Compare commits

..

1 Commits

Author SHA1 Message Date
Esben Sparre Andreasen
5f41b3afb6 make boosted version of js/code-injection
Changes generated with:

```
tb boost src/Security/CWE-094/CodeInjection.ql CodeInjectionSink
```
2022-01-31 10:57:18 +01:00
2108 changed files with 129409 additions and 364198 deletions

View File

@@ -4,10 +4,8 @@
"*/ql/lib/qlpack.yml",
"*/ql/test/qlpack.yml",
"*/ql/examples/qlpack.yml",
"*/ql/consistency-queries/qlpack.yml",
"cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/qlpack.yml",
"javascript/ql/experimental/adaptivethreatmodeling/lib/qlpack.yml",
"javascript/ql/experimental/adaptivethreatmodeling/modelbuilding/qlpack.yml",
"javascript/ql/experimental/adaptivethreatmodeling/src/qlpack.yml",
"csharp/ql/campaigns/Solorigate/lib/qlpack.yml",
"csharp/ql/campaigns/Solorigate/src/qlpack.yml",
@@ -15,6 +13,8 @@
"misc/legacy-support/*/qlpack.yml",
"misc/suite-helpers/qlpack.yml",
"ruby/extractor-pack/codeql-extractor.yml",
"ruby/ql/consistency-queries/qlpack.yml",
"ql/ql/consistency-queries/qlpack.yml",
"ql/extractor-pack/codeql-extractor.yml"
],
"versionPolicies": {

13
.gitattributes vendored
View File

@@ -50,15 +50,4 @@
*.pdb -text
java/ql/test/stubs/**/*.java linguist-generated=true
java/ql/test/experimental/stubs/**/*.java linguist-generated=true
# For some languages, upgrade script testing references really old dbscheme
# files from legacy upgrades that have CRLF line endings. Since upgrade
# resolution relies on object hashes, we must suppress line ending conversion
# for those testing dbscheme files.
*/ql/lib/upgrades/initial/*.dbscheme -text
# Generated test files - these are synced from the standard JavaScript libraries using
# `javascript/ql/experimental/adaptivethreatmodeling/test/update_endpoint_test_files.py`.
javascript/ql/experimental/adaptivethreatmodeling/test/endpoint_large_scale/autogenerated/**/*.js linguist-generated=true -merge
javascript/ql/experimental/adaptivethreatmodeling/test/endpoint_large_scale/autogenerated/**/*.ts linguist-generated=true -merge
java/ql/test/experimental/stubs/**/*.java linguist-generated=true

View File

@@ -6,11 +6,8 @@ on:
paths:
- "*/ql/src/**/*.ql"
- "*/ql/src/**/*.qll"
- "*/ql/lib/**/*.ql"
- "*/ql/lib/**/*.qll"
- "!**/experimental/**"
- "!ql/**"
- ".github/workflows/check-change-note.yml"
jobs:
check-change-note:

View File

@@ -27,11 +27,6 @@ jobs:
pull-requests: read
steps:
- name: Setup dotnet
uses: actions/setup-dotnet@v1
with:
dotnet-version: 6.0.101
- name: Checkout repository
uses: actions/checkout@v2
@@ -56,7 +51,7 @@ jobs:
# uses a compiled language
- run: |
dotnet build csharp /p:UseSharedCompilation=false
dotnet build csharp
- name: Perform CodeQL Analysis
uses: github/codeql-action/analyze@main

View File

@@ -1,43 +0,0 @@
name: "Publish framework coverage as metrics"
on:
schedule:
- cron: '5 0 * * *'
push:
branches:
- main
workflow_dispatch:
pull_request:
branches:
- main
paths:
- ".github/workflows/csv-coverage-metrics.yml"
jobs:
publish:
runs-on: ubuntu-latest
steps:
- name: Checkout repository
uses: actions/checkout@v2
- name: Setup CodeQL
uses: ./.github/actions/fetch-codeql
- name: Create empty database
run: |
DATABASE="${{ runner.temp }}/java-database"
PROJECT="${{ runner.temp }}/java-project"
mkdir -p "$PROJECT/src/tmp/empty"
echo "class Empty {}" >> "$PROJECT/src/tmp/empty/Empty.java"
codeql database create "$DATABASE" --language=java --source-root="$PROJECT" --command 'javac src/tmp/empty/Empty.java'
- name: Capture coverage information
run: |
DATABASE="${{ runner.temp }}/java-database"
codeql database analyze --format=sarif-latest --output=metrics.sarif -- "$DATABASE" ./java/ql/src/Metrics/Summaries/FrameworkCoverage.ql
- uses: actions/upload-artifact@v2
with:
name: metrics.sarif
path: metrics.sarif
retention-days: 20
- name: Upload SARIF file
uses: github/codeql-action/upload-sarif@v1
with:
sarif_file: metrics.sarif

View File

@@ -1,76 +0,0 @@
name: JS ML-powered queries tests
on:
push:
paths:
- "javascript/ql/experimental/adaptivethreatmodeling/**"
- .github/workflows/js-ml-tests.yml
branches:
- main
- "rc/*"
pull_request:
paths:
- "javascript/ql/experimental/adaptivethreatmodeling/**"
- .github/workflows/js-ml-tests.yml
defaults:
run:
working-directory: javascript/ql/experimental/adaptivethreatmodeling
jobs:
qlformat:
name: Check QL formatting
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Check QL formatting
run: |
find . "(" -name "*.ql" -or -name "*.qll" ")" -print0 | \
xargs -0 codeql query format --check-only
qlcompile:
name: Check QL compilation
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Install pack dependencies
run: |
for pack in modelbuilding src; do
codeql pack install --mode verify -- "${pack}"
done
- name: Check QL compilation
run: |
codeql query compile \
--check-only \
--ram 5120 \
--additional-packs "${{ github.workspace }}" \
--threads=0 \
-- \
lib modelbuilding src
qltest:
name: Run QL tests
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Install pack dependencies
run: codeql pack install -- test
- name: Run QL tests
run: |
codeql test run \
--threads=0 \
--ram 5120 \
--additional-packs "${{ github.workspace }}" \
-- \
test

View File

@@ -1,103 +0,0 @@
name: Models as Data - Diff
on:
workflow_dispatch:
inputs:
projects:
description: "The projects to generate models for"
required: true
default: '["netty/netty"]'
pull_request:
branches:
- main
paths:
- "java/ql/src/utils/model-generator/**/*.*"
- ".github/workflows/mad_modelDiff.yml"
permissions:
contents: read
jobs:
model-diff:
name: Model Difference
runs-on: ubuntu-latest
if: github.repository == 'github/codeql'
strategy:
matrix:
slug: ${{fromJson(github.event.inputs.projects || '["apache/commons-codec", "apache/commons-io", "apache/commons-beanutils", "apache/commons-logging", "apache/commons-fileupload", "apache/commons-lang", "apache/commons-validator", "apache/commons-csv", "apache/dubbo"]' )}}
steps:
- name: Clone github/codeql from PR
uses: actions/checkout@v2
if: github.event.pull_request
with:
path: codeql-pr
- name: Clone github/codeql from main
uses: actions/checkout@v2
with:
path: codeql-main
ref: main
- uses: ./codeql-main/.github/actions/fetch-codeql
- name: Download database
env:
SLUG: ${{ matrix.slug }}
run: |
set -x
mkdir lib-dbs
SHORTNAME=${SLUG//[^a-zA-Z0-9_]/}
projectId=`curl -s https://lgtm.com/api/v1.0/projects/g/${SLUG} | jq .id`
curl -L "https://lgtm.com/api/v1.0/snapshots/$projectId/java" -o "$SHORTNAME.zip"
unzip -q -d "${SHORTNAME}-db" "${SHORTNAME}.zip"
mkdir "lib-dbs/$SHORTNAME/"
mv "${SHORTNAME}-db/"$(ls -1 "${SHORTNAME}"-db)/* "lib-dbs/${SHORTNAME}/"
- name: Generate Models (PR and main)
run: |
set -x
mkdir tmp-models
MODELS=`pwd`/tmp-models
DATABASES=`pwd`/lib-dbs
analyzeDatabaseWithCheckout() {
QL_VARIANT=$1
DATABASE=$2
cd codeql-$QL_VARIANT
SHORTNAME=`basename $DATABASE`
python java/ql/src/utils/model-generator/GenerateFlowModel.py $DATABASE $MODELS/${SHORTNAME}.qll
mv $MODELS/${SHORTNAME}.qll $MODELS/${SHORTNAME}Generated_${QL_VARIANT}.qll
cd ..
}
for d in $DATABASES/*/ ; do
ls -1 "$d"
analyzeDatabaseWithCheckout "main" $d
if [[ "$GITHUB_EVENT_NAME" == "pull_request" ]]
then
analyzeDatabaseWithCheckout "pr" $d
fi
done
- name: Install diff2html
if: github.event.pull_request
run: |
npm install -g diff2html-cli
- name: Generate Model Diff
if: github.event.pull_request
run: |
set -x
MODELS=`pwd`/tmp-models
ls -1 tmp-models/
for m in $MODELS/*_main.qll ; do
t="${m/main/"pr"}"
basename=`basename $m`
name="diff_${basename/_main.qll/""}"
(diff -w -u $m $t | diff2html -i stdin -F $MODELS/$name.html) || true
done
- uses: actions/upload-artifact@v2
with:
name: models
path: tmp-models/*.qll
retention-days: 20
- uses: actions/upload-artifact@v2
with:
name: diffs
path: tmp-models/*.html
retention-days: 20

View File

@@ -1,62 +0,0 @@
name: Regenerate framework models
on:
workflow_dispatch:
schedule:
- cron: "30 2 * * *"
pull_request:
branches:
- main
paths:
- ".github/workflows/mad_regenerate-models.yml"
jobs:
regenerate-models:
runs-on: ubuntu-latest
strategy:
matrix:
# placeholder required for each axis, excluded below, replaced by the actual combinations (see include)
slug: ["placeholder"]
ref: ["placeholder"]
include:
- slug: "apache/commons-io"
ref: "8985de8fe74f6622a419b37a6eed0dbc484dc128"
exclude:
- slug: "placeholder"
ref: "placeholder"
steps:
- name: Clone self (github/codeql)
uses: actions/checkout@v2
- name: Setup CodeQL binaries
uses: ./.github/actions/fetch-codeql
- name: Clone repositories
uses: actions/checkout@v2
with:
path: repos/${{ matrix.ref }}
ref: ${{ matrix.ref }}
repository: ${{ matrix.slug }}
- name: Build database
env:
SLUG: ${{ matrix.slug }}
REF: ${{ matrix.ref }}
run: |
mkdir dbs
cd repos/${REF}
SHORTNAME=${SLUG//[^a-zA-Z0-9_]/}
codeql database create --language=java ../../dbs/${SHORTNAME}
- name: Regenerate models in-place
env:
SLUG: ${{ matrix.slug }}
run: |
SHORTNAME=${SLUG//[^a-zA-Z0-9_]/}
java/ql/src/utils/model-generator/RegenerateModels.py "${SLUG}" dbs/${SHORTNAME}
- name: Stage changes
run: |
find java -name "*.qll" -print0 | xargs -0 git add
git status
git diff --cached > models.patch
- uses: actions/upload-artifact@v2
with:
name: patch
path: models.patch
retention-days: 7

View File

@@ -31,13 +31,13 @@ jobs:
uses: actions/cache@v2
with:
path: ${{ runner.temp }}/query-pack.zip
key: queries-${{ hashFiles('ql/**/*.ql*') }}-${{ hashFiles('ql/**/qlpack.yml') }}-${{ hashFiles('ql/ql/src/ql.dbscheme*') }}-${{ steps.get-codeql-version.outputs.version }}
key: queries-${{ hashFiles('ql/**/*.ql*') }}-${{ hashFiles('ql/ql/src/ql.dbscheme*') }}-${{ steps.get-codeql-version.outputs.version }}
- name: Build query pack
if: steps.cache-queries.outputs.cache-hit != 'true'
run: |
cd ql/ql/src
"${CODEQL}" pack create
cd .codeql/pack/codeql/ql/0.0.0
cd .codeql/pack/codeql/ql-all/0.0.0
zip "${PACKZIP}" -r .
env:
CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
@@ -189,11 +189,4 @@ jobs:
uses: github/codeql-action/analyze@erik-krogh/ql
with:
category: "ql-for-ql-${{ matrix.folder }}"
- name: Copy sarif file to CWD
run: cp ../results/ql.sarif ./${{ matrix.folder }}.sarif
- name: Sarif as artifact
uses: actions/upload-artifact@v2
with:
name: ${{ matrix.folder }}.sarif
path: ${{ matrix.folder }}.sarif

View File

@@ -17,7 +17,7 @@ jobs:
CODEQL_THREADS: 4 # TODO: remove this once it's set by the CLI
strategy:
matrix:
repo:
repo:
- github/codeql
- github/codeql-go
runs-on: ubuntu-latest
@@ -35,7 +35,7 @@ jobs:
~/.cargo/registry
~/.cargo/git
ql/target
key: ${{ runner.os }}-qltest-cargo-${{ hashFiles('ql/**/Cargo.lock') }}
key: ${{ runner.os }}-qltest-cargo-${{ hashFiles('**/Cargo.lock') }}
- name: Build Extractor
run: cd ql; env "PATH=$PATH:`dirname ${CODEQL}`" ./create-extractor-pack.sh
env:

View File

@@ -29,24 +29,24 @@ jobs:
~/.cargo/registry
~/.cargo/git
ql/target
key: ${{ runner.os }}-qltest-cargo-${{ hashFiles('ql/**/Cargo.lock') }}
key: ${{ runner.os }}-qltest-cargo-${{ hashFiles('**/Cargo.lock') }}
- name: Build extractor
run: |
cd ql;
codeqlpath=$(dirname ${{ steps.find-codeql.outputs.codeql-path }});
env "PATH=$PATH:$codeqlpath" ./create-extractor-pack.sh
- name: Run QL tests
run: |
run: |
"${CODEQL}" test run --check-databases --check-unused-labels --check-repeated-labels --check-redefined-labels --check-use-before-definition --search-path "${{ github.workspace }}/ql/extractor-pack" --consistency-queries ql/ql/consistency-queries ql/ql/test
env:
CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
- name: Check QL formatting
run: |
run: |
find ql/ql "(" -name "*.ql" -or -name "*.qll" ")" -print0 | xargs -0 "${CODEQL}" query format --check-only
env:
CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
- name: Check QL compilation
run: |
run: |
"${CODEQL}" query compile --check-only --threads=4 --warnings=error --search-path "${{ github.workspace }}/ql/extractor-pack" "ql/ql/src" "ql/ql/examples"
env:
CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}

View File

@@ -50,7 +50,7 @@ jobs:
~/.cargo/registry
~/.cargo/git
ruby/target
key: ${{ runner.os }}-ruby-rust-cargo-${{ hashFiles('ruby/rust-toolchain.toml', 'ruby/**/Cargo.lock') }}
key: ${{ runner.os }}-rust-cargo-${{ hashFiles('**/Cargo.lock') }}
- name: Check formatting
run: cargo fmt --all -- --check
- name: Build

View File

@@ -24,54 +24,27 @@ defaults:
working-directory: ruby
jobs:
qlformat:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Check QL formatting
run: find ql "(" -name "*.ql" -or -name "*.qll" ")" -print0 | xargs -0 codeql query format --check-only
qlcompile:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Check QL compilation
run: |
codeql query compile --check-only --threads=0 --ram 5000 --warnings=error "ql/src" "ql/examples"
env:
GITHUB_TOKEN: ${{ github.token }}
qlupgrade:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- name: Check DB upgrade scripts
run: |
echo >empty.trap
codeql dataset import -S ql/lib/upgrades/initial/ruby.dbscheme testdb empty.trap
codeql dataset upgrade testdb --additional-packs ql/lib
diff -q testdb/ruby.dbscheme ql/lib/ruby.dbscheme
- name: Check DB downgrade scripts
run: |
echo >empty.trap
rm -rf testdb; codeql dataset import -S ql/lib/ruby.dbscheme testdb empty.trap
codeql resolve upgrades --format=lines --allow-downgrades --additional-packs downgrades \
--dbscheme=ql/lib/ruby.dbscheme --target-dbscheme=downgrades/initial/ruby.dbscheme |
xargs codeql execute upgrades testdb
diff -q testdb/ruby.dbscheme downgrades/initial/ruby.dbscheme
qltest:
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
slice: ["1/2", "2/2"]
steps:
- uses: actions/checkout@v2
- uses: ./.github/actions/fetch-codeql
- uses: ./ruby/actions/create-extractor-pack
- name: Run QL tests
run: |
codeql test run --threads=0 --ram 5000 --slice ${{ matrix.slice }} --search-path "${{ github.workspace }}/ruby/extractor-pack" --check-databases --check-unused-labels --check-repeated-labels --check-redefined-labels --check-use-before-definition --consistency-queries ql/consistency-queries ql/test
codeql test run --search-path "${{ github.workspace }}/ruby/extractor-pack" --check-databases --check-unused-labels --check-repeated-labels --check-redefined-labels --check-use-before-definition --consistency-queries ql/consistency-queries ql/test
env:
GITHUB_TOKEN: ${{ github.token }}
- name: Check QL formatting
run: find ql "(" -name "*.ql" -or -name "*.qll" ")" -print0 | xargs -0 codeql query format --check-only
- name: Check QL compilation
run: |
codeql query compile --check-only --threads=4 --warnings=error "ql/src" "ql/examples"
env:
GITHUB_TOKEN: ${{ github.token }}
- name: Check DB upgrade scripts
run: |
echo >empty.trap
codeql dataset import -S ql/lib/upgrades/initial/ruby.dbscheme testdb empty.trap
codeql dataset upgrade testdb --additional-packs ql/lib
diff -q testdb/ruby.dbscheme ql/lib/ruby.dbscheme

View File

@@ -1,29 +0,0 @@
name: Validate change notes
on:
push:
paths:
- "*/ql/*/change-notes/**/*"
- ".github/workflows/validate-change-notes.yml"
branches:
- main
- "rc/*"
pull_request:
paths:
- "*/ql/*/change-notes/**/*"
- ".github/workflows/validate-change-notes.yml"
jobs:
check-change-note:
runs-on: ubuntu-latest
steps:
- name: Checkout repository
uses: actions/checkout@v2
- name: Setup CodeQL
uses: ./.github/actions/fetch-codeql
- name: Fail if there are any errors with existing change notes
run: |
codeql pack release --groups cpp,csharp,java,javascript,python,ruby,-examples,-test,-experimental

View File

@@ -1,29 +0,0 @@
# See https://pre-commit.com for more information
# See https://pre-commit.com/hooks.html for more hooks
exclude: /test/.*$(?<!\.ql)(?<!\.qll)(?<!\.qlref)
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v3.2.0
hooks:
- id: trailing-whitespace
- id: end-of-file-fixer
- repo: local
hooks:
- id: codeql-format
name: Fix QL file formatting
files: \.qll?$
language: system
entry: codeql query format --in-place
- id: sync-files
name: Fix files required to be identical
language: system
entry: python3 config/sync-files.py --latest
pass_filenames: false
- id: qhelp
name: Check query help generation
files: \.qhelp$
language: system
entry: python3 misc/scripts/check-qhelp.py

View File

@@ -13,9 +13,6 @@
/python/**/experimental/**/* @github/codeql-python @xcorail
/ruby/**/experimental/**/* @github/codeql-ruby @xcorail
# ML-powered queries
/javascript/ql/experimental/adaptivethreatmodeling/ @github/codeql-ml-powered-queries-reviewers
# Notify members of codeql-go about PRs to the shared data-flow library files
/java/ql/src/semmle/code/java/dataflow/internal/DataFlowImpl.qll @github/codeql-java @github/codeql-go
/java/ql/src/semmle/code/java/dataflow/internal/DataFlowImpl2.qll @github/codeql-java @github/codeql-go
@@ -30,4 +27,4 @@
/docs/query-*-style-guide.md @github/codeql-analysis-reviewers
# QL for QL reviewers
/ql/ @github/codeql-ql-for-ql-reviewers
/ql/ @github/codeql-ql-for-ql-reviewers

View File

@@ -42,11 +42,7 @@ If you have an idea for a query that you would like to share with other CodeQL u
- The queries and libraries must be autoformatted, for example using the "Format Document" command in [CodeQL for Visual Studio Code](https://help.semmle.com/codeql/codeql-for-vscode/procedures/about-codeql-for-vscode.html).
If you prefer, you can either:
1. install the [pre-commit framework](https://pre-commit.com/) and install the configured hooks on this repo via `pre-commit install`, or
2. use this [pre-commit hook](misc/scripts/pre-commit) that automatically checks whether your files are correctly formatted.
See the [pre-commit hook installation guide](docs/pre-commit-hook-setup.md) for instructions on the two approaches.
If you prefer, you can use this [pre-commit hook](misc/scripts/pre-commit) that automatically checks whether your files are correctly formatted. See the [pre-commit hook installation guide](docs/pre-commit-hook-setup.md) for instructions on how to install the hook.
4. **Compilation**
@@ -67,6 +63,6 @@ After the experimental query is merged, we welcome pull requests to improve it.
## Using your personal data
If you contribute to this project, we will record your name and email address (as provided by you with your contributions) as part of the code repositories, which are public. We might also use this information to contact you in relation to your contributions, as well as in the normal course of software development. We also store records of CLA agreements signed in the past, but no longer require contributors to sign a CLA. Under GDPR legislation, we do this on the basis of our legitimate interest in creating the CodeQL product.
If you contribute to this project, we will record your name and email address (as provided by you with your contributions) as part of the code repositories, which are public. We might also use this information to contact you in relation to your contributions, as well as in the normal course of software development. We also store records of CLA agreements signed in the past, but no longer require contributors to sign a CLA. Under GDPR legislation, we do this on the basis of our legitimate interest in creating the CodeQL product.
Please do get in touch (privacy@github.com) if you have any questions about this or our data protection policies.

View File

@@ -426,6 +426,7 @@
"python/ql/src/Lexical/CommentedOutCodeMetricOverview.inc.qhelp"
],
"FLinesOfDuplicatedCodeCommon.inc.qhelp": [
"cpp/ql/src/Metrics/Files/FLinesOfDuplicatedCodeCommon.inc.qhelp",
"java/ql/src/Metrics/Files/FLinesOfDuplicatedCodeCommon.inc.qhelp",
"javascript/ql/src/Metrics/FLinesOfDuplicatedCodeCommon.inc.qhelp",
"python/ql/src/Metrics/FLinesOfDuplicatedCodeCommon.inc.qhelp"
@@ -464,8 +465,7 @@
],
"SensitiveDataHeuristics Python/JS": [
"javascript/ql/lib/semmle/javascript/security/internal/SensitiveDataHeuristics.qll",
"python/ql/lib/semmle/python/security/internal/SensitiveDataHeuristics.qll",
"ruby/ql/lib/codeql/ruby/security/internal/SensitiveDataHeuristics.qll"
"python/ql/lib/semmle/python/security/internal/SensitiveDataHeuristics.qll"
],
"ReDoS Util Python/JS/Ruby": [
"javascript/ql/lib/semmle/javascript/security/performance/ReDoSUtil.qll",
@@ -501,11 +501,5 @@
"javascript/ql/lib/tutorial.qll",
"python/ql/lib/tutorial.qll",
"ruby/ql/lib/tutorial.qll"
],
"AccessPathSyntax": [
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/AccessPathSyntax.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/AccessPathSyntax.qll",
"javascript/ql/lib/semmle/javascript/frameworks/data/internal/AccessPathSyntax.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/AccessPathSyntax.qll"
]
}

View File

@@ -2,7 +2,7 @@
<PropertyGroup>
<OutputType>Exe</OutputType>
<TargetFramework>net6.0</TargetFramework>
<TargetFramework>net5.0</TargetFramework>
<GenerateAssemblyInfo>false</GenerateAssemblyInfo>
<RuntimeIdentifiers>win-x64;linux-x64;osx-x64</RuntimeIdentifiers>
<Nullable>enable</Nullable>

View File

@@ -1,7 +1,7 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFramework>net6.0</TargetFramework>
<TargetFramework>net5.0</TargetFramework>
<AssemblyName>Semmle.Autobuild.Cpp</AssemblyName>
<RootNamespace>Semmle.Autobuild.Cpp</RootNamespace>
<ApplicationIcon />

View File

@@ -31,7 +31,6 @@
+ semmlecode-cpp-queries/Critical/NewArrayDeleteMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Critical/NewDeleteArrayMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Critical/NewFreeMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Likely Bugs/Memory Management/UsingExpiredStackAddress.ql: /Correctness/Common Errors
# Use of Libraries
+ semmlecode-cpp-queries/Likely Bugs/Memory Management/SuspiciousCallToMemset.ql: /Correctness/Use of Libraries
+ semmlecode-cpp-queries/Likely Bugs/Memory Management/SuspiciousSizeof.ql: /Correctness/Use of Libraries

View File

@@ -34,7 +34,6 @@
+ semmlecode-cpp-queries/Critical/NewArrayDeleteMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Critical/NewDeleteArrayMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Critical/NewFreeMismatch.ql: /Correctness/Common Errors
+ semmlecode-cpp-queries/Likely Bugs/Memory Management/UsingExpiredStackAddress.ql: /Correctness/Common Errors
# Exceptions
+ semmlecode-cpp-queries/Best Practices/Exceptions/AccidentalRethrow.ql: /Correctness/Exceptions
+ semmlecode-cpp-queries/Best Practices/Exceptions/CatchingByValue.ql: /Correctness/Exceptions

View File

@@ -1,2 +0,0 @@
description: Remove unused legacy relations
compatibility: backwards

View File

@@ -1,3 +0,0 @@
description: Add relation for tracking variables from structured binding declarations
compatibility: full
is_structured_binding.rel: delete

File diff suppressed because it is too large Load Diff

View File

@@ -1,4 +0,0 @@
name: codeql/cpp-downgrades
groups: cpp
downgrades: .
library: true

View File

@@ -1,6 +1,4 @@
name: codeql/cpp-examples
groups:
- cpp
- examples
version: 0.0.2
dependencies:
codeql/cpp-all: "*"

View File

@@ -1,22 +1,3 @@
## 0.0.10
### New Features
* Added a `isStructuredBinding` predicate to the `Variable` class which holds when the variable is declared as part of a structured binding declaration.
## 0.0.9
## 0.0.8
### Deprecated APIs
* The `codeql/cpp-upgrades` CodeQL pack has been removed. All upgrades scripts have been merged into the `codeql/cpp-all` CodeQL pack.
### Minor Analysis Improvements
* `FormatLiteral::getMaxConvertedLength` now uses range analysis to provide a
more accurate length for integers formatted with `%x`
## 0.0.7
## 0.0.6

View File

@@ -0,0 +1,4 @@
---
category: deprecated
---
* The `codeql/cpp-upgrades` CodeQL pack has been removed. All upgrades scripts have been merged into the `codeql/cpp-all` CodeQL pack.

View File

@@ -0,0 +1,5 @@
---
category: minorAnalysis
---
* `FormatLiteral::getMaxConvertedLength` now uses range analysis to provide a
more accurate length for integers formatted with `%x`

View File

@@ -1,4 +0,0 @@
---
category: minorAnalysis
---
* Many queries now support structured bindings, as structured bindings are now handled in the IR translation.

View File

@@ -1,5 +0,0 @@
## 0.0.10
### New Features
* Added a `isStructuredBinding` predicate to the `Variable` class which holds when the variable is declared as part of a structured binding declaration.

View File

@@ -1,10 +0,0 @@
## 0.0.8
### Deprecated APIs
* The `codeql/cpp-upgrades` CodeQL pack has been removed. All upgrades scripts have been merged into the `codeql/cpp-all` CodeQL pack.
### Minor Analysis Improvements
* `FormatLiteral::getMaxConvertedLength` now uses range analysis to provide a
more accurate length for integers formatted with `%x`

View File

@@ -1,2 +0,0 @@
## 0.0.9

View File

@@ -1,2 +1,2 @@
---
lastReleaseVersion: 0.0.10
lastReleaseVersion: 0.0.7

View File

@@ -1,5 +1,5 @@
name: codeql/cpp-all
version: 0.0.11-dev
version: 0.0.8-dev
groups: cpp
dbscheme: semmlecode.cpp.dbscheme
extractor: cpp

View File

@@ -169,12 +169,6 @@ class Variable extends Declaration, @variable {
variable_instantiation(underlyingElement(this), unresolveElement(v))
}
/**
* Holds if this variable is declated as part of a structured binding
* declaration. For example, `x` in `auto [x, y] = ...`.
*/
predicate isStructuredBinding() { is_structured_binding(underlyingElement(this)) }
/**
* Holds if this is a compiler-generated variable. For example, a
* [range-based for loop](http://en.cppreference.com/w/cpp/language/range-for)

View File

@@ -11,10 +11,10 @@ import cpp
*/
bindingset[input]
int parseOctal(string input) {
input.regexpMatch("0[0-7]+") and
input.charAt(0) = "0" and
result =
strictsum(int ix |
ix in [1 .. input.length()]
ix in [0 .. input.length()]
|
8.pow(input.length() - (ix + 1)) * input.charAt(ix).toInt()
)

View File

@@ -1290,7 +1290,7 @@ class DataFlowCallOption extends TDataFlowCallOption {
}
}
/** A `Content` tagged with the type of a containing object. */
/** Content tagged with the type of a containing object. */
class TypedContent extends MkTypedContent {
private Content c;
private DataFlowType t;

View File

@@ -592,14 +592,12 @@ predicate simpleLocalFlowStep(Node nodeFrom, Node nodeTo) {
* Holds if data flows from `source` to `sink` in zero or more local
* (intra-procedural) steps.
*/
pragma[inline]
predicate localFlow(Node source, Node sink) { localFlowStep*(source, sink) }
/**
* Holds if data can flow from `e1` to `e2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localExprFlow(Expr e1, Expr e2) { localFlow(exprNode(e1), exprNode(e2)) }
/**

View File

@@ -353,9 +353,9 @@ module FlowVar_internal {
// indirection.
result = def.getAUse(v)
or
exists(SsaDefinition descendantDef |
this.getASuccessorSsaVar+() = TSsaVar(descendantDef, _) and
result = descendantDef.getAUse(v)
exists(SsaDefinition descendentDef |
this.getASuccessorSsaVar+() = TSsaVar(descendentDef, _) and
result = descendentDef.getAUse(v)
)
)
or

View File

@@ -124,14 +124,12 @@ predicate localAdditionalTaintStep(DataFlow::Node nodeFrom, DataFlow::Node nodeT
* Holds if taint may propagate from `source` to `sink` in zero or more local
* (intra-procedural) steps.
*/
pragma[inline]
predicate localTaint(DataFlow::Node source, DataFlow::Node sink) { localTaintStep*(source, sink) }
/**
* Holds if taint can flow from `e1` to `e2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localExprTaint(Expr e1, Expr e2) {
localTaint(DataFlow::exprNode(e1), DataFlow::exprNode(e2))
}

View File

@@ -1,8 +1,3 @@
/**
* An IR taint tracking library that uses an IR DataFlow configuration to track
* taint from user inputs as defined by `semmle.code.cpp.security.Security`.
*/
import cpp
import semmle.code.cpp.security.Security
private import semmle.code.cpp.ir.dataflow.DataFlow
@@ -241,8 +236,8 @@ private module Cached {
// For compatibility, send flow from arguments to parameters, even for
// functions with no body.
exists(FunctionCall call, int i |
sink.asExpr() = call.getArgument(pragma[only_bind_into](i)) and
result = resolveCall(call).getParameter(pragma[only_bind_into](i))
sink.asExpr() = call.getArgument(i) and
result = resolveCall(call).getParameter(i)
)
or
// For compatibility, send flow into a `Variable` if there is flow to any

View File

@@ -1,270 +0,0 @@
/**
* This file provides a library for inter-procedural must-flow data flow analysis.
* Unlike `DataFlow.qll`, the analysis provided by this file checks whether data _must_ flow
* from a source to a _sink_.
*/
private import cpp
import semmle.code.cpp.ir.dataflow.DataFlow
private import semmle.code.cpp.ir.IR
/**
* A configuration of a data flow analysis that performs must-flow analysis. This is different
* from `DataFlow.qll` which performs may-flow analysis (i.e., it finds paths where the source _may_
* flow to the sink).
*
* Like in `DataFlow.qll`, each use of the `MustFlow.qll` library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string and override `isSource`, `isSink` (and
* `isAdditionalFlowStep` if additional steps are required).
*/
abstract class MustFlowConfiguration extends string {
bindingset[this]
MustFlowConfiguration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
abstract predicate isSource(DataFlow::Node source);
/**
* Holds if `sink` is a relevant data flow sink.
*/
abstract predicate isSink(DataFlow::Node sink);
/**
* Holds if the additional flow step from `node1` to `node2` must be taken
* into account in the analysis.
*/
predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
/**
* Holds if data must flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
final predicate hasFlowPath(MustFlowPathNode source, MustFlowPathSink sink) {
this.isSource(source.getNode()) and
source.getASuccessor+() = sink
}
}
/** Holds if `node` flows from a source. */
pragma[nomagic]
private predicate flowsFromSource(DataFlow::Node node, MustFlowConfiguration config) {
config.isSource(node)
or
exists(DataFlow::Node mid |
step(mid, node, config) and
flowsFromSource(mid, pragma[only_bind_into](config))
)
}
/** Holds if `node` flows to a sink. */
pragma[nomagic]
private predicate flowsToSink(DataFlow::Node node, MustFlowConfiguration config) {
flowsFromSource(node, pragma[only_bind_into](config)) and
(
config.isSink(node)
or
exists(DataFlow::Node mid |
step(node, mid, config) and
flowsToSink(mid, pragma[only_bind_into](config))
)
)
}
cached
private module Cached {
/** Holds if `p` is the `n`'th parameter of the non-virtual function `f`. */
private predicate parameterOf(Parameter p, Function f, int n) {
not f.isVirtual() and f.getParameter(n) = p
}
/**
* Holds if `instr` is the `n`'th argument to a call to the non-virtual function `f`, and
* `init` is the corresponding initialization instruction that receives the value of `instr` in `f`.
*/
private predicate flowIntoParameter(
Function f, int n, CallInstruction call, Instruction instr, InitializeParameterInstruction init
) {
not f.isVirtual() and
call.getPositionalArgument(n) = instr and
f = call.getStaticCallTarget() and
getEnclosingNonVirtualFunctionInitializeParameter(init, f) and
init.getParameter().getIndex() = pragma[only_bind_into](pragma[only_bind_out](n))
}
/**
* Holds if `instr` is an argument to a call to the function `f`, and `init` is the
* corresponding initialization instruction that receives the value of `instr` in `f`.
*/
pragma[noinline]
private predicate getPositionalArgumentInitParam(
CallInstruction call, Instruction instr, InitializeParameterInstruction init, Function f
) {
exists(int n |
parameterOf(_, f, n) and
flowIntoParameter(f, pragma[only_bind_into](pragma[only_bind_out](n)), call, instr, init)
)
}
/**
* Holds if `instr` is the qualifier to a call to the non-virtual function `f`, and
* `init` is the corresponding initialization instruction that receives the value of
* `instr` in `f`.
*/
pragma[noinline]
private predicate getThisArgumentInitParam(
CallInstruction call, Instruction instr, InitializeParameterInstruction init, Function f
) {
not f.isVirtual() and
call.getStaticCallTarget() = f and
getEnclosingNonVirtualFunctionInitializeParameter(init, f) and
call.getThisArgument() = instr and
init.getIRVariable() instanceof IRThisVariable
}
/** Holds if `f` is the enclosing non-virtual function of `init`. */
private predicate getEnclosingNonVirtualFunctionInitializeParameter(
InitializeParameterInstruction init, Function f
) {
not f.isVirtual() and
init.getEnclosingFunction() = f
}
/** Holds if `f` is the enclosing non-virtual function of `init`. */
private predicate getEnclosingNonVirtualFunctionInitializeIndirection(
InitializeIndirectionInstruction init, Function f
) {
not f.isVirtual() and
init.getEnclosingFunction() = f
}
/**
* Holds if `instr` is an argument (or argument indirection) to a call, and
* `succ` is the corresponding initialization instruction in the call target.
*/
private predicate flowThroughCallable(Instruction argument, Instruction parameter) {
// Flow from an argument to a parameter
exists(CallInstruction call, InitializeParameterInstruction init | init = parameter |
getPositionalArgumentInitParam(call, argument, init, call.getStaticCallTarget())
or
getThisArgumentInitParam(call, argument, init, call.getStaticCallTarget())
)
or
// Flow from argument indirection to parameter indirection
exists(
CallInstruction call, ReadSideEffectInstruction read, InitializeIndirectionInstruction init
|
init = parameter and
read.getPrimaryInstruction() = call and
getEnclosingNonVirtualFunctionInitializeIndirection(init, call.getStaticCallTarget())
|
exists(int n |
read.getSideEffectOperand().getAnyDef() = argument and
read.getIndex() = pragma[only_bind_into](n) and
init.getParameter().getIndex() = pragma[only_bind_into](n)
)
or
call.getThisArgument() = argument and
init.getIRVariable() instanceof IRThisVariable
)
}
private predicate instructionToOperandStep(Instruction instr, Operand operand) {
operand.getDef() = instr
}
/**
* Holds if data flows from `operand` to `instr`.
*
* This predicate ignores flow through `PhiInstruction`s to create a 'must flow' relation.
*/
private predicate operandToInstructionStep(Operand operand, Instruction instr) {
instr.(CopyInstruction).getSourceValueOperand() = operand
or
instr.(ConvertInstruction).getUnaryOperand() = operand
or
instr.(CheckedConvertOrNullInstruction).getUnaryOperand() = operand
or
instr.(InheritanceConversionInstruction).getUnaryOperand() = operand
or
instr.(ChiInstruction).getTotalOperand() = operand
}
cached
predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) {
instructionToOperandStep(nodeFrom.asInstruction(), nodeTo.asOperand())
or
flowThroughCallable(nodeFrom.asInstruction(), nodeTo.asInstruction())
or
operandToInstructionStep(nodeFrom.asOperand(), nodeTo.asInstruction())
}
}
/** Holds if `nodeFrom` flows to `nodeTo`. */
private predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo, MustFlowConfiguration config) {
exists(config) and
Cached::step(nodeFrom, nodeTo)
or
config.isAdditionalFlowStep(nodeFrom, nodeTo)
}
private newtype TLocalPathNode =
MkLocalPathNode(DataFlow::Node n, MustFlowConfiguration config) {
flowsToSink(n, config) and
(
config.isSource(n)
or
exists(MustFlowPathNode mid | step(mid.getNode(), n, config))
)
}
/** A `Node` that is in a path from a source to a sink. */
class MustFlowPathNode extends TLocalPathNode {
DataFlow::Node n;
MustFlowPathNode() { this = MkLocalPathNode(n, _) }
/** Gets the underlying node. */
DataFlow::Node getNode() { result = n }
/** Gets a textual representation of this node. */
string toString() { result = n.toString() }
/** Gets the location of this element. */
Location getLocation() { result = n.getLocation() }
/** Gets a successor node, if any. */
MustFlowPathNode getASuccessor() {
step(this.getNode(), result.getNode(), this.getConfiguration())
}
/** Gets the associated configuration. */
MustFlowConfiguration getConfiguration() { this = MkLocalPathNode(_, result) }
}
private class MustFlowPathSink extends MustFlowPathNode {
MustFlowPathSink() { this.getConfiguration().isSink(this.getNode()) }
}
/**
* Provides the query predicates needed to include a graph in a path-problem query.
*/
module PathGraph {
private predicate reach(MustFlowPathNode n) {
n instanceof MustFlowPathSink or reach(n.getASuccessor())
}
/** Holds if `(a,b)` is an edge in the graph of data flow path explanations. */
query predicate edges(MustFlowPathNode a, MustFlowPathNode b) {
a.getASuccessor() = b and reach(b)
}
/** Holds if `n` is a node in the graph of data flow path explanations. */
query predicate nodes(MustFlowPathNode n, string key, string val) {
reach(n) and key = "semmle.label" and val = n.toString()
}
}

View File

@@ -1290,7 +1290,7 @@ class DataFlowCallOption extends TDataFlowCallOption {
}
}
/** A `Content` tagged with the type of a containing object. */
/** Content tagged with the type of a containing object. */
class TypedContent extends MkTypedContent {
private Content c;
private DataFlowType t;

View File

@@ -1032,14 +1032,12 @@ SideEffectInstruction getSideEffectFor(CallInstruction call, int argument) {
* Holds if data flows from `source` to `sink` in zero or more local
* (intra-procedural) steps.
*/
pragma[inline]
predicate localFlow(Node source, Node sink) { localFlowStep*(source, sink) }
/**
* Holds if data can flow from `i1` to `i2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localInstructionFlow(Instruction e1, Instruction e2) {
localFlow(instructionNode(e1), instructionNode(e2))
}
@@ -1048,7 +1046,6 @@ predicate localInstructionFlow(Instruction e1, Instruction e2) {
* Holds if data can flow from `e1` to `e2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localExprFlow(Expr e1, Expr e2) { localFlow(exprNode(e1), exprNode(e2)) }
private newtype TContent =

View File

@@ -121,14 +121,12 @@ private predicate operandToInstructionTaintStep(Operand opFrom, Instruction inst
* Holds if taint may propagate from `source` to `sink` in zero or more local
* (intra-procedural) steps.
*/
pragma[inline]
predicate localTaint(DataFlow::Node source, DataFlow::Node sink) { localTaintStep*(source, sink) }
/**
* Holds if taint can flow from `i1` to `i2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localInstructionTaint(Instruction i1, Instruction i2) {
localTaint(DataFlow::instructionNode(i1), DataFlow::instructionNode(i2))
}
@@ -137,7 +135,6 @@ predicate localInstructionTaint(Instruction i1, Instruction i2) {
* Holds if taint can flow from `e1` to `e2` in zero or more
* local (intra-procedural) steps.
*/
pragma[inline]
predicate localExprTaint(Expr e1, Expr e2) {
localTaint(DataFlow::exprNode(e1), DataFlow::exprNode(e2))
}

View File

@@ -200,7 +200,7 @@ class IRBlock extends IRBlockBase {
* post-dominate block `B`, but block `A` does post-dominate an immediate successor of block `B`.
*/
pragma[noinline]
final IRBlock postDominanceFrontier() {
final IRBlock postPominanceFrontier() {
this.postDominates(result.getASuccessor()) and
not this.strictlyPostDominates(result)
}

View File

@@ -106,12 +106,6 @@ private predicate filteredNumberableInstruction(Instruction instr) {
or
instr instanceof FieldAddressInstruction and
count(instr.(FieldAddressInstruction).getField()) != 1
or
instr instanceof InheritanceConversionInstruction and
(
count(instr.(InheritanceConversionInstruction).getBaseClass()) != 1 or
count(instr.(InheritanceConversionInstruction).getDerivedClass()) != 1
)
}
private predicate variableAddressValueNumber(
@@ -121,7 +115,8 @@ private predicate variableAddressValueNumber(
// The underlying AST element is used as value-numbering key instead of the
// `IRVariable` to work around a problem where a variable or expression with
// multiple types gives rise to multiple `IRVariable`s.
unique( | | instr.getIRVariable().getAST()) = ast
instr.getIRVariable().getAST() = ast and
strictcount(instr.getIRVariable().getAST()) = 1
}
private predicate initializeParameterValueNumber(
@@ -138,7 +133,8 @@ private predicate constantValueNumber(
ConstantInstruction instr, IRFunction irFunc, IRType type, string value
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getResultIRType()) = type and
strictcount(instr.getResultIRType()) = 1 and
instr.getResultIRType() = type and
instr.getValue() = value
}
@@ -155,7 +151,8 @@ private predicate fieldAddressValueNumber(
TValueNumber objectAddress
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getField()) = field and
instr.getField() = field and
strictcount(instr.getField()) = 1 and
tvalueNumber(instr.getObjectAddress()) = objectAddress
}
@@ -198,9 +195,9 @@ private predicate inheritanceConversionValueNumber(
) {
instr.getEnclosingIRFunction() = irFunc and
instr.getOpcode() = opcode and
tvalueNumber(instr.getUnary()) = operand and
unique( | | instr.getBaseClass()) = baseClass and
unique( | | instr.getDerivedClass()) = derivedClass
instr.getBaseClass() = baseClass and
instr.getDerivedClass() = derivedClass and
tvalueNumber(instr.getUnary()) = operand
}
private predicate loadTotalOverlapValueNumber(

View File

@@ -200,7 +200,7 @@ class IRBlock extends IRBlockBase {
* post-dominate block `B`, but block `A` does post-dominate an immediate successor of block `B`.
*/
pragma[noinline]
final IRBlock postDominanceFrontier() {
final IRBlock postPominanceFrontier() {
this.postDominates(result.getASuccessor()) and
not this.strictlyPostDominates(result)
}

View File

@@ -106,12 +106,6 @@ private predicate filteredNumberableInstruction(Instruction instr) {
or
instr instanceof FieldAddressInstruction and
count(instr.(FieldAddressInstruction).getField()) != 1
or
instr instanceof InheritanceConversionInstruction and
(
count(instr.(InheritanceConversionInstruction).getBaseClass()) != 1 or
count(instr.(InheritanceConversionInstruction).getDerivedClass()) != 1
)
}
private predicate variableAddressValueNumber(
@@ -121,7 +115,8 @@ private predicate variableAddressValueNumber(
// The underlying AST element is used as value-numbering key instead of the
// `IRVariable` to work around a problem where a variable or expression with
// multiple types gives rise to multiple `IRVariable`s.
unique( | | instr.getIRVariable().getAST()) = ast
instr.getIRVariable().getAST() = ast and
strictcount(instr.getIRVariable().getAST()) = 1
}
private predicate initializeParameterValueNumber(
@@ -138,7 +133,8 @@ private predicate constantValueNumber(
ConstantInstruction instr, IRFunction irFunc, IRType type, string value
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getResultIRType()) = type and
strictcount(instr.getResultIRType()) = 1 and
instr.getResultIRType() = type and
instr.getValue() = value
}
@@ -155,7 +151,8 @@ private predicate fieldAddressValueNumber(
TValueNumber objectAddress
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getField()) = field and
instr.getField() = field and
strictcount(instr.getField()) = 1 and
tvalueNumber(instr.getObjectAddress()) = objectAddress
}
@@ -198,9 +195,9 @@ private predicate inheritanceConversionValueNumber(
) {
instr.getEnclosingIRFunction() = irFunc and
instr.getOpcode() = opcode and
tvalueNumber(instr.getUnary()) = operand and
unique( | | instr.getBaseClass()) = baseClass and
unique( | | instr.getDerivedClass()) = derivedClass
instr.getBaseClass() = baseClass and
instr.getDerivedClass() = derivedClass and
tvalueNumber(instr.getUnary()) = operand
}
private predicate loadTotalOverlapValueNumber(

View File

@@ -71,8 +71,7 @@ newtype TInstructionTag =
AsmTag() or
AsmInputTag(int elementIndex) { exists(AsmStmt asm | exists(asm.getChild(elementIndex))) } or
ThisAddressTag() or
ThisLoadTag() or
StructuredBindingAccessTag()
ThisLoadTag()
class InstructionTag extends TInstructionTag {
final string toString() { result = "Tag" }
@@ -222,6 +221,4 @@ string getInstructionTagId(TInstructionTag tag) {
tag = ThisAddressTag() and result = "ThisAddress"
or
tag = ThisLoadTag() and result = "ThisLoad"
or
tag = StructuredBindingAccessTag() and result = "StructuredBindingAccess"
}

View File

@@ -111,45 +111,6 @@ private predicate hasDefaultSideEffect(Call call, ParameterIndex i, boolean buff
)
}
/**
* A `Call` or `NewOrNewArrayExpr`.
*
* Both kinds of expression invoke a function as part of their evaluation. This class provides a
* way to treat both kinds of function similarly, and to get the invoked `Function`.
*/
class CallOrAllocationExpr extends Expr {
CallOrAllocationExpr() {
this instanceof Call
or
this instanceof NewOrNewArrayExpr
}
/** Gets the `Function` invoked by this expression, if known. */
final Function getTarget() {
result = this.(Call).getTarget()
or
result = this.(NewOrNewArrayExpr).getAllocator()
}
}
/**
* Returns the side effect opcode, if any, that represents any side effects not specifically modeled
* by an argument side effect.
*/
Opcode getCallSideEffectOpcode(CallOrAllocationExpr expr) {
not exists(expr.getTarget().(SideEffectFunction)) and result instanceof Opcode::CallSideEffect
or
exists(SideEffectFunction sideEffectFunction |
sideEffectFunction = expr.getTarget() and
if not sideEffectFunction.hasOnlySpecificWriteSideEffects()
then result instanceof Opcode::CallSideEffect
else (
not sideEffectFunction.hasOnlySpecificReadSideEffects() and
result instanceof Opcode::CallReadSideEffect
)
)
}
/**
* Returns a side effect opcode for parameter index `i` of the specified call.
*

View File

@@ -49,6 +49,19 @@ abstract class TranslatedCall extends TranslatedExpr {
tag = CallTag() and
opcode instanceof Opcode::Call and
resultType = getTypeForPRValue(getCallResultType())
or
hasSideEffect() and
tag = CallSideEffectTag() and
(
if hasWriteSideEffect()
then (
opcode instanceof Opcode::CallSideEffect and
resultType = getUnknownType()
) else (
opcode instanceof Opcode::CallReadSideEffect and
resultType = getVoidType()
)
)
}
override Instruction getChildSuccessor(TranslatedElement child) {
@@ -71,8 +84,25 @@ abstract class TranslatedCall extends TranslatedExpr {
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
kind instanceof GotoEdge and
tag = CallTag() and
result = getSideEffects().getFirstInstruction()
(
(
tag = CallTag() and
if hasSideEffect()
then result = getInstruction(CallSideEffectTag())
else
if hasPreciseSideEffect()
then result = getSideEffects().getFirstInstruction()
else result = getParent().getChildSuccessor(this)
)
or
(
hasSideEffect() and
tag = CallSideEffectTag() and
if hasPreciseSideEffect()
then result = getSideEffects().getFirstInstruction()
else result = getParent().getChildSuccessor(this)
)
)
}
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
@@ -91,6 +121,15 @@ abstract class TranslatedCall extends TranslatedExpr {
)
}
final override CppType getInstructionMemoryOperandType(
InstructionTag tag, TypedOperandTag operandTag
) {
tag = CallSideEffectTag() and
hasSideEffect() and
operandTag instanceof SideEffectOperandTag and
result = getUnknownType()
}
final override Instruction getResult() { result = getInstruction(CallTag()) }
/**
@@ -161,31 +200,40 @@ abstract class TranslatedCall extends TranslatedExpr {
*/
abstract predicate hasArguments();
predicate hasReadSideEffect() { any() }
predicate hasWriteSideEffect() { any() }
private predicate hasSideEffect() { hasReadSideEffect() or hasWriteSideEffect() }
override Instruction getPrimaryInstructionForSideEffect(InstructionTag tag) {
hasSideEffect() and
tag = CallSideEffectTag() and
result = getResult()
}
predicate hasPreciseSideEffect() { exists(getSideEffects()) }
final TranslatedSideEffects getSideEffects() { result.getExpr() = expr }
}
/**
* The IR translation of the side effects of the parent `TranslatedElement`.
*
* This object does not itself generate the side effect instructions. Instead, its children provide
* the actual side effects, with this object acting as a placeholder so the parent only needs to
* insert this one element at the point where all the side effects are supposed to occur.
*/
abstract class TranslatedSideEffects extends TranslatedElement {
/** Gets the expression whose side effects are being modeled. */
abstract Expr getExpr();
final override Locatable getAST() { result = getExpr() }
final override Function getFunction() { result = getExpr().getEnclosingFunction() }
final override TranslatedElement getChild(int i) {
override TranslatedElement getChild(int i) {
result =
rank[i + 1](TranslatedSideEffect tse, int group, int indexInGroup |
tse.getPrimaryExpr() = getExpr() and
tse.sortOrder(group, indexInGroup)
rank[i + 1](TranslatedSideEffect tse, int isWrite, int index |
(
tse.getCall() = getExpr() and
tse.getArgumentIndex() = index and
if tse.isWrite() then isWrite = 1 else isWrite = 0
)
|
tse order by group, indexInGroup
tse order by isWrite, index
)
}
@@ -198,21 +246,12 @@ abstract class TranslatedSideEffects extends TranslatedElement {
)
}
final override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType type) {
none()
/**
* Gets the `TranslatedFunction` containing this expression.
*/
final TranslatedFunction getEnclosingFunction() {
result = getTranslatedFunction(getExpr().getEnclosingFunction())
}
final override Instruction getFirstInstruction() {
result = getChild(0).getFirstInstruction()
or
// Some functions, like `std::move()`, have no side effects whatsoever.
not exists(getChild(0)) and result = getParent().getChildSuccessor(this)
}
final override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) { none() }
/** Gets the primary instruction to be associated with each side effect instruction. */
abstract Instruction getPrimaryInstruction();
}
/**
@@ -286,6 +325,14 @@ class TranslatedFunctionCall extends TranslatedCallExpr, TranslatedDirectCall {
tag = CallTargetTag() and result = expr.getTarget()
}
override predicate hasReadSideEffect() {
not expr.getTarget().(SideEffectFunction).hasOnlySpecificReadSideEffects()
}
override predicate hasWriteSideEffect() {
not expr.getTarget().(SideEffectFunction).hasOnlySpecificWriteSideEffects()
}
override Instruction getQualifierResult() {
hasQualifier() and
result = getQualifier().getResult()
@@ -316,116 +363,209 @@ class TranslatedStructorCall extends TranslatedFunctionCall {
override predicate hasQualifier() { any() }
}
/**
* The IR translation of the side effects of a function call, including the implicit allocator
* call in a `new` or `new[]` expression.
*/
class TranslatedAllocationSideEffects extends TranslatedSideEffects,
TTranslatedAllocationSideEffects {
AllocationExpr expr;
TranslatedAllocationSideEffects() { this = TTranslatedAllocationSideEffects(expr) }
final override AllocationExpr getExpr() { result = expr }
override string toString() { result = "(allocation side effects for " + expr.toString() + ")" }
override Instruction getFirstInstruction() { result = getInstruction(OnlyInstructionTag()) }
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType type) {
opcode instanceof Opcode::InitializeDynamicAllocation and
tag = OnlyInstructionTag() and
type = getUnknownType()
}
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
tag = OnlyInstructionTag() and
kind = EdgeKind::gotoEdge() and
if exists(getChild(0))
then result = getChild(0).getFirstInstruction()
else result = getParent().getChildSuccessor(this)
}
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag = OnlyInstructionTag() and
operandTag = addressOperand() and
result = getPrimaryInstructionForSideEffect(OnlyInstructionTag())
}
override Instruction getPrimaryInstructionForSideEffect(InstructionTag tag) {
tag = OnlyInstructionTag() and
if expr instanceof NewOrNewArrayExpr
then result = getTranslatedAllocatorCall(expr).getInstruction(CallTag())
else result = getTranslatedCallInstruction(expr)
}
}
class TranslatedCallSideEffects extends TranslatedSideEffects, TTranslatedCallSideEffects {
Expr expr;
Call expr;
TranslatedCallSideEffects() { this = TTranslatedCallSideEffects(expr) }
final override string toString() { result = "(side effects for " + expr.toString() + ")" }
override string toString() { result = "(side effects for " + expr.toString() + ")" }
final override Expr getExpr() { result = expr }
override Call getExpr() { result = expr }
final override Instruction getPrimaryInstruction() {
expr instanceof Call and result = getTranslatedCallInstruction(expr)
or
expr instanceof NewOrNewArrayExpr and
result = getTranslatedAllocatorCall(expr).getInstruction(CallTag())
}
}
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType type) { none() }
/** Returns the sort group index for argument read side effects. */
private int argumentReadGroup() { result = 1 }
override Instruction getFirstInstruction() { result = getChild(0).getFirstInstruction() }
/** Returns the sort group index for conservative call side effects. */
private int callSideEffectGroup() {
result = 0 // Make this group first for now to preserve the existing ordering
}
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) { none() }
/** Returns the sort group index for argument write side effects. */
private int argumentWriteGroup() { result = 2 }
/** Returns the sort group index for dynamic allocation side effects. */
private int initializeAllocationGroup() { result = 3 }
/**
* The IR translation of a single side effect of a call.
*/
abstract class TranslatedSideEffect extends TranslatedElement {
final override TranslatedElement getChild(int n) { none() }
final override Instruction getChildSuccessor(TranslatedElement child) { none() }
final override Instruction getFirstInstruction() { result = getInstruction(OnlyInstructionTag()) }
final override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType type) {
override Instruction getPrimaryInstructionForSideEffect(InstructionTag tag) {
tag = OnlyInstructionTag() and
sideEffectInstruction(opcode, type)
result = getTranslatedCallInstruction(expr)
}
}
class TranslatedStructorCallSideEffects extends TranslatedCallSideEffects {
TranslatedStructorCallSideEffects() {
getParent().(TranslatedStructorCall).hasQualifier() and
getASideEffectOpcode(expr, -1) instanceof WriteSideEffectOpcode
}
final override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType t) {
tag instanceof OnlyInstructionTag and
t = getTypeForPRValue(expr.getTarget().getDeclaringType()) and
opcode = getASideEffectOpcode(expr, -1).(WriteSideEffectOpcode)
}
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
(
if exists(getChild(0))
then result = getChild(0).getFirstInstruction()
else result = getParent().getChildSuccessor(this)
) and
tag = OnlyInstructionTag() and
kind instanceof GotoEdge
}
override Instruction getFirstInstruction() { result = getInstruction(OnlyInstructionTag()) }
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag instanceof OnlyInstructionTag and
operandTag instanceof AddressOperandTag and
result = getParent().(TranslatedStructorCall).getQualifierResult()
}
final override int getInstructionIndex(InstructionTag tag) {
tag = OnlyInstructionTag() and
result = -1
}
}
class TranslatedSideEffect extends TranslatedElement, TTranslatedArgumentSideEffect {
Call call;
Expr arg;
int index;
SideEffectOpcode sideEffectOpcode;
TranslatedSideEffect() {
this = TTranslatedArgumentSideEffect(call, arg, index, sideEffectOpcode)
}
override Locatable getAST() { result = arg }
Expr getExpr() { result = arg }
Call getCall() { result = call }
int getArgumentIndex() { result = index }
predicate isWrite() { sideEffectOpcode instanceof WriteSideEffectOpcode }
override string toString() {
isWrite() and
result = "(write side effect for " + arg.toString() + ")"
or
not isWrite() and
result = "(read side effect for " + arg.toString() + ")"
}
override TranslatedElement getChild(int n) { none() }
override Instruction getChildSuccessor(TranslatedElement child) { none() }
override Instruction getFirstInstruction() { result = getInstruction(OnlyInstructionTag()) }
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType type) {
tag = OnlyInstructionTag() and
opcode = sideEffectOpcode and
(
isWrite() and
(
opcode instanceof BufferAccessOpcode and
type = getUnknownType()
or
not opcode instanceof BufferAccessOpcode and
exists(Type baseType | baseType = arg.getUnspecifiedType().(DerivedType).getBaseType() |
if baseType instanceof VoidType
then type = getUnknownType()
else type = getTypeForPRValueOrUnknown(baseType)
)
or
index = -1 and
not arg.getUnspecifiedType() instanceof DerivedType and
type = getTypeForPRValueOrUnknown(arg.getUnspecifiedType())
)
or
not isWrite() and
type = getVoidType()
)
}
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
result = getParent().getChildSuccessor(this) and
tag = OnlyInstructionTag() and
kind instanceof GotoEdge
}
final override Function getFunction() { result = getParent().getFunction() }
final override Instruction getPrimaryInstructionForSideEffect(InstructionTag tag) {
tag = OnlyInstructionTag() and
result = getParent().(TranslatedSideEffects).getPrimaryInstruction()
}
/**
* Gets the expression that caused this side effect.
*
* All side effects with the same `getPrimaryExpr()` will appear in the same contiguous sequence
* in the IR.
*/
abstract Expr getPrimaryExpr();
/**
* Gets the order in which this side effect should be sorted with respect to other side effects
* for the same expression.
*
* Side effects are sorted first by `group`, and then by `indexInGroup`.
*/
abstract predicate sortOrder(int group, int indexInGroup);
/**
* Gets the opcode and result type for the side effect instruction.
*/
abstract predicate sideEffectInstruction(Opcode opcode, CppType type);
}
/**
* The IR translation of a single argument side effect for a call.
*/
abstract class TranslatedArgumentSideEffect extends TranslatedSideEffect {
Call call;
int index;
SideEffectOpcode sideEffectOpcode;
// All subclass charpreds must bind the `index` field.
bindingset[index]
TranslatedArgumentSideEffect() { any() }
override string toString() {
isWrite() and
result = "(write side effect for " + getArgString() + ")"
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag instanceof OnlyInstructionTag and
operandTag instanceof AddressOperandTag and
result = getTranslatedExpr(arg).getResult()
or
not isWrite() and
result = "(read side effect for " + getArgString() + ")"
tag instanceof OnlyInstructionTag and
operandTag instanceof BufferSizeOperandTag and
result =
getTranslatedExpr(call.getArgument(call.getTarget()
.(SideEffectFunction)
.getParameterSizeIndex(index)).getFullyConverted()).getResult()
}
override Call getPrimaryExpr() { result = call }
override CppType getInstructionMemoryOperandType(InstructionTag tag, TypedOperandTag operandTag) {
not isWrite() and
if sideEffectOpcode instanceof BufferAccessOpcode
then
result = getUnknownType() and
tag instanceof OnlyInstructionTag and
operandTag instanceof SideEffectOperandTag
else
exists(Type operandType |
tag instanceof OnlyInstructionTag and
operandType = arg.getType().getUnspecifiedType().(DerivedType).getBaseType() and
operandTag instanceof SideEffectOperandTag
or
tag instanceof OnlyInstructionTag and
operandType = arg.getType().getUnspecifiedType() and
not operandType instanceof DerivedType and
operandTag instanceof SideEffectOperandTag
|
// If the type we select is an incomplete type (e.g. a forward-declared `struct`), there will
// not be a `CppType` that represents that type. In that case, fall back to `UnknownCppType`.
result = getTypeForPRValueOrUnknown(operandType)
)
}
override predicate sortOrder(int group, int indexInGroup) {
indexInGroup = index and
if isWrite() then group = argumentWriteGroup() else group = argumentReadGroup()
override Instruction getPrimaryInstructionForSideEffect(InstructionTag tag) {
tag = OnlyInstructionTag() and
result = getTranslatedCallInstruction(call)
}
final override int getInstructionIndex(InstructionTag tag) {
@@ -437,199 +577,11 @@ abstract class TranslatedArgumentSideEffect extends TranslatedSideEffect {
* Gets the `TranslatedFunction` containing this expression.
*/
final TranslatedFunction getEnclosingFunction() {
result = getTranslatedFunction(call.getEnclosingFunction())
result = getTranslatedFunction(arg.getEnclosingFunction())
}
final override predicate sideEffectInstruction(Opcode opcode, CppType type) {
opcode = sideEffectOpcode and
(
isWrite() and
(
opcode instanceof BufferAccessOpcode and
type = getUnknownType()
or
not opcode instanceof BufferAccessOpcode and
exists(Type indirectionType | indirectionType = getIndirectionType() |
if indirectionType instanceof VoidType
then type = getUnknownType()
else type = getTypeForPRValueOrUnknown(indirectionType)
)
)
or
not isWrite() and
type = getVoidType()
)
}
final override CppType getInstructionMemoryOperandType(
InstructionTag tag, TypedOperandTag operandTag
) {
not isWrite() and
if sideEffectOpcode instanceof BufferAccessOpcode
then
result = getUnknownType() and
tag instanceof OnlyInstructionTag and
operandTag instanceof SideEffectOperandTag
else
exists(Type operandType |
tag instanceof OnlyInstructionTag and
operandType = getIndirectionType() and
operandTag instanceof SideEffectOperandTag
|
// If the type we select is an incomplete type (e.g. a forward-declared `struct`), there will
// not be a `CppType` that represents that type. In that case, fall back to `UnknownCppType`.
result = getTypeForPRValueOrUnknown(operandType)
)
}
final override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag instanceof OnlyInstructionTag and
operandTag instanceof AddressOperandTag and
result = getArgInstruction()
or
tag instanceof OnlyInstructionTag and
operandTag instanceof BufferSizeOperandTag and
result =
getTranslatedExpr(call.getArgument(call.getTarget()
.(SideEffectFunction)
.getParameterSizeIndex(index)).getFullyConverted()).getResult()
}
/** Holds if this side effect is a write side effect, rather than a read side effect. */
final predicate isWrite() { sideEffectOpcode instanceof WriteSideEffectOpcode }
/** Gets a text representation of the argument. */
abstract string getArgString();
/** Gets the `Instruction` whose result is the value of the argument. */
abstract Instruction getArgInstruction();
/** Gets the type pointed to by the argument. */
abstract Type getIndirectionType();
}
/**
* The IR translation of an argument side effect where the argument has an `Expr` object in the AST.
*
* This generally applies to all positional arguments, as well as qualifier (`this`) arguments for
* calls other than constructor calls.
*/
class TranslatedArgumentExprSideEffect extends TranslatedArgumentSideEffect,
TTranslatedArgumentExprSideEffect {
Expr arg;
TranslatedArgumentExprSideEffect() {
this = TTranslatedArgumentExprSideEffect(call, arg, index, sideEffectOpcode)
}
final override Locatable getAST() { result = arg }
final override Type getIndirectionType() {
result = arg.getUnspecifiedType().(DerivedType).getBaseType()
or
// Sometimes the qualifier type gets the type of the class itself, rather than a pointer to the
// class.
index = -1 and
not arg.getUnspecifiedType() instanceof DerivedType and
result = arg.getUnspecifiedType()
}
final override string getArgString() { result = arg.toString() }
final override Instruction getArgInstruction() { result = getTranslatedExpr(arg).getResult() }
}
/**
* The IR translation of an argument side effect for `*this` on a call, where there is no `Expr`
* object that represents the `this` argument.
*
* The applies only to constructor calls, as the AST has explioit qualifier `Expr`s for all other
* calls to non-static member functions.
*/
class TranslatedStructorQualifierSideEffect extends TranslatedArgumentSideEffect,
TTranslatedStructorQualifierSideEffect {
TranslatedStructorQualifierSideEffect() {
this = TTranslatedStructorQualifierSideEffect(call, sideEffectOpcode) and
index = -1
}
final override Locatable getAST() { result = call }
final override Type getIndirectionType() { result = call.getTarget().getDeclaringType() }
final override string getArgString() { result = "this" }
final override Instruction getArgInstruction() {
exists(TranslatedStructorCall structorCall |
structorCall.getExpr() = call and
result = structorCall.getQualifierResult()
)
}
}
/** The IR translation of the non-argument-specific side effect of a call. */
class TranslatedCallSideEffect extends TranslatedSideEffect, TTranslatedCallSideEffect {
Expr expr;
SideEffectOpcode sideEffectOpcode;
TranslatedCallSideEffect() { this = TTranslatedCallSideEffect(expr, sideEffectOpcode) }
override Locatable getAST() { result = expr }
override Expr getPrimaryExpr() { result = expr }
override predicate sortOrder(int group, int indexInGroup) {
group = callSideEffectGroup() and indexInGroup = 0
}
override string toString() { result = "(call side effect for '" + expr.toString() + "')" }
override predicate sideEffectInstruction(Opcode opcode, CppType type) {
opcode = sideEffectOpcode and
(
opcode instanceof Opcode::CallSideEffect and
type = getUnknownType()
or
opcode instanceof Opcode::CallReadSideEffect and
type = getVoidType()
)
}
override CppType getInstructionMemoryOperandType(InstructionTag tag, TypedOperandTag operandTag) {
tag instanceof OnlyInstructionTag and
operandTag instanceof SideEffectOperandTag and
result = getUnknownType()
}
}
/**
* The IR translation of the allocation side effect of a call to a memory allocation function.
*
* This side effect provides a definition for the newly-allocated memory.
*/
class TranslatedAllocationSideEffect extends TranslatedSideEffect, TTranslatedAllocationSideEffect {
AllocationExpr expr;
TranslatedAllocationSideEffect() { this = TTranslatedAllocationSideEffect(expr) }
override Locatable getAST() { result = expr }
override Expr getPrimaryExpr() { result = expr }
override predicate sortOrder(int group, int indexInGroup) {
group = initializeAllocationGroup() and indexInGroup = 0
}
override string toString() { result = "(allocation side effect for '" + expr.toString() + "')" }
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag = OnlyInstructionTag() and
operandTag = addressOperand() and
result = getPrimaryInstructionForSideEffect(OnlyInstructionTag())
}
override predicate sideEffectInstruction(Opcode opcode, CppType type) {
opcode instanceof Opcode::InitializeDynamicAllocation and
type = getUnknownType()
}
/**
* Gets the `Function` containing this expression.
*/
override Function getFunction() { result = arg.getEnclosingFunction() }
}

View File

@@ -135,20 +135,6 @@ private predicate ignoreExpr(Expr expr) {
ignoreExprAndDescendants(expr)
}
/**
* Holds if the side effects of `expr` should be ignoredf for the purposes of IR generation.
*
* In cases involving `constexpr`, a call can wind up as a constant expression. `ignoreExpr()` will
* not hold for such a call, since we do need to translate the call (as a constant), but we need to
* ignore all of the side effects of that call, since we will not actually be generating a `Call`
* instruction.
*/
private predicate ignoreSideEffects(Expr expr) {
ignoreExpr(expr)
or
isIRConstant(expr)
}
/**
* Holds if `func` contains an AST that cannot be translated into IR. This is mostly used to work
* around extractor bugs. Once the relevant extractor bugs are fixed, this predicate can be removed.
@@ -567,13 +553,6 @@ newtype TTranslatedElement =
} or
// The initialization of a base class from within a constructor.
TTranslatedConstructorBaseInit(ConstructorBaseInit init) { not ignoreExpr(init) } or
// Workaround for a case where no base constructor is generated but a targetless base
// constructor call is present.
TTranslatedConstructorBareInit(ConstructorInit init) {
not ignoreExpr(init) and
not init instanceof ConstructorBaseInit and
not init instanceof ConstructorFieldInit
} or
// The destruction of a base class from within a destructor.
TTranslatedDestructorBaseDestruction(DestructorBaseDestruction destruction) {
not ignoreExpr(destruction)
@@ -642,34 +621,32 @@ newtype TTranslatedElement =
// The declaration/initialization part of a `ConditionDeclExpr`
TTranslatedConditionDecl(ConditionDeclExpr expr) { not ignoreExpr(expr) } or
// The side effects of a `Call`
TTranslatedCallSideEffects(CallOrAllocationExpr expr) { not ignoreSideEffects(expr) } or
// The non-argument-specific side effect of a `Call`
TTranslatedCallSideEffect(Expr expr, SideEffectOpcode opcode) {
not ignoreSideEffects(expr) and
opcode = getCallSideEffectOpcode(expr)
TTranslatedCallSideEffects(Call expr) {
// Exclude allocations such as `malloc` (which happen to also be function calls).
// Both `TranslatedCallSideEffects` and `TranslatedAllocationSideEffects` generate
// the same side effects for its children as they both extend the `TranslatedSideEffects`
// class.
// Note: We can separate allocation side effects and call side effects into two
// translated elements as no call can be both a `ConstructorCall` and an `AllocationExpr`.
not expr instanceof AllocationExpr and
(
exists(TTranslatedArgumentSideEffect(expr, _, _, _)) or
expr instanceof ConstructorCall
)
} or
// The side effects of an allocation, i.e. `new`, `new[]` or `malloc`
TTranslatedAllocationSideEffects(AllocationExpr expr) { not ignoreExpr(expr) } or
// A precise side effect of an argument to a `Call`
TTranslatedArgumentExprSideEffect(Call call, Expr expr, int n, SideEffectOpcode opcode) {
TTranslatedArgumentSideEffect(Call call, Expr expr, int n, SideEffectOpcode opcode) {
not ignoreExpr(expr) and
not ignoreSideEffects(call) and
not ignoreExpr(call) and
(
n >= 0 and expr = call.getArgument(n).getFullyConverted()
or
n = -1 and expr = call.getQualifier().getFullyConverted()
) and
opcode = getASideEffectOpcode(call, n)
} or
// Constructor calls lack a qualifier (`this`) expression, so we need to handle the side effects
// on `*this` without an `Expr`.
TTranslatedStructorQualifierSideEffect(Call call, SideEffectOpcode opcode) {
not ignoreSideEffects(call) and
// Don't bother with destructor calls for now, since we won't see very many of them in the IR
// until we start injecting implicit destructor calls.
call instanceof ConstructorCall and
opcode = getASideEffectOpcode(call, -1)
} or
// The side effect that initializes newly-allocated memory.
TTranslatedAllocationSideEffect(AllocationExpr expr) { not ignoreSideEffects(expr) }
}
/**
* Gets the index of the first explicitly initialized element in `initList`

View File

@@ -3,7 +3,6 @@ private import semmle.code.cpp.ir.implementation.IRType
private import semmle.code.cpp.ir.implementation.Opcode
private import semmle.code.cpp.ir.implementation.internal.OperandTag
private import semmle.code.cpp.ir.internal.CppType
private import semmle.code.cpp.ir.internal.IRUtilities
private import semmle.code.cpp.ir.internal.TempVariableTag
private import InstructionTag
private import TranslatedCondition
@@ -814,9 +813,7 @@ abstract class TranslatedVariableAccess extends TranslatedNonConstantExpr {
}
class TranslatedNonFieldVariableAccess extends TranslatedVariableAccess {
TranslatedNonFieldVariableAccess() {
not expr instanceof FieldAccess and not isNonReferenceStructuredBinding(expr.getTarget())
}
TranslatedNonFieldVariableAccess() { not expr instanceof FieldAccess }
override Instruction getFirstInstruction() {
if exists(this.getQualifier())
@@ -863,71 +860,6 @@ class TranslatedFieldAccess extends TranslatedVariableAccess {
}
}
/**
* The IR translation of a variable access of a structured binding, where the type
* of the structured binding is not of a reference type, e.g., `x0` and `x1`
* in `auto [x0, x1] = xs` where `xs` is an array. Although the type of the
* structured binding is a non-reference type, the structured binding behaves
* like a reference. Hence, the translation requires a `VariableAddress` followed
* by a `Load` instead of only a `VariableAddress` as produced by
* `TranslatedVariableAccess`.
*/
class TranslatedStructuredBindingVariableAccess extends TranslatedNonConstantExpr {
override VariableAccess expr;
TranslatedStructuredBindingVariableAccess() { isNonReferenceStructuredBinding(expr.getTarget()) }
override Instruction getFirstInstruction() {
// Structured bindings cannot be qualified.
result = this.getInstruction(StructuredBindingAccessTag())
}
override TranslatedElement getChild(int id) {
// Structured bindings cannot be qualified.
none()
}
override Instruction getResult() { result = this.getInstruction(LoadTag()) }
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) {
tag = StructuredBindingAccessTag() and
kind instanceof GotoEdge and
result = this.getInstruction(LoadTag())
or
tag = LoadTag() and
kind instanceof GotoEdge and
result = this.getParent().getChildSuccessor(this)
}
override Instruction getChildSuccessor(TranslatedElement child) { none() }
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType resultType) {
tag = StructuredBindingAccessTag() and
opcode instanceof Opcode::VariableAddress and
resultType = getTypeForGLValue(this.getLValueReferenceType())
or
tag = LoadTag() and
opcode instanceof Opcode::Load and
resultType = getTypeForPRValue(this.getLValueReferenceType())
}
private LValueReferenceType getLValueReferenceType() {
// The extractor ensures `result` exists when `isNonReferenceStructuredBinding(expr.getTarget())` holds.
result.getBaseType() = expr.getUnspecifiedType()
}
override Instruction getInstructionRegisterOperand(InstructionTag tag, OperandTag operandTag) {
tag = LoadTag() and
operandTag instanceof AddressOperandTag and
result = this.getInstruction(StructuredBindingAccessTag())
}
override IRVariable getInstructionVariable(InstructionTag tag) {
tag = StructuredBindingAccessTag() and
result = getIRUserVariable(expr.getEnclosingFunction(), expr.getTarget())
}
}
class TranslatedFunctionAccess extends TranslatedNonConstantExpr {
override FunctionAccess expr;

View File

@@ -573,11 +573,6 @@ class TranslatedConstructorInitList extends TranslatedElement, InitializationCon
baseInit = func.(Constructor).getInitializer(id) and
result = getTranslatedConstructorBaseInit(baseInit)
)
or
exists(ConstructorInit bareInit |
bareInit = func.(Constructor).getInitializer(id) and
result = getTranslatedConstructorBareInit(bareInit)
)
}
override Instruction getFirstInstruction() {

View File

@@ -917,36 +917,3 @@ class TranslatedDestructorBaseDestruction extends TranslatedBaseStructorCall,
final override string toString() { result = "destroy base: " + call.toString() }
}
/**
* A constructor base init call where no base constructor has been generated.
*
* Workaround for an extractor issue.
*/
class TranslatedConstructorBareInit extends TranslatedElement, TTranslatedConstructorBareInit {
ConstructorInit init;
TranslatedConstructorBareInit() { this = TTranslatedConstructorBareInit(init) }
override Locatable getAST() { result = init }
final override string toString() { result = "construct base (no constructor)" }
override Instruction getFirstInstruction() { result = getParent().getChildSuccessor(this) }
override predicate hasInstruction(Opcode opcode, InstructionTag tag, CppType resultType) {
none()
}
override TranslatedElement getChild(int id) { none() }
override Function getFunction() { result = getParent().getFunction() }
override Instruction getInstructionSuccessor(InstructionTag tag, EdgeKind kind) { none() }
override Instruction getChildSuccessor(TranslatedElement child) { none() }
}
TranslatedConstructorBareInit getTranslatedConstructorBareInit(ConstructorInit init) {
result.getAST() = init
}

View File

@@ -200,7 +200,7 @@ class IRBlock extends IRBlockBase {
* post-dominate block `B`, but block `A` does post-dominate an immediate successor of block `B`.
*/
pragma[noinline]
final IRBlock postDominanceFrontier() {
final IRBlock postPominanceFrontier() {
this.postDominates(result.getASuccessor()) and
not this.strictlyPostDominates(result)
}

View File

@@ -106,12 +106,6 @@ private predicate filteredNumberableInstruction(Instruction instr) {
or
instr instanceof FieldAddressInstruction and
count(instr.(FieldAddressInstruction).getField()) != 1
or
instr instanceof InheritanceConversionInstruction and
(
count(instr.(InheritanceConversionInstruction).getBaseClass()) != 1 or
count(instr.(InheritanceConversionInstruction).getDerivedClass()) != 1
)
}
private predicate variableAddressValueNumber(
@@ -121,7 +115,8 @@ private predicate variableAddressValueNumber(
// The underlying AST element is used as value-numbering key instead of the
// `IRVariable` to work around a problem where a variable or expression with
// multiple types gives rise to multiple `IRVariable`s.
unique( | | instr.getIRVariable().getAST()) = ast
instr.getIRVariable().getAST() = ast and
strictcount(instr.getIRVariable().getAST()) = 1
}
private predicate initializeParameterValueNumber(
@@ -138,7 +133,8 @@ private predicate constantValueNumber(
ConstantInstruction instr, IRFunction irFunc, IRType type, string value
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getResultIRType()) = type and
strictcount(instr.getResultIRType()) = 1 and
instr.getResultIRType() = type and
instr.getValue() = value
}
@@ -155,7 +151,8 @@ private predicate fieldAddressValueNumber(
TValueNumber objectAddress
) {
instr.getEnclosingIRFunction() = irFunc and
unique( | | instr.getField()) = field and
instr.getField() = field and
strictcount(instr.getField()) = 1 and
tvalueNumber(instr.getObjectAddress()) = objectAddress
}
@@ -198,9 +195,9 @@ private predicate inheritanceConversionValueNumber(
) {
instr.getEnclosingIRFunction() = irFunc and
instr.getOpcode() = opcode and
tvalueNumber(instr.getUnary()) = operand and
unique( | | instr.getBaseClass()) = baseClass and
unique( | | instr.getDerivedClass()) = derivedClass
instr.getBaseClass() = baseClass and
instr.getDerivedClass() = derivedClass and
tvalueNumber(instr.getUnary()) = operand
}
private predicate loadTotalOverlapValueNumber(

View File

@@ -11,15 +11,6 @@ private Type getDecayedType(Type type) {
result.(PointerType).getBaseType() = type.(ArrayType).getBaseType()
}
/**
* Holds if the sepcified variable is a structured binding with a non-reference
* type.
*/
predicate isNonReferenceStructuredBinding(Variable v) {
v.isStructuredBinding() and
not v.getUnspecifiedType() instanceof ReferenceType
}
/**
* Get the actual type of the specified variable, as opposed to the declared type.
* This returns the type of the variable after any pointer decay is applied, and
@@ -39,12 +30,7 @@ Type getVariableType(Variable v) {
result = v.getInitializer().getExpr().getType()
or
not exists(v.getInitializer()) and result = v.getType()
else
if isNonReferenceStructuredBinding(v)
then
// The extractor ensures `r` exists when `isNonReferenceStructuredBinding(v)` holds.
exists(LValueReferenceType r | r.getBaseType() = v.getUnspecifiedType() | result = r)
else result = v.getType()
else result = v.getType()
)
}

View File

@@ -11,14 +11,15 @@ import semmle.code.cpp.models.interfaces.SideEffect
import semmle.code.cpp.models.interfaces.FlowSource
/**
* The standard functions `fgets` and `fgetws`.
* The standard functions `gets` and `fgets`.
*/
private class FgetsFunction extends DataFlowFunction, TaintFunction, ArrayFunction, AliasFunction,
private class GetsFunction extends DataFlowFunction, TaintFunction, ArrayFunction, AliasFunction,
SideEffectFunction, RemoteFlowSourceFunction {
FgetsFunction() {
GetsFunction() {
// gets(str)
// fgets(str, num, stream)
// fgetws(wstr, num, stream)
this.hasGlobalOrStdOrBslName(["fgets", "fgetws"])
this.hasGlobalOrStdOrBslName(["gets", "fgets", "fgetws"])
}
override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
@@ -50,61 +51,20 @@ private class FgetsFunction extends DataFlowFunction, TaintFunction, ArrayFuncti
override predicate hasRemoteFlowSource(FunctionOutput output, string description) {
output.isParameterDeref(0) and
description = "String read by " + this.getName()
or
output.isReturnValue() and
description = "String read by " + this.getName()
}
override predicate hasArrayWithVariableSize(int bufParam, int countParam) {
not this.hasName("gets") and
bufParam = 0 and
countParam = 1
}
override predicate hasArrayWithUnknownSize(int bufParam) {
this.hasName("gets") and
bufParam = 0
}
override predicate hasArrayOutput(int bufParam) { bufParam = 0 }
override predicate hasSocketInput(FunctionInput input) { input.isParameterDeref(2) }
}
/**
* The standard functions `gets`.
*/
private class GetsFunction extends DataFlowFunction, ArrayFunction, AliasFunction,
SideEffectFunction, LocalFlowSourceFunction {
GetsFunction() {
// gets(str)
this.hasGlobalOrStdOrBslName("gets")
}
override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
input.isParameter(0) and
output.isReturnValue()
}
override predicate parameterNeverEscapes(int index) { none() }
override predicate parameterEscapesOnlyViaReturn(int index) { index = 0 }
override predicate parameterIsAlwaysReturned(int index) { index = 0 }
override predicate hasOnlySpecificReadSideEffects() { any() }
override predicate hasOnlySpecificWriteSideEffects() { any() }
override predicate hasSpecificWriteSideEffect(ParameterIndex i, boolean buffer, boolean mustWrite) {
i = 0 and
buffer = true and
mustWrite = true
}
override predicate hasLocalFlowSource(FunctionOutput output, string description) {
output.isParameterDeref(0) and
description = "String read by " + this.getName()
or
output.isReturnValue() and
description = "String read by " + this.getName()
}
override predicate hasArrayWithUnknownSize(int bufParam) { bufParam = 0 }
override predicate hasArrayOutput(int bufParam) { bufParam = 0 }
override predicate hasSocketInput(FunctionInput input) { input.isParameter(2) }
}

View File

@@ -1,10 +1,7 @@
/*
* Support for tracking tainted data through the program. This is an alias for
* `semmle.code.cpp.ir.dataflow.DefaultTaintTracking` provided for backwards
* compatibility.
* Support for tracking tainted data through the program.
*
* Prefer to use `semmle.code.cpp.dataflow.TaintTracking` or
* `semmle.code.cpp.ir.dataflow.TaintTracking` when designing new queries.
* Prefer to use `semmle.code.cpp.dataflow.TaintTracking` when designing new queries.
*/
import semmle.code.cpp.ir.dataflow.DefaultTaintTracking

View File

@@ -135,11 +135,52 @@ externalData(
string value : string ref
);
/**
* The date of the snapshot.
*/
snapshotDate(unique date snapshotDate : date ref);
/**
* The source location of the snapshot.
*/
sourceLocationPrefix(string prefix : string ref);
/**
* Data used by the 'duplicate code' detection.
*/
duplicateCode(
unique int id : @duplication,
string relativePath : string ref,
int equivClass : int ref
);
/**
* Data used by the 'similar code' detection.
*/
similarCode(
unique int id : @similarity,
string relativePath : string ref,
int equivClass : int ref
);
/**
* Data used by the 'duplicate code' and 'similar code' detection.
*/
@duplication_or_similarity = @duplication | @similarity
/**
* Data used by the 'duplicate code' and 'similar code' detection.
*/
#keyset[id, offset]
tokens(
int id : @duplication_or_similarity ref,
int offset : int ref,
int beginLine : int ref,
int beginColumn : int ref,
int endLine : int ref,
int endColumn : int ref
);
/**
* Information about packages that provide code used during compilation.
* The `id` is just a unique identifier.
@@ -446,7 +487,6 @@ var_decl_specifiers(
int id: @var_decl ref,
string name: string ref
)
is_structured_binding(unique int id: @variable ref);
type_decls(
unique int id: @type_decl,

File diff suppressed because it is too large Load Diff

View File

@@ -6,22 +6,122 @@
*/
class Person extends string {
Person() {
this =
[
"Ronil", "Dina", "Ravi", "Bruce", "Jo", "Aida", "Esme", "Charlie", "Fred", "Meera", "Maya",
"Chad", "Tiana", "Laura", "George", "Will", "Mary", "Almira", "Susannah", "Rhoda",
"Cynthia", "Eunice", "Olive", "Virginia", "Angeline", "Helen", "Cornelia", "Harriet",
"Mahala", "Abby", "Margaret", "Deb", "Minerva", "Severus", "Lavina", "Adeline", "Cath",
"Elisa", "Lucretia", "Anne", "Eleanor", "Joanna", "Adam", "Agnes", "Rosanna", "Clara",
"Melissa", "Amy", "Isabel", "Jemima", "Cordelia", "Melinda", "Delila", "Jeremiah", "Elijah",
"Hester", "Walter", "Oliver", "Hugh", "Aaron", "Reuben", "Eli", "Amos", "Augustus",
"Theodore", "Ira", "Timothy", "Cyrus", "Horace", "Simon", "Asa", "Frank", "Nelson",
"Leonard", "Harrison", "Anthony", "Louis", "Milton", "Noah", "Cornelius", "Abdul", "Warren",
"Harvey", "Dennis", "Wesley", "Sylvester", "Gilbert", "Sullivan", "Edmund", "Wilson",
"Perry", "Matthew", "Simba", "Nala", "Rafiki", "Shenzi", "Ernest", "Gertrude", "Oscar",
"Lilian", "Raymond", "Elgar", "Elmer", "Herbert", "Maude", "Mae", "Otto", "Edwin",
"Ophelia", "Parsley", "Sage", "Rosemary", "Thyme", "Garfunkel", "King Basil", "Stephen"
]
this = "Ronil" or
this = "Dina" or
this = "Ravi" or
this = "Bruce" or
this = "Jo" or
this = "Aida" or
this = "Esme" or
this = "Charlie" or
this = "Fred" or
this = "Meera" or
this = "Maya" or
this = "Chad" or
this = "Tiana" or
this = "Laura" or
this = "George" or
this = "Will" or
this = "Mary" or
this = "Almira" or
this = "Susannah" or
this = "Rhoda" or
this = "Cynthia" or
this = "Eunice" or
this = "Olive" or
this = "Virginia" or
this = "Angeline" or
this = "Helen" or
this = "Cornelia" or
this = "Harriet" or
this = "Mahala" or
this = "Abby" or
this = "Margaret" or
this = "Deb" or
this = "Minerva" or
this = "Severus" or
this = "Lavina" or
this = "Adeline" or
this = "Cath" or
this = "Elisa" or
this = "Lucretia" or
this = "Anne" or
this = "Eleanor" or
this = "Joanna" or
this = "Adam" or
this = "Agnes" or
this = "Rosanna" or
this = "Clara" or
this = "Melissa" or
this = "Amy" or
this = "Isabel" or
this = "Jemima" or
this = "Cordelia" or
this = "Melinda" or
this = "Delila" or
this = "Jeremiah" or
this = "Elijah" or
this = "Hester" or
this = "Walter" or
this = "Oliver" or
this = "Hugh" or
this = "Aaron" or
this = "Reuben" or
this = "Eli" or
this = "Amos" or
this = "Augustus" or
this = "Theodore" or
this = "Ira" or
this = "Timothy" or
this = "Cyrus" or
this = "Horace" or
this = "Simon" or
this = "Asa" or
this = "Frank" or
this = "Nelson" or
this = "Leonard" or
this = "Harrison" or
this = "Anthony" or
this = "Louis" or
this = "Milton" or
this = "Noah" or
this = "Cornelius" or
this = "Abdul" or
this = "Warren" or
this = "Harvey" or
this = "Dennis" or
this = "Wesley" or
this = "Sylvester" or
this = "Gilbert" or
this = "Sullivan" or
this = "Edmund" or
this = "Wilson" or
this = "Perry" or
this = "Matthew" or
this = "Simba" or
this = "Nala" or
this = "Rafiki" or
this = "Shenzi" or
this = "Ernest" or
this = "Gertrude" or
this = "Oscar" or
this = "Lilian" or
this = "Raymond" or
this = "Elgar" or
this = "Elmer" or
this = "Herbert" or
this = "Maude" or
this = "Mae" or
this = "Otto" or
this = "Edwin" or
this = "Ophelia" or
this = "Parsley" or
this = "Sage" or
this = "Rosemary" or
this = "Thyme" or
this = "Garfunkel" or
this = "King Basil" or
this = "Stephen"
}
/** Gets the hair color of the person. If the person is bald, there is no result. */
@@ -836,12 +936,25 @@ class Person extends string {
/** Holds if the person is deceased. */
predicate isDeceased() {
this =
[
"Ernest", "Gertrude", "Oscar", "Lilian", "Edwin", "Raymond", "Elgar", "Elmer", "Herbert",
"Maude", "Mae", "Otto", "Ophelia", "Parsley", "Sage", "Rosemary", "Thyme", "Garfunkel",
"King Basil"
]
this = "Ernest" or
this = "Gertrude" or
this = "Oscar" or
this = "Lilian" or
this = "Edwin" or
this = "Raymond" or
this = "Elgar" or
this = "Elmer" or
this = "Herbert" or
this = "Maude" or
this = "Mae" or
this = "Otto" or
this = "Ophelia" or
this = "Parsley" or
this = "Sage" or
this = "Rosemary" or
this = "Thyme" or
this = "Garfunkel" or
this = "King Basil"
}
/** Gets a parent of the person (alive or deceased). */
@@ -1082,7 +1195,12 @@ class Person extends string {
}
/** Holds if the person is allowed in the region. Initially, all villagers are allowed in every region. */
predicate isAllowedIn(string region) { region = ["north", "south", "east", "west"] }
predicate isAllowedIn(string region) {
region = "north" or
region = "south" or
region = "east" or
region = "west"
}
}
/** Returns a parent of the person. */

View File

@@ -1,6 +0,0 @@
description: Remove unused legacy relations
compatibility: full
snapshotDate.rel: delete
duplicateCode.rel: delete
similarCode.rel: delete
tokens.rel: delete

View File

@@ -1,2 +0,0 @@
description: Add relation for tracking variables from structured binding declarations
compatibility: backwards

File diff suppressed because it is too large Load Diff

View File

@@ -1,42 +1,3 @@
## 0.0.10
### Deprecated Classes
* The `CodeDuplication.Copy`, `CodeDuplication.DuplicateBlock`, and `CodeDuplication.SimilarBlock` classes have been deprecated.
## 0.0.9
### New Queries
* Added a new query, `cpp/open-call-with-mode-argument`, to detect when `open` or `openat` is called with the `O_CREAT` or `O_TMPFILE` flag but when the `mode` argument is omitted.
### Minor Analysis Improvements
* The "Cleartext transmission of sensitive information" (`cpp/cleartext-transmission`) query has been further improved to reduce false positive results, and upgraded from `medium` to `high` precision.
* The "Cleartext transmission of sensitive information" (`cpp/cleartext-transmission`) query now finds more results, where a password is stored in a struct field or class member variable.
* The `cpp/cleartext-storage-file` query has been improved, removing false positives where data is written to a standard output stream.
* The `cpp/cleartext-storage-buffer` query has been updated to use the `semmle.code.cpp.dataflow.TaintTracking` library.
* The `cpp/world-writable-file-creation` query now only detects `open` and `openat` calls with the `O_CREAT` or `O_TMPFILE` flag.
## 0.0.8
### New Queries
* The `security` tag has been added to the `cpp/return-stack-allocated-memory` query. As a result, its results will now appear by default.
* The "Uncontrolled data in arithmetic expression" (cpp/uncontrolled-arithmetic) query has been enhanced to reduce false positive results and its @precision increased to high.
* A new `cpp/very-likely-overruning-write` query has been added to the default query suite for C/C++. The query reports some results that were formerly flagged by `cpp/overruning-write`.
### Minor Analysis Improvements
* Fix an issue with the `cpp/declaration-hides-variable` query where it would report variables that are unnamed in a database.
* The `cpp/cleartext-storage-file` query has been upgraded with non-local taint flow and has been converted to a `path-problem` query.
* The `cpp/return-stack-allocated-memory` query has been improved to produce fewer false positives. The
query has also been converted to a `path-problem` query.
* The "Cleartext transmission of sensitive information" (`cpp/cleartext-transmission`) query has been improved in several ways to reduce false positive results.
* The "Potential improper null termination" (`cpp/improper-null-termination`) query now produces fewer false positive results around control flow branches and loops.
* Added exception for GLib's gboolean to cpp/ambiguously-signed-bit-field.
This change reduces the number of false positives in the query.
## 0.0.7
## 0.0.6

View File

@@ -14,75 +14,158 @@
*/
import cpp
// We don't actually use the global value numbering library in this query, but without it we end up
// recomputing the IR.
private import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import semmle.code.cpp.ir.IR
import semmle.code.cpp.ir.dataflow.MustFlow
import PathGraph
import semmle.code.cpp.ir.dataflow.DataFlow::DataFlow
/** Holds if `f` has a name that we intrepret as evidence of intentionally returning the value of the stack pointer. */
predicate intentionallyReturnsStackPointer(Function f) {
f.getName().toLowerCase().matches(["%stack%", "%sp%"])
}
class ReturnStackAllocatedMemoryConfig extends MustFlowConfiguration {
ReturnStackAllocatedMemoryConfig() { this = "ReturnStackAllocatedMemoryConfig" }
override predicate isSource(DataFlow::Node source) {
// Holds if `source` is a node that represents the use of a stack variable
exists(VariableAddressInstruction var, Function func |
var = source.asInstruction() and
func = var.getEnclosingFunction() and
var.getASTVariable() instanceof StackVariable and
// Pointer-to-member types aren't properly handled in the dbscheme.
not var.getResultType() instanceof PointerToMemberType and
// Rule out FPs caused by extraction errors.
not any(ErrorExpr e).getEnclosingFunction() = func and
not intentionallyReturnsStackPointer(func)
)
}
override predicate isSink(DataFlow::Node sink) {
// Holds if `sink` is a node that represents the `StoreInstruction` that is subsequently used in
// a `ReturnValueInstruction`.
// We use the `StoreInstruction` instead of the instruction that defines the
// `ReturnValueInstruction`'s source value oprand because the former has better location information.
exists(StoreInstruction store |
store.getDestinationAddress().(VariableAddressInstruction).getIRVariable() instanceof
IRReturnVariable and
sink.asOperand() = store.getSourceValueOperand()
)
}
/**
* This configuration intentionally conflates addresses of fields and their object, and pointer offsets
* with their base pointer as this allows us to detect cases where an object's address flows to a
* return statement via a field. For example:
*
* ```cpp
* struct S { int x, y };
* int* test() {
* S s;
* return &s.x; // BAD: &s.x is an address of a variable on the stack.
* }
* ```
*/
override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
node2.asInstruction().(FieldAddressInstruction).getObjectAddressOperand() = node1.asOperand()
or
node2.asInstruction().(PointerOffsetInstruction).getLeftOperand() = node1.asOperand()
}
/**
* Holds if `source` is a node that represents the use of a stack variable
*/
predicate isSource(Node source) {
exists(VariableAddressInstruction var, Function func |
var = source.asInstruction() and
func = var.getEnclosingFunction() and
var.getASTVariable() instanceof StackVariable and
// Pointer-to-member types aren't properly handled in the dbscheme.
not var.getResultType() instanceof PointerToMemberType and
// Rule out FPs caused by extraction errors.
not any(ErrorExpr e).getEnclosingFunction() = func and
not intentionallyReturnsStackPointer(func)
)
}
from
MustFlowPathNode source, MustFlowPathNode sink, VariableAddressInstruction var,
ReturnStackAllocatedMemoryConfig conf
/**
* Holds if `sink` is a node that represents the `StoreInstruction` that is subsequently used in
* a `ReturnValueInstruction`. We use the `StoreInstruction` instead of the instruction that defines the
* `ReturnValueInstruction`'s source value oprand because the former has better location information.
*/
predicate isSink(Node sink) {
exists(StoreInstruction store |
store.getDestinationAddress().(VariableAddressInstruction).getIRVariable() instanceof
IRReturnVariable and
sink.asOperand() = store.getSourceValueOperand()
)
}
/** Holds if `node1` _must_ flow to `node2`. */
predicate step(Node node1, Node node2) {
instructionToOperandStep(node1.asInstruction(), node2.asOperand())
or
operandToInstructionStep(node1.asOperand(), node2.asInstruction())
}
predicate instructionToOperandStep(Instruction instr, Operand operand) { operand.getDef() = instr }
/**
* Holds if `operand` flows to the result of `instr`.
*
* This predicate ignores flow through `PhiInstruction`s to create a 'must flow' relation. It also
* intentionally conflates addresses of fields and their object, and pointer offsets with their
* base pointer as this allows us to detect cases where an object's address flows to a return statement
* via a field. For example:
*
* ```cpp
* struct S { int x, y };
* int* test() {
* S s;
* return &s.x; // BAD: &s.x is an address of a variable on the stack.
* }
* ```
*/
predicate operandToInstructionStep(Operand operand, Instruction instr) {
instr.(CopyInstruction).getSourceValueOperand() = operand
or
instr.(ConvertInstruction).getUnaryOperand() = operand
or
instr.(CheckedConvertOrNullInstruction).getUnaryOperand() = operand
or
instr.(InheritanceConversionInstruction).getUnaryOperand() = operand
or
instr.(FieldAddressInstruction).getObjectAddressOperand() = operand
or
instr.(PointerOffsetInstruction).getLeftOperand() = operand
}
/** Holds if a source node flows to `n`. */
predicate branchlessLocalFlow0(Node n) {
isSource(n)
or
exists(Node mid |
branchlessLocalFlow0(mid) and
step(mid, n)
)
}
/** Holds if `n` is reachable through some source node, and `n` also eventually reaches a sink. */
predicate branchlessLocalFlow1(Node n) {
branchlessLocalFlow0(n) and
(
isSink(n)
or
exists(Node mid |
branchlessLocalFlow1(mid) and
step(n, mid)
)
)
}
newtype TLocalPathNode =
TLocalPathNodeMid(Node n) {
branchlessLocalFlow1(n) and
(
isSource(n) or
exists(LocalPathNodeMid mid | step(mid.getNode(), n))
)
}
abstract class LocalPathNode extends TLocalPathNode {
Node n;
/** Gets the underlying node. */
Node getNode() { result = n }
/** Gets a textual representation of this node. */
string toString() { result = n.toString() }
/** Gets the location of this element. */
Location getLocation() { result = n.getLocation() }
/** Gets a successor `LocalPathNode`, if any. */
LocalPathNode getASuccessor() { step(this.getNode(), result.getNode()) }
}
class LocalPathNodeMid extends LocalPathNode, TLocalPathNodeMid {
LocalPathNodeMid() { this = TLocalPathNodeMid(n) }
}
class LocalPathNodeSink extends LocalPathNodeMid {
LocalPathNodeSink() { isSink(this.getNode()) }
}
/**
* Holds if `source` is a source node, `sink` is a sink node, and there's flow
* from `source` to `sink` using `step` relation.
*/
predicate hasFlow(LocalPathNode source, LocalPathNodeSink sink) {
isSource(source.getNode()) and
source.getASuccessor+() = sink
}
predicate reach(LocalPathNode n) { n instanceof LocalPathNodeSink or reach(n.getASuccessor()) }
query predicate edges(LocalPathNode a, LocalPathNode b) { a.getASuccessor() = b and reach(b) }
query predicate nodes(LocalPathNode n, string key, string val) {
reach(n) and key = "semmle.label" and val = n.toString()
}
from LocalPathNode source, LocalPathNodeSink sink, VariableAddressInstruction var
where
conf.hasFlowPath(source, sink) and
source.getNode().asInstruction() = var and
// Only raise an alert if we're returning from the _same_ callable as the on that
// declared the stack variable.
var.getEnclosingFunction() = sink.getNode().getEnclosingCallable()
hasFlow(source, sink) and
source.getNode().asInstruction() = var
select sink.getNode(), source, sink, "May return stack-allocated memory from $@.", var.getAST(),
var.getAST().toString()

View File

@@ -1,25 +0,0 @@
static const int* xptr;
void localAddressEscapes() {
int x = 0;
xptr = &x;
}
void example1() {
localAddressEscapes();
const int* x = xptr; // BAD: This pointer points to expired stack allocated memory.
}
void localAddressDoesNotEscape() {
int x = 0;
xptr = &x;
// ...
// use `xptr`
// ...
xptr = nullptr;
}
void example2() {
localAddressDoesNotEscape();
const int* x = xptr; // GOOD: This pointer does not point to expired memory.
}

View File

@@ -1,49 +0,0 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>
<overview>
<p>
This rule finds uses of pointers that likely point to local variables in
expired stack frames. A pointer to a local variable is only valid
until the function returns, after which it becomes a dangling pointer.
</p>
</overview>
<recommendation>
<ol>
<li>
If it is necessary to take the address of a local variable, then make
sure that the address is only stored in memory that does not outlive
the local variable. For example, it is safe to store the address in
another local variable. Similarly, it is also safe to pass the address
of a local variable to another function provided that the other
function only uses it locally and does not store it in non-local
memory.
</li>
<li>
If it is necessary to store an address which will outlive the
current function scope, then it should be allocated on the heap. Care
should be taken to make sure that the memory is deallocated when it is
no longer needed, particularly when using low-level memory management
routines such as <tt>malloc</tt>/<tt>free</tt> or
<tt>new</tt>/<tt>delete</tt>. Modern C++ applications often use smart
pointers, such as <tt>std::shared_ptr</tt>, to reduce the chance of
a memory leak.
</li>
</ol>
</recommendation>
<example>
<sample src="UsingExpiredStackAddress.cpp" />
</example>
<references>
<li>Wikipedia: <a href="https://en.wikipedia.org/wiki/Dangling_pointer">Dangling pointer</a>.</li>
</references>
</qhelp>

View File

@@ -1,348 +0,0 @@
/**
* @name Use of expired stack-address
* @description Accessing the stack-allocated memory of a function
* after it has returned can lead to memory corruption.
* @kind path-problem
* @problem.severity error
* @security-severity 9.3
* @precision high
* @id cpp/using-expired-stack-address
* @tags reliability
* security
* external/cwe/cwe-825
*/
import cpp
// We don't actually use the global value numbering library in this query, but without it we end up
// recomputing the IR.
import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import semmle.code.cpp.ir.IR
predicate instructionHasVariable(VariableAddressInstruction vai, StackVariable var, Function f) {
var = vai.getASTVariable() and
f = vai.getEnclosingFunction() and
// Pointer-to-member types aren't properly handled in the dbscheme.
not vai.getResultType() instanceof PointerToMemberType and
// Rule out FPs caused by extraction errors.
not any(ErrorExpr e).getEnclosingFunction() = f
}
/**
* Holds if `source` is the base address of an address computation whose
* result is stored in `address`.
*/
predicate stackPointerFlowsToUse(Instruction address, VariableAddressInstruction source) {
address = source and
instructionHasVariable(source, _, _)
or
stackPointerFlowsToUse(address.(CopyInstruction).getSourceValue(), source)
or
stackPointerFlowsToUse(address.(ConvertInstruction).getUnary(), source)
or
stackPointerFlowsToUse(address.(CheckedConvertOrNullInstruction).getUnary(), source)
or
stackPointerFlowsToUse(address.(InheritanceConversionInstruction).getUnary(), source)
or
stackPointerFlowsToUse(address.(FieldAddressInstruction).getObjectAddress(), source)
or
stackPointerFlowsToUse(address.(PointerOffsetInstruction).getLeft(), source)
}
/**
* A HashCons-like table for comparing addresses that are
* computed relative to some global variable.
*/
newtype TGlobalAddress =
TGlobalVariable(GlobalOrNamespaceVariable v) {
// Pointer-to-member types aren't properly handled in the dbscheme.
not v.getUnspecifiedType() instanceof PointerToMemberType
} or
TLoad(TGlobalAddress address) {
address = globalAddress(any(LoadInstruction load).getSourceAddress())
} or
TConversion(string kind, TGlobalAddress address, Type fromType, Type toType) {
kind = "unchecked" and
exists(ConvertInstruction convert |
uncheckedConversionTypes(convert, fromType, toType) and
address = globalAddress(convert.getUnary())
)
or
kind = "checked" and
exists(CheckedConvertOrNullInstruction convert |
checkedConversionTypes(convert, fromType, toType) and
address = globalAddress(convert.getUnary())
)
or
kind = "inheritance" and
exists(InheritanceConversionInstruction convert |
inheritanceConversionTypes(convert, fromType, toType) and
address = globalAddress(convert.getUnary())
)
} or
TFieldAddress(TGlobalAddress address, Field f) {
exists(FieldAddressInstruction fai |
fai.getField() = f and
address = globalAddress(fai.getObjectAddress())
)
}
pragma[noinline]
predicate uncheckedConversionTypes(ConvertInstruction convert, Type fromType, Type toType) {
fromType = convert.getUnary().getResultType() and
toType = convert.getResultType()
}
pragma[noinline]
predicate checkedConversionTypes(CheckedConvertOrNullInstruction convert, Type fromType, Type toType) {
fromType = convert.getUnary().getResultType() and
toType = convert.getResultType()
}
pragma[noinline]
predicate inheritanceConversionTypes(
InheritanceConversionInstruction convert, Type fromType, Type toType
) {
fromType = convert.getUnary().getResultType() and
toType = convert.getResultType()
}
/** Gets the HashCons value of an address computed by `instr`, if any. */
TGlobalAddress globalAddress(Instruction instr) {
result = TGlobalVariable(instr.(VariableAddressInstruction).getASTVariable())
or
not instr instanceof LoadInstruction and
result = globalAddress(instr.(CopyInstruction).getSourceValue())
or
exists(LoadInstruction load | instr = load |
result = TLoad(globalAddress(load.getSourceAddress()))
)
or
exists(ConvertInstruction convert, Type fromType, Type toType | instr = convert |
uncheckedConversionTypes(convert, fromType, toType) and
result = TConversion("unchecked", globalAddress(convert.getUnary()), fromType, toType)
)
or
exists(CheckedConvertOrNullInstruction convert, Type fromType, Type toType | instr = convert |
checkedConversionTypes(convert, fromType, toType) and
result = TConversion("checked", globalAddress(convert.getUnary()), fromType, toType)
)
or
exists(InheritanceConversionInstruction convert, Type fromType, Type toType | instr = convert |
inheritanceConversionTypes(convert, fromType, toType) and
result = TConversion("inheritance", globalAddress(convert.getUnary()), fromType, toType)
)
or
exists(FieldAddressInstruction fai | instr = fai |
result = TFieldAddress(globalAddress(fai.getObjectAddress()), fai.getField())
)
or
result = globalAddress(instr.(PointerOffsetInstruction).getLeft())
}
/** Gets a `StoreInstruction` that may be executed after executing `store`. */
pragma[inline]
StoreInstruction getAStoreStrictlyAfter(StoreInstruction store) {
exists(IRBlock block, int index1, int index2 |
block.getInstruction(index1) = store and
block.getInstruction(index2) = result and
index2 > index1
)
or
exists(IRBlock block1, IRBlock block2 |
store.getBlock() = block1 and
result.getBlock() = block2 and
block1.getASuccessor+() = block2
)
}
/**
* Holds if `store` copies the address of `f`'s local variable `var`
* into the address `globalAddress`.
*/
predicate stackAddressEscapes(
StoreInstruction store, StackVariable var, TGlobalAddress globalAddress, Function f
) {
globalAddress = globalAddress(store.getDestinationAddress()) and
exists(VariableAddressInstruction vai |
instructionHasVariable(pragma[only_bind_into](vai), var, f) and
stackPointerFlowsToUse(store.getSourceValue(), vai)
) and
// Ensure there's no subsequent store that overrides the global address.
not globalAddress = globalAddress(getAStoreStrictlyAfter(store).getDestinationAddress())
}
predicate blockStoresToAddress(
IRBlock block, int index, StoreInstruction store, TGlobalAddress globalAddress
) {
block.getInstruction(index) = store and
globalAddress = globalAddress(store.getDestinationAddress())
}
/**
* Holds if `globalAddress` evaluates to the address of `var` (which escaped through `store` before
* returning through `call`) when control reaches `block`.
*/
predicate globalAddressPointsToStack(
StoreInstruction store, StackVariable var, CallInstruction call, IRBlock block,
TGlobalAddress globalAddress, boolean isCallBlock, boolean isStoreBlock
) {
(
if blockStoresToAddress(block, _, _, globalAddress)
then isStoreBlock = true
else isStoreBlock = false
) and
(
isCallBlock = true and
exists(Function f |
stackAddressEscapes(store, var, globalAddress, f) and
call.getStaticCallTarget() = f and
call.getBlock() = block
)
or
isCallBlock = false and
step(store, var, call, globalAddress, _, block)
)
}
pragma[inline]
int getInstructionIndex(Instruction instr, IRBlock block) { block.getInstruction(result) = instr }
predicate step(
StoreInstruction store, StackVariable var, CallInstruction call, TGlobalAddress globalAddress,
IRBlock pred, IRBlock succ
) {
exists(boolean isCallBlock, boolean isStoreBlock |
// Only recurse if there is no store to `globalAddress` in `mid`.
globalAddressPointsToStack(store, var, call, pred, globalAddress, isCallBlock, isStoreBlock)
|
// Post domination ensures that `block` is always executed after `mid`
// Domination ensures that `mid` is always executed before `block`
isStoreBlock = false and
succ.immediatelyPostDominates(pred) and
pred.immediatelyDominates(succ)
or
exists(CallInstruction anotherCall, int anotherCallIndex |
anotherCall = pred.getInstruction(anotherCallIndex) and
succ.getFirstInstruction() instanceof EnterFunctionInstruction and
succ.getEnclosingFunction() = anotherCall.getStaticCallTarget() and
(if isCallBlock = true then getInstructionIndex(call, _) < anotherCallIndex else any()) and
(
if isStoreBlock = true
then
forex(int storeIndex | blockStoresToAddress(pred, storeIndex, _, globalAddress) |
anotherCallIndex < storeIndex
)
else any()
)
)
)
}
newtype TPathElement =
TStore(StoreInstruction store) { globalAddressPointsToStack(store, _, _, _, _, _, _) } or
TCall(CallInstruction call, IRBlock block) {
globalAddressPointsToStack(_, _, call, block, _, _, _)
} or
TMid(IRBlock block) { step(_, _, _, _, _, block) } or
TSink(LoadInstruction load, IRBlock block) {
exists(TGlobalAddress address |
globalAddressPointsToStack(_, _, _, block, address, _, _) and
block.getAnInstruction() = load and
globalAddress(load.getSourceAddress()) = address
)
}
class PathElement extends TPathElement {
StoreInstruction asStore() { this = TStore(result) }
CallInstruction asCall(IRBlock block) { this = TCall(result, block) }
predicate isCall(IRBlock block) { exists(this.asCall(block)) }
IRBlock asMid() { this = TMid(result) }
LoadInstruction asSink(IRBlock block) { this = TSink(result, block) }
predicate isSink(IRBlock block) { exists(this.asSink(block)) }
string toString() {
result = [asStore().toString(), asCall(_).toString(), asMid().toString(), asSink(_).toString()]
}
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
this.asStore()
.getLocation()
.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
or
this.asCall(_)
.getLocation()
.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
or
this.asMid().getLocation().hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
or
this.asSink(_)
.getLocation()
.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
}
predicate isSink(LoadInstruction load, IRBlock block, int index, TGlobalAddress globalAddress) {
block.getInstruction(index) = load and
globalAddress(load.getSourceAddress()) = globalAddress
}
query predicate edges(PathElement pred, PathElement succ) {
// Store -> caller
globalAddressPointsToStack(pred.asStore(), _, succ.asCall(_), _, _, _, _)
or
// Call -> basic block
pred.isCall(succ.asMid())
or
// Special case for when the caller goes directly to the load with no steps
// across basic blocks (i.e., caller -> sink)
exists(IRBlock block |
pred.isCall(block) and
succ.isSink(block)
)
or
// Basic block -> basic block
step(_, _, _, _, pred.asMid(), succ.asMid())
or
// Basic block -> load
succ.isSink(pred.asMid())
}
from
StoreInstruction store, StackVariable var, LoadInstruction load, CallInstruction call,
IRBlock block, boolean isCallBlock, TGlobalAddress address, boolean isStoreBlock,
PathElement source, PathElement sink, int loadIndex
where
globalAddressPointsToStack(store, var, call, block, address, isCallBlock, isStoreBlock) and
isSink(load, block, loadIndex, address) and
(
// We know that we have a sequence:
// (1) store to `address` -> (2) return from `f` -> (3) load from `address`.
// But if (2) and (3) happen in the sam block we need to check the
// block indices to ensure that (3) happens after (2).
if isCallBlock = true
then
// If so, the load must happen after the call.
getInstructionIndex(call, _) < loadIndex
else any()
) and
(
// If there is a store to the address we need to make sure that the load we found was
// before that store (So that the load doesn't read an overwritten value).
if isStoreBlock = true
then
forex(int storeIndex | blockStoresToAddress(block, storeIndex, _, address) |
loadIndex < storeIndex
)
else any()
) and
source.asStore() = store and
sink.asSink(_) = load
select sink, source, sink, "Stack variable $@ escapes $@ and is used after it has expired.", var,
var.toString(), store, "here"

View File

@@ -17,47 +17,169 @@
import cpp
// We don't actually use the global value numbering library in this query, but without it we end up
// recomputing the IR.
import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import semmle.code.cpp.ir.IR
import semmle.code.cpp.ir.dataflow.MustFlow
import PathGraph
private import semmle.code.cpp.valuenumbering.GlobalValueNumbering
private import semmle.code.cpp.ir.IR
class UnsafeUseOfThisConfig extends MustFlowConfiguration {
UnsafeUseOfThisConfig() { this = "UnsafeUseOfThisConfig" }
bindingset[n, result]
int unbind(int n) { result >= n and result <= n }
override predicate isSource(DataFlow::Node source) { isSource(source, _, _) }
/** Holds if `p` is the `n`'th parameter of the non-virtual function `f`. */
predicate parameterOf(Parameter p, Function f, int n) {
not f.isVirtual() and f.getParameter(n) = p
}
override predicate isSink(DataFlow::Node sink) { isSink(sink, _) }
/**
* Holds if `instr` is the `n`'th argument to a call to the non-virtual function `f`, and
* `init` is the corresponding initiazation instruction that receives the value of `instr` in `f`.
*/
predicate flowIntoParameter(
CallInstruction call, Instruction instr, Function f, int n, InitializeParameterInstruction init
) {
not f.isVirtual() and
call.getPositionalArgument(n) = instr and
f = call.getStaticCallTarget() and
getEnclosingNonVirtualFunctionInitializeParameter(init, f) and
init.getParameter().getIndex() = unbind(n)
}
/**
* Holds if `instr` is an argument to a call to the function `f`, and `init` is the
* corresponding initialization instruction that receives the value of `instr` in `f`.
*/
pragma[noinline]
predicate getPositionalArgumentInitParam(
CallInstruction call, Instruction instr, InitializeParameterInstruction init, Function f
) {
exists(int n |
parameterOf(_, f, n) and
flowIntoParameter(call, instr, f, unbind(n), init)
)
}
/**
* Holds if `instr` is the qualifier to a call to the non-virtual function `f`, and
* `init` is the corresponding initiazation instruction that receives the value of
* `instr` in `f`.
*/
pragma[noinline]
predicate getThisArgumentInitParam(
CallInstruction call, Instruction instr, InitializeParameterInstruction init, Function f
) {
not f.isVirtual() and
call.getStaticCallTarget() = f and
getEnclosingNonVirtualFunctionInitializeParameter(init, f) and
call.getThisArgument() = instr and
init.getIRVariable() instanceof IRThisVariable
}
/** Holds if `instr` is a `this` pointer used by the call instruction `call`. */
predicate isSink(DataFlow::Node sink, CallInstruction call) {
predicate isSink(Instruction instr, CallInstruction call) {
exists(PureVirtualFunction func |
call.getStaticCallTarget() = func and
call.getThisArgument() = sink.asInstruction() and
call.getThisArgument() = instr and
// Weed out implicit calls to destructors of a base class
not func instanceof Destructor
)
}
/** Holds if `init` initializes the `this` pointer in class `c`. */
predicate isSource(DataFlow::Node source, string msg, Class c) {
exists(InitializeParameterInstruction init | init = source.asInstruction() |
(
exists(Constructor func |
not func instanceof CopyConstructor and
not func instanceof MoveConstructor and
func = init.getEnclosingFunction() and
msg = "construction"
)
or
init.getEnclosingFunction() instanceof Destructor and msg = "destruction"
) and
init.getIRVariable() instanceof IRThisVariable and
init.getEnclosingFunction().getDeclaringType() = c
predicate isSource(InitializeParameterInstruction init, string msg, Class c) {
(
exists(Constructor func |
not func instanceof CopyConstructor and
not func instanceof MoveConstructor and
func = init.getEnclosingFunction() and
msg = "construction"
)
or
init.getEnclosingFunction() instanceof Destructor and msg = "destruction"
) and
init.getIRVariable() instanceof IRThisVariable and
init.getEnclosingFunction().getDeclaringType() = c
}
/**
* Holds if `instr` flows to a sink (which is a use of the value of `instr` as a `this` pointer).
*/
predicate flowsToSink(Instruction instr, Instruction sink) {
flowsFromSource(instr) and
(
isSink(instr, _) and instr = sink
or
exists(Instruction mid |
successor(instr, mid) and
flowsToSink(mid, sink)
)
)
}
/** Holds if `instr` flows from a source. */
predicate flowsFromSource(Instruction instr) {
isSource(instr, _, _)
or
exists(Instruction mid |
successor(mid, instr) and
flowsFromSource(mid)
)
}
/** Holds if `f` is the enclosing non-virtual function of `init`. */
predicate getEnclosingNonVirtualFunctionInitializeParameter(
InitializeParameterInstruction init, Function f
) {
not f.isVirtual() and
init.getEnclosingFunction() = f
}
/** Holds if `f` is the enclosing non-virtual function of `init`. */
predicate getEnclosingNonVirtualFunctionInitializeIndirection(
InitializeIndirectionInstruction init, Function f
) {
not f.isVirtual() and
init.getEnclosingFunction() = f
}
/**
* Holds if `instr` is an argument (or argument indirection) to a call, and
* `succ` is the corresponding initialization instruction in the call target.
*/
predicate flowThroughCallable(Instruction instr, Instruction succ) {
// Flow from an argument to a parameter
exists(CallInstruction call, InitializeParameterInstruction init | init = succ |
getPositionalArgumentInitParam(call, instr, init, call.getStaticCallTarget())
or
getThisArgumentInitParam(call, instr, init, call.getStaticCallTarget())
)
or
// Flow from argument indirection to parameter indirection
exists(
CallInstruction call, ReadSideEffectInstruction read, InitializeIndirectionInstruction init
|
init = succ and
read.getPrimaryInstruction() = call and
getEnclosingNonVirtualFunctionInitializeIndirection(init, call.getStaticCallTarget())
|
exists(int n |
read.getSideEffectOperand().getAnyDef() = instr and
read.getIndex() = n and
init.getParameter().getIndex() = unbind(n)
)
or
call.getThisArgument() = instr and
init.getIRVariable() instanceof IRThisVariable
)
}
/** Holds if `instr` flows to `succ`. */
predicate successor(Instruction instr, Instruction succ) {
succ.(CopyInstruction).getSourceValue() = instr or
succ.(CheckedConvertOrNullInstruction).getUnary() = instr or
succ.(ChiInstruction).getTotal() = instr or
succ.(ConvertInstruction).getUnary() = instr or
succ.(InheritanceConversionInstruction).getUnary() = instr or
flowThroughCallable(instr, succ)
}
/**
* Holds if:
* - `source` is an initialization of a `this` pointer of type `sourceClass`, and
@@ -66,19 +188,22 @@ predicate isSource(DataFlow::Node source, string msg, Class c) {
* - `msg` is a string describing whether `source` is from a constructor or destructor.
*/
predicate flows(
MustFlowPathNode source, string msg, Class sourceClass, MustFlowPathNode sink,
CallInstruction call
Instruction source, string msg, Class sourceClass, Instruction sink, CallInstruction call
) {
exists(UnsafeUseOfThisConfig conf |
conf.hasFlowPath(source, sink) and
isSource(source.getNode(), msg, sourceClass) and
isSink(sink.getNode(), call)
)
isSource(source, msg, sourceClass) and
flowsToSink(source, sink) and
isSink(sink, call)
}
from
MustFlowPathNode source, MustFlowPathNode sink, CallInstruction call, string msg,
Class sourceClass
query predicate edges(Instruction a, Instruction b) { successor(a, b) and flowsToSink(b, _) }
query predicate nodes(Instruction n, string key, string val) {
flowsToSink(n, _) and
key = "semmle.label" and
val = n.toString()
}
from Instruction source, Instruction sink, CallInstruction call, string msg, Class sourceClass
where
flows(source, msg, sourceClass, sink, call) and
// Only raise an alert if there is no override of the pure virtual function in any base class.

View File

@@ -0,0 +1,6 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>
<include src="FLinesOfDuplicatedCodeCommon.inc.qhelp" />
</qhelp>

View File

@@ -0,0 +1,27 @@
/**
* @deprecated
* @name Duplicated lines in files
* @description The number of lines in a file, including code, comment
* and whitespace lines, which are duplicated in at least
* one other place.
* @kind treemap
* @treemap.warnOn highValues
* @metricType file
* @metricAggregate avg sum max
* @id cpp/duplicated-lines-in-files
* @tags testability
* modularity
*/
import external.CodeDuplication
from File f, int n
where
n =
count(int line |
exists(DuplicateBlock d | d.sourceFile() = f |
line in [d.sourceStartLine() .. d.sourceEndLine()]
) and
not whitelistedLineForDuplication(f, line)
)
select f, n order by n desc

View File

@@ -0,0 +1,35 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>
<overview>
<p>
This metric measures the number of lines in a file that are contained within a block that is duplicated elsewhere. These lines may include code, comments and whitespace, and the duplicate block may be in this file or in another file.
</p>
<p>
A file that contains many lines that are duplicated within the code base is problematic
for a number of reasons.
</p>
</overview>
<include src="DuplicationProblems.inc.qhelp" />
<recommendation>
<p>
Refactor files with lots of duplicated code to extract the common code into
a shared library or module.
</p>
</recommendation>
<references>
<li>Wikipedia: <a href="http://en.wikipedia.org/wiki/Duplicate_code">Duplicate code</a>.</li>
<li>M. Fowler, <em>Refactoring</em>. Addison-Wesley, 1999.</li>
</references>
</qhelp>

View File

@@ -12,33 +12,23 @@
*/
import cpp
import semmle.code.cpp.security.BufferWrite as BufferWrite
import semmle.code.cpp.security.BufferWrite
import semmle.code.cpp.security.TaintTracking
import semmle.code.cpp.security.SensitiveExprs
import semmle.code.cpp.security.FlowSources
import semmle.code.cpp.ir.dataflow.TaintTracking
import DataFlow::PathGraph
import TaintedWithPath
/**
* Taint flow from user input to a buffer write.
*/
class ToBufferConfiguration extends TaintTracking::Configuration {
ToBufferConfiguration() { this = "ToBufferConfiguration" }
override predicate isSource(DataFlow::Node source) { source instanceof FlowSource }
override predicate isSink(DataFlow::Node sink) {
exists(BufferWrite::BufferWrite w | w.getASource() = sink.asExpr())
}
class Configuration extends TaintTrackingConfiguration {
override predicate isSink(Element tainted) { exists(BufferWrite w | w.getASource() = tainted) }
}
from
ToBufferConfiguration config, BufferWrite::BufferWrite w, DataFlow::PathNode sourceNode,
DataFlow::PathNode sinkNode, FlowSource source, SensitiveExpr dest
BufferWrite w, Expr taintedArg, Expr taintSource, PathNode sourceNode, PathNode sinkNode,
string taintCause, SensitiveExpr dest
where
config.hasFlowPath(sourceNode, sinkNode) and
sourceNode.getNode() = source and
w.getASource() = sinkNode.getNode().asExpr() and
taintedWithPath(taintSource, taintedArg, sourceNode, sinkNode) and
isUserInput(taintSource, taintCause) and
w.getASource() = taintedArg and
dest = w.getDest()
select w, sourceNode, sinkNode,
"This write into buffer '" + dest.toString() + "' may contain unencrypted data from $@", source,
"user input (" + source.getSourceType() + ")"
"This write into buffer '" + dest.toString() + "' may contain unencrypted data from $@",
taintSource, "user input (" + taintCause + ")"

View File

@@ -65,7 +65,6 @@ where
midNode.getNode().asExpr() = mid and
mid = w.getASource() and
dest = w.getDest() and
not dest.(VariableAccess).getTarget().getName() = ["stdin", "stdout", "stderr"] and // exclude calls with standard streams
not isFileName(globalValueNumber(source)) and // file names are not passwords
not exists(string convChar | convChar = w.getSourceConvChar(mid) | not convChar = ["s", "S"]) // ignore things written with other conversion characters
select w, sourceNode, midNode,

View File

@@ -5,7 +5,7 @@
* @kind path-problem
* @problem.severity warning
* @security-severity 7.5
* @precision high
* @precision medium
* @id cpp/cleartext-transmission
* @tags security
* external/cwe/cwe-319
@@ -14,8 +14,8 @@
import cpp
import semmle.code.cpp.security.SensitiveExprs
import semmle.code.cpp.dataflow.TaintTracking
import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import semmle.code.cpp.models.interfaces.FlowSource
import semmle.code.cpp.commons.File
import DataFlow::PathGraph
/**
@@ -27,7 +27,6 @@ class SensitiveNode extends DataFlow::Node {
this.asExpr() = any(SensitiveVariable sv).getInitializer().getExpr() or
this.asExpr().(VariableAccess).getTarget() =
any(SensitiveVariable sv).(GlobalOrNamespaceVariable) or
this.asExpr().(VariableAccess).getTarget() = any(SensitiveVariable v | v instanceof Field) or
this.asUninitialized() instanceof SensitiveVariable or
this.asParameter() instanceof SensitiveVariable or
this.asExpr().(FunctionCall).getTarget() instanceof SensitiveFunction
@@ -59,10 +58,7 @@ class Send extends SendRecv instanceof RemoteFlowSinkFunction {
call.getTarget() = this and
exists(FunctionInput input, int arg |
super.hasSocketInput(input) and
(
input.isParameter(arg) or
input.isParameterDeref(arg)
) and
input.isParameter(arg) and
result = call.getArgument(arg)
)
}
@@ -85,10 +81,7 @@ class Recv extends SendRecv instanceof RemoteFlowSourceFunction {
call.getTarget() = this and
exists(FunctionInput input, int arg |
super.hasSocketInput(input) and
(
input.isParameter(arg) or
input.isParameterDeref(arg)
) and
input.isParameter(arg) and
result = call.getArgument(arg)
)
}
@@ -121,32 +114,24 @@ abstract class NetworkSendRecv extends FunctionCall {
NetworkSendRecv() {
this.getTarget() = target and
// exclude calls based on the socket...
not exists(DataFlow::Node src, DataFlow::Node dest |
DataFlow::localFlow(src, dest) and
dest.asExpr() = target.getSocketExpr(this) and
not exists(GVN g |
g = globalValueNumber(target.getSocketExpr(this)) and
(
// literal constant
src.asExpr() instanceof Literal
globalValueNumber(any(Literal l)) = g
or
// variable (such as a global) initialized to a literal constant
exists(Variable v |
v.getInitializer().getExpr() instanceof Literal and
src.asExpr() = v.getAnAccess()
g = globalValueNumber(v.getAnAccess())
)
or
// result of a function call with literal inputs (likely constant)
forex(Expr arg | arg = src.asExpr().(FunctionCall).getAnArgument() | arg instanceof Literal)
or
// variable called `stdin`, `stdout` or `stderr`
src.asExpr().(VariableAccess).getTarget().getName() = ["stdin", "stdout", "stderr"]
or
// open of `"/dev/tty"`
exists(FunctionCall fc |
fopenCall(fc) and
fc.getAnArgument().getValue() = "/dev/tty" and
src.asExpr() = fc
forex(Expr arg | arg = fc.getAnArgument() | arg instanceof Literal) and
g = globalValueNumber(fc)
)
// (this is not exhaustive)
// (this is far from exhaustive)
)
)
}
@@ -168,16 +153,6 @@ class NetworkRecv extends NetworkSendRecv {
override Recv target;
}
pragma[noinline]
predicate encryptionFunction(Function f) {
f.getName().toLowerCase().regexpMatch(".*(crypt|encode|decode|hash|securezero).*")
}
pragma[noinline]
predicate encryptionType(UserType t) {
t.getName().toLowerCase().regexpMatch(".*(crypt|encode|decode|hash|securezero).*")
}
/**
* An expression that is an argument or return value from an encryption /
* decryption call. This is quite inclusive to minimize false positives, for
@@ -187,7 +162,10 @@ predicate encryptionType(UserType t) {
class Encrypted extends Expr {
Encrypted() {
exists(FunctionCall fc |
encryptionFunction(fc.getTarget()) and
fc.getTarget()
.getName()
.toLowerCase()
.regexpMatch(".*(crypt|encode|decode|hash|securezero).*") and
(
this = fc or
this = fc.getAnArgument()
@@ -196,7 +174,7 @@ class Encrypted extends Expr {
or
exists(Type t |
this.getType().refersTo(t) and
encryptionType(t)
t.getName().toLowerCase().regexpMatch(".*(crypt|encode|decode|hash|securezero).*")
)
}
}

View File

@@ -3,7 +3,7 @@
* @description Non-HTTPS connections can be intercepted by third parties.
* @kind path-problem
* @problem.severity warning
* @precision high
* @precision medium
* @id cpp/non-https-url
* @tags security
* external/cwe/cwe-319
@@ -12,7 +12,6 @@
import cpp
import semmle.code.cpp.dataflow.TaintTracking
import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import DataFlow::PathGraph
/**
@@ -58,12 +57,7 @@ class HttpStringToUrlOpenConfig extends TaintTracking::Configuration {
override predicate isSource(DataFlow::Node src) {
// Sources are strings containing an HTTP URL not in a private domain.
src.asExpr() instanceof HttpStringLiteral and
// block taint starting at `strstr`, which is likely testing an existing URL, rather than constructing an HTTP URL.
not exists(FunctionCall fc |
fc.getTarget().getName() = ["strstr", "strcasestr"] and
fc.getArgument(1) = globalValueNumber(src.asExpr()).getAnExpr()
)
src.asExpr() instanceof HttpStringLiteral
}
override predicate isSink(DataFlow::Node sink) {

View File

@@ -1,8 +0,0 @@
void encrypt_with_openssl(EVP_PKEY_CTX *ctx) {
// BAD: only 1024 bits for an RSA key
EVP_PKEY_CTX_set_rsa_keygen_bits(ctx, 1024);
// GOOD: 2048 bits for an RSA key
EVP_PKEY_CTX_set_rsa_keygen_bits(ctx, 2048);
}

View File

@@ -1,38 +0,0 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>
<overview>
<p>Using cryptographic algorithms with a small key size can leave data vulnerable to being decrypted.</p>
<p>Many cryptographic algorithms provided by cryptography libraries can be configured with key sizes that are
vulnerable to brute force attacks. Using such a key size means that an attacker may be able to easily decrypt the
encrypted data.</p>
</overview>
<recommendation>
<p>Ensure that you use a strong, modern cryptographic algorithm. Use at least AES-128 or RSA-2048.</p>
</recommendation>
<example>
<p>The following code shows an example of using the <code>openssl</code> library to generate an RSA key.
When creating a key, you must specify which key size to use. The first example uses 1024 bits, which is not
considered sufficient. The second example uses 2048 bits, which is currently considered sufficient.</p>
<sample src="InsufficientKeySize.c" />
</example>
<references>
<li>NIST, FIPS 140 Annex a: <a href="http://csrc.nist.gov/publications/fips/fips140-2/fips1402annexa.pdf">
Approved Security Functions</a>.</li>
<li>NIST, SP 800-131A: <a href="https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-131Ar2.pdf">
Transitions: Recommendation for Transitioning the Use of Cryptographic Algorithms and Key Lengths</a>.</li>
<!-- LocalWords: CWE
-->
</references>
</qhelp>

View File

@@ -1,63 +0,0 @@
/**
* @name Use of a cryptographic algorithm with insufficient key size
* @description Using cryptographic algorithms with too small a key size can
* allow an attacker to compromise security.
* @kind path-problem
* @problem.severity error
* @precision high
* @id cpp/insufficient-key-size
* @tags security
* external/cwe/cwe-326
*/
import cpp
import semmle.code.cpp.ir.dataflow.DataFlow
import semmle.code.cpp.ir.IR
import DataFlow::PathGraph
// Gets the recommended minimum key size (in bits) of `func`, the name of an encryption function that accepts a key size as parameter `paramIndex`
int getMinimumKeyStrength(string func, int paramIndex) {
func =
[
"EVP_PKEY_CTX_set_dsa_paramgen_bits", "DSA_generate_parameters_ex",
"EVP_PKEY_CTX_set_rsa_keygen_bits", "RSA_generate_key_ex", "RSA_generate_key_fips",
"EVP_PKEY_CTX_set_dh_paramgen_prime_len", "DH_generate_parameters_ex"
] and
paramIndex = 1 and
result = 2048
}
class KeyStrengthFlow extends DataFlow::Configuration {
KeyStrengthFlow() { this = "KeyStrengthFlow" }
override predicate isSource(DataFlow::Node node) {
exists(int bits |
node.asInstruction().(IntegerConstantInstruction).getValue().toInt() = bits and
bits < getMinimumKeyStrength(_, _) and
bits > 0 // exclude sentinel values
)
}
override predicate isSink(DataFlow::Node node) {
exists(FunctionCall fc, string name, int param |
node.asExpr() = fc.getArgument(param) and
fc.getTarget().hasGlobalName(name) and
exists(getMinimumKeyStrength(name, param))
)
}
}
from
DataFlow::PathNode source, DataFlow::PathNode sink, KeyStrengthFlow conf, FunctionCall fc,
int param, string name, int minimumBits, int bits
where
conf.hasFlowPath(source, sink) and
sink.getNode().asExpr() = fc.getArgument(param) and
fc.getTarget().hasGlobalName(name) and
minimumBits = getMinimumKeyStrength(name, param) and
bits = source.getNode().asInstruction().(ConstantValueInstruction).getValue().toInt() and
bits < minimumBits and
bits != 0
select fc, source, sink,
"The key size $@ is less than the recommended key size of " + minimumBits.toString() + " bits.",
source, bits.toString()

View File

@@ -3,7 +3,7 @@
* @description Exposing system data or debugging information helps
* an adversary learn about the system and form an
* attack plan.
* @kind path-problem
* @kind problem
* @problem.severity warning
* @security-severity 6.5
* @precision medium
@@ -14,9 +14,7 @@
import cpp
import semmle.code.cpp.commons.Environment
import semmle.code.cpp.ir.dataflow.TaintTracking
import semmle.code.cpp.models.interfaces.FlowSource
import DataFlow::PathGraph
import semmle.code.cpp.security.OutputWrite
/**
* An element that should not be exposed to an adversary.
@@ -26,19 +24,42 @@ abstract class SystemData extends Element {
* Gets an expression that is part of this `SystemData`.
*/
abstract Expr getAnExpr();
/**
* Gets an expression whose value originates from, or is used by,
* this `SystemData`.
*/
Expr getAnExprIndirect() {
// direct SystemData
result = this.getAnExpr() or
// flow via global or member variable (conservative approximation)
result = this.getAnAffectedVar().getAnAccess() or
// flow via stack variable
definitionUsePair(_, this.getAnExprIndirect(), result) or
useUsePair(_, this.getAnExprIndirect(), result) or
useUsePair(_, result, this.getAnExprIndirect()) or
// flow from assigned value to assignment expression
result.(AssignExpr).getRValue() = this.getAnExprIndirect()
}
/**
* Gets a global or member variable that may be affected by this system
* data (conservative approximation).
*/
private Variable getAnAffectedVar() {
(
result.getAnAssignedValue() = this.getAnExprIndirect() or
result.getAnAccess() = this.getAnExprIndirect()
) and
not result instanceof LocalScopeVariable
}
}
/**
* Data originating from the environment.
*/
class EnvData extends SystemData {
EnvData() {
// identify risky looking environment variables only
this.(EnvironmentRead)
.getEnvironmentVariable()
.toLowerCase()
.regexpMatch(".*(user|host|admin|root|home|path|http|ssl|snmp|sock|port|proxy|pass|token|crypt|key).*")
}
EnvData() { this instanceof EnvironmentRead }
override Expr getAnExpr() { result = this }
}
@@ -70,6 +91,11 @@ class SQLConnectInfo extends SystemData {
}
private predicate posixSystemInfo(FunctionCall source, Element use) {
// long sysconf(int name)
// - various OS / system values and limits
source.getTarget().hasName("sysconf") and
use = source
or
// size_t confstr(int name, char *buf, size_t len)
// - various OS / system strings, such as the libc version
// int statvfs(const char *__path, struct statvfs *__buf)
@@ -285,31 +311,70 @@ class RegQuery extends SystemData {
override Expr getAnExpr() { regQuery(this, result) }
}
class ExposedSystemDataConfiguration extends TaintTracking::Configuration {
ExposedSystemDataConfiguration() { this = "ExposedSystemDataConfiguration" }
override predicate isSource(DataFlow::Node source) {
source.asConvertedExpr() = any(SystemData sd).getAnExpr()
}
override predicate isSink(DataFlow::Node sink) {
exists(FunctionCall fc, FunctionInput input, int arg |
fc.getTarget().(RemoteFlowSinkFunction).hasRemoteFlowSink(input, _) and
input.isParameterDeref(arg) and
fc.getArgument(arg).getAChild*() = sink.asExpr()
)
}
/**
* Somewhere data is output.
*/
abstract class DataOutput extends Element {
/**
* Get an expression containing data that is output.
*/
abstract Expr getASource();
}
from ExposedSystemDataConfiguration config, DataFlow::PathNode source, DataFlow::PathNode sink
where
config.hasFlowPath(source, sink) and
not exists(
DataFlow::Node alt // remove duplicate results on conversions
|
config.hasFlow(source.getNode(), alt) and
alt.asConvertedExpr() = sink.getNode().asExpr() and
alt != sink.getNode()
/**
* Data that is output via standard output or standard error.
*/
class StandardOutput extends DataOutput instanceof OutputWrite {
override Expr getASource() { result = OutputWrite.super.getASource() }
}
private predicate socketCallOrIndirect(FunctionCall call) {
// direct socket call
// int socket(int domain, int type, int protocol);
call.getTarget().getName() = "socket"
or
exists(ReturnStmt rtn |
// indirect socket call
call.getTarget() = rtn.getEnclosingFunction() and
(
socketCallOrIndirect(rtn.getExpr()) or
socketCallOrIndirect(rtn.getExpr().(VariableAccess).getTarget().getAnAssignedValue())
)
)
select sink, source, sink, "This operation exposes system data from $@.", source,
source.getNode().toString()
}
private predicate socketFileDescriptor(Expr e) {
exists(Variable var, FunctionCall socket |
socketCallOrIndirect(socket) and
var.getAnAssignedValue() = socket and
e = var.getAnAccess()
)
}
private predicate socketOutput(FunctionCall call, Expr data) {
(
// ssize_t send(int sockfd, const void *buf, size_t len, int flags);
// ssize_t sendto(int sockfd, const void *buf, size_t len, int flags,
// const struct sockaddr *dest_addr, socklen_t addrlen);
// ssize_t sendmsg(int sockfd, const struct msghdr *msg, int flags);
// int write(int handle, void *buffer, int nbyte);
call.getTarget().hasGlobalName(["send", "sendto", "sendmsg", "write"]) and
data = call.getArgument(1) and
socketFileDescriptor(call.getArgument(0))
)
}
/**
* Data that is output via a socket.
*/
class SocketOutput extends DataOutput {
SocketOutput() { socketOutput(this, _) }
override Expr getASource() { socketOutput(this, result) }
}
from SystemData sd, DataOutput ow
where
sd.getAnExprIndirect() = ow.getASource() or
sd.getAnExprIndirect() = ow.getASource().getAChild*()
select ow, "This operation exposes system data from $@.", sd, sd.toString()

View File

@@ -12,16 +12,17 @@
import cpp
import FilePermissions
import semmle.code.cpp.commons.unix.Constants
predicate worldWritableCreation(FileCreationExpr fc, int mode) {
mode = localUmask(fc).mask(fc.getMode()) and
setsAnyBits(mode, UnixConstants::s_iwoth())
sets(mode, s_iwoth())
}
predicate setWorldWritable(FunctionCall fc, int mode) {
fc.getTarget().getName() = ["chmod", "fchmod", "_chmod", "_wchmod"] and
mode = fc.getArgument(1).getValue().toInt() and
setsAnyBits(mode, UnixConstants::s_iwoth())
sets(mode, s_iwoth())
}
from Expr fc, int mode, string message

View File

@@ -1,49 +1,5 @@
import cpp
import semmle.code.cpp.commons.unix.Constants as UnixConstants
/**
* Gets the number corresponding to the contents of `input` in base-16.
* Note: the first two characters of `input` must be `0x`. For example:
* `parseHex("0x123abc") = 1194684`.
*/
bindingset[input]
int parseHex(string input) {
exists(string lowerCaseInput | lowerCaseInput = input.toLowerCase() |
lowerCaseInput.regexpMatch("0x[0-9a-f]+") and
result =
strictsum(int ix |
ix in [2 .. input.length()]
|
16.pow(input.length() - (ix + 1)) * "0123456789abcdef".indexOf(lowerCaseInput.charAt(ix))
)
)
}
/**
* Gets the value defined by the `O_CREAT` macro if the macro
* exists and if every definition defines the same value.
*/
int o_creat() {
result =
unique(int v |
exists(Macro m | m.getName() = "O_CREAT" |
v = parseHex(m.getBody()) or v = UnixConstants::parseOctal(m.getBody())
)
)
}
/**
* Gets the value defined by the `O_TMPFILE` macro if the macro
* exists and if every definition defines the same value.
*/
int o_tmpfile() {
result =
unique(int v |
exists(Macro m | m.getName() = "O_TMPFILE" |
v = parseHex(m.getBody()) or v = UnixConstants::parseOctal(m.getBody())
)
)
}
import semmle.code.cpp.commons.unix.Constants
bindingset[n, digit]
private string octalDigit(int n, int digit) {
@@ -64,17 +20,11 @@ string octalFileMode(int mode) {
else result = "[non-standard mode: decimal " + mode + "]"
}
/**
* Holds if the bitmask `value` sets the bits in `flag`.
*/
bindingset[value, flag]
predicate setsFlag(int value, int flag) { value.bitAnd(flag) = flag }
/**
* Holds if the bitmask `mask` sets any of the bit fields in `fields`.
*/
bindingset[mask, fields]
predicate setsAnyBits(int mask, int fields) { mask.bitAnd(fields) != 0 }
predicate sets(int mask, int fields) { mask.bitAnd(fields) != 0 }
/**
* Gets the value that `fc` sets the umask to, if `fc` is a call to
@@ -133,24 +83,16 @@ abstract class FileCreationExpr extends FunctionCall {
abstract int getMode();
}
abstract class FileCreationWithOptionalModeExpr extends FileCreationExpr {
abstract predicate hasModeArgument();
}
class OpenCreationExpr extends FileCreationWithOptionalModeExpr {
class OpenCreationExpr extends FileCreationExpr {
OpenCreationExpr() {
this.getTarget().hasGlobalOrStdName(["open", "_open", "_wopen"]) and
exists(int flag | flag = this.getArgument(1).getValue().toInt() |
setsFlag(flag, o_creat()) or setsFlag(flag, o_tmpfile())
)
this.getTarget().getName() = ["open", "_open", "_wopen"] and
sets(this.getArgument(1).getValue().toInt(), o_creat())
}
override Expr getPath() { result = this.getArgument(0) }
override predicate hasModeArgument() { exists(this.getArgument(2)) }
override int getMode() {
if this.hasModeArgument()
if exists(this.getArgument(2))
then result = this.getArgument(2).getValue().toInt()
else
// assume anything is permitted
@@ -166,35 +108,20 @@ class CreatCreationExpr extends FileCreationExpr {
override int getMode() { result = this.getArgument(1).getValue().toInt() }
}
class OpenatCreationExpr extends FileCreationWithOptionalModeExpr {
class OpenatCreationExpr extends FileCreationExpr {
OpenatCreationExpr() {
this.getTarget().hasGlobalOrStdName("openat") and
exists(int flag | flag = this.getArgument(2).getValue().toInt() |
setsFlag(flag, o_creat()) or setsFlag(flag, o_tmpfile())
)
this.getTarget().getName() = "openat" and
this.getNumberOfArguments() = 4
}
override Expr getPath() { result = this.getArgument(1) }
override predicate hasModeArgument() { exists(this.getArgument(3)) }
override int getMode() {
if this.hasModeArgument()
then result = this.getArgument(3).getValue().toInt()
else
// assume anything is permitted
result = 0.bitNot()
}
override int getMode() { result = this.getArgument(3).getValue().toInt() }
}
private int fopenMode() {
result =
UnixConstants::s_irusr()
.bitOr(UnixConstants::s_irgrp())
.bitOr(UnixConstants::s_iroth())
.bitOr(UnixConstants::s_iwusr())
.bitOr(UnixConstants::s_iwgrp())
.bitOr(UnixConstants::s_iwoth())
s_irusr().bitOr(s_irgrp()).bitOr(s_iroth()).bitOr(s_iwusr()).bitOr(s_iwgrp()).bitOr(s_iwoth())
}
class FopenCreationExpr extends FileCreationExpr {
@@ -226,6 +153,6 @@ class FopensCreationExpr extends FileCreationExpr {
// fopen_s has restrictive permissions unless you have "u" in the mode
if this.getArgument(2).getValue().charAt(_) = "u"
then result = fopenMode()
else result = UnixConstants::s_irusr().bitOr(UnixConstants::s_iwusr())
else result = s_irusr().bitOr(s_iwusr())
}
}

View File

@@ -1,9 +0,0 @@
int open_file_bad() {
// BAD - this uses arbitrary bytes from the stack as mode argument
return open(FILE, O_CREAT)
}
int open_file_good() {
// GOOD - the mode argument is supplied
return open(FILE, O_CREAT, S_IRUSR | S_IWUSR)
}

View File

@@ -1,31 +0,0 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>
<overview>
<p>
When opening a file with the <code>O_CREAT</code> or <code>O_TMPFILE</code> flag, the <code>mode</code> must
be supplied. If the <code>mode</code> argument is omitted, some arbitrary bytes from the stack will be used
as the file mode. This leaks some bits from the stack into the permissions of the file.
</p>
</overview>
<recommendation>
<p>
The <code>mode</code> must be supplied when <code>O_CREAT</code> or <code>O_TMPFILE</code> is specified.
</p>
</recommendation>
<example>
<p>
The first example opens a file with the <code>O_CREAT</code> flag without supplying the <code>mode</code>
argument. In this case arbitrary bytes from the stack will be used as <code>mode</code> argument. The
second example correctly supplies the <code>mode</code> argument and creates a file that is user readable
and writable.
</p>
<sample src="OpenCallMissingModeArgument.c" />
</example>
</qhelp>

View File

@@ -1,19 +0,0 @@
/**
* @name File opened with O_CREAT flag but without mode argument
* @description Opening a file with the O_CREAT flag but without mode argument reads arbitrary bytes from the stack.
* @kind problem
* @problem.severity error
* @security-severity 7.8
* @precision high
* @id cpp/open-call-with-mode-argument
* @tags security
* external/cwe/cwe-732
*/
import cpp
import FilePermissions
from FileCreationWithOptionalModeExpr fc
where not fc.hasModeArgument()
select fc,
"A file is created here without providing a mode argument, which may leak bits from the stack."

View File

@@ -0,0 +1,4 @@
---
category: newQuery
---
* A new `cpp/very-likely-overruning-write` query has been added to the default query suite for C/C++. The query reports some results that were formerly flagged by `cpp/overruning-write`.

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