ind) }
override int getIndirection() { isUse(_, operand, base, result, ind) }
override int getIndirection() { isUse(_, operand, _, result, ind) }
override predicate isCertain() { isUse(true, operand, base, _, ind) }
override BaseSourceVariableInstruction getBase() { isUse(_, operand, result, _, ind) }
override predicate isCertain() { isUse(true, operand, _, _, ind) }
override Node getNode() { nodeHasOperand(result, operand, ind) }
}
private class IteratorUse extends OperandBasedUse, TIteratorUse {
IteratorUse() { this = TIteratorUse(operand, base, ind) }
BaseSourceVariableInstruction container;
override int getIndirection() { isIteratorUse(base, operand, result, ind) }
IteratorUse() { this = TIteratorUse(operand, container, ind) }
override int getIndirection() { isIteratorUse(container, operand, result, ind) }
override BaseSourceVariableInstruction getBase() { result = container }
override predicate isCertain() { none() }

View File

@@ -6,7 +6,6 @@ private import DataFlowImplCommon as DataFlowImplCommon
private import DataFlowUtil
private import semmle.code.cpp.models.interfaces.PointerWrapper
private import DataFlowPrivate
private import semmle.code.cpp.ir.ValueNumbering
/**
* Holds if `operand` is an operand that is not used by the dataflow library.
@@ -118,16 +117,6 @@ private int countIndirections(Type t) {
else (
result = any(Indirection ind | ind.getType() = t).getNumberOfIndirections()
or
// If there is an indirection for the type, but we cannot count the number of indirections
// it means we couldn't reach a non-indirection type by stripping off indirections. This
// can occur if an iterator specifies itself as the value type. In this case we default to
// 1 indirection fore the type.
exists(Indirection ind |
ind.getType() = t and
not exists(ind.getNumberOfIndirections()) and
result = 1
)
or
not exists(Indirection ind | ind.getType() = t) and
result = 0
)
@@ -147,6 +136,14 @@ int countIndirectionsForCppType(LanguageType langType) {
)
}
/**
* A `CallInstruction` that calls an allocation function such
* as `malloc` or `operator new`.
*/
class AllocationInstruction extends CallInstruction {
AllocationInstruction() { this.getStaticCallTarget() instanceof Cpp::AllocationFunction }
}
private predicate isIndirectionType(Type t) { t instanceof Indirection }
private predicate hasUnspecifiedBaseType(Indirection t, Type base) {
@@ -266,7 +263,7 @@ private module IteratorIndirections {
// Taint through `operator+=` and `operator-=` on iterators.
call.getStaticCallTarget() instanceof Iterator::IteratorAssignArithmeticOperator and
node2.(IndirectArgumentOutNode).getPreUpdateNode() = node1 and
node1.(IndirectOperand).hasOperandAndIndirectionIndex(call.getArgumentOperand(0), _) and
node1.(IndirectOperand).getOperand() = call.getArgumentOperand(0) and
node1.getType().getUnspecifiedType() = this
)
}
@@ -361,22 +358,17 @@ newtype TBaseSourceVariable =
// Each IR variable gets its own source variable
TBaseIRVariable(IRVariable var) or
// Each allocation gets its own source variable
TBaseCallVariable(CallInstruction call) { not call.getResultIRType() instanceof IRVoidType }
TBaseCallVariable(AllocationInstruction call)
abstract private class AbstractBaseSourceVariable extends TBaseSourceVariable {
abstract class BaseSourceVariable extends TBaseSourceVariable {
/** Gets a textual representation of this element. */
abstract string toString();
/** Gets the type of this base source variable. */
final DataFlowType getType() { this.getLanguageType().hasUnspecifiedType(result, _) }
/** Gets the `CppType` of this base source variable. */
abstract CppType getLanguageType();
abstract DataFlowType getType();
}
final class BaseSourceVariable = AbstractBaseSourceVariable;
class BaseIRVariable extends AbstractBaseSourceVariable, TBaseIRVariable {
class BaseIRVariable extends BaseSourceVariable, TBaseIRVariable {
IRVariable var;
IRVariable getIRVariable() { result = var }
@@ -385,19 +377,19 @@ class BaseIRVariable extends AbstractBaseSourceVariable, TBaseIRVariable {
override string toString() { result = var.toString() }
override CppType getLanguageType() { result = var.getLanguageType() }
override DataFlowType getType() { result = var.getType() }
}
class BaseCallVariable extends AbstractBaseSourceVariable, TBaseCallVariable {
CallInstruction call;
class BaseCallVariable extends BaseSourceVariable, TBaseCallVariable {
AllocationInstruction call;
BaseCallVariable() { this = TBaseCallVariable(call) }
CallInstruction getCallInstruction() { result = call }
AllocationInstruction getCallInstruction() { result = call }
override string toString() { result = call.toString() }
override CppType getLanguageType() { result = getResultLanguageType(call) }
override DataFlowType getType() { result = call.getResultType() }
}
/**
@@ -497,7 +489,8 @@ private class BaseIRVariableInstruction extends BaseSourceVariableInstruction,
override BaseIRVariable getBaseSourceVariable() { result.getIRVariable() = this.getIRVariable() }
}
private class BaseCallInstruction extends BaseSourceVariableInstruction, CallInstruction {
private class BaseAllocationInstruction extends BaseSourceVariableInstruction, AllocationInstruction
{
override BaseCallVariable getBaseSourceVariable() { result.getCallInstruction() = this }
}
@@ -585,6 +578,7 @@ private module Cached {
)
}
pragma[assume_small_delta]
private predicate convertsIntoArgumentRev(Instruction instr) {
convertsIntoArgumentFwd(instr) and
(
@@ -802,7 +796,7 @@ private module Cached {
address.getDef() = instr and
isDereference(load, address) and
isUseImpl(address, _, indirectionIndex - 1) and
result = load
result = instr
)
}
@@ -865,7 +859,7 @@ private module Cached {
* to a specific address.
*/
private predicate isCertainAddress(Operand operand) {
valueNumberOfOperand(operand).getAnInstruction() instanceof VariableAddressInstruction
operand.getDef() instanceof VariableAddressInstruction
or
operand.getType() instanceof Cpp::ReferenceType
}

View File

@@ -1,10 +0,0 @@
/**
* Provides C++-specific definitions for use in the taint tracking library.
*/
private import codeql.dataflow.TaintTracking
private import DataFlowImplSpecific
module CppTaintTracking implements InputSig<CppDataFlow> {
import TaintTrackingUtil
}

View File

@@ -112,7 +112,7 @@ predicate defaultAdditionalTaintStep(DataFlow::Node src, DataFlow::Node sink) {
* of `c` at sinks and inputs to additional taint steps.
*/
bindingset[node]
predicate defaultImplicitTaintRead(DataFlow::Node node, DataFlow::ContentSet c) { none() }
predicate defaultImplicitTaintRead(DataFlow::Node node, DataFlow::Content c) { none() }
/**
* Holds if `node` should be a sanitizer in all global taint flow configurations
@@ -160,7 +160,7 @@ predicate modeledTaintStep(DataFlow::Node nodeIn, DataFlow::Node nodeOut) {
FunctionInput modelIn, FunctionOutput modelOut
|
indirectArgument = callInput(call, modelIn) and
indirectArgument.hasAddressOperandAndIndirectionIndex(nodeIn.asOperand(), _) and
indirectArgument.getAddressOperand() = nodeIn.asOperand() and
call.getStaticCallTarget() = func and
(
func.(DataFlowFunction).hasDataFlow(modelIn, modelOut)

View File

@@ -15,12 +15,15 @@ private import semmle.code.cpp.ir.dataflow.internal.DataFlowUtil
private import semmle.code.cpp.ir.dataflow.internal.SsaInternalsCommon
private module SourceVariables {
class SourceVariable extends BaseSourceVariable {
/**
* Gets the base source variable of this `SourceVariable`.
*/
class SourceVariable instanceof BaseSourceVariable {
string toString() { result = BaseSourceVariable.super.toString() }
BaseSourceVariable getBaseVariable() { result = this }
}
class SourceIRVariable = BaseIRVariable;
class CallVariable = BaseCallVariable;
}
import SourceVariables

View File

@@ -0,0 +1,74 @@
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
private module AddTaintDefaults<DataFlowInternal::FullStateConfigSig Config> implements
DataFlowInternal::FullStateConfigSig
{
import Config
predicate isBarrier(DataFlow::Node node) {
Config::isBarrier(node) or defaultTaintSanitizer(node)
}
predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
Config::isAdditionalFlowStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2)
}
predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
Config::allowImplicitRead(node, c)
or
(
Config::isSink(node, _) or
Config::isAdditionalFlowStep(node, _) or
Config::isAdditionalFlowStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
}
/**
* Constructs a global taint tracking computation.
*/
module Global<DataFlow::ConfigSig Config> implements DataFlow::GlobalFlowSig {
private module Config0 implements DataFlowInternal::FullStateConfigSig {
import DataFlowInternal::DefaultState<Config>
import Config
}
private module C implements DataFlowInternal::FullStateConfigSig {
import AddTaintDefaults<Config0>
}
import DataFlowInternal::Impl<C>
}
/** DEPRECATED: Use `Global` instead. */
deprecated module Make<DataFlow::ConfigSig Config> implements DataFlow::GlobalFlowSig {
import Global<Config>
}
/**
* Constructs a global taint tracking computation using flow state.
*/
module GlobalWithState<DataFlow::StateConfigSig Config> implements DataFlow::GlobalFlowSig {
private module Config0 implements DataFlowInternal::FullStateConfigSig {
import Config
}
private module C implements DataFlowInternal::FullStateConfigSig {
import AddTaintDefaults<Config0>
}
import DataFlowInternal::Impl<C>
}
/** DEPRECATED: Use `GlobalWithState` instead. */
deprecated module MakeWithState<DataFlow::StateConfigSig Config> implements DataFlow::GlobalFlowSig {
import GlobalWithState<Config>
}

View File

@@ -4,8 +4,8 @@
* This file contains the actual implementation of `PrintIR.ql`. For test cases and very small
* databases, `PrintIR.ql` can be run directly to dump the IR for the entire database. For most
* uses, however, it is better to write a query that imports `PrintIR.qll`, extends
* `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
* to dump.
* `PrintIRConfiguration`, and overrides `shouldPrintFunction()` to select a subset of functions to
* dump.
*/
private import internal.IRInternal
@@ -16,7 +16,7 @@ import Imports::IRConfiguration
private newtype TPrintIRConfiguration = MkPrintIRConfiguration()
/**
* The query can extend this class to control which declarations are printed.
* The query can extend this class to control which functions are printed.
*/
class PrintIRConfiguration extends TPrintIRConfiguration {
/** Gets a textual representation of this configuration. */
@@ -24,9 +24,9 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
/**
* Holds if the IR for `func` should be printed. By default, holds for all
* functions, global and namespace variables, and static local variables.
* functions.
*/
predicate shouldPrintDeclaration(Language::Declaration decl) { any() }
predicate shouldPrintFunction(Language::Declaration decl) { any() }
}
/**
@@ -34,12 +34,12 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
*/
private class FilteredIRConfiguration extends IRConfiguration {
override predicate shouldEvaluateDebugStringsForFunction(Language::Declaration func) {
shouldPrintDeclaration(func)
shouldPrintFunction(func)
}
}
private predicate shouldPrintDeclaration(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintDeclaration(decl))
private predicate shouldPrintFunction(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintFunction(decl))
}
private predicate shouldPrintInstruction(Instruction i) {
@@ -90,10 +90,10 @@ private string getOperandPropertyString(Operand operand) {
}
private newtype TPrintableIRNode =
TPrintableIRFunction(IRFunction irFunc) { shouldPrintDeclaration(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintDeclaration(block.getEnclosingFunction()) } or
TPrintableIRFunction(IRFunction irFunc) { shouldPrintFunction(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintFunction(block.getEnclosingFunction()) } or
TPrintableInstruction(Instruction instr) {
shouldPrintInstruction(instr) and shouldPrintDeclaration(instr.getEnclosingFunction())
shouldPrintInstruction(instr) and shouldPrintFunction(instr.getEnclosingFunction())
}
/**

View File

@@ -176,6 +176,7 @@ private predicate binaryValueNumber0(
)
}
pragma[assume_small_delta]
private predicate binaryValueNumber(
BinaryInstruction instr, IRFunction irFunc, Opcode opcode, TValueNumber leftOperand,
TValueNumber rightOperand
@@ -201,6 +202,7 @@ private predicate pointerArithmeticValueNumber0(
)
}
pragma[assume_small_delta]
private predicate pointerArithmeticValueNumber(
PointerArithmeticInstruction instr, IRFunction irFunc, Opcode opcode, int elementSize,
TValueNumber leftOperand, TValueNumber rightOperand
@@ -247,6 +249,7 @@ private predicate loadTotalOverlapValueNumber0(
)
}
pragma[assume_small_delta]
private predicate loadTotalOverlapValueNumber(
LoadTotalOverlapInstruction instr, IRFunction irFunc, IRType type, TValueNumber memOperand,
TValueNumber operand

View File

@@ -4,8 +4,8 @@
* This file contains the actual implementation of `PrintIR.ql`. For test cases and very small
* databases, `PrintIR.ql` can be run directly to dump the IR for the entire database. For most
* uses, however, it is better to write a query that imports `PrintIR.qll`, extends
* `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
* to dump.
* `PrintIRConfiguration`, and overrides `shouldPrintFunction()` to select a subset of functions to
* dump.
*/
private import internal.IRInternal
@@ -16,7 +16,7 @@ import Imports::IRConfiguration
private newtype TPrintIRConfiguration = MkPrintIRConfiguration()
/**
* The query can extend this class to control which declarations are printed.
* The query can extend this class to control which functions are printed.
*/
class PrintIRConfiguration extends TPrintIRConfiguration {
/** Gets a textual representation of this configuration. */
@@ -24,9 +24,9 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
/**
* Holds if the IR for `func` should be printed. By default, holds for all
* functions, global and namespace variables, and static local variables.
* functions.
*/
predicate shouldPrintDeclaration(Language::Declaration decl) { any() }
predicate shouldPrintFunction(Language::Declaration decl) { any() }
}
/**
@@ -34,12 +34,12 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
*/
private class FilteredIRConfiguration extends IRConfiguration {
override predicate shouldEvaluateDebugStringsForFunction(Language::Declaration func) {
shouldPrintDeclaration(func)
shouldPrintFunction(func)
}
}
private predicate shouldPrintDeclaration(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintDeclaration(decl))
private predicate shouldPrintFunction(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintFunction(decl))
}
private predicate shouldPrintInstruction(Instruction i) {
@@ -90,10 +90,10 @@ private string getOperandPropertyString(Operand operand) {
}
private newtype TPrintableIRNode =
TPrintableIRFunction(IRFunction irFunc) { shouldPrintDeclaration(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintDeclaration(block.getEnclosingFunction()) } or
TPrintableIRFunction(IRFunction irFunc) { shouldPrintFunction(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintFunction(block.getEnclosingFunction()) } or
TPrintableInstruction(Instruction instr) {
shouldPrintInstruction(instr) and shouldPrintDeclaration(instr.getEnclosingFunction())
shouldPrintInstruction(instr) and shouldPrintFunction(instr.getEnclosingFunction())
}
/**

View File

@@ -176,6 +176,7 @@ private predicate binaryValueNumber0(
)
}
pragma[assume_small_delta]
private predicate binaryValueNumber(
BinaryInstruction instr, IRFunction irFunc, Opcode opcode, TValueNumber leftOperand,
TValueNumber rightOperand
@@ -201,6 +202,7 @@ private predicate pointerArithmeticValueNumber0(
)
}
pragma[assume_small_delta]
private predicate pointerArithmeticValueNumber(
PointerArithmeticInstruction instr, IRFunction irFunc, Opcode opcode, int elementSize,
TValueNumber leftOperand, TValueNumber rightOperand
@@ -247,6 +249,7 @@ private predicate loadTotalOverlapValueNumber0(
)
}
pragma[assume_small_delta]
private predicate loadTotalOverlapValueNumber(
LoadTotalOverlapInstruction instr, IRFunction irFunc, IRType type, TValueNumber memOperand,
TValueNumber operand

View File

@@ -991,19 +991,9 @@ class TranslatedStructuredBindingVariableAccess extends TranslatedNonConstantExp
class TranslatedFunctionAccess extends TranslatedNonConstantExpr {
override FunctionAccess expr;
override TranslatedElement getChild(int id) {
id = 0 and result = this.getQualifier() // Might not exist
}
override TranslatedElement getChild(int id) { none() }
final TranslatedExpr getQualifier() {
result = getTranslatedExpr(expr.getQualifier().getFullyConverted())
}
override Instruction getFirstInstruction() {
if exists(this.getQualifier())
then result = this.getQualifier().getFirstInstruction()
else result = this.getInstruction(OnlyInstructionTag())
}
override Instruction getFirstInstruction() { result = this.getInstruction(OnlyInstructionTag()) }
override Instruction getResult() { result = this.getInstruction(OnlyInstructionTag()) }
@@ -1024,9 +1014,7 @@ class TranslatedFunctionAccess extends TranslatedNonConstantExpr {
result = expr.getTarget()
}
override Instruction getChildSuccessor(TranslatedElement child) {
child = this.getQualifier() and result = this.getInstruction(OnlyInstructionTag())
}
override Instruction getChildSuccessor(TranslatedElement child) { none() }
}
/**

View File

@@ -4,8 +4,8 @@
* This file contains the actual implementation of `PrintIR.ql`. For test cases and very small
* databases, `PrintIR.ql` can be run directly to dump the IR for the entire database. For most
* uses, however, it is better to write a query that imports `PrintIR.qll`, extends
* `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
* to dump.
* `PrintIRConfiguration`, and overrides `shouldPrintFunction()` to select a subset of functions to
* dump.
*/
private import internal.IRInternal
@@ -16,7 +16,7 @@ import Imports::IRConfiguration
private newtype TPrintIRConfiguration = MkPrintIRConfiguration()
/**
* The query can extend this class to control which declarations are printed.
* The query can extend this class to control which functions are printed.
*/
class PrintIRConfiguration extends TPrintIRConfiguration {
/** Gets a textual representation of this configuration. */
@@ -24,9 +24,9 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
/**
* Holds if the IR for `func` should be printed. By default, holds for all
* functions, global and namespace variables, and static local variables.
* functions.
*/
predicate shouldPrintDeclaration(Language::Declaration decl) { any() }
predicate shouldPrintFunction(Language::Declaration decl) { any() }
}
/**
@@ -34,12 +34,12 @@ class PrintIRConfiguration extends TPrintIRConfiguration {
*/
private class FilteredIRConfiguration extends IRConfiguration {
override predicate shouldEvaluateDebugStringsForFunction(Language::Declaration func) {
shouldPrintDeclaration(func)
shouldPrintFunction(func)
}
}
private predicate shouldPrintDeclaration(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintDeclaration(decl))
private predicate shouldPrintFunction(Language::Declaration decl) {
exists(PrintIRConfiguration config | config.shouldPrintFunction(decl))
}
private predicate shouldPrintInstruction(Instruction i) {
@@ -90,10 +90,10 @@ private string getOperandPropertyString(Operand operand) {
}
private newtype TPrintableIRNode =
TPrintableIRFunction(IRFunction irFunc) { shouldPrintDeclaration(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintDeclaration(block.getEnclosingFunction()) } or
TPrintableIRFunction(IRFunction irFunc) { shouldPrintFunction(irFunc.getFunction()) } or
TPrintableIRBlock(IRBlock block) { shouldPrintFunction(block.getEnclosingFunction()) } or
TPrintableInstruction(Instruction instr) {
shouldPrintInstruction(instr) and shouldPrintDeclaration(instr.getEnclosingFunction())
shouldPrintInstruction(instr) and shouldPrintFunction(instr.getEnclosingFunction())
}
/**

View File

@@ -176,6 +176,7 @@ private predicate binaryValueNumber0(
)
}
pragma[assume_small_delta]
private predicate binaryValueNumber(
BinaryInstruction instr, IRFunction irFunc, Opcode opcode, TValueNumber leftOperand,
TValueNumber rightOperand
@@ -201,6 +202,7 @@ private predicate pointerArithmeticValueNumber0(
)
}
pragma[assume_small_delta]
private predicate pointerArithmeticValueNumber(
PointerArithmeticInstruction instr, IRFunction irFunc, Opcode opcode, int elementSize,
TValueNumber leftOperand, TValueNumber rightOperand
@@ -247,6 +249,7 @@ private predicate loadTotalOverlapValueNumber0(
)
}
pragma[assume_small_delta]
private predicate loadTotalOverlapValueNumber(
LoadTotalOverlapInstruction instr, IRFunction irFunc, IRType type, TValueNumber memOperand,
TValueNumber operand

View File

@@ -108,7 +108,7 @@ class StrcpyFunction extends ArrayFunction, DataFlowFunction, TaintFunction, Sid
// these may do only a partial copy of the input buffer to the output
// buffer
exists(this.getParamSize()) and
input.isParameterDeref(this.getParamSrc()) and
input.isParameter(this.getParamSrc()) and
(
output.isParameterDeref(this.getParamDest()) or
output.isReturnValueDeref()

View File

@@ -188,9 +188,6 @@ module SemanticExprConfig {
none()
}
/** Holds if no range analysis should be performed on the phi edges in `f`. */
private predicate excludeFunction(Cpp::Function f) { count(f.getEntryPoint()) > 1 }
SemType getUnknownExprType(Expr expr) { result = getSemanticType(expr.getResultIRType()) }
class BasicBlock = IR::IRBlock;
@@ -273,13 +270,7 @@ module SemanticExprConfig {
getSemanticExpr(v.asInstruction()) = sourceExpr
}
predicate phi(SsaVariable v) {
exists(IR::PhiInstruction phi, Cpp::Function f |
phi = v.asInstruction() and
f = phi.getEnclosingFunction() and
not excludeFunction(f)
)
}
predicate phi(SsaVariable v) { v.asInstruction() instanceof IR::PhiInstruction }
SsaVariable getAPhiInput(SsaVariable v) {
exists(IR::PhiInstruction instr | v.asInstruction() = instr |

View File

@@ -70,27 +70,6 @@ predicate semBackEdge(SemSsaPhiNode phi, SemSsaVariable inp, SemSsaReadPositionP
// Conservatively assume that every edge is a back edge if we don't have dominance information.
(
phi.getBasicBlock().bbDominates(edge.getOrigBlock()) or
irreducibleSccEdge(edge.getOrigBlock(), phi.getBasicBlock()) or
not edge.getOrigBlock().hasDominanceInformation()
)
}
/**
* Holds if the edge from b1 to b2 is part of a multiple-entry cycle in an irreducible control flow
* graph.
*
* An ireducible control flow graph is one where the usual dominance-based back edge detection does
* not work, because there is a cycle with multiple entry points, meaning there are
* mutually-reachable basic blocks where neither dominates the other. For such a graph, we first
* remove all detectable back-edges using the normal condition that the predecessor block is
* dominated by the successor block, then mark all edges in a cycle in the resulting graph as back
* edges.
*/
private predicate irreducibleSccEdge(SemBasicBlock b1, SemBasicBlock b2) {
trimmedEdge(b1, b2) and trimmedEdge+(b2, b1)
}
private predicate trimmedEdge(SemBasicBlock pred, SemBasicBlock succ) {
pred.getASuccessor() = succ and
not succ.bbDominates(pred)
}

View File

@@ -574,6 +574,16 @@ module RangeStage<
)
}
/** Holds if `e >= 1` as determined by sign analysis. */
private predicate strictlyPositiveIntegralExpr(SemExpr e) {
semStrictlyPositive(e) and getTrackedType(e) instanceof SemIntegerType
}
/** Holds if `e <= -1` as determined by sign analysis. */
private predicate strictlyNegativeIntegralExpr(SemExpr e) {
semStrictlyNegative(e) and getTrackedType(e) instanceof SemIntegerType
}
/**
* Holds if `e1 + delta` is a valid bound for `e2`.
* - `upper = true` : `e2 <= e1 + delta`
@@ -587,6 +597,27 @@ module RangeStage<
delta = D::fromInt(0) and
(upper = true or upper = false)
or
exists(SemExpr x, SemSubExpr sub |
e2 = sub and
sub.getLeftOperand() = e1 and
sub.getRightOperand() = x
|
// `x instanceof ConstantIntegerExpr` is covered by valueFlowStep
not x instanceof SemConstantIntegerExpr and
if strictlyPositiveIntegralExpr(x)
then upper = true and delta = D::fromInt(-1)
else
if semPositive(x)
then upper = true and delta = D::fromInt(0)
else
if strictlyNegativeIntegralExpr(x)
then upper = false and delta = D::fromInt(1)
else
if semNegative(x)
then upper = false and delta = D::fromInt(0)
else none()
)
or
e2.(SemRemExpr).getRightOperand() = e1 and
semPositive(e1) and
delta = D::fromInt(-1) and
@@ -846,6 +877,7 @@ module RangeStage<
)
}
pragma[assume_small_delta]
pragma[nomagic]
private predicate boundedPhiRankStep(
SemSsaPhiNode phi, SemBound b, D::Delta delta, boolean upper, boolean fromBackEdge,
@@ -1106,23 +1138,6 @@ module RangeStage<
b = bRight and origdelta = odRight and reason = rRight and bLeft instanceof SemZeroBound
)
or
exists(D::Delta dLeft, D::Delta dRight, boolean fbeLeft, boolean fbeRight |
boundedSubOperandLeft(e, upper, b, dLeft, fbeLeft, origdelta, reason) and
boundedSubOperandRight(e, upper, dRight, fbeRight) and
// when `upper` is `true` we have:
// left <= b + dLeft
// right >= 0 + dRight
// left - right <= b + dLeft - (0 + dRight)
// = b + (dLeft - dRight)
// and when `upper` is `false` we have:
// left >= b + dLeft
// right <= 0 + dRight
// left - right >= b + dLeft - (0 + dRight)
// = b + (dLeft - dRight)
delta = D::fromFloat(D::toFloat(dLeft) - D::toFloat(dRight)) and
fromBackEdge = fbeLeft.booleanOr(fbeRight)
)
or
exists(
SemRemExpr rem, D::Delta d_max, D::Delta d1, D::Delta d2, boolean fbe1, boolean fbe2,
D::Delta od1, D::Delta od2, SemReason r1, SemReason r2
@@ -1187,37 +1202,6 @@ module RangeStage<
)
}
/**
* Holds if `sub = left - right` and `left <= b + delta` if `upper` is `true`
* and `left >= b + delta` is `upper` is `false`.
*/
pragma[nomagic]
private predicate boundedSubOperandLeft(
SemSubExpr sub, boolean upper, SemBound b, D::Delta delta, boolean fromBackEdge,
D::Delta origdelta, SemReason reason
) {
// `semValueFlowStep` already handles the case where one of the operands is a constant.
not semValueFlowStep(sub, _, _) and
bounded(sub.getLeftOperand(), b, delta, upper, fromBackEdge, origdelta, reason)
}
/**
* Holds if `sub = left - right` and `right <= 0 + delta` if `upper` is `false`
* and `right >= 0 + delta` is `upper` is `true`.
*
* Note that the boolean value of `upper` is flipped compared to many other predicates in
* this file. This ensures a clean join at the call-site.
*/
pragma[nomagic]
private predicate boundedSubOperandRight(
SemSubExpr sub, boolean upper, D::Delta delta, boolean fromBackEdge
) {
// `semValueFlowStep` already handles the case where one of the operands is a constant.
not semValueFlowStep(sub, _, _) and
bounded(sub.getRightOperand(), any(SemZeroBound zb), delta, upper.booleanNot(), fromBackEdge, _,
_)
}
pragma[nomagic]
private predicate boundedRemExpr(
SemRemExpr rem, boolean upper, D::Delta delta, boolean fromBackEdge, D::Delta origdelta,

View File

@@ -53,7 +53,7 @@ private class ArgvSource extends LocalFlowSource {
exists(Function main, Parameter argv |
main.hasGlobalName("main") and
main.getParameter(1) = argv and
this.asParameter(2) = argv
this.asParameter(_) = argv
)
}

View File

@@ -1,312 +0,0 @@
/**
* This file provides the first phase of the `cpp/invalid-pointer-deref` query that identifies flow
* from an allocation to a pointer-arithmetic instruction that constructs a pointer that is out of bounds.
*
* Consider the following snippet:
* ```cpp
* 1. char* base = (char*)malloc(size);
* 2. char* end = base + size;
* 3. for(int *p = base; p <= end; p++) {
* 4. use(*p); // BUG: Should have been bounded by `p < end`.
* 5. }
* ```
* this file identifies the flow from `new int[size]` to `base + size`.
*
* This is done using the product-flow library. The configuration tracks flow from the pair
* `(allocation, size of allocation)` to a pair `(a, b)` where there exists a pointer-arithmetic instruction
* `pai = a + r` such that `b` is a dataflow node where `b <= r`. Because there will be a dataflow-path from
* `allocation` to `a` this means that the `pai` will compute a pointer that is some number of elements beyond
* the end position of the allocation. See `pointerAddInstructionHasBounds` for the implementation of this.
*
* In the above example, the pair `(a, b)` is `(base, size)` with `base` and `size` coming from the expression
* `base + size` on line 2, which is also the pointer-arithmetic instruction. In general, the pair does not necessarily
* correspond directly to the operands of the pointer-arithmetic instruction.
* In the following example, the pair is again `(base, size)`, but with `base` coming from line 3 and `size` from line 2,
* and the pointer-arithmetic instruction being `base + n` on line 3:
* ```cpp
* 1. int* base = new int[size];
* 2. if(n <= size) {
* 3. int* end = base + n;
* 4. for(int* p = base; p <= end; ++p) {
* 5. *p = 0; // BUG: Should have been bounded by `p < end`.
* 6. }
* 7. }
* ```
*
* Handling false positives:
*
* Consider a snippet such as:
* ```cpp
* 1. int* base = new int[size];
* 2. int n = condition() ? size : 0;
* 3. if(n >= size) return;
* 4. int* end = base + n;
* 5. for(int* p = base; p <= end; ++p) {
* 6. *p = 0; // This is fine since `end < base + size`
* 7. }
* ```
* In order to remove this false positive we define a barrier (see `SizeBarrier::SizeBarrierConfig`) that finds the
* possible guards that compares a value to the size of the allocation. In the above example, this is the `(n >= size)`
* guard on line 3. `SizeBarrier::getABarrierNode` then defines any node that is guarded by such a guard as a barrier
* in the dataflow configuration.
*/
private import cpp
private import semmle.code.cpp.ir.dataflow.internal.ProductFlow
private import semmle.code.cpp.ir.ValueNumbering
private import semmle.code.cpp.controlflow.IRGuards
private import codeql.util.Unit
private import RangeAnalysisUtil
private VariableAccess getAVariableAccess(Expr e) { e.getAChild*() = result }
/**
* Holds if the `(n, state)` pair represents the source of flow for the size
* expression associated with `alloc`.
*/
predicate hasSize(HeuristicAllocationExpr alloc, DataFlow::Node n, int state) {
exists(VariableAccess va, Expr size, int delta |
size = alloc.getSizeExpr() and
// Get the unique variable in a size expression like `x` in `malloc(x + 1)`.
va = unique( | | getAVariableAccess(size)) and
// Compute `delta` as the constant difference between `x` and `x + 1`.
bounded1(any(Instruction instr | instr.getUnconvertedResultExpression() = size),
any(LoadInstruction load | load.getUnconvertedResultExpression() = va), delta) and
n.asConvertedExpr() = va.getFullyConverted() and
state = delta
)
}
/**
* A module that encapsulates a barrier guard to remove false positives from flow like:
* ```cpp
* char *p = new char[size];
* // ...
* unsigned n = size;
* // ...
* if(n < size) {
* use(*p[n]);
* }
* ```
* In this case, the sink pair identified by the product flow library (without any additional barriers)
* would be `(p, n)` (where `n` is the `n` in `p[n]`), because there exists a pointer-arithmetic
* instruction `pai = a + b` such that:
* 1. the allocation flows to `a`, and
* 2. `b <= n` where `n` is the `n` in `p[n]`
* but because there's a strict comparison that compares `n` against the size of the allocation this
* snippet is fine.
*/
private module SizeBarrier {
private module SizeBarrierConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node source) {
// The sources is the same as in the sources for the second
// projection in the `AllocToInvalidPointerConfig` module.
hasSize(_, source, _)
}
/**
* Holds if `small <= large + k` holds if `g` evaluates to `testIsTrue`.
*/
additional predicate isSink(
DataFlow::Node small, DataFlow::Node large, IRGuardCondition g, int k, boolean testIsTrue
) {
// The sink is any "large" side of a relational comparison. i.e., the `large` expression
// in a guard such as `small <= large + k`.
g.comparesLt(small.asOperand(), large.asOperand(), k + 1, true, testIsTrue)
}
predicate isSink(DataFlow::Node sink) { isSink(_, sink, _, _, _) }
}
module SizeBarrierFlow = DataFlow::Global<SizeBarrierConfig>;
private int getASizeAddend(DataFlow::Node node) {
exists(DataFlow::Node source |
SizeBarrierFlow::flow(source, node) and
hasSize(_, source, result)
)
}
/**
* Holds if `small <= large + k` holds if `g` evaluates to `edge`.
*/
private predicate operandGuardChecks(
IRGuardCondition g, Operand small, DataFlow::Node large, int k, boolean edge
) {
SizeBarrierFlow::flowTo(large) and
SizeBarrierConfig::isSink(DataFlow::operandNode(small), large, g, k, edge)
}
/**
* Gets an instruction `instr` that is guarded by a check such as `instr <= small + delta` where
* `small <= _ + k` and `small` is the "small side" of of a relational comparison that checks
* whether `small <= size` where `size` is the size of an allocation.
*/
Instruction getABarrierInstruction0(int delta, int k) {
exists(
IRGuardCondition g, ValueNumber value, Operand small, boolean edge, DataFlow::Node large
|
// We know:
// 1. result <= value + delta (by `bounded`)
// 2. value <= large + k (by `operandGuardChecks`).
// So:
// result <= value + delta (by 1.)
// <= large + k + delta (by 2.)
small = value.getAUse() and
operandGuardChecks(pragma[only_bind_into](g), pragma[only_bind_into](small), large,
pragma[only_bind_into](k), pragma[only_bind_into](edge)) and
bounded(result, value.getAnInstruction(), delta) and
g.controls(result.getBlock(), edge) and
k < getASizeAddend(large)
)
}
/**
* Gets an instruction that is guarded by a guard condition which ensures that
* the value of the instruction is upper-bounded by size of some allocation.
*/
bindingset[state]
pragma[inline_late]
Instruction getABarrierInstruction(int state) {
exists(int delta, int k |
state > k + delta and
// result <= "size of allocation" + delta + k
// < "size of allocation" + state
result = getABarrierInstruction0(delta, k)
)
}
/**
* Gets a `DataFlow::Node` that is guarded by a guard condition which ensures that
* the value of the node is upper-bounded by size of some allocation.
*/
DataFlow::Node getABarrierNode(int state) {
exists(DataFlow::Node source, int delta, int k |
SizeBarrierFlow::flow(source, result) and
hasSize(_, source, state) and
result.asInstruction() = SizeBarrier::getABarrierInstruction0(delta, k) and
state > k + delta
// so now we have:
// result <= "size of allocation" + delta + k
// < "size of allocation" + state
)
}
}
private module InterestingPointerAddInstruction {
private module PointerAddInstructionConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node source) {
// The sources is the same as in the sources for the second
// projection in the `AllocToInvalidPointerConfig` module.
hasSize(source.asConvertedExpr(), _, _)
}
predicate isSink(DataFlow::Node sink) {
sink.asInstruction() = any(PointerAddInstruction pai).getLeft()
}
}
private import DataFlow::Global<PointerAddInstructionConfig>
/**
* Holds if `pai` is a pointer-arithmetic instruction such that the
* result of an allocation flows to the left-hand side of `pai`.
*
* This predicate is used to reduce the set of tuples in `isSinkPair`.
*/
predicate isInteresting(PointerAddInstruction pai) {
exists(DataFlow::Node n |
n.asInstruction() = pai.getLeft() and
flowTo(n)
)
}
}
/**
* A product-flow configuration for flow from an `(allocation, size)` pair to a
* pointer-arithmetic operation `pai` such that `pai <= allocation + size`.
*/
private module Config implements ProductFlow::StateConfigSig {
class FlowState1 = Unit;
class FlowState2 = int;
predicate isSourcePair(
DataFlow::Node allocSource, FlowState1 unit, DataFlow::Node sizeSource, FlowState2 sizeAddend
) {
// In the case of an allocation like
// ```cpp
// malloc(size + 1);
// ```
// we use `state2` to remember that there was an offset (in this case an offset of `1`) added
// to the size of the allocation. This state is then checked in `isSinkPair`.
exists(unit) and
hasSize(allocSource.asConvertedExpr(), sizeSource, sizeAddend)
}
predicate isSinkPair(
DataFlow::Node allocSink, FlowState1 unit, DataFlow::Node sizeSink, FlowState2 sizeAddend
) {
exists(unit) and
// We check that the delta computed by the range analysis matches the
// state value that we set in `isSourcePair`.
pointerAddInstructionHasBounds0(_, allocSink, sizeSink, sizeAddend)
}
predicate isBarrier2(DataFlow::Node node, FlowState2 state) {
node = SizeBarrier::getABarrierNode(state)
}
predicate isBarrierIn1(DataFlow::Node node) { isSourcePair(node, _, _, _) }
predicate isBarrierOut2(DataFlow::Node node) {
node = any(DataFlow::SsaPhiNode phi).getAnInput(true)
}
}
private module AllocToInvalidPointerFlow = ProductFlow::GlobalWithState<Config>;
/**
* Holds if `pai` is non-strictly upper bounded by `sizeSink + delta` and `allocSink` is the
* left operand of the pointer-arithmetic operation.
*
* For example in,
* ```cpp
* char* end = p + (size + 1);
* ```
* We will have:
* - `pai` is `p + (size + 1)`,
* - `allocSink` is `p`
* - `sizeSink` is `size`
* - `delta` is `1`.
*/
pragma[nomagic]
private predicate pointerAddInstructionHasBounds0(
PointerAddInstruction pai, DataFlow::Node allocSink, DataFlow::Node sizeSink, int delta
) {
InterestingPointerAddInstruction::isInteresting(pragma[only_bind_into](pai)) and
exists(Instruction right, Instruction sizeInstr |
pai.getRight() = right and
pai.getLeft() = allocSink.asInstruction() and
sizeInstr = sizeSink.asInstruction() and
// pai.getRight() <= sizeSink + delta
bounded1(right, sizeInstr, delta) and
not right = SizeBarrier::getABarrierInstruction(delta) and
not sizeInstr = SizeBarrier::getABarrierInstruction(delta)
)
}
/**
* Holds if `allocation` flows to `allocSink` and `allocSink` represents the left operand
* of the pointer-arithmetic instruction `pai = a + b` (i.e., `allocSink = a`), and
* `b <= allocation + delta`.
*/
pragma[nomagic]
predicate pointerAddInstructionHasBounds(
DataFlow::Node allocation, PointerAddInstruction pai, DataFlow::Node allocSink, int delta
) {
exists(DataFlow::Node sizeSink |
AllocToInvalidPointerFlow::flow(allocation, _, allocSink, sizeSink) and
pointerAddInstructionHasBounds0(pai, allocSink, sizeSink, delta)
)
}

View File

@@ -1,294 +0,0 @@
/**
* This file provides the second phase of the `cpp/invalid-pointer-deref` query that identifies flow
* from the out-of-bounds pointer identified by the `AllocationToInvalidPointer.qll` library to
* a dereference of the out-of-bounds pointer.
*
* Consider the following snippet:
* ```cpp
* 1. char* base = (char*)malloc(size);
* 2. char* end = base + size;
* 3. for(char *p = base; p <= end; p++) {
* 4. use(*p); // BUG: Should have been bounded by `p < end`.
* 5. }
* ```
* this file identifies the flow from `base + size` to `end`. We call `base + size` the "dereference source" and `end`
* the "dereference sink" (even though `end` is not actually dereferenced we will use this term because we will perform
* dataflow to find a use of a pointer `x` such that `x <= end` which is dereferenced. In the above example, `x` is `p`
* on line 4).
*
* Merely _constructing_ a pointer that's out-of-bounds is fine if the pointer is never dereferenced (in reality, the
* standard only guarantees that it is safe to move the pointer one element past the last element, but we ignore that
* here). So this step is about identifying which of the out-of-bounds pointers found by `pointerAddInstructionHasBounds`
* in `AllocationToInvalidPointer.qll` are actually being dereferenced. We do this using a regular dataflow
* configuration (see `InvalidPointerToDerefConfig`).
*
* The dataflow traversal defines the set of sources as any dataflow node `n` such that there exists a pointer-arithmetic
* instruction `pai` found by `AllocationToInvalidPointer.qll` and a `n.asInstruction() >= pai + deltaDerefSourceAndPai`.
* Here, `deltaDerefSourceAndPai` is the constant difference between the source we track for finding a dereference and the
* pointer-arithmetic instruction.
*
* The set of sinks is defined as any dataflow node `n` such that `addr <= n.asInstruction() + deltaDerefSinkAndDerefAddress`
* for some address operand `addr` and constant difference `deltaDerefSinkAndDerefAddress`. Since an address operand is
* always consumed by an instruction that performs a dereference this lets us identify a "bad dereference". We call the
* instruction that consumes the address operand the "operation".
*
* For example, consider the flow from `base + size` to `end` above. The sink is `end` on line 3 because
* `p <= end.asInstruction() + deltaDerefSinkAndDerefAddress`, where `p` is the address operand in `use(*p)` and
* `deltaDerefSinkAndDerefAddress >= 0`. The load attached to `*p` is the "operation". To ensure that the path makes
* intuitive sense, we only pick operations that are control-flow reachable from the dereference sink.
*
* To compute how many elements the dereference is beyond the end position of the allocation, we sum the two deltas
* `deltaDerefSourceAndPai` and `deltaDerefSinkAndDerefAddress`. This is done in the `operationIsOffBy` predicate
* (which is the only predicate exposed by this file).
*
* Handling false positives:
*
* Consider the following snippet:
* ```cpp
* 1. char *p = new char[size];
* 2. char *end = p + size;
* 3. if (p < end) {
* 4. p += 1;
* 5. }
* 6. if (p < end) {
* 7. int val = *p; // GOOD
* 8. }
* ```
* this is safe because `p` is guarded to be strictly less than `end` on line 6 before the dereference on line 7. However, if we
* run the query on the above without further modifications we would see an alert on line 7. This is because range analysis infers
* that `p <= end` after the increment on line 4, and thus the result of `p += 1` is seen as a valid dereference source. This
* node then flows to `p` on line 6 (which is a valid dereference sink since it non-strictly upper bounds an address operand), and
* range analysis then infers that the address operand of `*p` (i.e., `p`) is non-strictly upper bounded by `p`, and thus reports
* an alert on line 7.
*
* In order to handle the above false positive, we define a barrier that identifies guards such as `p < end` that ensures that a value
* is less than the pointer-arithmetic instruction that computed the invalid pointer. This is done in the `InvalidPointerToDerefBarrier`
* module. Since the node we are tracking is not necessarily _equal_ to the pointer-arithmetic instruction, but rather satisfies
* `node.asInstruction() <= pai + deltaDerefSourceAndPai`, we need to account for the delta when checking if a guard is sufficiently
* strong to infer that a future dereference is safe. To do this, we check that the guard guarantees that a node `n` satisfies
* `n < node + k` where `node` is a node such that `node <= pai`. Thus, we know that any node `m` such that `m <= n + delta` where
* `delta + k <= 0` will be safe because:
* ```
* m <= n + delta
* < node + k + delta
* <= pai + k + delta
* <= pai
* ```
*/
private import cpp
private import semmle.code.cpp.dataflow.new.DataFlow
private import semmle.code.cpp.ir.ValueNumbering
private import semmle.code.cpp.controlflow.IRGuards
private import AllocationToInvalidPointer as AllocToInvalidPointer
private import RangeAnalysisUtil
private module InvalidPointerToDerefBarrier {
private module BarrierConfig implements DataFlow::ConfigSig {
additional predicate isSource(DataFlow::Node source, PointerArithmeticInstruction pai) {
invalidPointerToDerefSource(_, pai, _, _) and
// source <= pai
bounded2(source.asInstruction(), pai, any(int d | d <= 0))
}
predicate isSource(DataFlow::Node source) { isSource(source, _) }
additional predicate isSink(
DataFlow::Node small, DataFlow::Node large, IRGuardCondition g, int k, boolean testIsTrue
) {
// The sink is any "large" side of a relational comparison.
g.comparesLt(small.asOperand(), large.asOperand(), k, true, testIsTrue)
}
predicate isSink(DataFlow::Node sink) { isSink(_, sink, _, _, _) }
}
private module BarrierFlow = DataFlow::Global<BarrierConfig>;
/**
* Holds if `g` ensures that `small < large + k` if `g` evaluates to `edge`.
*
* Additionally, it also holds that `large <= pai`. Thus, when `g` evaluates to `edge`
* it holds that `small < pai + k`.
*/
private predicate operandGuardChecks(
PointerArithmeticInstruction pai, IRGuardCondition g, Operand small, int k, boolean edge
) {
exists(DataFlow::Node source, DataFlow::Node nSmall, DataFlow::Node nLarge |
nSmall.asOperand() = small and
BarrierConfig::isSource(source, pai) and
BarrierFlow::flow(source, nLarge) and
BarrierConfig::isSink(nSmall, nLarge, g, k, edge)
)
}
/**
* Gets an instruction `instr` such that `instr < pai`.
*/
Instruction getABarrierInstruction(PointerArithmeticInstruction pai) {
exists(IRGuardCondition g, ValueNumber value, Operand use, boolean edge, int delta, int k |
use = value.getAUse() and
// value < pai + k
operandGuardChecks(pai, pragma[only_bind_into](g), pragma[only_bind_into](use),
pragma[only_bind_into](k), pragma[only_bind_into](edge)) and
// result <= value + delta
bounded(result, value.getAnInstruction(), delta) and
g.controls(result.getBlock(), edge) and
delta + k <= 0
// combining the above we have: result < pai + k + delta <= pai
)
}
DataFlow::Node getABarrierNode(PointerArithmeticInstruction pai) {
result.asOperand() = getABarrierInstruction(pai).getAUse()
}
/**
* Gets an address operand whose definition `instr` satisfies `instr < pai`.
*/
AddressOperand getABarrierAddressOperand(PointerArithmeticInstruction pai) {
result.getDef() = getABarrierInstruction(pai)
}
}
/**
* A configuration to track flow from a pointer-arithmetic operation found
* by `AllocToInvalidPointerConfig` to a dereference of the pointer.
*/
private module InvalidPointerToDerefConfig implements DataFlow::StateConfigSig {
class FlowState extends PointerArithmeticInstruction {
FlowState() { invalidPointerToDerefSource(_, this, _, _) }
}
predicate isSource(DataFlow::Node source, FlowState pai) {
invalidPointerToDerefSource(_, pai, source, _)
}
pragma[inline]
predicate isSink(DataFlow::Node sink) { isInvalidPointerDerefSink(sink, _, _, _, _) }
predicate isSink(DataFlow::Node sink, FlowState pai) { none() }
predicate isBarrier(DataFlow::Node node) {
node = any(DataFlow::SsaPhiNode phi | not phi.isPhiRead()).getAnInput(true)
}
predicate isBarrier(DataFlow::Node node, FlowState pai) {
// `node = getABarrierNode(pai)` ensures that node < pai, so this node is safe to dereference.
// Note that this is the only place where the `FlowState` is used in this configuration.
node = InvalidPointerToDerefBarrier::getABarrierNode(pai)
}
}
private import DataFlow::GlobalWithState<InvalidPointerToDerefConfig>
/**
* Holds if `allocSource` is dataflow node that represents an allocation that flows to the
* left-hand side of the pointer-arithmetic `pai`, and `derefSource <= pai + derefSourcePaiDelta`.
*
* For example, if `pai` is a pointer-arithmetic operation `p + size` in an expression such
* as `(p + size) + 1` and `derefSource` is the node representing `(p + size) + 1`. In this
* case `derefSourcePaiDelta` is 1.
*/
private predicate invalidPointerToDerefSource(
DataFlow::Node allocSource, PointerArithmeticInstruction pai, DataFlow::Node derefSource,
int deltaDerefSourceAndPai
) {
// Note that `deltaDerefSourceAndPai` is not necessarily equal to `rhsSizeDelta`:
// `rhsSizeDelta` is the constant offset added to the size of the allocation, and
// `deltaDerefSourceAndPai` is the constant difference between the pointer-arithmetic instruction
// and the instruction computing the address for which we will search for a dereference.
AllocToInvalidPointer::pointerAddInstructionHasBounds(allocSource, pai, _, _) and
// derefSource <= pai + deltaDerefSourceAndPai
bounded2(derefSource.asInstruction(), pai, deltaDerefSourceAndPai) and
deltaDerefSourceAndPai >= 0
}
/**
* Holds if `sink` is a sink for `InvalidPointerToDerefConfig` and `i` is a `StoreInstruction` that
* writes to an address `addr` such that `addr <= sink`, or `i` is a `LoadInstruction` that
* reads from an address `addr` such that `addr <= sink`.
*/
pragma[inline]
private predicate isInvalidPointerDerefSink(
DataFlow::Node sink, AddressOperand addr, Instruction i, string operation,
int deltaDerefSinkAndDerefAddress
) {
exists(Instruction s |
s = sink.asInstruction() and
bounded(addr.getDef(), s, deltaDerefSinkAndDerefAddress) and
deltaDerefSinkAndDerefAddress >= 0 and
i.getAnOperand() = addr
|
i instanceof StoreInstruction and
operation = "write"
or
i instanceof LoadInstruction and
operation = "read"
)
}
/**
* Yields any instruction that is control-flow reachable from `instr`.
*/
bindingset[instr, result]
pragma[inline_late]
private Instruction getASuccessor(Instruction instr) {
exists(IRBlock b, int instrIndex, int resultIndex |
b.getInstruction(instrIndex) = instr and
b.getInstruction(resultIndex) = result
|
resultIndex >= instrIndex
)
or
instr.getBlock().getASuccessor+() = result.getBlock()
}
private predicate paiForDereferenceSink(
PointerArithmeticInstruction pai, DataFlow::Node derefSink, int deltaDerefSourceAndPai
) {
exists(DataFlow::Node derefSource |
invalidPointerToDerefSource(_, pai, derefSource, deltaDerefSourceAndPai) and
flow(derefSource, derefSink)
)
}
/**
* Holds if `derefSink` is a dataflow node that represents an out-of-bounds address that is about to
* be dereferenced by `operation` (which is either a `StoreInstruction` or `LoadInstruction`), and
* `pai` is the pointer-arithmetic operation that caused the `derefSink` to be out-of-bounds.
*/
private predicate derefSinkToOperation(
DataFlow::Node derefSink, PointerArithmeticInstruction pai, DataFlow::Node operation,
string description, int deltaDerefSourceAndPai, int deltaDerefSinkAndDerefAddress
) {
exists(Instruction operationInstr, AddressOperand addr |
paiForDereferenceSink(pai, pragma[only_bind_into](derefSink), deltaDerefSourceAndPai) and
isInvalidPointerDerefSink(derefSink, addr, operationInstr, description,
deltaDerefSinkAndDerefAddress) and
operationInstr = getASuccessor(derefSink.asInstruction()) and
operation.asInstruction() = operationInstr and
not addr = InvalidPointerToDerefBarrier::getABarrierAddressOperand(pai)
)
}
/**
* Holds if `allocation` is the result of an allocation that flows to the left-hand side of `pai`, and where
* the right-hand side of `pai` is an offset such that the result of `pai` points to an out-of-bounds pointer.
*
* Furthermore, `derefSource` is at least as large as `pai` and flows to `derefSink` before being dereferenced
* by `operation` (which is either a `StoreInstruction` or `LoadInstruction`). The result is that `operation`
* dereferences a pointer that's "off by `delta`" number of elements.
*/
predicate operationIsOffBy(
DataFlow::Node allocation, PointerArithmeticInstruction pai, DataFlow::Node derefSource,
DataFlow::Node derefSink, string description, DataFlow::Node operation, int delta
) {
exists(int deltaDerefSourceAndPai, int deltaDerefSinkAndDerefAddress |
invalidPointerToDerefSource(allocation, pai, derefSource, deltaDerefSourceAndPai) and
flow(derefSource, derefSink) and
derefSinkToOperation(derefSink, pai, operation, description, deltaDerefSourceAndPai,
deltaDerefSinkAndDerefAddress) and
delta = deltaDerefSourceAndPai + deltaDerefSinkAndDerefAddress
)
}

View File

@@ -1,57 +0,0 @@
/**
* This file contains the range-analysis specific parts of the `cpp/invalid-pointer-deref` query
* that is used by both `AllocationToInvalidPointer.qll` and `InvalidPointerToDereference.qll`.
*/
private import cpp
private import semmle.code.cpp.rangeanalysis.new.internal.semantic.analysis.RangeAnalysis
private import semmle.code.cpp.rangeanalysis.new.internal.semantic.SemanticExprSpecific
private import semmle.code.cpp.ir.IR
pragma[nomagic]
private Instruction getABoundIn(SemBound b, IRFunction func) {
getSemanticExpr(result) = b.getExpr(0) and
result.getEnclosingIRFunction() = func
}
/**
* Holds if `i <= b + delta`.
*/
pragma[inline]
private predicate boundedImplCand(Instruction i, Instruction b, int delta) {
exists(SemBound bound, IRFunction func |
semBounded(getSemanticExpr(i), bound, delta, true, _) and
b = getABoundIn(bound, func) and
i.getEnclosingIRFunction() = func
)
}
/**
* Holds if `i <= b + delta` and `delta` is the smallest integer that satisfies
* this condition.
*/
pragma[inline]
private predicate boundedImpl(Instruction i, Instruction b, int delta) {
delta = min(int cand | boundedImplCand(i, b, cand))
}
/**
* Holds if `i <= b + delta`.
*
* This predicate enforces a join-order that ensures that `i` has already been bound.
*/
bindingset[i]
pragma[inline_late]
predicate bounded1(Instruction i, Instruction b, int delta) { boundedImpl(i, b, delta) }
/**
* Holds if `i <= b + delta`.
*
* This predicate enforces a join-order that ensures that `b` has already been bound.
*/
bindingset[b]
pragma[inline_late]
predicate bounded2(Instruction i, Instruction b, int delta) { boundedImpl(i, b, delta) }
/** Holds if `i <= b + delta`. */
predicate bounded = boundedImpl/3;

View File

@@ -608,7 +608,7 @@ case @builtintype.kind of
| 47 = @std_float64 // _Float64
| 48 = @float64x // _Float64x
| 49 = @std_float128 // _Float128
// ... 50 _Float128x
| 50 = @float128x // _Float128x
| 51 = @char8_t
| 52 = @float16 // _Float16
| 53 = @complex_float16 // _Complex _Float16

File diff suppressed because it is too large Load Diff

View File

@@ -1,13 +0,0 @@
class BuiltinType extends @builtintype {
string toString() { none() }
}
predicate isFloat128xBuiltinType(BuiltinType type) {
exists(int kind | builtintypes(type, _, kind, _, _, _) | kind = 50)
}
from BuiltinType type, string name, int kind, int kind_new, int size, int sign, int alignment
where
builtintypes(type, name, kind, size, sign, alignment) and
if isFloat128xBuiltinType(type) then kind_new = 1 else kind_new = kind
select type, name, kind_new, size, sign, alignment

View File

@@ -1,3 +0,0 @@
description: Remove _Float128 type
compatibility: partial
builtintypes.rel: run builtintypes.qlo

Some files were not shown because too many files have changed in this diff Show More