Better words (#1497)

* Fix recv_error receipt limit allowance

* backport #1463 recv_error behavior changes

---------

Co-authored-by: JackDoan <me@jackdoan.com>

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -0,0 +1,69 @@
+name: "\U0001F41B Bug Report"
+description: Report an issue or possible bug
+title: "\U0001F41B BUG:"
+labels: []
+assignees: []
+body:
+  - type: markdown
+    attributes:
+      value: |
+        ### Thank you for taking the time to file a bug report!
+
+        Please fill out this form as completely as possible.
+
+  - type: input
+    id: version
+    attributes:
+      label: What version of `nebula` are you using? (`nebula -version`)
+      placeholder: 0.0.0
+    validations:
+      required: true
+
+  - type: input
+    id: os
+    attributes:
+      label: What operating system are you using?
+      description: iOS and Android specific issues belong in the [mobile_nebula](https://github.com/DefinedNet/mobile_nebula) repo.
+      placeholder: Linux, Mac, Windows
+    validations:
+      required: true
+
+  - type: textarea
+    id: description
+    attributes:
+      label: Describe the Bug
+      description: A clear and concise description of what the bug is.
+    validations:
+      required: true
+
+  - type: textarea
+    id: logs
+    attributes:
+      label: Logs from affected hosts
+      description: |
+        Please provide logs from ALL affected hosts during the time of the issue. If you do not provide logs we will be unable to assist you!
+
+        [Learn how to find Nebula logs here.](https://nebula.defined.net/docs/guides/viewing-nebula-logs/)
+
+        Improve formatting by using <code>```</code> at the beginning and end of each log block.
+      value: |
+        ```
+
+        ```
+    validations:
+      required: true
+
+  - type: textarea
+    id: configs
+    attributes:
+      label: Config files from affected hosts
+      description: |
+        Provide config files for all affected hosts.
+
+        Improve formatting by using <code>```</code> at the beginning and end of each config file.
+      value: |
+        ```
+
+        ```
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,13 @@
+blank_issues_enabled: true
+contact_links:
+  - name: 📘 Documentation
+    url: https://nebula.defined.net/docs/
+    about: Review documentation.
+
+  - name: 💁 Support/Chat
+    url: https://join.slack.com/t/nebulaoss/shared_invite/enQtOTA5MDI4NDg3MTg4LTkwY2EwNTI4NzQyMzc0M2ZlODBjNWI3NTY1MzhiOThiMmZlZjVkMTI0NGY4YTMyNjUwMWEyNzNkZTJmYzQxOGU
+    about: 'This issue tracker is not for support questions. Join us on Slack for assistance!'
+
+  - name: 📱 Mobile Nebula
+    url: https://github.com/definednet/mobile_nebula
+    about: 'This issue tracker is not for mobile support. Try the Mobile Nebula repo instead!'

.github/dependabot.yml

@@ -0,0 +1,22 @@
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"
+
+  - package-ecosystem: "gomod"
+    directory: "/"
+    schedule:
+      interval: "weekly"
+    groups:
+      golang-x-dependencies:
+        patterns:
+          - "golang.org/x/*"
+      zx2c4-dependencies:
+        patterns:
+          - "golang.zx2c4.com/*"
+      protobuf-dependencies:
+        patterns:
+          - "github.com/golang/protobuf"
+          - "google.golang.org/protobuf"

.github/workflows/gofmt.yml

@@ -14,19 +14,21 @@ jobs:
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.14
-      uses: actions/setup-go@v1
-      with:
-        go-version: 1.14
-      id: go
+    - uses: actions/checkout@v4
 
-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
+    - uses: actions/setup-go@v5
+      with:
+        go-version: '1.22'
+        check-latest: true
+
+    - name: Install goimports
+      run: |
+        go install golang.org/x/tools/cmd/goimports@latest
 
     - name: gofmt
       run: |
-        if [ "$(find . -iname '*.go' | xargs gofmt -l)" ]
+        if [ "$(find . -iname '*.go' | grep -v '\.pb\.go$' | xargs goimports -l)" ]
         then
-          find . -iname '*.go' | xargs gofmt -d
+          find . -iname '*.go' | grep -v '\.pb\.go$' | xargs goimports -d
           exit 1
         fi

.github/workflows/release.yml

@@ -7,274 +7,212 @@ name: Create release and upload binaries
 
 jobs:
   build-linux:
-    name: Build Linux All
+    name: Build Linux/BSD All
     runs-on: ubuntu-latest
     steps:
-      - name: Set up Go 1.14
-        uses: actions/setup-go@v1
-        with:
-          go-version: 1.14
+      - uses: actions/checkout@v4
 
-      - name: Checkout code
-        uses: actions/checkout@v2
+      - uses: actions/setup-go@v5
+        with:
+          go-version: '1.22'
+          check-latest: true
 
       - name: Build
         run: |
-          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" release-linux
+          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" release-linux release-freebsd release-openbsd release-netbsd
           mkdir release
           mv build/*.tar.gz release
 
       - name: Upload artifacts
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v4
         with:
           name: linux-latest
           path: release
 
   build-windows:
-    name: Build Windows amd64
+    name: Build Windows
     runs-on: windows-latest
     steps:
-      - name: Set up Go 1.14
-        uses: actions/setup-go@v1
-        with:
-          go-version: 1.14
+      - uses: actions/checkout@v4
 
-      - name: Checkout code
-        uses: actions/checkout@v2
+      - uses: actions/setup-go@v5
+        with:
+          go-version: '1.22'
+          check-latest: true
 
       - name: Build
         run: |
           echo $Env:GITHUB_REF.Substring(11)
-          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\nebula.exe ./cmd/nebula-service
-          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\nebula-cert.exe ./cmd/nebula-cert
+          mkdir build\windows-amd64
+          $Env:GOARCH = "amd64"
+          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\windows-amd64\nebula.exe ./cmd/nebula-service
+          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\windows-amd64\nebula-cert.exe ./cmd/nebula-cert
+          mkdir build\windows-arm64
+          $Env:GOARCH = "arm64"
+          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\windows-arm64\nebula.exe ./cmd/nebula-service
+          go build -trimpath -ldflags "-X main.Build=$($Env:GITHUB_REF.Substring(11))" -o build\windows-arm64\nebula-cert.exe ./cmd/nebula-cert
+          mkdir build\dist\windows
+          mv dist\windows\wintun build\dist\windows\
 
       - name: Upload artifacts
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v4
         with:
           name: windows-latest
           path: build
 
   build-darwin:
-    name: Build Darwin amd64
-    runs-on: macOS-latest
+    name: Build Universal Darwin
+    env:
+      HAS_SIGNING_CREDS: ${{ secrets.AC_USERNAME != '' }}
+    runs-on: macos-latest
     steps:
-      - name: Set up Go 1.14
-        uses: actions/setup-go@v1
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-go@v5
         with:
-          go-version: 1.14
+          go-version: '1.22'
+          check-latest: true
 
-      - name: Checkout code
-        uses: actions/checkout@v2
+      - name: Import certificates
+        if: env.HAS_SIGNING_CREDS == 'true'
+        uses: Apple-Actions/import-codesign-certs@v3
+        with:
+          p12-file-base64: ${{ secrets.APPLE_DEVELOPER_CERTIFICATE_P12_BASE64 }}
+          p12-password: ${{ secrets.APPLE_DEVELOPER_CERTIFICATE_PASSWORD }}
 
-      - name: Build
+      - name: Build, sign, and notarize
+        env:
+          AC_USERNAME: ${{ secrets.AC_USERNAME }}
+          AC_PASSWORD: ${{ secrets.AC_PASSWORD }}
         run: |
-          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" service build/nebula-darwin-amd64.tar.gz
+          rm -rf release
           mkdir release
-          mv build/*.tar.gz release
+          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" service build/darwin-amd64/nebula build/darwin-amd64/nebula-cert
+          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" service build/darwin-arm64/nebula build/darwin-arm64/nebula-cert
+          lipo -create -output ./release/nebula ./build/darwin-amd64/nebula ./build/darwin-arm64/nebula
+          lipo -create -output ./release/nebula-cert ./build/darwin-amd64/nebula-cert ./build/darwin-arm64/nebula-cert
+
+          if [ -n "$AC_USERNAME" ]; then
+            codesign -s "10BC1FDDEB6CE753550156C0669109FAC49E4D1E" -f -v --timestamp --options=runtime -i "net.defined.nebula" ./release/nebula
+            codesign -s "10BC1FDDEB6CE753550156C0669109FAC49E4D1E" -f -v --timestamp --options=runtime -i "net.defined.nebula-cert" ./release/nebula-cert
+          fi
+
+          zip -j release/nebula-darwin.zip release/nebula-cert release/nebula
+
+          if [ -n "$AC_USERNAME" ]; then
+            xcrun notarytool submit ./release/nebula-darwin.zip --team-id "576H3XS7FP" --apple-id "$AC_USERNAME" --password "$AC_PASSWORD" --wait
+          fi
 
       - name: Upload artifacts
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v4
         with:
           name: darwin-latest
-          path: release
+          path: ./release/*
+
+  build-docker:
+    name: Create and Upload Docker Images
+    # Technically we only need build-linux to succeed, but if any platforms fail we'll
+    # want to investigate and restart the build
+    needs: [build-linux, build-darwin, build-windows]
+    runs-on: ubuntu-latest
+    env:
+      HAS_DOCKER_CREDS: ${{ vars.DOCKERHUB_USERNAME != '' && secrets.DOCKERHUB_TOKEN != '' }}
+    # XXX It's not possible to write a conditional here, so instead we do it on every step
+    #if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+    steps:
+      # Be sure to checkout the code before downloading artifacts, or they will
+      # be overwritten
+      - name: Checkout code
+        if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+        uses: actions/checkout@v4
+
+      - name: Download artifacts
+        if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+        uses: actions/download-artifact@v4
+        with:
+          name: linux-latest
+          path: artifacts
+
+      - name: Login to Docker Hub
+        if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+        uses: docker/login-action@v3
+        with:
+          username: ${{ vars.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Set up Docker Buildx
+        if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build and push images
+        if: ${{ env.HAS_DOCKER_CREDS == 'true' }}
+        env:
+          DOCKER_IMAGE_REPO: ${{ vars.DOCKER_IMAGE_REPO || 'nebulaoss/nebula' }}
+          DOCKER_IMAGE_TAG: ${{ vars.DOCKER_IMAGE_TAG || 'latest' }}
+        run: |
+          mkdir -p build/linux-{amd64,arm64}
+          tar -zxvf artifacts/nebula-linux-amd64.tar.gz -C build/linux-amd64/
+          tar -zxvf artifacts/nebula-linux-arm64.tar.gz -C build/linux-arm64/
+          docker buildx build . --push -f docker/Dockerfile --platform linux/amd64,linux/arm64 --tag "${DOCKER_IMAGE_REPO}:${DOCKER_IMAGE_TAG}" --tag "${DOCKER_IMAGE_REPO}:${GITHUB_REF#refs/tags/v}"
 
   release:
     name: Create and Upload Release
     needs: [build-linux, build-darwin, build-windows]
     runs-on: ubuntu-latest
     steps:
-      - name: Download Linux artifacts
-        uses: actions/download-artifact@v1
-        with:
-          name: linux-latest
+      - uses: actions/checkout@v4
 
-      - name: Download Darwin artifacts
-        uses: actions/download-artifact@v1
+      - name: Download artifacts
+        uses: actions/download-artifact@v4
         with:
-          name: darwin-latest
-
-      - name: Download Windows artifacts
-        uses: actions/download-artifact@v1
-        with:
-          name: windows-latest
+          path: artifacts
 
       - name: Zip Windows
         run: |
-          cd windows-latest
-          zip nebula-windows-amd64.zip nebula.exe nebula-cert.exe
+          cd artifacts/windows-latest
+          cp windows-amd64/* .
+          zip -r nebula-windows-amd64.zip nebula.exe nebula-cert.exe dist
+          cp windows-arm64/* .
+          zip -r nebula-windows-arm64.zip nebula.exe nebula-cert.exe dist
 
       - name: Create sha256sum
         run: |
+          cd artifacts
           for dir in linux-latest darwin-latest windows-latest
           do
             (
               cd $dir
               if [ "$dir" = windows-latest ]
               then
-                sha256sum <nebula.exe | sed 's=-$=nebula-windows-amd64.zip/nebula.exe='
-                sha256sum <nebula-cert.exe | sed 's=-$=nebula-windows-amd64.zip/nebula-cert.exe='
+                sha256sum <windows-amd64/nebula.exe | sed 's=-$=nebula-windows-amd64.zip/nebula.exe='
+                sha256sum <windows-amd64/nebula-cert.exe | sed 's=-$=nebula-windows-amd64.zip/nebula-cert.exe='
+                sha256sum <windows-arm64/nebula.exe | sed 's=-$=nebula-windows-arm64.zip/nebula.exe='
+                sha256sum <windows-arm64/nebula-cert.exe | sed 's=-$=nebula-windows-arm64.zip/nebula-cert.exe='
                 sha256sum nebula-windows-amd64.zip
+                sha256sum nebula-windows-arm64.zip
+              elif [ "$dir" = darwin-latest ]
+              then
+                sha256sum <nebula-darwin.zip | sed 's=-$=nebula-darwin.zip='
+                sha256sum <nebula | sed 's=-$=nebula-darwin.zip/nebula='
+                sha256sum <nebula-cert | sed 's=-$=nebula-darwin.zip/nebula-cert='
               else
-                  for v in *.tar.gz
-                  do
-                    sha256sum $v
-                    tar zxf $v --to-command='sh -c "sha256sum | sed s=-$='$v'/$TAR_FILENAME="'
-                  done
+                for v in *.tar.gz
+                do
+                  sha256sum $v
+                  tar zxf $v --to-command='sh -c "sha256sum | sed s=-$='$v'/$TAR_FILENAME="'
+                done
               fi
             )
           done | sort -k 2 >SHASUM256.txt
 
       - name: Create Release
         id: create_release
-        uses: actions/create-release@v1
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          tag_name: ${{ github.ref }}
-          release_name: Release ${{ github.ref }}
-          draft: false
-          prerelease: false
-
-      ##
-      ## Upload assets (I wish we could just upload the whole folder at once...
-      ##
-
-      - name: Upload SHASUM256.txt
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./SHASUM256.txt
-          asset_name: SHASUM256.txt
-          asset_content_type: text/plain
-
-      - name: Upload darwin-amd64
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./darwin-latest/nebula-darwin-amd64.tar.gz
-          asset_name: nebula-darwin-amd64.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload windows-amd64
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./windows-latest/nebula-windows-amd64.zip
-          asset_name: nebula-windows-amd64.zip
-          asset_content_type: application/zip
-
-      - name: Upload linux-amd64
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-amd64.tar.gz
-          asset_name: nebula-linux-amd64.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-386
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-386.tar.gz
-          asset_name: nebula-linux-386.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-ppc64le
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-ppc64le.tar.gz
-          asset_name: nebula-linux-ppc64le.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-arm-5
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-arm-5.tar.gz
-          asset_name: nebula-linux-arm-5.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-arm-6
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-arm-6.tar.gz
-          asset_name: nebula-linux-arm-6.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-arm-7
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-arm-7.tar.gz
-          asset_name: nebula-linux-arm-7.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-arm64
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-arm64.tar.gz
-          asset_name: nebula-linux-arm64.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-mips
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-mips.tar.gz
-          asset_name: nebula-linux-mips.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-mipsle
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-mipsle.tar.gz
-          asset_name: nebula-linux-mipsle.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-mips64
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-mips64.tar.gz
-          asset_name: nebula-linux-mips64.tar.gz
-          asset_content_type: application/gzip
-
-      - name: Upload linux-mips64le
-        uses: actions/upload-release-asset@v1.0.1
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          upload_url: ${{ steps.create_release.outputs.upload_url }}
-          asset_path: ./linux-latest/nebula-linux-mips64le.tar.gz
-          asset_name: nebula-linux-mips64le.tar.gz
-          asset_content_type: application/gzip
+        run: |
+          cd artifacts
+          gh release create \
+            --verify-tag \
+            --title "Release ${{ github.ref_name }}" \
+            "${{ github.ref_name }}" \
+            SHASUM256.txt *-latest/*.zip *-latest/*.tar.gz

.github/workflows/smoke-extra.yml

@@ -0,0 +1,48 @@
+name: smoke-extra
+on:
+  push:
+    branches:
+      - master
+  pull_request:
+    types: [opened, synchronize, labeled, reopened]
+    paths:
+      - '.github/workflows/smoke**'
+      - '**Makefile'
+      - '**.go'
+      - '**.proto'
+      - 'go.mod'
+      - 'go.sum'
+jobs:
+
+  smoke-extra:
+    if: github.ref == 'refs/heads/master' || contains(github.event.pull_request.labels.*.name, 'smoke-test-extra')
+    name: Run extra smoke tests
+    runs-on: ubuntu-latest
+    steps:
+
+    - uses: actions/checkout@v4
+
+    - uses: actions/setup-go@v5
+      with:
+        go-version-file: 'go.mod'
+        check-latest: true
+
+    - name: install vagrant
+      run: sudo apt-get update && sudo apt-get install -y vagrant virtualbox
+
+    - name: freebsd-amd64
+      run: make smoke-vagrant/freebsd-amd64
+
+    - name: openbsd-amd64
+      run: make smoke-vagrant/openbsd-amd64
+
+    - name: netbsd-amd64
+      run: make smoke-vagrant/netbsd-amd64
+
+    - name: linux-386
+      run: make smoke-vagrant/linux-386
+
+    - name: linux-amd64-ipv6disable
+      run: make smoke-vagrant/linux-amd64-ipv6disable
+
+    timeout-minutes: 30

.github/workflows/smoke.yml

@@ -14,28 +14,19 @@ on:
 jobs:
 
   smoke:
-    name: Run 3 node smoke test
+    name: Run multi node smoke test
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.14
-      uses: actions/setup-go@v1
-      with:
-        go-version: 1.14
-      id: go
+    - uses: actions/checkout@v4
 
-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
-
-    - uses: actions/cache@v1
+    - uses: actions/setup-go@v5
       with:
-        path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-        restore-keys: |
-          ${{ runner.os }}-go-
+        go-version: '1.22'
+        check-latest: true
 
     - name: build
-      run: make
+      run: make bin-docker CGO_ENABLED=1 BUILD_ARGS=-race
 
     - name: setup docker image
       working-directory: ./.github/workflows/smoke
@@ -45,4 +36,20 @@ jobs:
       working-directory: ./.github/workflows/smoke
       run: ./smoke.sh
 
+    - name: setup relay docker image
+      working-directory: ./.github/workflows/smoke
+      run: ./build-relay.sh
+
+    - name: run smoke relay
+      working-directory: ./.github/workflows/smoke
+      run: ./smoke-relay.sh
+
+    - name: setup docker image for P256
+      working-directory: ./.github/workflows/smoke
+      run: NAME="smoke-p256" CURVE=P256 ./build.sh
+
+    - name: run smoke-p256
+      working-directory: ./.github/workflows/smoke
+      run: NAME="smoke-p256" ./smoke.sh
+
     timeout-minutes: 10

.github/workflows/smoke/Dockerfile

@@ -1,5 +1,9 @@
-FROM debian:buster
+FROM ubuntu:jammy
 
-ADD ./build /
+RUN apt-get update && apt-get install -y iputils-ping ncat tcpdump
 
-ENTRYPOINT ["/nebula"]
+ADD ./build /nebula
+
+WORKDIR /nebula
+
+ENTRYPOINT ["/nebula/nebula"]

.github/workflows/smoke/build-relay.sh

@@ -0,0 +1,44 @@
+#!/bin/sh
+
+set -e -x
+
+rm -rf ./build
+mkdir ./build
+
+(
+    cd build
+
+    cp ../../../../build/linux-amd64/nebula .
+    cp ../../../../build/linux-amd64/nebula-cert .
+
+    HOST="lighthouse1" AM_LIGHTHOUSE=true ../genconfig.sh >lighthouse1.yml <<EOF
+relay:
+  am_relay: true
+EOF
+
+    export LIGHTHOUSES="192.168.100.1 172.17.0.2:4242"
+    export REMOTE_ALLOW_LIST='{"172.17.0.4/32": false, "172.17.0.5/32": false}'
+
+    HOST="host2" ../genconfig.sh >host2.yml <<EOF
+relay:
+  relays:
+    - 192.168.100.1
+EOF
+
+    export REMOTE_ALLOW_LIST='{"172.17.0.3/32": false}'
+
+    HOST="host3" ../genconfig.sh >host3.yml
+
+    HOST="host4" ../genconfig.sh >host4.yml <<EOF
+relay:
+  use_relays: false
+EOF
+
+    ../../../../nebula-cert ca -name "Smoke Test"
+    ../../../../nebula-cert sign -name "lighthouse1" -groups "lighthouse,lighthouse1" -ip "192.168.100.1/24"
+    ../../../../nebula-cert sign -name "host2" -groups "host,host2" -ip "192.168.100.2/24"
+    ../../../../nebula-cert sign -name "host3" -groups "host,host3" -ip "192.168.100.3/24"
+    ../../../../nebula-cert sign -name "host4" -groups "host,host4" -ip "192.168.100.4/24"
+)
+
+docker build -t nebula:smoke-relay .

.github/workflows/smoke/build.sh

@@ -8,17 +8,37 @@ mkdir ./build
 (
     cd build
 
-    cp ../../../../nebula .
-    cp ../../../../nebula-cert .
+    cp ../../../../build/linux-amd64/nebula .
+    cp ../../../../build/linux-amd64/nebula-cert .
 
-    HOST="lighthouse1" AM_LIGHTHOUSE=true ../genconfig.sh >lighthouse1.yml
-    HOST="host2" LIGHTHOUSES="192.168.100.1 172.17.0.2:4242" ../genconfig.sh >host2.yml
-    HOST="host3" LIGHTHOUSES="192.168.100.1 172.17.0.2:4242" ../genconfig.sh >host3.yml
+    if [ "$1" ]
+    then
+        cp "../../../../build/$1/nebula" "$1-nebula"
+    fi
 
-    ./nebula-cert ca -name "Smoke Test"
-    ./nebula-cert sign -name "lighthouse1" -ip "192.168.100.1/24"
-    ./nebula-cert sign -name "host2" -ip "192.168.100.2/24"
-    ./nebula-cert sign -name "host3" -ip "192.168.100.3/24"
+    HOST="lighthouse1" \
+        AM_LIGHTHOUSE=true \
+        ../genconfig.sh >lighthouse1.yml
+
+    HOST="host2" \
+        LIGHTHOUSES="192.168.100.1 172.17.0.2:4242" \
+        ../genconfig.sh >host2.yml
+
+    HOST="host3" \
+        LIGHTHOUSES="192.168.100.1 172.17.0.2:4242" \
+        INBOUND='[{"port": "any", "proto": "icmp", "group": "lighthouse"}]' \
+        ../genconfig.sh >host3.yml
+
+    HOST="host4" \
+        LIGHTHOUSES="192.168.100.1 172.17.0.2:4242" \
+        OUTBOUND='[{"port": "any", "proto": "icmp", "group": "lighthouse"}]' \
+        ../genconfig.sh >host4.yml
+
+    ../../../../nebula-cert ca -curve "${CURVE:-25519}" -name "Smoke Test"
+    ../../../../nebula-cert sign -name "lighthouse1" -groups "lighthouse,lighthouse1" -ip "192.168.100.1/24"
+    ../../../../nebula-cert sign -name "host2" -groups "host,host2" -ip "192.168.100.2/24"
+    ../../../../nebula-cert sign -name "host3" -groups "host,host3" -ip "192.168.100.3/24"
+    ../../../../nebula-cert sign -name "host4" -groups "host,host4" -ip "192.168.100.4/24"
 )
 
-docker build -t nebula:smoke .
+docker build -t "nebula:${NAME:-smoke}" .

.github/workflows/smoke/genconfig.sh

@@ -2,6 +2,7 @@
 
 set -e
 
+FIREWALL_ALL='[{"port": "any", "proto": "any", "host": "any"}]'
 
 if [ "$STATIC_HOSTS" ] || [ "$LIGHTHOUSES" ]
 then
@@ -32,29 +33,27 @@ lighthouse_hosts() {
 
 cat <<EOF
 pki:
-  ca: /ca.crt
-  cert: /${HOST}.crt
-  key: /${HOST}.key
+  ca: ca.crt
+  cert: ${HOST}.crt
+  key: ${HOST}.key
 
 lighthouse:
   am_lighthouse: ${AM_LIGHTHOUSE:-false}
   hosts: $(lighthouse_hosts)
+  remote_allow_list: ${REMOTE_ALLOW_LIST}
 
 listen:
   host: 0.0.0.0
   port: ${LISTEN_PORT:-4242}
 
 tun:
-  dev: ${TUN_DEV:-nebula1}
+  dev: ${TUN_DEV:-tun0}
 
 firewall:
-  outbound:
-    - port: any
-      proto: any
-      host: any
+  inbound_action: reject
+  outbound_action: reject
+  outbound: ${OUTBOUND:-$FIREWALL_ALL}
+  inbound: ${INBOUND:-$FIREWALL_ALL}
 
-  inbound:
-    - port: any
-      proto: any
-      host: any
+$(test -t 0 || cat)
 EOF

.github/workflows/smoke/smoke-relay.sh

@@ -0,0 +1,85 @@
+#!/bin/bash
+
+set -e -x
+
+set -o pipefail
+
+mkdir -p logs
+
+cleanup() {
+    echo
+    echo " *** cleanup"
+    echo
+
+    set +e
+    if [ "$(jobs -r)" ]
+    then
+        docker kill lighthouse1 host2 host3 host4
+    fi
+}
+
+trap cleanup EXIT
+
+docker run --name lighthouse1 --rm nebula:smoke-relay -config lighthouse1.yml -test
+docker run --name host2 --rm nebula:smoke-relay -config host2.yml -test
+docker run --name host3 --rm nebula:smoke-relay -config host3.yml -test
+docker run --name host4 --rm nebula:smoke-relay -config host4.yml -test
+
+docker run --name lighthouse1 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke-relay -config lighthouse1.yml 2>&1 | tee logs/lighthouse1 | sed -u 's/^/  [lighthouse1]  /' &
+sleep 1
+docker run --name host2 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke-relay -config host2.yml 2>&1 | tee logs/host2 | sed -u 's/^/  [host2]  /' &
+sleep 1
+docker run --name host3 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke-relay -config host3.yml 2>&1 | tee logs/host3 | sed -u 's/^/  [host3]  /' &
+sleep 1
+docker run --name host4 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke-relay -config host4.yml 2>&1 | tee logs/host4 | sed -u 's/^/  [host4]  /' &
+sleep 1
+
+set +x
+echo
+echo " *** Testing ping from lighthouse1"
+echo
+set -x
+docker exec lighthouse1 ping -c1 192.168.100.2
+docker exec lighthouse1 ping -c1 192.168.100.3
+docker exec lighthouse1 ping -c1 192.168.100.4
+
+set +x
+echo
+echo " *** Testing ping from host2"
+echo
+set -x
+docker exec host2 ping -c1 192.168.100.1
+# Should fail because no relay configured in this direction
+! docker exec host2 ping -c1 192.168.100.3 -w5 || exit 1
+! docker exec host2 ping -c1 192.168.100.4 -w5 || exit 1
+
+set +x
+echo
+echo " *** Testing ping from host3"
+echo
+set -x
+docker exec host3 ping -c1 192.168.100.1
+docker exec host3 ping -c1 192.168.100.2
+docker exec host3 ping -c1 192.168.100.4
+
+set +x
+echo
+echo " *** Testing ping from host4"
+echo
+set -x
+docker exec host4 ping -c1 192.168.100.1
+# Should fail because relays not allowed
+! docker exec host4 ping -c1 192.168.100.2 -w5 || exit 1
+docker exec host4 ping -c1 192.168.100.3
+
+docker exec host4 sh -c 'kill 1'
+docker exec host3 sh -c 'kill 1'
+docker exec host2 sh -c 'kill 1'
+docker exec lighthouse1 sh -c 'kill 1'
+sleep 5
+
+if [ "$(jobs -r)" ]
+then
+    echo "nebula still running after SIGTERM sent" >&2
+    exit 1
+fi

.github/workflows/smoke/smoke-vagrant.sh

@@ -0,0 +1,105 @@
+#!/bin/bash
+
+set -e -x
+
+set -o pipefail
+
+export VAGRANT_CWD="$PWD/vagrant-$1"
+
+mkdir -p logs
+
+cleanup() {
+    echo
+    echo " *** cleanup"
+    echo
+
+    set +e
+    if [ "$(jobs -r)" ]
+    then
+        docker kill lighthouse1 host2
+    fi
+    vagrant destroy -f
+}
+
+trap cleanup EXIT
+
+CONTAINER="nebula:${NAME:-smoke}"
+
+docker run --name lighthouse1 --rm "$CONTAINER" -config lighthouse1.yml -test
+docker run --name host2 --rm "$CONTAINER" -config host2.yml -test
+
+vagrant up
+vagrant ssh -c "cd /nebula && /nebula/$1-nebula -config host3.yml -test"
+
+docker run --name lighthouse1 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config lighthouse1.yml 2>&1 | tee logs/lighthouse1 | sed -u 's/^/  [lighthouse1]  /' &
+sleep 1
+docker run --name host2 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config host2.yml 2>&1 | tee logs/host2 | sed -u 's/^/  [host2]  /' &
+sleep 1
+vagrant ssh -c "cd /nebula && sudo sh -c 'echo \$\$ >/nebula/pid && exec /nebula/$1-nebula -config host3.yml'" &
+sleep 15
+
+# grab tcpdump pcaps for debugging
+docker exec lighthouse1 tcpdump -i nebula1 -q -w - -U 2>logs/lighthouse1.inside.log >logs/lighthouse1.inside.pcap &
+docker exec lighthouse1 tcpdump -i eth0 -q -w - -U 2>logs/lighthouse1.outside.log >logs/lighthouse1.outside.pcap &
+docker exec host2 tcpdump -i nebula1 -q -w - -U 2>logs/host2.inside.log >logs/host2.inside.pcap &
+docker exec host2 tcpdump -i eth0 -q -w - -U 2>logs/host2.outside.log >logs/host2.outside.pcap &
+# vagrant ssh -c "tcpdump -i nebula1 -q -w - -U" 2>logs/host3.inside.log >logs/host3.inside.pcap &
+# vagrant ssh -c "tcpdump -i eth0 -q -w - -U" 2>logs/host3.outside.log >logs/host3.outside.pcap &
+
+docker exec host2 ncat -nklv 0.0.0.0 2000 &
+vagrant ssh -c "ncat -nklv 0.0.0.0 2000" &
+#docker exec host2 ncat -e '/usr/bin/echo host2' -nkluv 0.0.0.0 3000 &
+#vagrant ssh -c "ncat -e '/usr/bin/echo host3' -nkluv 0.0.0.0 3000" &
+
+set +x
+echo
+echo " *** Testing ping from lighthouse1"
+echo
+set -x
+docker exec lighthouse1 ping -c1 192.168.100.2
+docker exec lighthouse1 ping -c1 192.168.100.3
+
+set +x
+echo
+echo " *** Testing ping from host2"
+echo
+set -x
+docker exec host2 ping -c1 192.168.100.1
+# Should fail because not allowed by host3 inbound firewall
+! docker exec host2 ping -c1 192.168.100.3 -w5 || exit 1
+
+set +x
+echo
+echo " *** Testing ncat from host2"
+echo
+set -x
+# Should fail because not allowed by host3 inbound firewall
+#! docker exec host2 ncat -nzv -w5 192.168.100.3 2000 || exit 1
+#! docker exec host2 ncat -nzuv -w5 192.168.100.3 3000 | grep -q host3 || exit 1
+
+set +x
+echo
+echo " *** Testing ping from host3"
+echo
+set -x
+vagrant ssh -c "ping -c1 192.168.100.1"
+vagrant ssh -c "ping -c1 192.168.100.2"
+
+set +x
+echo
+echo " *** Testing ncat from host3"
+echo
+set -x
+#vagrant ssh -c "ncat -nzv -w5 192.168.100.2 2000"
+#vagrant ssh -c "ncat -nzuv -w5 192.168.100.2 3000" | grep -q host2
+
+vagrant ssh -c "sudo xargs kill </nebula/pid"
+docker exec host2 sh -c 'kill 1'
+docker exec lighthouse1 sh -c 'kill 1'
+sleep 1
+
+if [ "$(jobs -r)" ]
+then
+    echo "nebula still running after SIGTERM sent" >&2
+    exit 1
+fi

.github/workflows/smoke/smoke.sh

@@ -1,17 +1,55 @@
-#!/bin/sh
+#!/bin/bash
 
 set -e -x
 
-docker run --name lighthouse1 --rm nebula:smoke -config lighthouse1.yml -test
-docker run --name host2 --rm nebula:smoke -config host2.yml -test
-docker run --name host3 --rm nebula:smoke -config host3.yml -test
+set -o pipefail
 
-docker run --name lighthouse1 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke -config lighthouse1.yml &
+mkdir -p logs
+
+cleanup() {
+    echo
+    echo " *** cleanup"
+    echo
+
+    set +e
+    if [ "$(jobs -r)" ]
+    then
+        docker kill lighthouse1 host2 host3 host4
+    fi
+}
+
+trap cleanup EXIT
+
+CONTAINER="nebula:${NAME:-smoke}"
+
+docker run --name lighthouse1 --rm "$CONTAINER" -config lighthouse1.yml -test
+docker run --name host2 --rm "$CONTAINER" -config host2.yml -test
+docker run --name host3 --rm "$CONTAINER" -config host3.yml -test
+docker run --name host4 --rm "$CONTAINER" -config host4.yml -test
+
+docker run --name lighthouse1 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config lighthouse1.yml 2>&1 | tee logs/lighthouse1 | sed -u 's/^/  [lighthouse1]  /' &
 sleep 1
-docker run --name host2 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke -config host2.yml &
+docker run --name host2 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config host2.yml 2>&1 | tee logs/host2 | sed -u 's/^/  [host2]  /' &
 sleep 1
-docker run --name host3 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm nebula:smoke -config host3.yml &
+docker run --name host3 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config host3.yml 2>&1 | tee logs/host3 | sed -u 's/^/  [host3]  /' &
 sleep 1
+docker run --name host4 --device /dev/net/tun:/dev/net/tun --cap-add NET_ADMIN --rm "$CONTAINER" -config host4.yml 2>&1 | tee logs/host4 | sed -u 's/^/  [host4]  /' &
+sleep 1
+
+# grab tcpdump pcaps for debugging
+docker exec lighthouse1 tcpdump -i nebula1 -q -w - -U 2>logs/lighthouse1.inside.log >logs/lighthouse1.inside.pcap &
+docker exec lighthouse1 tcpdump -i eth0 -q -w - -U 2>logs/lighthouse1.outside.log >logs/lighthouse1.outside.pcap &
+docker exec host2 tcpdump -i nebula1 -q -w - -U 2>logs/host2.inside.log >logs/host2.inside.pcap &
+docker exec host2 tcpdump -i eth0 -q -w - -U 2>logs/host2.outside.log >logs/host2.outside.pcap &
+docker exec host3 tcpdump -i nebula1 -q -w - -U 2>logs/host3.inside.log >logs/host3.inside.pcap &
+docker exec host3 tcpdump -i eth0 -q -w - -U 2>logs/host3.outside.log >logs/host3.outside.pcap &
+docker exec host4 tcpdump -i nebula1 -q -w - -U 2>logs/host4.inside.log >logs/host4.inside.pcap &
+docker exec host4 tcpdump -i eth0 -q -w - -U 2>logs/host4.outside.log >logs/host4.outside.pcap &
+
+docker exec host2 ncat -nklv 0.0.0.0 2000 &
+docker exec host3 ncat -nklv 0.0.0.0 2000 &
+docker exec host2 ncat -e '/usr/bin/echo host2' -nkluv 0.0.0.0 3000 &
+docker exec host3 ncat -e '/usr/bin/echo host3' -nkluv 0.0.0.0 3000 &
 
 set +x
 echo
@@ -27,7 +65,17 @@ echo " *** Testing ping from host2"
 echo
 set -x
 docker exec host2 ping -c1 192.168.100.1
-docker exec host2 ping -c1 192.168.100.3
+# Should fail because not allowed by host3 inbound firewall
+! docker exec host2 ping -c1 192.168.100.3 -w5 || exit 1
+
+set +x
+echo
+echo " *** Testing ncat from host2"
+echo
+set -x
+# Should fail because not allowed by host3 inbound firewall
+! docker exec host2 ncat -nzv -w5 192.168.100.3 2000 || exit 1
+! docker exec host2 ncat -nzuv -w5 192.168.100.3 3000 | grep -q host3 || exit 1
 
 set +x
 echo
@@ -36,3 +84,55 @@ echo
 set -x
 docker exec host3 ping -c1 192.168.100.1
 docker exec host3 ping -c1 192.168.100.2
+
+set +x
+echo
+echo " *** Testing ncat from host3"
+echo
+set -x
+docker exec host3 ncat -nzv -w5 192.168.100.2 2000
+docker exec host3 ncat -nzuv -w5 192.168.100.2 3000 | grep -q host2
+
+set +x
+echo
+echo " *** Testing ping from host4"
+echo
+set -x
+docker exec host4 ping -c1 192.168.100.1
+# Should fail because not allowed by host4 outbound firewall
+! docker exec host4 ping -c1 192.168.100.2 -w5 || exit 1
+! docker exec host4 ping -c1 192.168.100.3 -w5 || exit 1
+
+set +x
+echo
+echo " *** Testing ncat from host4"
+echo
+set -x
+# Should fail because not allowed by host4 outbound firewall
+! docker exec host4 ncat -nzv -w5 192.168.100.2 2000 || exit 1
+! docker exec host4 ncat -nzv -w5 192.168.100.3 2000 || exit 1
+! docker exec host4 ncat -nzuv -w5 192.168.100.2 3000 | grep -q host2 || exit 1
+! docker exec host4 ncat -nzuv -w5 192.168.100.3 3000 | grep -q host3 || exit 1
+
+set +x
+echo
+echo " *** Testing conntrack"
+echo
+set -x
+# host2 can ping host3 now that host3 pinged it first
+docker exec host2 ping -c1 192.168.100.3
+# host4 can ping host2 once conntrack established
+docker exec host2 ping -c1 192.168.100.4
+docker exec host4 ping -c1 192.168.100.2
+
+docker exec host4 sh -c 'kill 1'
+docker exec host3 sh -c 'kill 1'
+docker exec host2 sh -c 'kill 1'
+docker exec lighthouse1 sh -c 'kill 1'
+sleep 5
+
+if [ "$(jobs -r)" ]
+then
+    echo "nebula still running after SIGTERM sent" >&2
+    exit 1
+fi

.github/workflows/smoke/vagrant-freebsd-amd64/Vagrantfile

@@ -0,0 +1,7 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+Vagrant.configure("2") do |config|
+  config.vm.box = "generic/freebsd14"
+
+  config.vm.synced_folder "../build", "/nebula", type: "rsync"
+end

.github/workflows/smoke/vagrant-linux-386/Vagrantfile

@@ -0,0 +1,7 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+Vagrant.configure("2") do |config|
+  config.vm.box = "ubuntu/xenial32"
+
+  config.vm.synced_folder "../build", "/nebula"
+end

.github/workflows/smoke/vagrant-linux-amd64-ipv6disable/Vagrantfile

@@ -0,0 +1,16 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+Vagrant.configure("2") do |config|
+  config.vm.box = "ubuntu/jammy64"
+
+  config.vm.synced_folder "../build", "/nebula"
+
+  config.vm.provision :shell do |shell|
+    shell.inline = <<-EOF
+      sed -i 's/GRUB_CMDLINE_LINUX=""/GRUB_CMDLINE_LINUX="ipv6.disable=1"/' /etc/default/grub
+      update-grub
+    EOF
+    shell.privileged = true
+    shell.reboot = true
+  end
+end

.github/workflows/smoke/vagrant-netbsd-amd64/Vagrantfile

@@ -0,0 +1,7 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+Vagrant.configure("2") do |config|
+  config.vm.box = "generic/netbsd9"
+
+  config.vm.synced_folder "../build", "/nebula", type: "rsync"
+end

.github/workflows/smoke/vagrant-openbsd-amd64/Vagrantfile

@@ -0,0 +1,7 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+Vagrant.configure("2") do |config|
+  config.vm.box = "generic/openbsd7"
+
+  config.vm.synced_folder "../build", "/nebula", type: "rsync"
+end

.github/workflows/test.yml

@@ -18,51 +18,69 @@ jobs:
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.14
-      uses: actions/setup-go@v1
-      with:
-        go-version: 1.14
-      id: go
+    - uses: actions/checkout@v4
 
-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
-
-    - uses: actions/cache@v1
+    - uses: actions/setup-go@v5
       with:
-        path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-        restore-keys: |
-          ${{ runner.os }}-go-
+        go-version: '1.22'
+        check-latest: true
 
     - name: Build
       run: make all
 
+    - name: Vet
+      run: make vet
+
     - name: Test
       run: make test
 
+    - name: End 2 end
+      run: make e2evv
+
+    - name: Build test mobile
+      run: make build-test-mobile
+
+    - uses: actions/upload-artifact@v4
+      with:
+        name: e2e packet flow linux-latest
+        path: e2e/mermaid/linux-latest
+        if-no-files-found: warn
+
+  test-linux-boringcrypto:
+    name: Build and test on linux with boringcrypto
+    runs-on: ubuntu-latest
+    steps:
+
+    - uses: actions/checkout@v4
+
+    - uses: actions/setup-go@v5
+      with:
+        go-version: '1.22'
+        check-latest: true
+
+    - name: Build
+      run: make bin-boringcrypto
+
+    - name: Test
+      run: make test-boringcrypto
+
+    - name: End 2 end
+      run: make e2evv GOEXPERIMENT=boringcrypto CGO_ENABLED=1
+
   test:
     name: Build and test on ${{ matrix.os }}
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [windows-latest, macOS-latest]
+        os: [windows-latest, macos-latest]
     steps:
 
-    - name: Set up Go 1.14
-      uses: actions/setup-go@v1
-      with:
-        go-version: 1.14
-      id: go
+    - uses: actions/checkout@v4
 
-    - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
-
-    - uses: actions/cache@v1
+    - uses: actions/setup-go@v5
       with:
-        path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-        restore-keys: |
-          ${{ runner.os }}-go-
+        go-version: '1.22'
+        check-latest: true
 
     - name: Build nebula
       run: go build ./cmd/nebula
@@ -70,5 +88,17 @@ jobs:
     - name: Build nebula-cert
       run: go build ./cmd/nebula-cert
 
+    - name: Vet
+      run: make vet
+
     - name: Test
-      run: go test -v ./...
+      run: make test
+
+    - name: End 2 end
+      run: make e2evv
+
+    - uses: actions/upload-artifact@v4
+      with:
+        name: e2e packet flow ${{ matrix.os }}
+        path: e2e/mermaid/${{ matrix.os }}
+        if-no-files-found: warn

.gitignore

@@ -4,9 +4,14 @@
 /nebula-arm6
 /nebula-darwin
 /nebula.exe
-/cert/*.crt
-/cert/*.key
+/nebula-cert.exe
 /coverage.out
 /cpu.pprof
 /build
 /*.tar.gz
+/e2e/mermaid/
+**.crt
+**.key
+**.pem
+!/examples/quickstart-vagrant/ansible/roles/nebula/files/vagrant-test-ca.key
+!/examples/quickstart-vagrant/ansible/roles/nebula/files/vagrant-test-ca.crt

CHANGELOG.md

@@ -7,6 +7,587 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.9.7] - 2025-10-10
+
+### Security
+
+- Fix an issue where Nebula could incorrectly accept and process a packet from an erroneous source IP when the sender's
+  certificate is configured with unsafe_routes (cert v1/v2) or multiple IPs (cert v2). (#1494)
+
+### Changed
+
+- Disable sending `recv_error` messages when a packet is received outside the allowable counter window. (#1459)
+- Improve error messages and remove some unnecessary fatal conditions in the Windows and generic udp listener. (#1543)
+
+## [1.9.6] - 2025-7-15
+
+### Added
+
+- Support dropping inactive tunnels. This is disabled by default in this release but can be enabled with `tunnels.drop_inactive`. See example config for more details. (#1413)
+
+### Fixed
+
+- Fix Darwin freeze due to presence of some Network Extensions (#1426)
+- Ensure the same relay tunnel is always used when multiple relay tunnels are present (#1422)
+- Fix Windows freeze due to ICMP error handling (#1412)
+- Fix relay migration panic (#1403)
+
+## [1.9.5] - 2024-12-05
+
+### Added
+
+- Gracefully ignore v2 certificates. (#1282)
+
+### Fixed
+
+- Fix relays that refuse to re-establish after one of the remote tunnel pairs breaks. (#1277)
+
+## [1.9.4] - 2024-09-09
+
+### Added
+
+- Support UDP dialing with gVisor. (#1181)
+
+### Changed
+
+- Make some Nebula state programmatically available via control object. (#1188)
+- Switch internal representation of IPs to netip, to prepare for IPv6 support
+  in the overlay. (#1173)
+- Minor build and cleanup changes. (#1171, #1164, #1162)
+- Various dependency updates. (#1195, #1190, #1174, #1168, #1167, #1161, #1147, #1146)
+
+### Fixed
+
+- Fix a bug on big endian hosts, like mips. (#1194)
+- Fix a rare panic if a local index collision happens. (#1191)
+- Fix integer wraparound in the calculation of handshake timeouts on 32-bit targets. (#1185)
+
+## [1.9.3] - 2024-06-06
+
+### Fixed
+
+- Initialize messageCounter to 2 instead of verifying later. (#1156)
+
+## [1.9.2] - 2024-06-03
+
+### Fixed
+
+- Ensure messageCounter is set before handshake is complete. (#1154)
+
+## [1.9.1] - 2024-05-29
+
+### Fixed
+
+- Fixed a potential deadlock in GetOrHandshake. (#1151)
+
+## [1.9.0] - 2024-05-07
+
+### Deprecated
+
+- This release adds a new setting `default_local_cidr_any` that defaults to
+  true to match previous behavior, but will default to false in the next
+  release (1.10). When set to false, `local_cidr` is matched correctly for
+  firewall rules on hosts acting as unsafe routers, and should be set for any
+  firewall rules you want to allow unsafe route hosts to access. See the issue
+  and example config for more details. (#1071, #1099)
+
+### Added
+
+- Nebula now has an official Docker image `nebulaoss/nebula` that is
+  distroless and contains just the `nebula` and `nebula-cert` binaries. You
+  can find it here: https://hub.docker.com/r/nebulaoss/nebula (#1037)
+
+- Experimental binaries for `loong64` are now provided. (#1003)
+
+- Added example service script for OpenRC. (#711)
+
+- The SSH daemon now supports inlined host keys. (#1054)
+
+- The SSH daemon now supports certificates with `sshd.trusted_cas`. (#1098)
+
+### Changed
+
+- Config setting `tun.unsafe_routes` is now reloadable. (#1083)
+
+- Small documentation and internal improvements. (#1065, #1067, #1069, #1108,
+  #1109, #1111, #1135)
+
+- Various dependency updates. (#1139, #1138, #1134, #1133, #1126, #1123, #1110,
+  #1094, #1092, #1087, #1086, #1085, #1072, #1063, #1059, #1055, #1053, #1047,
+  #1046, #1034, #1022)
+
+### Removed
+
+- Support for the deprecated `local_range` option has been removed. Please
+  change to `preferred_ranges` (which is also now reloadable). (#1043)
+
+- We are now building with go1.22, which means that for Windows you need at
+  least Windows 10 or Windows Server 2016. This is because support for earlier
+  versions was removed in Go 1.21. See https://go.dev/doc/go1.21#windows (#981)
+
+- Removed vagrant example, as it was unmaintained. (#1129)
+
+- Removed Fedora and Arch nebula.service files, as they are maintained in the
+  upstream repos. (#1128, #1132)
+
+- Remove the TCP round trip tracking metrics, as they never had correct data
+  and were an experiment to begin with. (#1114)
+
+### Fixed
+
+- Fixed a potential deadlock introduced in 1.8.1. (#1112)
+
+- Fixed support for Linux when IPv6 has been disabled at the OS level. (#787)
+
+- DNS will return NXDOMAIN now when there are no results. (#845)
+
+- Allow `::` in `lighthouse.dns.host`. (#1115)
+
+- Capitalization of `NotAfter` fixed in DNS TXT response. (#1127)
+
+- Don't log invalid certificates. It is untrusted data and can cause a large
+  volume of logs. (#1116)
+
+## [1.8.2] - 2024-01-08
+
+### Fixed
+
+- Fix multiple routines when listen.port is zero. This was a regression
+  introduced in v1.6.0. (#1057)
+
+### Changed
+
+- Small dependency update for Noise. (#1038)
+
+## [1.8.1] - 2023-12-19
+
+### Security
+
+- Update `golang.org/x/crypto`, which includes a fix for CVE-2023-48795. (#1048)
+
+### Fixed
+
+- Fix a deadlock introduced in v1.8.0 that could occur during handshakes. (#1044)
+
+- Fix mobile builds. (#1035)
+
+## [1.8.0] - 2023-12-06
+
+### Deprecated
+
+- The next minor release of Nebula, 1.9.0, will require at least Windows 10 or
+  Windows Server 2016. This is because support for earlier versions was removed
+  in Go 1.21. See https://go.dev/doc/go1.21#windows
+
+### Added
+
+- Linux: Notify systemd of service readiness. This should resolve timing issues
+  with services that depend on Nebula being active. For an example of how to
+  enable this, see: `examples/service_scripts/nebula.service`. (#929)
+
+- Windows: Use Registered IO (RIO) when possible. Testing on a Windows 11
+  machine shows ~50x improvement in throughput. (#905)
+
+- NetBSD, OpenBSD: Added rudimentary support. (#916, #812)
+
+- FreeBSD: Add support for naming tun devices. (#903)
+
+### Changed
+
+- `pki.disconnect_invalid` will now default to true. This means that once a
+  certificate expires, the tunnel will be disconnected. If you use SIGHUP to
+  reload certificates without restarting Nebula, you should ensure all of your
+  clients are on 1.7.0 or newer before you enable this feature. (#859)
+
+- Limit how often a busy tunnel can requery the lighthouse. The new config
+  option `timers.requery_wait_duration` defaults to `60s`. (#940)
+
+- The internal structures for hostmaps were refactored to reduce memory usage
+  and the potential for subtle bugs. (#843, #938, #953, #954, #955)
+
+- Lots of dependency updates.
+
+### Fixed
+
+- Windows: Retry wintun device creation if it fails the first time. (#985)
+
+- Fix issues with firewall reject packets that could cause panics. (#957)
+
+- Fix relay migration during re-handshakes. (#964)
+
+- Various other refactors and fixes. (#935, #952, #972, #961, #996, #1002,
+  #987, #1004, #1030, #1032, ...)
+
+## [1.7.2] - 2023-06-01
+
+### Fixed
+
+- Fix a freeze during config reload if the `static_host_map` config was changed. (#886)
+
+## [1.7.1] - 2023-05-18
+
+### Fixed
+
+- Fix IPv4 addresses returned by `static_host_map` DNS lookup queries being
+  treated as IPv6 addresses. (#877)
+
+## [1.7.0] - 2023-05-17
+
+### Added
+
+- `nebula-cert ca` now supports encrypting the CA's private key with a
+  passphrase. Pass `-encrypt` in order to be prompted for a passphrase.
+  Encryption is performed using AES-256-GCM and Argon2id for KDF. KDF
+  parameters default to RFC recommendations, but can be overridden via CLI
+  flags `-argon-memory`, `-argon-parallelism`, and `-argon-iterations`. (#386)
+
+- Support for curve P256 and BoringCrypto has been added. See README section
+  "Curve P256 and BoringCrypto" for more details. (#865, #861, #769, #856, #803)
+
+- New firewall rule `local_cidr`. This could be used to filter destinations
+  when using `unsafe_routes`. (#507)
+
+- Add `unsafe_route` option `install`. This controls whether the route is
+  installed in the systems routing table. (#831)
+
+- Add `tun.use_system_route_table` option. Set to true to manage unsafe routes
+  directly on the system route table with gateway routes instead of in Nebula
+  configuration files. This is only supported on Linux. (#839)
+
+- The metric `certificate.ttl_seconds` is now exposed via stats. (#782)
+
+- Add `punchy.respond_delay` option. This allows you to change the delay
+  before attempting punchy.respond. Default is 5 seconds. (#721)
+
+- Added SSH commands to allow the capture of a mutex profile. (#737)
+
+- You can now set `lighthouse.calculated_remotes` to make it possible to do
+  handshakes without a lighthouse in certain configurations. (#759)
+
+- The firewall can be configured to send REJECT replies instead of the default
+  DROP behavior. (#738)
+
+- For macOS, an example launchd configuration file is now provided. (#762)
+
+### Changed
+
+- Lighthouses and other `static_host_map` entries that use DNS names will now
+  be automatically refreshed to detect when the IP address changes. (#796)
+
+- Lighthouses send ACK replies back to clients so that they do not fall into
+  connection testing as often by clients. (#851, #408)
+
+- Allow the `listen.host` option to contain a hostname. (#825)
+
+- When Nebula switches to a new certificate (such as via SIGHUP), we now
+  rehandshake with all existing tunnels. This allows firewall groups to be
+  updated and `pki.disconnect_invalid` to know about the new certificate
+  expiration time. (#838, #857, #842, #840, #835, #828, #820, #807)
+
+### Fixed
+
+- Always disconnect blocklisted hosts, even if `pki.disconnect_invalid` is
+  not set. (#858)
+
+- Dependencies updated and go1.20 required. (#780, #824, #855, #854)
+
+- Fix possible race condition with relays. (#827)
+
+- FreeBSD: Fix connection to the localhost's own Nebula IP. (#808)
+
+- Normalize and document some common log field values. (#837, #811)
+
+- Fix crash if you set unlucky values for the firewall timeout configuration
+  options. (#802)
+
+- Make DNS queries case insensitive. (#793)
+
+- Update example systemd configurations to want `nss-lookup`. (#791)
+
+- Errors with SSH commands now go to the SSH tunnel instead of stderr. (#757)
+
+- Fix a hang when shutting down Android. (#772)
+
+## [1.6.1] - 2022-09-26
+
+### Fixed
+
+- Refuse to process underlay packets received from overlay IPs. This prevents
+  confusion on hosts that have unsafe routes configured. (#741)
+
+- The ssh `reload` command did not work on Windows, since it relied on sending
+  a SIGHUP signal internally. This has been fixed. (#725)
+
+- A regression in v1.5.2 that broke unsafe routes on Mobile clients has been
+  fixed. (#729)
+
+## [1.6.0] - 2022-06-30
+
+### Added
+
+- Experimental: nebula clients can be configured to act as relays for other nebula clients.
+  Primarily useful when stubborn NATs make a direct tunnel impossible. (#678)
+
+- Configuration option to report manually specified `ip:port`s to lighthouses. (#650)
+
+- Windows arm64 build. (#638)
+
+- `punchy` and most `lighthouse` config options now support hot reloading. (#649)
+
+### Changed
+
+- Build against go 1.18. (#656)
+
+- Promoted `routines` config from experimental to supported feature. (#702)
+
+- Dependencies updated. (#664)
+
+### Fixed
+
+- Packets destined for the same host that sent it will be returned on MacOS.
+  This matches the default behavior of other operating systems. (#501)
+
+- `unsafe_route` configuration will no longer crash on Windows. (#648)
+
+- A few panics that were introduced in 1.5.x. (#657, #658, #675)
+
+### Security
+
+- You can set `listen.send_recv_error` to control the conditions in which
+  `recv_error` messages are sent. Sending these messages can expose the fact
+  that Nebula is running on a host, but it speeds up re-handshaking. (#670)
+
+### Removed
+
+- `x509` config stanza support has been removed. (#685)
+
+## [1.5.2] - 2021-12-14
+
+### Added
+
+- Warn when a non lighthouse node does not have lighthouse hosts configured. (#587)
+
+### Changed
+
+- No longer fatals if expired CA certificates are present in `pki.ca`, as long as 1 valid CA is present. (#599)
+
+- `nebula-cert` will now enforce ipv4 addresses. (#604)
+
+- Warn on macOS if an unsafe route cannot be created due to a collision with an
+  existing route. (#610)
+
+- Warn if you set a route MTU on platforms where we don't support it. (#611)
+
+### Fixed
+
+- Rare race condition when tearing down a tunnel due to `recv_error` and sending packets on another thread. (#590)
+
+- Bug in `routes` and `unsafe_routes` handling that was introduced in 1.5.0. (#595)
+
+- `-test` mode no longer results in a crash. (#602)
+
+### Removed
+
+- `x509.ca` config alias for `pki.ca`. (#604)
+
+### Security
+
+- Upgraded `golang.org/x/crypto` to address an issue which allowed unauthenticated clients to cause a panic in SSH
+  servers. (#603)
+
+## 1.5.1 - 2021-12-13
+
+(This release was skipped due to discovering #610 and #611 after the tag was
+created.)
+
+## [1.5.0] - 2021-11-11
+
+### Added
+
+- SSH `print-cert` has a new `-raw` flag to get the PEM representation of a certificate. (#483)
+
+- New build architecture: Linux `riscv64`. (#542)
+
+- New experimental config option `remote_allow_ranges`. (#540)
+
+- New config option `pki.disconnect_invalid` that will tear down tunnels when they become invalid (through expiry or
+  removal of root trust). Default is `false`. Note, this will not currently recognize if a remote has changed
+  certificates since the last handshake. (#370)
+
+- New config option `unsafe_routes.<route>.metric` will set a metric for a specific unsafe route. It's useful if you have
+  more than one identical route and want to prefer one against the other. (#353)
+
+### Changed
+
+- Build against go 1.17. (#553)
+
+- Build with `CGO_ENABLED=0` set, to create more portable binaries. This could
+  have an effect on DNS resolution if you rely on anything non-standard. (#421)
+
+- Windows now uses the [wintun](https://www.wintun.net/) driver which does not require installation. This driver
+  is a large improvement over the TAP driver that was used in previous versions. If you had a previous version
+  of `nebula` running, you will want to disable the tap driver in Control Panel, or uninstall the `tap0901` driver
+  before running this version. (#289)
+
+- Darwin binaries are now universal (works on both amd64 and arm64), signed, and shipped in a notarized zip file.
+  `nebula-darwin.zip` will be the only darwin release artifact. (#571)
+
+- Darwin uses syscalls and AF_ROUTE to configure the routing table, instead of
+  using `/sbin/route`. Setting `tun.dev` is now allowed on Darwin as well, it
+  must be in the format `utun[0-9]+` or it will be ignored. (#163)
+
+### Deprecated
+
+- The `preferred_ranges` option has been supported as a replacement for
+  `local_range` since v1.0.0. It has now been documented and `local_range`
+  has been officially deprecated. (#541)
+
+### Fixed
+
+- Valid recv_error packets were incorrectly marked as "spoofing" and ignored. (#482)
+
+- SSH server handles single `exec` requests correctly. (#483)
+
+- Signing a certificate with `nebula-cert sign` now verifies that the supplied
+  ca-key matches the ca-crt. (#503)
+
+- If `preferred_ranges` (or the deprecated `local_range`) is configured, we
+  will immediately switch to a preferred remote address after the reception of
+  a handshake packet (instead of waiting until 1,000 packets have been sent).
+  (#532)
+
+- A race condition when `punchy.respond` is enabled and ensures the correct
+  vpn ip is sent a punch back response in highly queried node. (#566)
+
+- Fix a rare crash during handshake due to a race condition. (#535)
+
+## [1.4.0] - 2021-05-11
+
+### Added
+
+- Ability to output qr code images in `print`, `ca`, and `sign` modes for `nebula-cert`.
+  This is useful when configuring mobile clients. (#297)
+
+- Experimental: Nebula can now do work on more than 2 cpu cores in send and receive paths via
+  the new `routines` config option. (#382, #391, #395)
+
+- ICMP ping requests can be responded to when the `tun.disabled` is `true`.
+  This is useful so that you can "ping" a lighthouse running in this mode. (#342)
+
+- Run smoke tests via `make smoke-docker`. (#287)
+
+- More reported stats, udp memory use on linux, build version (when using Prometheus), firewall,
+  handshake, and cached packet stats. (#390, #405, #450, #453)
+
+- IPv6 support for the underlay network. (#369)
+
+- End to end testing, run with `make e2e`. (#425, #427, #428)
+
+### Changed
+
+- Darwin will now log stdout/stderr to a file when using `-service` mode. (#303)
+
+- Example systemd unit file now better arranged startup order when using `sshd`
+  and other fixes. (#317, #412, #438)
+
+- Reduced memory utilization/garbage collection. (#320, #323, #340)
+
+- Reduced CPU utilization. (#329)
+
+- Build against go 1.16. (#381)
+
+- Refactored handshakes to improve performance and correctness. (#401, #402, #404, #416, #451)
+
+- Improved roaming support for mobile clients. (#394, #457)
+
+- Lighthouse performance and correctness improvements. (#406, #418, #429, #433, #437, #442, #449)
+
+- Better ordered startup to enable `sshd`, `stats`, and `dns` subsystems to listen on
+  the nebula interface. (#375)
+
+### Fixed
+
+- No longer report handshake packets as `lost` in stats. (#331)
+
+- Error handling in the `cert` package. (#339, #373)
+
+- Orphaned pending hostmap entries are cleaned up. (#344)
+
+- Most known data races are now resolved. (#396, #400, #424)
+
+- Refuse to run a lighthouse on an ephemeral port. (#399)
+
+- Removed the global references. (#423, #426, #446)
+
+- Reloading via ssh command avoids a panic. (#447)
+
+- Shutdown is now performed in a cleaner way. (#448)
+
+- Logs will now find their way to Windows event viewer when running under `-service` mode
+  in Windows. (#443)
+
+## [1.3.0] - 2020-09-22
+
+### Added
+
+- You can emit statistics about non-message packets by setting the option
+  `stats.message_metrics`. You can similarly emit detailed statistics about
+  lighthouse packets by setting the option `stats.lighthouse_metrics`. See
+  the example config for more details. (#230)
+
+- We now support freebsd/amd64. This is experimental, please give us feedback.
+  (#103)
+
+- We now release a binary for `linux/mips-softfloat` which has also been
+  stripped to reduce filesize and hopefully have a better chance on running on
+  small mips devices. (#231)
+
+- You can set `tun.disabled` to true to run a standalone lighthouse without a
+  tun device (and thus, without root). (#269)
+
+- You can set `logging.disable_timestamp` to remove timestamps from log lines,
+  which is useful when output is redirected to a logging system that already
+  adds timestamps. (#288)
+
+### Changed
+
+- Handshakes should now trigger faster, as we try to be proactive with sending
+  them instead of waiting for the next timer tick in most cases. (#246, #265)
+
+- Previously, we would drop the conntrack table whenever firewall rules were
+  changed during a SIGHUP. Now, we will maintain the table and just validate
+  that an entry still matches with the new rule set. (#233)
+
+- Debug logs for firewall drops now include the reason. (#220, #239)
+
+- Logs for handshakes now include the fingerprint of the remote host. (#262)
+
+- Config item `pki.blacklist` is now `pki.blocklist`. (#272)
+
+- Better support for older Linux kernels. We now only set `SO_REUSEPORT` if
+  `tun.routines` is greater than 1 (default is 1). We also only use the
+  `recvmmsg` syscall if `listen.batch` is greater than 1 (default is 64).
+  (#275)
+
+- It is possible to run Nebula as a library inside of another process now.
+  Note that this is still experimental and the internal APIs around this might
+  change in minor version releases. (#279)
+
+### Deprecated
+
+- `pki.blacklist` is deprecated in favor of `pki.blocklist` with the same
+   functionality. Existing configs will continue to load for this release to
+   allow for migrations. (#272)
+
+### Fixed
+
+- `advmss` is now set correctly for each route table entry when `tun.routes`
+  is configured to have some routes with higher MTU. (#245)
+
+- Packets that arrive on the tun device with an unroutable destination IP are
+  now dropped correctly, instead of wasting time making queries to the
+  lighthouses for IP `0.0.0.0` (#267)
+
 ## [1.2.0] - 2020-04-08
 
 ### Added
@@ -118,7 +699,27 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Initial public release.
 
-[Unreleased]: https://github.com/slackhq/nebula/compare/v1.2.0...HEAD
+[Unreleased]: https://github.com/slackhq/nebula/compare/v1.9.7...HEAD
+[1.9.7]: https://github.com/slackhq/nebula/releases/tag/v1.9.7
+[1.9.6]: https://github.com/slackhq/nebula/releases/tag/v1.9.6
+[1.9.5]: https://github.com/slackhq/nebula/releases/tag/v1.9.5
+[1.9.4]: https://github.com/slackhq/nebula/releases/tag/v1.9.4
+[1.9.3]: https://github.com/slackhq/nebula/releases/tag/v1.9.3
+[1.9.2]: https://github.com/slackhq/nebula/releases/tag/v1.9.2
+[1.9.1]: https://github.com/slackhq/nebula/releases/tag/v1.9.1
+[1.9.0]: https://github.com/slackhq/nebula/releases/tag/v1.9.0
+[1.8.2]: https://github.com/slackhq/nebula/releases/tag/v1.8.2
+[1.8.1]: https://github.com/slackhq/nebula/releases/tag/v1.8.1
+[1.8.0]: https://github.com/slackhq/nebula/releases/tag/v1.8.0
+[1.7.2]: https://github.com/slackhq/nebula/releases/tag/v1.7.2
+[1.7.1]: https://github.com/slackhq/nebula/releases/tag/v1.7.1
+[1.7.0]: https://github.com/slackhq/nebula/releases/tag/v1.7.0
+[1.6.1]: https://github.com/slackhq/nebula/releases/tag/v1.6.1
+[1.6.0]: https://github.com/slackhq/nebula/releases/tag/v1.6.0
+[1.5.2]: https://github.com/slackhq/nebula/releases/tag/v1.5.2
+[1.5.0]: https://github.com/slackhq/nebula/releases/tag/v1.5.0
+[1.4.0]: https://github.com/slackhq/nebula/releases/tag/v1.4.0
+[1.3.0]: https://github.com/slackhq/nebula/releases/tag/v1.3.0
 [1.2.0]: https://github.com/slackhq/nebula/releases/tag/v1.2.0
 [1.1.0]: https://github.com/slackhq/nebula/releases/tag/v1.1.0
 [1.0.0]: https://github.com/slackhq/nebula/releases/tag/v1.0.0

LOGGING.md

@@ -0,0 +1,37 @@
+### Logging conventions
+
+A log message (the string/format passed to `Info`, `Error`, `Debug` etc, as well as their `Sprintf` counterparts) should
+be a descriptive message about the event and may contain specific identifying characteristics. Regardless of the
+level of detail in the message identifying characteristics should always be included via `WithField`, `WithFields` or
+`WithError`
+
+If an error is being logged use `l.WithError(err)` so that there is better discoverability about the event as well
+as the specific error condition.
+
+#### Common fields
+
+- `cert` - a `cert.NebulaCertificate` object, do not `.String()` this manually, `logrus` will marshal objects properly
+  for the formatter it is using.
+- `fingerprint` - a single `NebeulaCertificate` hex encoded fingerprint
+- `fingerprints` - an array of `NebulaCertificate` hex encoded fingerprints
+- `fwPacket` - a FirewallPacket object
+- `handshake` - an object containing:
+    - `stage` - the current stage counter
+    - `style` - noise handshake style `ix_psk0`, `xx`, etc
+- `header` - a nebula header object
+- `udpAddr` - a `net.UDPAddr` object
+- `udpIp` - a udp ip address
+- `vpnIp` - vpn ip of the host (remote or local)
+- `relay` - the vpnIp of the relay host that is or should be handling the relay packet
+- `relayFrom` - The vpnIp of the initial sender of the relayed packet 
+- `relayTo` - The vpnIp of the final destination of a relayed packet
+
+#### Example:
+
+```
+l.WithError(err).
+    WithField("vpnIp", IntIp(hostinfo.hostId)).
+    WithField("udpAddr", addr).
+    WithField("handshake", m{"stage": 1, "style": "ix"}).
+    Info("Invalid certificate from host")
+```

Makefile

@@ -1,7 +1,31 @@
 NEBULA_CMD_PATH = "./cmd/nebula"
-BUILD_NUMBER ?= dev+$(shell date -u '+%Y%m%d%H%M%S')
-GO111MODULE = on
-export GO111MODULE
+CGO_ENABLED = 0
+export CGO_ENABLED
+
+# Set up OS specific bits
+ifeq ($(OS),Windows_NT)
+	NEBULA_CMD_SUFFIX = .exe
+	NULL_FILE = nul
+	# RIO on windows does pointer stuff that makes go vet angry
+	VET_FLAGS = -unsafeptr=false
+else
+	NEBULA_CMD_SUFFIX =
+	NULL_FILE = /dev/null
+endif
+
+# Only defined the build number if we haven't already
+ifndef BUILD_NUMBER
+	ifeq ($(shell git describe --exact-match 2>$(NULL_FILE)),)
+		BUILD_NUMBER = $(shell git describe --abbrev=0 --match "v*" | cut -dv -f2)-$(shell git branch --show-current)-$(shell git describe --long --dirty | cut -d- -f2-)
+	else
+		BUILD_NUMBER = $(shell git describe --exact-match --dirty | cut -dv -f2)
+	endif
+endif
+
+DOCKER_IMAGE_REPO ?= nebulaoss/nebula
+DOCKER_IMAGE_TAG ?= latest
+
+LDFLAGS = -X main.Build=$(BUILD_NUMBER)
 
 ALL_LINUX = linux-amd64 \
 	linux-386 \
@@ -13,43 +37,112 @@ ALL_LINUX = linux-amd64 \
 	linux-mips \
 	linux-mipsle \
 	linux-mips64 \
-	linux-mips64le
+	linux-mips64le \
+	linux-mips-softfloat \
+	linux-riscv64 \
+        linux-loong64
+
+ALL_FREEBSD = freebsd-amd64 \
+	freebsd-arm64
+
+ALL_OPENBSD = openbsd-amd64 \
+	openbsd-arm64
+
+ALL_NETBSD = netbsd-amd64 \
+ 	netbsd-arm64
 
 ALL = $(ALL_LINUX) \
+	$(ALL_FREEBSD) \
+	$(ALL_OPENBSD) \
+	$(ALL_NETBSD) \
 	darwin-amd64 \
-	windows-amd64
+	darwin-arm64 \
+	windows-amd64 \
+	windows-arm64
+
+e2e:
+	$(TEST_ENV) go test -tags=e2e_testing -count=1 $(TEST_FLAGS) ./e2e
+
+e2ev: TEST_FLAGS = -v
+e2ev: e2e
+
+e2evv: TEST_ENV += TEST_LOGS=1
+e2evv: e2ev
+
+e2evvv: TEST_ENV += TEST_LOGS=2
+e2evvv: e2ev
+
+e2evvvv: TEST_ENV += TEST_LOGS=3
+e2evvvv: e2ev
+
+e2e-bench: TEST_FLAGS = -bench=. -benchmem -run=^$
+e2e-bench: e2e
+
+DOCKER_BIN = build/linux-amd64/nebula build/linux-amd64/nebula-cert
 
 all: $(ALL:%=build/%/nebula) $(ALL:%=build/%/nebula-cert)
 
+docker: docker/linux-$(shell go env GOARCH)
+
 release: $(ALL:%=build/nebula-%.tar.gz)
 
 release-linux: $(ALL_LINUX:%=build/nebula-%.tar.gz)
 
+release-freebsd: $(ALL_FREEBSD:%=build/nebula-%.tar.gz)
+
+release-openbsd: $(ALL_OPENBSD:%=build/nebula-%.tar.gz)
+
+release-netbsd: $(ALL_NETBSD:%=build/nebula-%.tar.gz)
+
+release-boringcrypto: build/nebula-linux-$(shell go env GOARCH)-boringcrypto.tar.gz
+
+BUILD_ARGS = -trimpath
+
 bin-windows: build/windows-amd64/nebula.exe build/windows-amd64/nebula-cert.exe
 	mv $? .
 
+bin-windows-arm64: build/windows-arm64/nebula.exe build/windows-arm64/nebula-cert.exe
+	mv $? .
+
 bin-darwin: build/darwin-amd64/nebula build/darwin-amd64/nebula-cert
 	mv $? .
 
+bin-freebsd: build/freebsd-amd64/nebula build/freebsd-amd64/nebula-cert
+	mv $? .
+
+bin-freebsd-arm64: build/freebsd-arm64/nebula build/freebsd-arm64/nebula-cert
+	mv $? .
+
+bin-boringcrypto: build/linux-$(shell go env GOARCH)-boringcrypto/nebula build/linux-$(shell go env GOARCH)-boringcrypto/nebula-cert
+	mv $? .
+
 bin:
-	go build -trimpath -ldflags "-X main.Build=$(BUILD_NUMBER)" -o ./nebula ${NEBULA_CMD_PATH}
-	go build -trimpath -ldflags "-X main.Build=$(BUILD_NUMBER)" -o ./nebula-cert ./cmd/nebula-cert
+	go build $(BUILD_ARGS) -ldflags "$(LDFLAGS)" -o ./nebula${NEBULA_CMD_SUFFIX} ${NEBULA_CMD_PATH}
+	go build $(BUILD_ARGS) -ldflags "$(LDFLAGS)" -o ./nebula-cert${NEBULA_CMD_SUFFIX} ./cmd/nebula-cert
 
 install:
-	go install -trimpath -ldflags "-X main.Build=$(BUILD_NUMBER)" ${NEBULA_CMD_PATH}
-	go install -trimpath -ldflags "-X main.Build=$(BUILD_NUMBER)" ./cmd/nebula-cert
+	go install $(BUILD_ARGS) -ldflags "$(LDFLAGS)" ${NEBULA_CMD_PATH}
+	go install $(BUILD_ARGS) -ldflags "$(LDFLAGS)" ./cmd/nebula-cert
+
+build/linux-arm-%: GOENV += GOARM=$(word 3, $(subst -, ,$*))
+build/linux-mips-%: GOENV += GOMIPS=$(word 3, $(subst -, ,$*))
+
+# Build an extra small binary for mips-softfloat
+build/linux-mips-softfloat/%: LDFLAGS += -s -w
+
+# boringcrypto
+build/linux-amd64-boringcrypto/%: GOENV += GOEXPERIMENT=boringcrypto CGO_ENABLED=1
+build/linux-arm64-boringcrypto/%: GOENV += GOEXPERIMENT=boringcrypto CGO_ENABLED=1
 
 build/%/nebula: .FORCE
 	GOOS=$(firstword $(subst -, , $*)) \
-		GOARCH=$(word 2, $(subst -, ,$*)) \
-		GOARM=$(word 3, $(subst -, ,$*)) \
-		go build -trimpath -o $@ -ldflags "-X main.Build=$(BUILD_NUMBER)" ${NEBULA_CMD_PATH}
+		GOARCH=$(word 2, $(subst -, ,$*)) $(GOENV) \
+		go build $(BUILD_ARGS) -o $@ -ldflags "$(LDFLAGS)" ${NEBULA_CMD_PATH}
 
 build/%/nebula-cert: .FORCE
 	GOOS=$(firstword $(subst -, , $*)) \
-		GOARCH=$(word 2, $(subst -, ,$*)) \
-		GOARM=$(word 3, $(subst -, ,$*)) \
-		go build -trimpath -o $@ -ldflags "-X main.Build=$(BUILD_NUMBER)" ./cmd/nebula-cert
+		GOARCH=$(word 2, $(subst -, ,$*)) $(GOENV) \
+		go build $(BUILD_ARGS) -o $@ -ldflags "$(LDFLAGS)" ./cmd/nebula-cert
 
 build/%/nebula.exe: build/%/nebula
 	mv $< $@
@@ -63,16 +156,28 @@ build/nebula-%.tar.gz: build/%/nebula build/%/nebula-cert
 build/nebula-%.zip: build/%/nebula.exe build/%/nebula-cert.exe
 	cd build/$* && zip ../nebula-$*.zip nebula.exe nebula-cert.exe
 
+docker/%: build/%/nebula build/%/nebula-cert
+	docker build . $(DOCKER_BUILD_ARGS) -f docker/Dockerfile --platform "$(subst -,/,$*)" --tag "${DOCKER_IMAGE_REPO}:${DOCKER_IMAGE_TAG}" --tag "${DOCKER_IMAGE_REPO}:$(BUILD_NUMBER)"
+
 vet:
-	go vet -v ./...
+	go vet $(VET_FLAGS) -v ./...
 
 test:
 	go test -v ./...
 
+test-boringcrypto:
+	GOEXPERIMENT=boringcrypto CGO_ENABLED=1 go test -v ./...
+
 test-cov-html:
 	go test -coverprofile=coverage.out
 	go tool cover -html=coverage.out
 
+build-test-mobile:
+	GOARCH=amd64 GOOS=ios go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
+	GOARCH=arm64 GOOS=ios go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
+	GOARCH=amd64 GOOS=android go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
+	GOARCH=arm64 GOOS=android go build $(shell go list ./... | grep -v '/cmd/\|/examples/')
+
 bench:
 	go test -bench=.
 
@@ -87,20 +192,40 @@ bench-cpu-long:
 proto: nebula.pb.go cert/cert.pb.go
 
 nebula.pb.go: nebula.proto .FORCE
-	go build github.com/golang/protobuf/protoc-gen-go
-	PATH="$(PWD):$(PATH)" protoc --go_out=. $<
-	rm protoc-gen-go
+	go build github.com/gogo/protobuf/protoc-gen-gogofaster
+	PATH="$(CURDIR):$(PATH)" protoc --gogofaster_out=paths=source_relative:. $<
+	rm protoc-gen-gogofaster
 
 cert/cert.pb.go: cert/cert.proto .FORCE
 	$(MAKE) -C cert cert.pb.go
 
 service:
-	@echo > /dev/null
+	@echo > $(NULL_FILE)
 	$(eval NEBULA_CMD_PATH := "./cmd/nebula-service")
 ifeq ($(words $(MAKECMDGOALS)),1)
-	$(MAKE) service ${.DEFAULT_GOAL} --no-print-directory
+	@$(MAKE) service ${.DEFAULT_GOAL} --no-print-directory
 endif
 
+bin-docker: bin build/linux-amd64/nebula build/linux-amd64/nebula-cert
+
+smoke-docker: bin-docker
+	cd .github/workflows/smoke/ && ./build.sh
+	cd .github/workflows/smoke/ && ./smoke.sh
+	cd .github/workflows/smoke/ && NAME="smoke-p256" CURVE="P256" ./build.sh
+	cd .github/workflows/smoke/ && NAME="smoke-p256" ./smoke.sh
+
+smoke-relay-docker: bin-docker
+	cd .github/workflows/smoke/ && ./build-relay.sh
+	cd .github/workflows/smoke/ && ./smoke-relay.sh
+
+smoke-docker-race: BUILD_ARGS = -race
+smoke-docker-race: CGO_ENABLED = 1
+smoke-docker-race: smoke-docker
+
+smoke-vagrant/%: bin-docker build/%/nebula
+	cd .github/workflows/smoke/ && ./build.sh $*
+	cd .github/workflows/smoke/ && ./smoke-vagrant.sh $*
+
 .FORCE:
-.PHONY: test test-cov-html bench bench-cpu bench-cpu-long bin proto release service
+.PHONY: bench bench-cpu bench-cpu-long bin build-test-mobile e2e e2ev e2evv e2evvv e2evvvv proto release service smoke-docker smoke-docker-race test test-cov-html smoke-vagrant/%
 .DEFAULT_GOAL := bin

README.md

@@ -1,7 +1,6 @@
-## What is Nebula?
+## What is Nebula?
 Nebula is a scalable overlay networking tool with a focus on performance, simplicity and security.
-It lets you seamlessly connect computers anywhere in the world. Nebula is portable, and runs on Linux, OSX, and Windows.
-(Also: keep this quiet, but we have an early prototype running on iOS).
+It lets you seamlessly connect computers anywhere in the world. Nebula is portable, and runs on Linux, OSX, Windows, iOS, and Android.
 It can be used to connect a small number of computers, but is also able to connect tens of thousands of computers.
 
 Nebula incorporates a number of existing concepts like encryption, security groups, certificates,
@@ -9,9 +8,59 @@ and tunneling, and each of those individual pieces existed before Nebula in vari
 What makes Nebula different to existing offerings is that it brings all of these ideas together,
 resulting in a sum that is greater than its individual parts.
 
+Further documentation can be found [here](https://nebula.defined.net/docs/).
+
 You can read more about Nebula [here](https://medium.com/p/884110a5579).
 
-You can also join the NebulaOSS Slack group [here](https://join.slack.com/t/nebulaoss/shared_invite/enQtOTA5MDI4NDg3MTg4LTkwY2EwNTI4NzQyMzc0M2ZlODBjNWI3NTY1MzhiOThiMmZlZjVkMTI0NGY4YTMyNjUwMWEyNzNkZTJmYzQxOGU)
+You can also join the NebulaOSS Slack group [here](https://join.slack.com/t/nebulaoss/shared_invite/enQtOTA5MDI4NDg3MTg4LTkwY2EwNTI4NzQyMzc0M2ZlODBjNWI3NTY1MzhiOThiMmZlZjVkMTI0NGY4YTMyNjUwMWEyNzNkZTJmYzQxOGU).
+
+## Supported Platforms
+
+#### Desktop and Server
+
+Check the [releases](https://github.com/slackhq/nebula/releases/latest) page for downloads or see the [Distribution Packages](https://github.com/slackhq/nebula#distribution-packages) section.
+
+- Linux - 64 and 32 bit, arm, and others
+- Windows
+- MacOS
+- Freebsd
+
+#### Distribution Packages
+
+- [Arch Linux](https://archlinux.org/packages/extra/x86_64/nebula/)
+    ```
+    $ sudo pacman -S nebula
+    ```
+
+- [Fedora Linux](https://src.fedoraproject.org/rpms/nebula)
+    ```
+    $ sudo dnf install nebula
+    ```
+
+- [Debian Linux](https://packages.debian.org/source/stable/nebula)
+    ```
+    $ sudo apt install nebula
+    ```
+
+- [Alpine Linux](https://pkgs.alpinelinux.org/packages?name=nebula)
+    ```
+    $ sudo apk add nebula
+    ```
+
+- [macOS Homebrew](https://github.com/Homebrew/homebrew-core/blob/HEAD/Formula/nebula.rb)
+    ```
+    $ brew install nebula
+    ```
+
+- [Docker](https://hub.docker.com/r/nebulaoss/nebula)
+    ```
+    $ docker pull nebulaoss/nebula
+    ```
+
+#### Mobile
+
+- [iOS](https://apps.apple.com/us/app/mobile-nebula/id1509587936?itsct=apps_box&itscg=30200)
+- [Android](https://play.google.com/store/apps/details?id=net.defined.mobile_nebula&pcampaignid=pcampaignidMKT-Other-global-all-co-prtnr-py-PartBadge-Mar2515-1)
 
 ## Technical Overview
 
@@ -21,15 +70,15 @@ Nebula's user-defined groups allow for provider agnostic traffic filtering betwe
 Discovery nodes allow individual peers to find each other and optionally use UDP hole punching to establish connections from behind most firewalls or NATs.
 Users can move data between nodes in any number of cloud service providers, datacenters, and endpoints, without needing to maintain a particular addressing scheme.
 
-Nebula uses elliptic curve Diffie-Hellman key exchange, and AES-256-GCM in its default configuration.
+Nebula uses Elliptic-curve Diffie-Hellman (`ECDH`) key exchange and `AES-256-GCM` in its default configuration.
 
-Nebula was created to provide a mechanism for groups hosts to communicate securely, even across the internet, while enabling expressive firewall definitions similar in style to cloud security groups.
+Nebula was created to provide a mechanism for groups of hosts to communicate securely, even across the internet, while enabling expressive firewall definitions similar in style to cloud security groups.
 
 ## Getting started (quickly)
 
 To set up a Nebula network, you'll need:
 
-#### 1. The [Nebula binaries](https://github.com/slackhq/nebula/releases) for your specific platform. Specifically you'll need `nebula-cert` and the specific nebula binary for each platform you use.
+#### 1. The [Nebula binaries](https://github.com/slackhq/nebula/releases) or [Distribution Packages](https://github.com/slackhq/nebula#distribution-packages) for your specific platform. Specifically you'll need `nebula-cert` and the specific nebula binary for each platform you use.
 
 #### 2. (Optional, but you really should..) At least one discovery node with a routable IP address, which we call a lighthouse.
 
@@ -64,18 +113,18 @@ Download a copy of the nebula [example configuration](https://github.com/slackhq
 
 #### 6. Copy nebula credentials, configuration, and binaries to each host
 
-For each host, copy the nebula binary to the host, along with `config.yaml` from step 5, and the files `ca.crt`, `{host}.crt`, and `{host}.key` from step 4.
+For each host, copy the nebula binary to the host, along with `config.yml` from step 5, and the files `ca.crt`, `{host}.crt`, and `{host}.key` from step 4.
 
 **DO NOT COPY `ca.key` TO INDIVIDUAL NODES.**
 
 #### 7. Run nebula on each host
 ```
-./nebula -config /path/to/config.yaml
+./nebula -config /path/to/config.yml
 ```
 
 ## Building Nebula from source
 
-Download go and clone this repo. Change to the nebula directory.
+Make sure you have [go](https://go.dev/doc/install) installed and clone this repo. Change to the nebula directory.
 
 To build nebula for all platforms:
 `make all`
@@ -85,6 +134,17 @@ To build nebula for a specific platform (ex, Windows):
 
 See the [Makefile](Makefile) for more details on build targets
 
+## Curve P256 and BoringCrypto
+
+The default curve used for cryptographic handshakes and signatures is Curve25519. This is the recommended setting for most users. If your deployment has certain compliance requirements, you have the option of creating your CA using `nebula-cert ca -curve P256` to use NIST Curve P256. The CA will then sign certificates using ECDSA P256, and any hosts using these certificates will use P256 for ECDH handshakes.
+
+In addition, Nebula can be built using the [BoringCrypto GOEXPERIMENT](https://github.com/golang/go/blob/go1.20/src/crypto/internal/boring/README.md) by running either of the following make targets:
+
+    make bin-boringcrypto
+    make release-boringcrypto
+
+This is not the recommended default deployment, but may be useful based on your compliance requirements.
+
 ## Credits
 
 Nebula was created at Slack Technologies, Inc by Nate Brown and Ryan Huber, with contributions from Oliver Fross, Alan Lam, Wade Simmons, and Lining Wang.

SECURITY.md

@@ -0,0 +1,12 @@
+Security Policy
+===============
+
+Reporting a Vulnerability
+-------------------------
+
+If you believe you have found a security vulnerability with Nebula, please let
+us know right away. We will investigate all reports and do our best to quickly
+fix valid issues.
+
+You can submit your report on [HackerOne](https://hackerone.com/slack) and our
+security team will respond as soon as possible.

allow_list.go

@@ -2,12 +2,28 @@ package nebula
 
 import (
 	"fmt"
+	"net/netip"
 	"regexp"
+
+	"github.com/gaissmai/bart"
+	"github.com/slackhq/nebula/config"
 )
 
 type AllowList struct {
 	// The values of this cidrTree are `bool`, signifying allow/deny
-	cidrTree *CIDRTree
+	cidrTree *bart.Table[bool]
+}
+
+type RemoteAllowList struct {
+	AllowList *AllowList
+
+	// Inside Range Specific, keys of this tree are inside CIDRs and values
+	// are *AllowList
+	insideAllowLists *bart.Table[*AllowList]
+}
+
+type LocalAllowList struct {
+	AllowList *AllowList
 
 	// To avoid ambiguity, all rules must be true, or all rules must be false.
 	nameRules []AllowListNameRule
@@ -18,21 +34,240 @@ type AllowListNameRule struct {
 	Allow bool
 }
 
-func (al *AllowList) Allow(ip uint32) bool {
+func NewLocalAllowListFromConfig(c *config.C, k string) (*LocalAllowList, error) {
+	var nameRules []AllowListNameRule
+	handleKey := func(key string, value interface{}) (bool, error) {
+		if key == "interfaces" {
+			var err error
+			nameRules, err = getAllowListInterfaces(k, value)
+			if err != nil {
+				return false, err
+			}
+
+			return true, nil
+		}
+		return false, nil
+	}
+
+	al, err := newAllowListFromConfig(c, k, handleKey)
+	if err != nil {
+		return nil, err
+	}
+	return &LocalAllowList{AllowList: al, nameRules: nameRules}, nil
+}
+
+func NewRemoteAllowListFromConfig(c *config.C, k, rangesKey string) (*RemoteAllowList, error) {
+	al, err := newAllowListFromConfig(c, k, nil)
+	if err != nil {
+		return nil, err
+	}
+	remoteAllowRanges, err := getRemoteAllowRanges(c, rangesKey)
+	if err != nil {
+		return nil, err
+	}
+	return &RemoteAllowList{AllowList: al, insideAllowLists: remoteAllowRanges}, nil
+}
+
+// If the handleKey func returns true, the rest of the parsing is skipped
+// for this key. This allows parsing of special values like `interfaces`.
+func newAllowListFromConfig(c *config.C, k string, handleKey func(key string, value interface{}) (bool, error)) (*AllowList, error) {
+	r := c.Get(k)
+	if r == nil {
+		return nil, nil
+	}
+
+	return newAllowList(k, r, handleKey)
+}
+
+// If the handleKey func returns true, the rest of the parsing is skipped
+// for this key. This allows parsing of special values like `interfaces`.
+func newAllowList(k string, raw interface{}, handleKey func(key string, value interface{}) (bool, error)) (*AllowList, error) {
+	rawMap, ok := raw.(map[interface{}]interface{})
+	if !ok {
+		return nil, fmt.Errorf("config `%s` has invalid type: %T", k, raw)
+	}
+
+	tree := new(bart.Table[bool])
+
+	// Keep track of the rules we have added for both ipv4 and ipv6
+	type allowListRules struct {
+		firstValue     bool
+		allValuesMatch bool
+		defaultSet     bool
+		allValues      bool
+	}
+
+	rules4 := allowListRules{firstValue: true, allValuesMatch: true, defaultSet: false}
+	rules6 := allowListRules{firstValue: true, allValuesMatch: true, defaultSet: false}
+
+	for rawKey, rawValue := range rawMap {
+		rawCIDR, ok := rawKey.(string)
+		if !ok {
+			return nil, fmt.Errorf("config `%s` has invalid key (type %T): %v", k, rawKey, rawKey)
+		}
+
+		if handleKey != nil {
+			handled, err := handleKey(rawCIDR, rawValue)
+			if err != nil {
+				return nil, err
+			}
+			if handled {
+				continue
+			}
+		}
+
+		value, ok := rawValue.(bool)
+		if !ok {
+			return nil, fmt.Errorf("config `%s` has invalid value (type %T): %v", k, rawValue, rawValue)
+		}
+
+		ipNet, err := netip.ParsePrefix(rawCIDR)
+		if err != nil {
+			return nil, fmt.Errorf("config `%s` has invalid CIDR: %s. %w", k, rawCIDR, err)
+		}
+
+		ipNet = netip.PrefixFrom(ipNet.Addr().Unmap(), ipNet.Bits())
+
+		// TODO: should we error on duplicate CIDRs in the config?
+		tree.Insert(ipNet, value)
+
+		maskBits := ipNet.Bits()
+
+		var rules *allowListRules
+		if ipNet.Addr().Is4() {
+			rules = &rules4
+		} else {
+			rules = &rules6
+		}
+
+		if rules.firstValue {
+			rules.allValues = value
+			rules.firstValue = false
+		} else {
+			if value != rules.allValues {
+				rules.allValuesMatch = false
+			}
+		}
+
+		// Check if this is 0.0.0.0/0 or ::/0
+		if maskBits == 0 {
+			rules.defaultSet = true
+		}
+	}
+
+	if !rules4.defaultSet {
+		if rules4.allValuesMatch {
+			tree.Insert(netip.PrefixFrom(netip.IPv4Unspecified(), 0), !rules4.allValues)
+		} else {
+			return nil, fmt.Errorf("config `%s` contains both true and false rules, but no default set for 0.0.0.0/0", k)
+		}
+	}
+
+	if !rules6.defaultSet {
+		if rules6.allValuesMatch {
+			tree.Insert(netip.PrefixFrom(netip.IPv6Unspecified(), 0), !rules6.allValues)
+		} else {
+			return nil, fmt.Errorf("config `%s` contains both true and false rules, but no default set for ::/0", k)
+		}
+	}
+
+	return &AllowList{cidrTree: tree}, nil
+}
+
+func getAllowListInterfaces(k string, v interface{}) ([]AllowListNameRule, error) {
+	var nameRules []AllowListNameRule
+
+	rawRules, ok := v.(map[interface{}]interface{})
+	if !ok {
+		return nil, fmt.Errorf("config `%s.interfaces` is invalid (type %T): %v", k, v, v)
+	}
+
+	firstEntry := true
+	var allValues bool
+	for rawName, rawAllow := range rawRules {
+		name, ok := rawName.(string)
+		if !ok {
+			return nil, fmt.Errorf("config `%s.interfaces` has invalid key (type %T): %v", k, rawName, rawName)
+		}
+		allow, ok := rawAllow.(bool)
+		if !ok {
+			return nil, fmt.Errorf("config `%s.interfaces` has invalid value (type %T): %v", k, rawAllow, rawAllow)
+		}
+
+		nameRE, err := regexp.Compile("^" + name + "$")
+		if err != nil {
+			return nil, fmt.Errorf("config `%s.interfaces` has invalid key: %s: %v", k, name, err)
+		}
+
+		nameRules = append(nameRules, AllowListNameRule{
+			Name:  nameRE,
+			Allow: allow,
+		})
+
+		if firstEntry {
+			allValues = allow
+			firstEntry = false
+		} else {
+			if allow != allValues {
+				return nil, fmt.Errorf("config `%s.interfaces` values must all be the same true/false value", k)
+			}
+		}
+	}
+
+	return nameRules, nil
+}
+
+func getRemoteAllowRanges(c *config.C, k string) (*bart.Table[*AllowList], error) {
+	value := c.Get(k)
+	if value == nil {
+		return nil, nil
+	}
+
+	remoteAllowRanges := new(bart.Table[*AllowList])
+
+	rawMap, ok := value.(map[interface{}]interface{})
+	if !ok {
+		return nil, fmt.Errorf("config `%s` has invalid type: %T", k, value)
+	}
+	for rawKey, rawValue := range rawMap {
+		rawCIDR, ok := rawKey.(string)
+		if !ok {
+			return nil, fmt.Errorf("config `%s` has invalid key (type %T): %v", k, rawKey, rawKey)
+		}
+
+		allowList, err := newAllowList(fmt.Sprintf("%s.%s", k, rawCIDR), rawValue, nil)
+		if err != nil {
+			return nil, err
+		}
+
+		ipNet, err := netip.ParsePrefix(rawCIDR)
+		if err != nil {
+			return nil, fmt.Errorf("config `%s` has invalid CIDR: %s. %w", k, rawCIDR, err)
+		}
+
+		remoteAllowRanges.Insert(netip.PrefixFrom(ipNet.Addr().Unmap(), ipNet.Bits()), allowList)
+	}
+
+	return remoteAllowRanges, nil
+}
+
+func (al *AllowList) Allow(ip netip.Addr) bool {
 	if al == nil {
 		return true
 	}
 
-	result := al.cidrTree.MostSpecificContains(ip)
-	switch v := result.(type) {
-	case bool:
-		return v
-	default:
-		panic(fmt.Errorf("invalid state, allowlist returned: %T %v", result, result))
-	}
+	result, _ := al.cidrTree.Lookup(ip)
+	return result
 }
 
-func (al *AllowList) AllowName(name string) bool {
+func (al *LocalAllowList) Allow(ip netip.Addr) bool {
+	if al == nil {
+		return true
+	}
+	return al.AllowList.Allow(ip)
+}
+
+func (al *LocalAllowList) AllowName(name string) bool {
 	if al == nil || len(al.nameRules) == 0 {
 		return true
 	}
@@ -46,3 +281,27 @@ func (al *AllowList) AllowName(name string) bool {
 	// If no rules match, return the default, which is the inverse of the rules
 	return !al.nameRules[0].Allow
 }
+
+func (al *RemoteAllowList) AllowUnknownVpnIp(ip netip.Addr) bool {
+	if al == nil {
+		return true
+	}
+	return al.AllowList.Allow(ip)
+}
+
+func (al *RemoteAllowList) Allow(vpnIp netip.Addr, ip netip.Addr) bool {
+	if !al.getInsideAllowList(vpnIp).Allow(ip) {
+		return false
+	}
+	return al.AllowList.Allow(ip)
+}
+
+func (al *RemoteAllowList) getInsideAllowList(vpnIp netip.Addr) *AllowList {
+	if al.insideAllowLists != nil {
+		inside, ok := al.insideAllowLists.Lookup(vpnIp)
+		if ok {
+			return inside
+		}
+	}
+	return nil
+}

allow_list_test.go

@@ -1,35 +1,133 @@
 package nebula
 
 import (
-	"net"
+	"net/netip"
 	"regexp"
 	"testing"
 
+	"github.com/gaissmai/bart"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/test"
 	"github.com/stretchr/testify/assert"
 )
 
-func TestAllowList_Allow(t *testing.T) {
-	assert.Equal(t, true, ((*AllowList)(nil)).Allow(ip2int(net.ParseIP("1.1.1.1"))))
+func TestNewAllowListFromConfig(t *testing.T) {
+	l := test.NewLogger()
+	c := config.NewC(l)
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"192.168.0.0": true,
+	}
+	r, err := newAllowListFromConfig(c, "allowlist", nil)
+	assert.EqualError(t, err, "config `allowlist` has invalid CIDR: 192.168.0.0. netip.ParsePrefix(\"192.168.0.0\"): no '/'")
+	assert.Nil(t, r)
 
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("0.0.0.0/0"), true)
-	tree.AddCIDR(getCIDR("10.0.0.0/8"), false)
-	tree.AddCIDR(getCIDR("10.42.42.0/24"), true)
-	al := &AllowList{cidrTree: tree}
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"192.168.0.0/16": "abc",
+	}
+	r, err = newAllowListFromConfig(c, "allowlist", nil)
+	assert.EqualError(t, err, "config `allowlist` has invalid value (type string): abc")
 
-	assert.Equal(t, true, al.Allow(ip2int(net.ParseIP("1.1.1.1"))))
-	assert.Equal(t, false, al.Allow(ip2int(net.ParseIP("10.0.0.4"))))
-	assert.Equal(t, true, al.Allow(ip2int(net.ParseIP("10.42.42.42"))))
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"192.168.0.0/16": true,
+		"10.0.0.0/8":     false,
+	}
+	r, err = newAllowListFromConfig(c, "allowlist", nil)
+	assert.EqualError(t, err, "config `allowlist` contains both true and false rules, but no default set for 0.0.0.0/0")
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"0.0.0.0/0":      true,
+		"10.0.0.0/8":     false,
+		"10.42.42.0/24":  true,
+		"fd00::/8":       true,
+		"fd00:fd00::/16": false,
+	}
+	r, err = newAllowListFromConfig(c, "allowlist", nil)
+	assert.EqualError(t, err, "config `allowlist` contains both true and false rules, but no default set for ::/0")
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"0.0.0.0/0":     true,
+		"10.0.0.0/8":    false,
+		"10.42.42.0/24": true,
+	}
+	r, err = newAllowListFromConfig(c, "allowlist", nil)
+	if assert.NoError(t, err) {
+		assert.NotNil(t, r)
+	}
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"0.0.0.0/0":      true,
+		"10.0.0.0/8":     false,
+		"10.42.42.0/24":  true,
+		"::/0":           false,
+		"fd00::/8":       true,
+		"fd00:fd00::/16": false,
+	}
+	r, err = newAllowListFromConfig(c, "allowlist", nil)
+	if assert.NoError(t, err) {
+		assert.NotNil(t, r)
+	}
+
+	// Test interface names
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"interfaces": map[interface{}]interface{}{
+			`docker.*`: "foo",
+		},
+	}
+	lr, err := NewLocalAllowListFromConfig(c, "allowlist")
+	assert.EqualError(t, err, "config `allowlist.interfaces` has invalid value (type string): foo")
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"interfaces": map[interface{}]interface{}{
+			`docker.*`: false,
+			`eth.*`:    true,
+		},
+	}
+	lr, err = NewLocalAllowListFromConfig(c, "allowlist")
+	assert.EqualError(t, err, "config `allowlist.interfaces` values must all be the same true/false value")
+
+	c.Settings["allowlist"] = map[interface{}]interface{}{
+		"interfaces": map[interface{}]interface{}{
+			`docker.*`: false,
+		},
+	}
+	lr, err = NewLocalAllowListFromConfig(c, "allowlist")
+	if assert.NoError(t, err) {
+		assert.NotNil(t, lr)
+	}
 }
 
-func TestAllowList_AllowName(t *testing.T) {
-	assert.Equal(t, true, ((*AllowList)(nil)).AllowName("docker0"))
+func TestAllowList_Allow(t *testing.T) {
+	assert.Equal(t, true, ((*AllowList)(nil)).Allow(netip.MustParseAddr("1.1.1.1")))
+
+	tree := new(bart.Table[bool])
+	tree.Insert(netip.MustParsePrefix("0.0.0.0/0"), true)
+	tree.Insert(netip.MustParsePrefix("10.0.0.0/8"), false)
+	tree.Insert(netip.MustParsePrefix("10.42.42.42/32"), true)
+	tree.Insert(netip.MustParsePrefix("10.42.0.0/16"), true)
+	tree.Insert(netip.MustParsePrefix("10.42.42.0/24"), true)
+	tree.Insert(netip.MustParsePrefix("10.42.42.0/24"), false)
+	tree.Insert(netip.MustParsePrefix("::1/128"), true)
+	tree.Insert(netip.MustParsePrefix("::2/128"), false)
+	al := &AllowList{cidrTree: tree}
+
+	assert.Equal(t, true, al.Allow(netip.MustParseAddr("1.1.1.1")))
+	assert.Equal(t, false, al.Allow(netip.MustParseAddr("10.0.0.4")))
+	assert.Equal(t, true, al.Allow(netip.MustParseAddr("10.42.42.42")))
+	assert.Equal(t, false, al.Allow(netip.MustParseAddr("10.42.42.41")))
+	assert.Equal(t, true, al.Allow(netip.MustParseAddr("10.42.0.1")))
+	assert.Equal(t, true, al.Allow(netip.MustParseAddr("::1")))
+	assert.Equal(t, false, al.Allow(netip.MustParseAddr("::2")))
+}
+
+func TestLocalAllowList_AllowName(t *testing.T) {
+	assert.Equal(t, true, ((*LocalAllowList)(nil)).AllowName("docker0"))
 
 	rules := []AllowListNameRule{
 		{Name: regexp.MustCompile("^docker.*$"), Allow: false},
 		{Name: regexp.MustCompile("^tun.*$"), Allow: false},
 	}
-	al := &AllowList{nameRules: rules}
+	al := &LocalAllowList{nameRules: rules}
 
 	assert.Equal(t, false, al.AllowName("docker0"))
 	assert.Equal(t, false, al.AllowName("tun0"))
@@ -39,7 +137,7 @@ func TestAllowList_AllowName(t *testing.T) {
 		{Name: regexp.MustCompile("^eth.*$"), Allow: true},
 		{Name: regexp.MustCompile("^ens.*$"), Allow: true},
 	}
-	al = &AllowList{nameRules: rules}
+	al = &LocalAllowList{nameRules: rules}
 
 	assert.Equal(t, false, al.AllowName("docker0"))
 	assert.Equal(t, true, al.AllowName("eth0"))

bits.go

@@ -26,7 +26,7 @@ func NewBits(bits uint64) *Bits {
 	}
 }
 
-func (b *Bits) Check(i uint64) bool {
+func (b *Bits) Check(l logrus.FieldLogger, i uint64) bool {
 	// If i is the next number, return true.
 	if i > b.current || (i == 0 && b.firstSeen == false && b.current < b.length) {
 		return true
@@ -47,7 +47,7 @@ func (b *Bits) Check(i uint64) bool {
 	return false
 }
 
-func (b *Bits) Update(i uint64) bool {
+func (b *Bits) Update(l *logrus.Logger, i uint64) bool {
 	// If i is the next number, return true and update current.
 	if i == b.current+1 {
 		// Report missed packets, we can only understand what was missed after the first window has been gone through

bits_test.go

@@ -3,10 +3,12 @@ package nebula
 import (
 	"testing"
 
+	"github.com/slackhq/nebula/test"
 	"github.com/stretchr/testify/assert"
 )
 
 func TestBits(t *testing.T) {
+	l := test.NewLogger()
 	b := NewBits(10)
 
 	// make sure it is the right size
@@ -14,46 +16,46 @@ func TestBits(t *testing.T) {
 
 	// This is initialized to zero - receive one. This should work.
 
-	assert.True(t, b.Check(1))
-	u := b.Update(1)
+	assert.True(t, b.Check(l, 1))
+	u := b.Update(l, 1)
 	assert.True(t, u)
 	assert.EqualValues(t, 1, b.current)
 	g := []bool{false, true, false, false, false, false, false, false, false, false}
 	assert.Equal(t, g, b.bits)
 
 	// Receive two
-	assert.True(t, b.Check(2))
-	u = b.Update(2)
+	assert.True(t, b.Check(l, 2))
+	u = b.Update(l, 2)
 	assert.True(t, u)
 	assert.EqualValues(t, 2, b.current)
 	g = []bool{false, true, true, false, false, false, false, false, false, false}
 	assert.Equal(t, g, b.bits)
 
 	// Receive two again - it will fail
-	assert.False(t, b.Check(2))
-	u = b.Update(2)
+	assert.False(t, b.Check(l, 2))
+	u = b.Update(l, 2)
 	assert.False(t, u)
 	assert.EqualValues(t, 2, b.current)
 
 	// Jump ahead to 15, which should clear everything and set the 6th element
-	assert.True(t, b.Check(15))
-	u = b.Update(15)
+	assert.True(t, b.Check(l, 15))
+	u = b.Update(l, 15)
 	assert.True(t, u)
 	assert.EqualValues(t, 15, b.current)
 	g = []bool{false, false, false, false, false, true, false, false, false, false}
 	assert.Equal(t, g, b.bits)
 
 	// Mark 14, which is allowed because it is in the window
-	assert.True(t, b.Check(14))
-	u = b.Update(14)
+	assert.True(t, b.Check(l, 14))
+	u = b.Update(l, 14)
 	assert.True(t, u)
 	assert.EqualValues(t, 15, b.current)
 	g = []bool{false, false, false, false, true, true, false, false, false, false}
 	assert.Equal(t, g, b.bits)
 
 	// Mark 5, which is not allowed because it is not in the window
-	assert.False(t, b.Check(5))
-	u = b.Update(5)
+	assert.False(t, b.Check(l, 5))
+	u = b.Update(l, 5)
 	assert.False(t, u)
 	assert.EqualValues(t, 15, b.current)
 	g = []bool{false, false, false, false, true, true, false, false, false, false}
@@ -61,63 +63,65 @@ func TestBits(t *testing.T) {
 
 	// make sure we handle wrapping around once to the current position
 	b = NewBits(10)
-	assert.True(t, b.Update(1))
-	assert.True(t, b.Update(11))
+	assert.True(t, b.Update(l, 1))
+	assert.True(t, b.Update(l, 11))
 	assert.Equal(t, []bool{false, true, false, false, false, false, false, false, false, false}, b.bits)
 
 	// Walk through a few windows in order
 	b = NewBits(10)
 	for i := uint64(0); i <= 100; i++ {
-		assert.True(t, b.Check(i), "Error while checking %v", i)
-		assert.True(t, b.Update(i), "Error while updating %v", i)
+		assert.True(t, b.Check(l, i), "Error while checking %v", i)
+		assert.True(t, b.Update(l, i), "Error while updating %v", i)
 	}
 }
 
 func TestBitsDupeCounter(t *testing.T) {
+	l := test.NewLogger()
 	b := NewBits(10)
 	b.lostCounter.Clear()
 	b.dupeCounter.Clear()
 	b.outOfWindowCounter.Clear()
 
-	assert.True(t, b.Update(1))
+	assert.True(t, b.Update(l, 1))
 	assert.Equal(t, int64(0), b.dupeCounter.Count())
 
-	assert.False(t, b.Update(1))
+	assert.False(t, b.Update(l, 1))
 	assert.Equal(t, int64(1), b.dupeCounter.Count())
 
-	assert.True(t, b.Update(2))
+	assert.True(t, b.Update(l, 2))
 	assert.Equal(t, int64(1), b.dupeCounter.Count())
 
-	assert.True(t, b.Update(3))
+	assert.True(t, b.Update(l, 3))
 	assert.Equal(t, int64(1), b.dupeCounter.Count())
 
-	assert.False(t, b.Update(1))
+	assert.False(t, b.Update(l, 1))
 	assert.Equal(t, int64(0), b.lostCounter.Count())
 	assert.Equal(t, int64(2), b.dupeCounter.Count())
 	assert.Equal(t, int64(0), b.outOfWindowCounter.Count())
 }
 
 func TestBitsOutOfWindowCounter(t *testing.T) {
+	l := test.NewLogger()
 	b := NewBits(10)
 	b.lostCounter.Clear()
 	b.dupeCounter.Clear()
 	b.outOfWindowCounter.Clear()
 
-	assert.True(t, b.Update(20))
+	assert.True(t, b.Update(l, 20))
 	assert.Equal(t, int64(0), b.outOfWindowCounter.Count())
 
-	assert.True(t, b.Update(21))
-	assert.True(t, b.Update(22))
-	assert.True(t, b.Update(23))
-	assert.True(t, b.Update(24))
-	assert.True(t, b.Update(25))
-	assert.True(t, b.Update(26))
-	assert.True(t, b.Update(27))
-	assert.True(t, b.Update(28))
-	assert.True(t, b.Update(29))
+	assert.True(t, b.Update(l, 21))
+	assert.True(t, b.Update(l, 22))
+	assert.True(t, b.Update(l, 23))
+	assert.True(t, b.Update(l, 24))
+	assert.True(t, b.Update(l, 25))
+	assert.True(t, b.Update(l, 26))
+	assert.True(t, b.Update(l, 27))
+	assert.True(t, b.Update(l, 28))
+	assert.True(t, b.Update(l, 29))
 	assert.Equal(t, int64(0), b.outOfWindowCounter.Count())
 
-	assert.False(t, b.Update(0))
+	assert.False(t, b.Update(l, 0))
 	assert.Equal(t, int64(1), b.outOfWindowCounter.Count())
 
 	//tODO: make sure lostcounter doesn't increase in orderly increment
@@ -127,23 +131,24 @@ func TestBitsOutOfWindowCounter(t *testing.T) {
 }
 
 func TestBitsLostCounter(t *testing.T) {
+	l := test.NewLogger()
 	b := NewBits(10)
 	b.lostCounter.Clear()
 	b.dupeCounter.Clear()
 	b.outOfWindowCounter.Clear()
 
 	//assert.True(t, b.Update(0))
-	assert.True(t, b.Update(0))
-	assert.True(t, b.Update(20))
-	assert.True(t, b.Update(21))
-	assert.True(t, b.Update(22))
-	assert.True(t, b.Update(23))
-	assert.True(t, b.Update(24))
-	assert.True(t, b.Update(25))
-	assert.True(t, b.Update(26))
-	assert.True(t, b.Update(27))
-	assert.True(t, b.Update(28))
-	assert.True(t, b.Update(29))
+	assert.True(t, b.Update(l, 0))
+	assert.True(t, b.Update(l, 20))
+	assert.True(t, b.Update(l, 21))
+	assert.True(t, b.Update(l, 22))
+	assert.True(t, b.Update(l, 23))
+	assert.True(t, b.Update(l, 24))
+	assert.True(t, b.Update(l, 25))
+	assert.True(t, b.Update(l, 26))
+	assert.True(t, b.Update(l, 27))
+	assert.True(t, b.Update(l, 28))
+	assert.True(t, b.Update(l, 29))
 	assert.Equal(t, int64(20), b.lostCounter.Count())
 	assert.Equal(t, int64(0), b.dupeCounter.Count())
 	assert.Equal(t, int64(0), b.outOfWindowCounter.Count())
@@ -153,56 +158,56 @@ func TestBitsLostCounter(t *testing.T) {
 	b.dupeCounter.Clear()
 	b.outOfWindowCounter.Clear()
 
-	assert.True(t, b.Update(0))
+	assert.True(t, b.Update(l, 0))
 	assert.Equal(t, int64(0), b.lostCounter.Count())
-	assert.True(t, b.Update(9))
+	assert.True(t, b.Update(l, 9))
 	assert.Equal(t, int64(0), b.lostCounter.Count())
 	// 10 will set 0 index, 0 was already set, no lost packets
-	assert.True(t, b.Update(10))
+	assert.True(t, b.Update(l, 10))
 	assert.Equal(t, int64(0), b.lostCounter.Count())
 	// 11 will set 1 index, 1 was missed, we should see 1 packet lost
-	assert.True(t, b.Update(11))
+	assert.True(t, b.Update(l, 11))
 	assert.Equal(t, int64(1), b.lostCounter.Count())
 	// Now let's fill in the window, should end up with 8 lost packets
-	assert.True(t, b.Update(12))
-	assert.True(t, b.Update(13))
-	assert.True(t, b.Update(14))
-	assert.True(t, b.Update(15))
-	assert.True(t, b.Update(16))
-	assert.True(t, b.Update(17))
-	assert.True(t, b.Update(18))
-	assert.True(t, b.Update(19))
+	assert.True(t, b.Update(l, 12))
+	assert.True(t, b.Update(l, 13))
+	assert.True(t, b.Update(l, 14))
+	assert.True(t, b.Update(l, 15))
+	assert.True(t, b.Update(l, 16))
+	assert.True(t, b.Update(l, 17))
+	assert.True(t, b.Update(l, 18))
+	assert.True(t, b.Update(l, 19))
 	assert.Equal(t, int64(8), b.lostCounter.Count())
 
 	// Jump ahead by a window size
-	assert.True(t, b.Update(29))
+	assert.True(t, b.Update(l, 29))
 	assert.Equal(t, int64(8), b.lostCounter.Count())
 	// Now lets walk ahead normally through the window, the missed packets should fill in
-	assert.True(t, b.Update(30))
-	assert.True(t, b.Update(31))
-	assert.True(t, b.Update(32))
-	assert.True(t, b.Update(33))
-	assert.True(t, b.Update(34))
-	assert.True(t, b.Update(35))
-	assert.True(t, b.Update(36))
-	assert.True(t, b.Update(37))
-	assert.True(t, b.Update(38))
+	assert.True(t, b.Update(l, 30))
+	assert.True(t, b.Update(l, 31))
+	assert.True(t, b.Update(l, 32))
+	assert.True(t, b.Update(l, 33))
+	assert.True(t, b.Update(l, 34))
+	assert.True(t, b.Update(l, 35))
+	assert.True(t, b.Update(l, 36))
+	assert.True(t, b.Update(l, 37))
+	assert.True(t, b.Update(l, 38))
 	// 39 packets tracked, 22 seen, 17 lost
 	assert.Equal(t, int64(17), b.lostCounter.Count())
 
 	// Jump ahead by 2 windows, should have recording 1 full window missing
-	assert.True(t, b.Update(58))
+	assert.True(t, b.Update(l, 58))
 	assert.Equal(t, int64(27), b.lostCounter.Count())
 	// Now lets walk ahead normally through the window, the missed packets should fill in from this window
-	assert.True(t, b.Update(59))
-	assert.True(t, b.Update(60))
-	assert.True(t, b.Update(61))
-	assert.True(t, b.Update(62))
-	assert.True(t, b.Update(63))
-	assert.True(t, b.Update(64))
-	assert.True(t, b.Update(65))
-	assert.True(t, b.Update(66))
-	assert.True(t, b.Update(67))
+	assert.True(t, b.Update(l, 59))
+	assert.True(t, b.Update(l, 60))
+	assert.True(t, b.Update(l, 61))
+	assert.True(t, b.Update(l, 62))
+	assert.True(t, b.Update(l, 63))
+	assert.True(t, b.Update(l, 64))
+	assert.True(t, b.Update(l, 65))
+	assert.True(t, b.Update(l, 66))
+	assert.True(t, b.Update(l, 67))
 	// 68 packets tracked, 32 seen, 36 missed
 	assert.Equal(t, int64(36), b.lostCounter.Count())
 	assert.Equal(t, int64(0), b.dupeCounter.Count())
@@ -212,10 +217,10 @@ func TestBitsLostCounter(t *testing.T) {
 func BenchmarkBits(b *testing.B) {
 	z := NewBits(10)
 	for n := 0; n < b.N; n++ {
-		for i, _ := range z.bits {
+		for i := range z.bits {
 			z.bits[i] = true
 		}
-		for i, _ := range z.bits {
+		for i := range z.bits {
 			z.bits[i] = false
 		}
 

boring.go

@@ -0,0 +1,8 @@
+//go:build boringcrypto
+// +build boringcrypto
+
+package nebula
+
+import "crypto/boring"
+
+var boringEnabled = boring.Enabled

calculated_remote.go

@@ -0,0 +1,159 @@
+package nebula
+
+import (
+	"encoding/binary"
+	"fmt"
+	"math"
+	"net"
+	"net/netip"
+	"strconv"
+
+	"github.com/gaissmai/bart"
+	"github.com/slackhq/nebula/config"
+)
+
+// This allows us to "guess" what the remote might be for a host while we wait
+// for the lighthouse response. See "lighthouse.calculated_remotes" in the
+// example config file.
+type calculatedRemote struct {
+	ipNet netip.Prefix
+	mask  netip.Prefix
+	port  uint32
+}
+
+func newCalculatedRemote(maskCidr netip.Prefix, port int) (*calculatedRemote, error) {
+	masked := maskCidr.Masked()
+	if port < 0 || port > math.MaxUint16 {
+		return nil, fmt.Errorf("invalid port: %d", port)
+	}
+
+	return &calculatedRemote{
+		ipNet: maskCidr,
+		mask:  masked,
+		port:  uint32(port),
+	}, nil
+}
+
+func (c *calculatedRemote) String() string {
+	return fmt.Sprintf("CalculatedRemote(mask=%v port=%d)", c.ipNet, c.port)
+}
+
+func (c *calculatedRemote) Apply(ip netip.Addr) *Ip4AndPort {
+	// Combine the masked bytes of the "mask" IP with the unmasked bytes
+	// of the overlay IP
+	if c.ipNet.Addr().Is4() {
+		return c.apply4(ip)
+	}
+	return c.apply6(ip)
+}
+
+func (c *calculatedRemote) apply4(ip netip.Addr) *Ip4AndPort {
+	//TODO: IPV6-WORK this can be less crappy
+	maskb := net.CIDRMask(c.mask.Bits(), c.mask.Addr().BitLen())
+	mask := binary.BigEndian.Uint32(maskb[:])
+
+	b := c.mask.Addr().As4()
+	maskIp := binary.BigEndian.Uint32(b[:])
+
+	b = ip.As4()
+	intIp := binary.BigEndian.Uint32(b[:])
+
+	return &Ip4AndPort{(maskIp & mask) | (intIp & ^mask), c.port}
+}
+
+func (c *calculatedRemote) apply6(ip netip.Addr) *Ip4AndPort {
+	//TODO: IPV6-WORK
+	panic("Can not calculate ipv6 remote addresses")
+}
+
+func NewCalculatedRemotesFromConfig(c *config.C, k string) (*bart.Table[[]*calculatedRemote], error) {
+	value := c.Get(k)
+	if value == nil {
+		return nil, nil
+	}
+
+	calculatedRemotes := new(bart.Table[[]*calculatedRemote])
+
+	rawMap, ok := value.(map[any]any)
+	if !ok {
+		return nil, fmt.Errorf("config `%s` has invalid type: %T", k, value)
+	}
+	for rawKey, rawValue := range rawMap {
+		rawCIDR, ok := rawKey.(string)
+		if !ok {
+			return nil, fmt.Errorf("config `%s` has invalid key (type %T): %v", k, rawKey, rawKey)
+		}
+
+		cidr, err := netip.ParsePrefix(rawCIDR)
+		if err != nil {
+			return nil, fmt.Errorf("config `%s` has invalid CIDR: %s", k, rawCIDR)
+		}
+
+		//TODO: IPV6-WORK this does not verify that rawValue contains the same bits as cidr here
+		entry, err := newCalculatedRemotesListFromConfig(rawValue)
+		if err != nil {
+			return nil, fmt.Errorf("config '%s.%s': %w", k, rawCIDR, err)
+		}
+
+		calculatedRemotes.Insert(cidr, entry)
+	}
+
+	return calculatedRemotes, nil
+}
+
+func newCalculatedRemotesListFromConfig(raw any) ([]*calculatedRemote, error) {
+	rawList, ok := raw.([]any)
+	if !ok {
+		return nil, fmt.Errorf("calculated_remotes entry has invalid type: %T", raw)
+	}
+
+	var l []*calculatedRemote
+	for _, e := range rawList {
+		c, err := newCalculatedRemotesEntryFromConfig(e)
+		if err != nil {
+			return nil, fmt.Errorf("calculated_remotes entry: %w", err)
+		}
+		l = append(l, c)
+	}
+
+	return l, nil
+}
+
+func newCalculatedRemotesEntryFromConfig(raw any) (*calculatedRemote, error) {
+	rawMap, ok := raw.(map[any]any)
+	if !ok {
+		return nil, fmt.Errorf("invalid type: %T", raw)
+	}
+
+	rawValue := rawMap["mask"]
+	if rawValue == nil {
+		return nil, fmt.Errorf("missing mask: %v", rawMap)
+	}
+	rawMask, ok := rawValue.(string)
+	if !ok {
+		return nil, fmt.Errorf("invalid mask (type %T): %v", rawValue, rawValue)
+	}
+	maskCidr, err := netip.ParsePrefix(rawMask)
+	if err != nil {
+		return nil, fmt.Errorf("invalid mask: %s", rawMask)
+	}
+
+	var port int
+	rawValue = rawMap["port"]
+	if rawValue == nil {
+		return nil, fmt.Errorf("missing port: %v", rawMap)
+	}
+	switch v := rawValue.(type) {
+	case int:
+		port = v
+	case string:
+		port, err = strconv.Atoi(v)
+		if err != nil {
+			return nil, fmt.Errorf("invalid port: %s: %w", v, err)
+		}
+	default:
+		return nil, fmt.Errorf("invalid port (type %T): %v", rawValue, rawValue)
+	}
+
+	return newCalculatedRemote(maskCidr, port)
+}

calculated_remote_test.go

@@ -0,0 +1,25 @@
+package nebula
+
+import (
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestCalculatedRemoteApply(t *testing.T) {
+	ipNet, err := netip.ParsePrefix("192.168.1.0/24")
+	require.NoError(t, err)
+
+	c, err := newCalculatedRemote(ipNet, 4242)
+	require.NoError(t, err)
+
+	input, err := netip.ParseAddr("10.0.10.182")
+	assert.NoError(t, err)
+
+	expected, err := netip.ParseAddr("192.168.1.182")
+	assert.NoError(t, err)
+
+	assert.Equal(t, NewIp4AndPortFromNetIP(expected, 4242), c.Apply(input))
+}

cert.go

@@ -1,159 +0,0 @@
-package nebula
-
-import (
-	"errors"
-	"fmt"
-	"io/ioutil"
-	"strings"
-	"time"
-
-	"github.com/slackhq/nebula/cert"
-)
-
-var trustedCAs *cert.NebulaCAPool
-
-type CertState struct {
-	certificate         *cert.NebulaCertificate
-	rawCertificate      []byte
-	rawCertificateNoKey []byte
-	publicKey           []byte
-	privateKey          []byte
-}
-
-func NewCertState(certificate *cert.NebulaCertificate, privateKey []byte) (*CertState, error) {
-	// Marshal the certificate to ensure it is valid
-	rawCertificate, err := certificate.Marshal()
-	if err != nil {
-		return nil, fmt.Errorf("invalid nebula certificate on interface: %s", err)
-	}
-
-	publicKey := certificate.Details.PublicKey
-	cs := &CertState{
-		rawCertificate: rawCertificate,
-		certificate:    certificate, // PublicKey has been set to nil above
-		privateKey:     privateKey,
-		publicKey:      publicKey,
-	}
-
-	cs.certificate.Details.PublicKey = nil
-	rawCertNoKey, err := cs.certificate.Marshal()
-	if err != nil {
-		return nil, fmt.Errorf("error marshalling certificate no key: %s", err)
-	}
-	cs.rawCertificateNoKey = rawCertNoKey
-	// put public key back
-	cs.certificate.Details.PublicKey = cs.publicKey
-	return cs, nil
-}
-
-func NewCertStateFromConfig(c *Config) (*CertState, error) {
-	var pemPrivateKey []byte
-	var err error
-
-	privPathOrPEM := c.GetString("pki.key", "")
-	if privPathOrPEM == "" {
-		// Support backwards compat with the old x509
-		//TODO: remove after this is rolled out everywhere - NB 2018/02/23
-		privPathOrPEM = c.GetString("x509.key", "")
-	}
-
-	if privPathOrPEM == "" {
-		return nil, errors.New("no pki.key path or PEM data provided")
-	}
-
-	if strings.Contains(privPathOrPEM, "-----BEGIN") {
-		pemPrivateKey = []byte(privPathOrPEM)
-		privPathOrPEM = "<inline>"
-	} else {
-		pemPrivateKey, err = ioutil.ReadFile(privPathOrPEM)
-		if err != nil {
-			return nil, fmt.Errorf("unable to read pki.key file %s: %s", privPathOrPEM, err)
-		}
-	}
-
-	rawKey, _, err := cert.UnmarshalX25519PrivateKey(pemPrivateKey)
-	if err != nil {
-		return nil, fmt.Errorf("error while unmarshaling pki.key %s: %s", privPathOrPEM, err)
-	}
-
-	var rawCert []byte
-
-	pubPathOrPEM := c.GetString("pki.cert", "")
-	if pubPathOrPEM == "" {
-		// Support backwards compat with the old x509
-		//TODO: remove after this is rolled out everywhere - NB 2018/02/23
-		pubPathOrPEM = c.GetString("x509.cert", "")
-	}
-
-	if pubPathOrPEM == "" {
-		return nil, errors.New("no pki.cert path or PEM data provided")
-	}
-
-	if strings.Contains(pubPathOrPEM, "-----BEGIN") {
-		rawCert = []byte(pubPathOrPEM)
-		pubPathOrPEM = "<inline>"
-	} else {
-		rawCert, err = ioutil.ReadFile(pubPathOrPEM)
-		if err != nil {
-			return nil, fmt.Errorf("unable to read pki.cert file %s: %s", pubPathOrPEM, err)
-		}
-	}
-
-	nebulaCert, _, err := cert.UnmarshalNebulaCertificateFromPEM(rawCert)
-	if err != nil {
-		return nil, fmt.Errorf("error while unmarshaling pki.cert %s: %s", pubPathOrPEM, err)
-	}
-
-	if nebulaCert.Expired(time.Now()) {
-		return nil, fmt.Errorf("nebula certificate for this host is expired")
-	}
-
-	if len(nebulaCert.Details.Ips) == 0 {
-		return nil, fmt.Errorf("no IPs encoded in certificate")
-	}
-
-	if err = nebulaCert.VerifyPrivateKey(rawKey); err != nil {
-		return nil, fmt.Errorf("private key is not a pair with public key in nebula cert")
-	}
-
-	return NewCertState(nebulaCert, rawKey)
-}
-
-func loadCAFromConfig(c *Config) (*cert.NebulaCAPool, error) {
-	var rawCA []byte
-	var err error
-
-	caPathOrPEM := c.GetString("pki.ca", "")
-	if caPathOrPEM == "" {
-		// Support backwards compat with the old x509
-		//TODO: remove after this is rolled out everywhere - NB 2018/02/23
-		caPathOrPEM = c.GetString("x509.ca", "")
-	}
-
-	if caPathOrPEM == "" {
-		return nil, errors.New("no pki.ca path or PEM data provided")
-	}
-
-	if strings.Contains(caPathOrPEM, "-----BEGIN") {
-		rawCA = []byte(caPathOrPEM)
-		caPathOrPEM = "<inline>"
-	} else {
-		rawCA, err = ioutil.ReadFile(caPathOrPEM)
-		if err != nil {
-			return nil, fmt.Errorf("unable to read pki.ca file %s: %s", caPathOrPEM, err)
-		}
-	}
-
-	CAs, err := cert.NewCAPoolFromBytes(rawCA)
-	if err != nil {
-		return nil, fmt.Errorf("error while adding CA certificate to CA trust store: %s", err)
-	}
-
-	// pki.blacklist entered the scene at about the same time we aliased x509 to pki, not supporting backwards compat
-	for _, fp := range c.GetStringSlice("pki.blacklist", []string{}) {
-		l.WithField("fingerprint", fp).Infof("Blacklisting cert")
-		CAs.BlacklistFingerprint(fp)
-	}
-
-	return CAs, nil
-}

cert/Makefile

@@ -2,8 +2,8 @@ GO111MODULE = on
 export GO111MODULE
 
 cert.pb.go: cert.proto .FORCE
-	go build github.com/golang/protobuf/protoc-gen-go
-	PATH="$(PWD):$(PATH)" protoc --go_out=. $<
+	go build google.golang.org/protobuf/cmd/protoc-gen-go
+	PATH="$(CURDIR):$(PATH)" protoc --go_out=. --go_opt=paths=source_relative $<
 	rm protoc-gen-go
 
 .FORCE:

cert/ca.go

@@ -1,6 +1,7 @@
 package cert
 
 import (
+	"errors"
 	"fmt"
 	"strings"
 	"time"
@@ -8,33 +9,54 @@ import (
 
 type NebulaCAPool struct {
 	CAs           map[string]*NebulaCertificate
-	certBlacklist map[string]struct{}
+	certBlocklist map[string]struct{}
 }
 
 // NewCAPool creates a CAPool
 func NewCAPool() *NebulaCAPool {
 	ca := NebulaCAPool{
 		CAs:           make(map[string]*NebulaCertificate),
-		certBlacklist: make(map[string]struct{}),
+		certBlocklist: make(map[string]struct{}),
 	}
 
 	return &ca
 }
 
-func NewCAPoolFromBytes(caPEMs []byte) (*NebulaCAPool, error) {
+// NewCAPoolFromBytes will create a new CA pool from the provided
+// input bytes, which must be a PEM-encoded set of nebula certificates.
+// If the pool contains unsupported certificates, they will generate warnings
+// in the []error return arg.
+// If the pool contains any expired certificates, an ErrExpired will be
+// returned along with the pool. The caller must handle any such errors.
+func NewCAPoolFromBytes(caPEMs []byte) (*NebulaCAPool, []error, error) {
 	pool := NewCAPool()
 	var err error
+	var warnings []error
+	good := 0
+
 	for {
 		caPEMs, err = pool.AddCACertificate(caPEMs)
-		if err != nil {
-			return nil, err
+		if errors.Is(err, ErrExpired) {
+			warnings = append(warnings, err)
+		} else if errors.Is(err, ErrInvalidPEMCertificateUnsupported) {
+			warnings = append(warnings, err)
+		} else if err != nil {
+			return nil, warnings, err
+		} else {
+			// Only consider a good certificate if there were no errors present
+			good++
 		}
-		if caPEMs == nil || len(caPEMs) == 0 || strings.TrimSpace(string(caPEMs)) == "" {
+
+		if len(caPEMs) == 0 || strings.TrimSpace(string(caPEMs)) == "" {
 			break
 		}
 	}
 
-	return pool, nil
+	if good == 0 {
+		return nil, warnings, errors.New("no valid CA certificates present")
+	}
+
+	return pool, warnings, nil
 }
 
 // AddCACertificate verifies a Nebula CA certificate and adds it to the pool
@@ -47,15 +69,11 @@ func (ncp *NebulaCAPool) AddCACertificate(pemBytes []byte) ([]byte, error) {
 	}
 
 	if !c.Details.IsCA {
-		return pemBytes, fmt.Errorf("provided certificate was not a CA; %s", c.Details.Name)
+		return pemBytes, fmt.Errorf("%s: %w", c.Details.Name, ErrNotCA)
 	}
 
 	if !c.CheckSignature(c.Details.PublicKey) {
-		return pemBytes, fmt.Errorf("provided certificate was not self signed; %s", c.Details.Name)
-	}
-
-	if c.Expired(time.Now()) {
-		return pemBytes, fmt.Errorf("provided CA certificate is expired; %s", c.Details.Name)
+		return pemBytes, fmt.Errorf("%s: %w", c.Details.Name, ErrNotSelfSigned)
 	}
 
 	sum, err := c.Sha256Sum()
@@ -64,27 +82,37 @@ func (ncp *NebulaCAPool) AddCACertificate(pemBytes []byte) ([]byte, error) {
 	}
 
 	ncp.CAs[sum] = c
+	if c.Expired(time.Now()) {
+		return pemBytes, fmt.Errorf("%s: %w", c.Details.Name, ErrExpired)
+	}
+
 	return pemBytes, nil
 }
 
-// BlacklistFingerprint adds a cert fingerprint to the blacklist
-func (ncp *NebulaCAPool) BlacklistFingerprint(f string) {
-	ncp.certBlacklist[f] = struct{}{}
+// BlocklistFingerprint adds a cert fingerprint to the blocklist
+func (ncp *NebulaCAPool) BlocklistFingerprint(f string) {
+	ncp.certBlocklist[f] = struct{}{}
 }
 
-// ResetCertBlacklist removes all previously blacklisted cert fingerprints
-func (ncp *NebulaCAPool) ResetCertBlacklist() {
-	ncp.certBlacklist = make(map[string]struct{})
+// ResetCertBlocklist removes all previously blocklisted cert fingerprints
+func (ncp *NebulaCAPool) ResetCertBlocklist() {
+	ncp.certBlocklist = make(map[string]struct{})
 }
 
-// IsBlacklisted returns true if the fingerprint fails to generate or has been explicitly blacklisted
-func (ncp *NebulaCAPool) IsBlacklisted(c *NebulaCertificate) bool {
-	h, err := c.Sha256Sum()
+// NOTE: This uses an internal cache for Sha256Sum() that will not be invalidated
+// automatically if you manually change any fields in the NebulaCertificate.
+func (ncp *NebulaCAPool) IsBlocklisted(c *NebulaCertificate) bool {
+	return ncp.isBlocklistedWithCache(c, false)
+}
+
+// IsBlocklisted returns true if the fingerprint fails to generate or has been explicitly blocklisted
+func (ncp *NebulaCAPool) isBlocklistedWithCache(c *NebulaCertificate, useCache bool) bool {
+	h, err := c.sha256SumWithCache(useCache)
 	if err != nil {
 		return true
 	}
 
-	if _, ok := ncp.certBlacklist[h]; ok {
+	if _, ok := ncp.certBlocklist[h]; ok {
 		return true
 	}
 

cert/cert.go

@@ -1,36 +1,57 @@
 package cert
 
 import (
-	"crypto"
+	"bytes"
+	"crypto/ecdh"
+	"crypto/ecdsa"
+	"crypto/ed25519"
+	"crypto/elliptic"
 	"crypto/rand"
 	"crypto/sha256"
 	"encoding/binary"
 	"encoding/hex"
+	"encoding/json"
 	"encoding/pem"
+	"errors"
 	"fmt"
+	"math"
+	"math/big"
 	"net"
+	"sync/atomic"
 	"time"
 
-	"bytes"
-	"encoding/json"
-	"github.com/golang/protobuf/proto"
 	"golang.org/x/crypto/curve25519"
-	"golang.org/x/crypto/ed25519"
+	"google.golang.org/protobuf/proto"
 )
 
 const publicKeyLen = 32
 
 const (
-	CertBanner              = "NEBULA CERTIFICATE"
-	X25519PrivateKeyBanner  = "NEBULA X25519 PRIVATE KEY"
-	X25519PublicKeyBanner   = "NEBULA X25519 PUBLIC KEY"
-	Ed25519PrivateKeyBanner = "NEBULA ED25519 PRIVATE KEY"
-	Ed25519PublicKeyBanner  = "NEBULA ED25519 PUBLIC KEY"
+	CertBanner                       = "NEBULA CERTIFICATE"
+	CertificateV2Banner              = "NEBULA CERTIFICATE V2"
+	X25519PrivateKeyBanner           = "NEBULA X25519 PRIVATE KEY"
+	X25519PublicKeyBanner            = "NEBULA X25519 PUBLIC KEY"
+	EncryptedEd25519PrivateKeyBanner = "NEBULA ED25519 ENCRYPTED PRIVATE KEY"
+	Ed25519PrivateKeyBanner          = "NEBULA ED25519 PRIVATE KEY"
+	Ed25519PublicKeyBanner           = "NEBULA ED25519 PUBLIC KEY"
+
+	P256PrivateKeyBanner               = "NEBULA P256 PRIVATE KEY"
+	P256PublicKeyBanner                = "NEBULA P256 PUBLIC KEY"
+	EncryptedECDSAP256PrivateKeyBanner = "NEBULA ECDSA P256 ENCRYPTED PRIVATE KEY"
+	ECDSAP256PrivateKeyBanner          = "NEBULA ECDSA P256 PRIVATE KEY"
 )
 
 type NebulaCertificate struct {
 	Details   NebulaCertificateDetails
 	Signature []byte
+
+	// the cached hex string of the calculated sha256sum
+	// for VerifyWithCache
+	sha256sum atomic.Pointer[string]
+
+	// the cached public key bytes if they were verified as the signer
+	// for VerifyWithCache
+	signatureVerified atomic.Pointer[[]byte]
 }
 
 type NebulaCertificateDetails struct {
@@ -46,10 +67,25 @@ type NebulaCertificateDetails struct {
 
 	// Map of groups for faster lookup
 	InvertedGroups map[string]struct{}
+
+	Curve Curve
+}
+
+type NebulaEncryptedData struct {
+	EncryptionMetadata NebulaEncryptionMetadata
+	Ciphertext         []byte
+}
+
+type NebulaEncryptionMetadata struct {
+	EncryptionAlgorithm string
+	Argon2Parameters    Argon2Parameters
 }
 
 type m map[string]interface{}
 
+// Returned if we try to unmarshal an encrypted private key without a passphrase
+var ErrPrivateKeyEncrypted = errors.New("private key must be decrypted")
+
 // UnmarshalNebulaCertificate will unmarshal a protobuf byte representation of a nebula cert
 func UnmarshalNebulaCertificate(b []byte) (*NebulaCertificate, error) {
 	if len(b) == 0 {
@@ -61,6 +97,10 @@ func UnmarshalNebulaCertificate(b []byte) (*NebulaCertificate, error) {
 		return nil, err
 	}
 
+	if rc.Details == nil {
+		return nil, fmt.Errorf("encoded Details was nil")
+	}
+
 	if len(rc.Details.Ips)%2 != 0 {
 		return nil, fmt.Errorf("encoded IPs should be in pairs, an odd number was found")
 	}
@@ -80,6 +120,7 @@ func UnmarshalNebulaCertificate(b []byte) (*NebulaCertificate, error) {
 			PublicKey:      make([]byte, len(rc.Details.PublicKey)),
 			IsCA:           rc.Details.IsCA,
 			InvertedGroups: make(map[string]struct{}),
+			Curve:          rc.Details.Curve,
 		},
 		Signature: make([]byte, len(rc.Signature)),
 	}
@@ -123,10 +164,38 @@ func UnmarshalNebulaCertificateFromPEM(b []byte) (*NebulaCertificate, []byte, er
 	if p == nil {
 		return nil, r, fmt.Errorf("input did not contain a valid PEM encoded block")
 	}
+	if p.Type == CertificateV2Banner {
+		return nil, r, fmt.Errorf("%w: %s", ErrInvalidPEMCertificateUnsupported, p.Type)
+	}
+	if p.Type != CertBanner {
+		return nil, r, fmt.Errorf("bytes did not contain a proper nebula certificate banner")
+	}
 	nc, err := UnmarshalNebulaCertificate(p.Bytes)
 	return nc, r, err
 }
 
+func MarshalPrivateKey(curve Curve, b []byte) []byte {
+	switch curve {
+	case Curve_CURVE25519:
+		return pem.EncodeToMemory(&pem.Block{Type: X25519PrivateKeyBanner, Bytes: b})
+	case Curve_P256:
+		return pem.EncodeToMemory(&pem.Block{Type: P256PrivateKeyBanner, Bytes: b})
+	default:
+		return nil
+	}
+}
+
+func MarshalSigningPrivateKey(curve Curve, b []byte) []byte {
+	switch curve {
+	case Curve_CURVE25519:
+		return pem.EncodeToMemory(&pem.Block{Type: Ed25519PrivateKeyBanner, Bytes: b})
+	case Curve_P256:
+		return pem.EncodeToMemory(&pem.Block{Type: ECDSAP256PrivateKeyBanner, Bytes: b})
+	default:
+		return nil
+	}
+}
+
 // MarshalX25519PrivateKey is a simple helper to PEM encode an X25519 private key
 func MarshalX25519PrivateKey(b []byte) []byte {
 	return pem.EncodeToMemory(&pem.Block{Type: X25519PrivateKeyBanner, Bytes: b})
@@ -137,6 +206,90 @@ func MarshalEd25519PrivateKey(key ed25519.PrivateKey) []byte {
 	return pem.EncodeToMemory(&pem.Block{Type: Ed25519PrivateKeyBanner, Bytes: key})
 }
 
+func UnmarshalPrivateKey(b []byte) ([]byte, []byte, Curve, error) {
+	k, r := pem.Decode(b)
+	if k == nil {
+		return nil, r, 0, fmt.Errorf("input did not contain a valid PEM encoded block")
+	}
+	var expectedLen int
+	var curve Curve
+	switch k.Type {
+	case X25519PrivateKeyBanner:
+		expectedLen = 32
+		curve = Curve_CURVE25519
+	case P256PrivateKeyBanner:
+		expectedLen = 32
+		curve = Curve_P256
+	default:
+		return nil, r, 0, fmt.Errorf("bytes did not contain a proper nebula private key banner")
+	}
+	if len(k.Bytes) != expectedLen {
+		return nil, r, 0, fmt.Errorf("key was not %d bytes, is invalid %s private key", expectedLen, curve)
+	}
+	return k.Bytes, r, curve, nil
+}
+
+func UnmarshalSigningPrivateKey(b []byte) ([]byte, []byte, Curve, error) {
+	k, r := pem.Decode(b)
+	if k == nil {
+		return nil, r, 0, fmt.Errorf("input did not contain a valid PEM encoded block")
+	}
+	var curve Curve
+	switch k.Type {
+	case EncryptedEd25519PrivateKeyBanner:
+		return nil, nil, Curve_CURVE25519, ErrPrivateKeyEncrypted
+	case EncryptedECDSAP256PrivateKeyBanner:
+		return nil, nil, Curve_P256, ErrPrivateKeyEncrypted
+	case Ed25519PrivateKeyBanner:
+		curve = Curve_CURVE25519
+		if len(k.Bytes) != ed25519.PrivateKeySize {
+			return nil, r, 0, fmt.Errorf("key was not %d bytes, is invalid Ed25519 private key", ed25519.PrivateKeySize)
+		}
+	case ECDSAP256PrivateKeyBanner:
+		curve = Curve_P256
+		if len(k.Bytes) != 32 {
+			return nil, r, 0, fmt.Errorf("key was not 32 bytes, is invalid ECDSA P256 private key")
+		}
+	default:
+		return nil, r, 0, fmt.Errorf("bytes did not contain a proper nebula Ed25519/ECDSA private key banner")
+	}
+	return k.Bytes, r, curve, nil
+}
+
+// EncryptAndMarshalSigningPrivateKey is a simple helper to encrypt and PEM encode a private key
+func EncryptAndMarshalSigningPrivateKey(curve Curve, b []byte, passphrase []byte, kdfParams *Argon2Parameters) ([]byte, error) {
+	ciphertext, err := aes256Encrypt(passphrase, kdfParams, b)
+	if err != nil {
+		return nil, err
+	}
+
+	b, err = proto.Marshal(&RawNebulaEncryptedData{
+		EncryptionMetadata: &RawNebulaEncryptionMetadata{
+			EncryptionAlgorithm: "AES-256-GCM",
+			Argon2Parameters: &RawNebulaArgon2Parameters{
+				Version:     kdfParams.version,
+				Memory:      kdfParams.Memory,
+				Parallelism: uint32(kdfParams.Parallelism),
+				Iterations:  kdfParams.Iterations,
+				Salt:        kdfParams.salt,
+			},
+		},
+		Ciphertext: ciphertext,
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	switch curve {
+	case Curve_CURVE25519:
+		return pem.EncodeToMemory(&pem.Block{Type: EncryptedEd25519PrivateKeyBanner, Bytes: b}), nil
+	case Curve_P256:
+		return pem.EncodeToMemory(&pem.Block{Type: EncryptedECDSAP256PrivateKeyBanner, Bytes: b}), nil
+	default:
+		return nil, fmt.Errorf("invalid curve: %v", curve)
+	}
+}
+
 // UnmarshalX25519PrivateKey will try to pem decode an X25519 private key, returning any other bytes b
 // or an error on failure
 func UnmarshalX25519PrivateKey(b []byte) ([]byte, []byte, error) {
@@ -161,9 +314,13 @@ func UnmarshalEd25519PrivateKey(b []byte) (ed25519.PrivateKey, []byte, error) {
 	if k == nil {
 		return nil, r, fmt.Errorf("input did not contain a valid PEM encoded block")
 	}
-	if k.Type != Ed25519PrivateKeyBanner {
+
+	if k.Type == EncryptedEd25519PrivateKeyBanner {
+		return nil, r, ErrPrivateKeyEncrypted
+	} else if k.Type != Ed25519PrivateKeyBanner {
 		return nil, r, fmt.Errorf("bytes did not contain a proper nebula Ed25519 private key banner")
 	}
+
 	if len(k.Bytes) != ed25519.PrivateKeySize {
 		return nil, r, fmt.Errorf("key was not 64 bytes, is invalid ed25519 private key")
 	}
@@ -171,6 +328,126 @@ func UnmarshalEd25519PrivateKey(b []byte) (ed25519.PrivateKey, []byte, error) {
 	return k.Bytes, r, nil
 }
 
+// UnmarshalNebulaEncryptedData will unmarshal a protobuf byte representation of a nebula cert into its
+// protobuf-generated struct.
+func UnmarshalNebulaEncryptedData(b []byte) (*NebulaEncryptedData, error) {
+	if len(b) == 0 {
+		return nil, fmt.Errorf("nil byte array")
+	}
+	var rned RawNebulaEncryptedData
+	err := proto.Unmarshal(b, &rned)
+	if err != nil {
+		return nil, err
+	}
+
+	if rned.EncryptionMetadata == nil {
+		return nil, fmt.Errorf("encoded EncryptionMetadata was nil")
+	}
+
+	if rned.EncryptionMetadata.Argon2Parameters == nil {
+		return nil, fmt.Errorf("encoded Argon2Parameters was nil")
+	}
+
+	params, err := unmarshalArgon2Parameters(rned.EncryptionMetadata.Argon2Parameters)
+	if err != nil {
+		return nil, err
+	}
+
+	ned := NebulaEncryptedData{
+		EncryptionMetadata: NebulaEncryptionMetadata{
+			EncryptionAlgorithm: rned.EncryptionMetadata.EncryptionAlgorithm,
+			Argon2Parameters:    *params,
+		},
+		Ciphertext: rned.Ciphertext,
+	}
+
+	return &ned, nil
+}
+
+func unmarshalArgon2Parameters(params *RawNebulaArgon2Parameters) (*Argon2Parameters, error) {
+	if params.Version < math.MinInt32 || params.Version > math.MaxInt32 {
+		return nil, fmt.Errorf("Argon2Parameters Version must be at least %d and no more than %d", math.MinInt32, math.MaxInt32)
+	}
+	if params.Memory <= 0 || params.Memory > math.MaxUint32 {
+		return nil, fmt.Errorf("Argon2Parameters Memory must be be greater than 0 and no more than %d KiB", uint32(math.MaxUint32))
+	}
+	if params.Parallelism <= 0 || params.Parallelism > math.MaxUint8 {
+		return nil, fmt.Errorf("Argon2Parameters Parallelism must be be greater than 0 and no more than %d", math.MaxUint8)
+	}
+	if params.Iterations <= 0 || params.Iterations > math.MaxUint32 {
+		return nil, fmt.Errorf("-argon-iterations must be be greater than 0 and no more than %d", uint32(math.MaxUint32))
+	}
+
+	return &Argon2Parameters{
+		version:     rune(params.Version),
+		Memory:      uint32(params.Memory),
+		Parallelism: uint8(params.Parallelism),
+		Iterations:  uint32(params.Iterations),
+		salt:        params.Salt,
+	}, nil
+
+}
+
+// DecryptAndUnmarshalSigningPrivateKey will try to pem decode and decrypt an Ed25519/ECDSA private key with
+// the given passphrase, returning any other bytes b or an error on failure
+func DecryptAndUnmarshalSigningPrivateKey(passphrase, b []byte) (Curve, []byte, []byte, error) {
+	var curve Curve
+
+	k, r := pem.Decode(b)
+	if k == nil {
+		return curve, nil, r, fmt.Errorf("input did not contain a valid PEM encoded block")
+	}
+
+	switch k.Type {
+	case EncryptedEd25519PrivateKeyBanner:
+		curve = Curve_CURVE25519
+	case EncryptedECDSAP256PrivateKeyBanner:
+		curve = Curve_P256
+	default:
+		return curve, nil, r, fmt.Errorf("bytes did not contain a proper nebula encrypted Ed25519/ECDSA private key banner")
+	}
+
+	ned, err := UnmarshalNebulaEncryptedData(k.Bytes)
+	if err != nil {
+		return curve, nil, r, err
+	}
+
+	var bytes []byte
+	switch ned.EncryptionMetadata.EncryptionAlgorithm {
+	case "AES-256-GCM":
+		bytes, err = aes256Decrypt(passphrase, &ned.EncryptionMetadata.Argon2Parameters, ned.Ciphertext)
+		if err != nil {
+			return curve, nil, r, err
+		}
+	default:
+		return curve, nil, r, fmt.Errorf("unsupported encryption algorithm: %s", ned.EncryptionMetadata.EncryptionAlgorithm)
+	}
+
+	switch curve {
+	case Curve_CURVE25519:
+		if len(bytes) != ed25519.PrivateKeySize {
+			return curve, nil, r, fmt.Errorf("key was not %d bytes, is invalid ed25519 private key", ed25519.PrivateKeySize)
+		}
+	case Curve_P256:
+		if len(bytes) != 32 {
+			return curve, nil, r, fmt.Errorf("key was not 32 bytes, is invalid ECDSA P256 private key")
+		}
+	}
+
+	return curve, bytes, r, nil
+}
+
+func MarshalPublicKey(curve Curve, b []byte) []byte {
+	switch curve {
+	case Curve_CURVE25519:
+		return pem.EncodeToMemory(&pem.Block{Type: X25519PublicKeyBanner, Bytes: b})
+	case Curve_P256:
+		return pem.EncodeToMemory(&pem.Block{Type: P256PublicKeyBanner, Bytes: b})
+	default:
+		return nil
+	}
+}
+
 // MarshalX25519PublicKey is a simple helper to PEM encode an X25519 public key
 func MarshalX25519PublicKey(b []byte) []byte {
 	return pem.EncodeToMemory(&pem.Block{Type: X25519PublicKeyBanner, Bytes: b})
@@ -181,6 +458,30 @@ func MarshalEd25519PublicKey(key ed25519.PublicKey) []byte {
 	return pem.EncodeToMemory(&pem.Block{Type: Ed25519PublicKeyBanner, Bytes: key})
 }
 
+func UnmarshalPublicKey(b []byte) ([]byte, []byte, Curve, error) {
+	k, r := pem.Decode(b)
+	if k == nil {
+		return nil, r, 0, fmt.Errorf("input did not contain a valid PEM encoded block")
+	}
+	var expectedLen int
+	var curve Curve
+	switch k.Type {
+	case X25519PublicKeyBanner:
+		expectedLen = 32
+		curve = Curve_CURVE25519
+	case P256PublicKeyBanner:
+		// Uncompressed
+		expectedLen = 65
+		curve = Curve_P256
+	default:
+		return nil, r, 0, fmt.Errorf("bytes did not contain a proper nebula public key banner")
+	}
+	if len(k.Bytes) != expectedLen {
+		return nil, r, 0, fmt.Errorf("key was not %d bytes, is invalid %s public key", expectedLen, curve)
+	}
+	return k.Bytes, r, curve, nil
+}
+
 // UnmarshalX25519PublicKey will try to pem decode an X25519 public key, returning any other bytes b
 // or an error on failure
 func UnmarshalX25519PublicKey(b []byte) ([]byte, []byte, error) {
@@ -216,27 +517,86 @@ func UnmarshalEd25519PublicKey(b []byte) (ed25519.PublicKey, []byte, error) {
 }
 
 // Sign signs a nebula cert with the provided private key
-func (nc *NebulaCertificate) Sign(key ed25519.PrivateKey) error {
+func (nc *NebulaCertificate) Sign(curve Curve, key []byte) error {
+	if curve != nc.Details.Curve {
+		return fmt.Errorf("curve in cert and private key supplied don't match")
+	}
+
 	b, err := proto.Marshal(nc.getRawDetails())
 	if err != nil {
 		return err
 	}
 
-	sig, err := key.Sign(rand.Reader, b, crypto.Hash(0))
-	if err != nil {
-		return err
+	var sig []byte
+
+	switch curve {
+	case Curve_CURVE25519:
+		signer := ed25519.PrivateKey(key)
+		sig = ed25519.Sign(signer, b)
+	case Curve_P256:
+		signer := &ecdsa.PrivateKey{
+			PublicKey: ecdsa.PublicKey{
+				Curve: elliptic.P256(),
+			},
+			// ref: https://github.com/golang/go/blob/go1.19/src/crypto/x509/sec1.go#L95
+			D: new(big.Int).SetBytes(key),
+		}
+		// ref: https://github.com/golang/go/blob/go1.19/src/crypto/x509/sec1.go#L119
+		signer.X, signer.Y = signer.Curve.ScalarBaseMult(key)
+
+		// We need to hash first for ECDSA
+		// - https://pkg.go.dev/crypto/ecdsa#SignASN1
+		hashed := sha256.Sum256(b)
+		sig, err = ecdsa.SignASN1(rand.Reader, signer, hashed[:])
+		if err != nil {
+			return err
+		}
+	default:
+		return fmt.Errorf("invalid curve: %s", nc.Details.Curve)
 	}
+
 	nc.Signature = sig
 	return nil
 }
 
 // CheckSignature verifies the signature against the provided public key
-func (nc *NebulaCertificate) CheckSignature(key ed25519.PublicKey) bool {
+func (nc *NebulaCertificate) CheckSignature(key []byte) bool {
 	b, err := proto.Marshal(nc.getRawDetails())
 	if err != nil {
 		return false
 	}
-	return ed25519.Verify(key, b, nc.Signature)
+	switch nc.Details.Curve {
+	case Curve_CURVE25519:
+		return ed25519.Verify(ed25519.PublicKey(key), b, nc.Signature)
+	case Curve_P256:
+		x, y := elliptic.Unmarshal(elliptic.P256(), key)
+		pubKey := &ecdsa.PublicKey{Curve: elliptic.P256(), X: x, Y: y}
+		hashed := sha256.Sum256(b)
+		return ecdsa.VerifyASN1(pubKey, hashed[:], nc.Signature)
+	default:
+		return false
+	}
+}
+
+// NOTE: This uses an internal cache that will not be invalidated automatically
+// if you manually change any fields in the NebulaCertificate.
+func (nc *NebulaCertificate) checkSignatureWithCache(key []byte, useCache bool) bool {
+	if !useCache {
+		return nc.CheckSignature(key)
+	}
+
+	if v := nc.signatureVerified.Load(); v != nil {
+		return bytes.Equal(*v, key)
+	}
+
+	verified := nc.CheckSignature(key)
+	if verified {
+		keyCopy := make([]byte, len(key))
+		copy(keyCopy, key)
+		nc.signatureVerified.Store(&keyCopy)
+	}
+
+	return verified
 }
 
 // Expired will return true if the nebula cert is too young or too old compared to the provided time, otherwise false
@@ -244,10 +604,29 @@ func (nc *NebulaCertificate) Expired(t time.Time) bool {
 	return nc.Details.NotBefore.After(t) || nc.Details.NotAfter.Before(t)
 }
 
-// Verify will ensure a certificate is good in all respects (expiry, group membership, signature, cert blacklist, etc)
+// Verify will ensure a certificate is good in all respects (expiry, group membership, signature, cert blocklist, etc)
 func (nc *NebulaCertificate) Verify(t time.Time, ncp *NebulaCAPool) (bool, error) {
-	if ncp.IsBlacklisted(nc) {
-		return false, fmt.Errorf("certificate has been blacklisted")
+	return nc.verify(t, ncp, false)
+}
+
+// VerifyWithCache will ensure a certificate is good in all respects (expiry, group membership, signature, cert blocklist, etc)
+//
+// NOTE: This uses an internal cache that will not be invalidated automatically
+// if you manually change any fields in the NebulaCertificate.
+func (nc *NebulaCertificate) VerifyWithCache(t time.Time, ncp *NebulaCAPool) (bool, error) {
+	return nc.verify(t, ncp, true)
+}
+
+// ResetCache resets the cache used by VerifyWithCache.
+func (nc *NebulaCertificate) ResetCache() {
+	nc.sha256sum.Store(nil)
+	nc.signatureVerified.Store(nil)
+}
+
+// Verify will ensure a certificate is good in all respects (expiry, group membership, signature, cert blocklist, etc)
+func (nc *NebulaCertificate) verify(t time.Time, ncp *NebulaCAPool, useCache bool) (bool, error) {
+	if ncp.isBlocklistedWithCache(nc, useCache) {
+		return false, ErrBlockListed
 	}
 
 	signer, err := ncp.GetCAForCert(nc)
@@ -256,15 +635,15 @@ func (nc *NebulaCertificate) Verify(t time.Time, ncp *NebulaCAPool) (bool, error
 	}
 
 	if signer.Expired(t) {
-		return false, fmt.Errorf("root certificate is expired")
+		return false, ErrRootExpired
 	}
 
 	if nc.Expired(t) {
-		return false, fmt.Errorf("certificate is expired")
+		return false, ErrExpired
 	}
 
-	if !nc.CheckSignature(signer.Details.PublicKey) {
-		return false, fmt.Errorf("certificate signature did not match")
+	if !nc.checkSignatureWithCache(signer.Details.PublicKey, useCache) {
+		return false, ErrSignatureMismatch
 	}
 
 	if err := nc.CheckRootConstrains(signer); err != nil {
@@ -317,13 +696,57 @@ func (nc *NebulaCertificate) CheckRootConstrains(signer *NebulaCertificate) erro
 }
 
 // VerifyPrivateKey checks that the public key in the Nebula certificate and a supplied private key match
-func (nc *NebulaCertificate) VerifyPrivateKey(key []byte) error {
-	var dst, key32 [32]byte
-	copy(key32[:], key)
-	curve25519.ScalarBaseMult(&dst, &key32)
-	if !bytes.Equal(dst[:], nc.Details.PublicKey) {
+func (nc *NebulaCertificate) VerifyPrivateKey(curve Curve, key []byte) error {
+	if curve != nc.Details.Curve {
+		return fmt.Errorf("curve in cert and private key supplied don't match")
+	}
+	if nc.Details.IsCA {
+		switch curve {
+		case Curve_CURVE25519:
+			// the call to PublicKey below will panic slice bounds out of range otherwise
+			if len(key) != ed25519.PrivateKeySize {
+				return fmt.Errorf("key was not 64 bytes, is invalid ed25519 private key")
+			}
+
+			if !ed25519.PublicKey(nc.Details.PublicKey).Equal(ed25519.PrivateKey(key).Public()) {
+				return fmt.Errorf("public key in cert and private key supplied don't match")
+			}
+		case Curve_P256:
+			privkey, err := ecdh.P256().NewPrivateKey(key)
+			if err != nil {
+				return fmt.Errorf("cannot parse private key as P256")
+			}
+			pub := privkey.PublicKey().Bytes()
+			if !bytes.Equal(pub, nc.Details.PublicKey) {
+				return fmt.Errorf("public key in cert and private key supplied don't match")
+			}
+		default:
+			return fmt.Errorf("invalid curve: %s", curve)
+		}
+		return nil
+	}
+
+	var pub []byte
+	switch curve {
+	case Curve_CURVE25519:
+		var err error
+		pub, err = curve25519.X25519(key, curve25519.Basepoint)
+		if err != nil {
+			return err
+		}
+	case Curve_P256:
+		privkey, err := ecdh.P256().NewPrivateKey(key)
+		if err != nil {
+			return err
+		}
+		pub = privkey.PublicKey().Bytes()
+	default:
+		return fmt.Errorf("invalid curve: %s", curve)
+	}
+	if !bytes.Equal(pub, nc.Details.PublicKey) {
 		return fmt.Errorf("public key in cert and private key supplied don't match")
 	}
+
 	return nil
 }
 
@@ -372,6 +795,7 @@ func (nc *NebulaCertificate) String() string {
 	s += fmt.Sprintf("\t\tIs CA: %v\n", nc.Details.IsCA)
 	s += fmt.Sprintf("\t\tIssuer: %s\n", nc.Details.Issuer)
 	s += fmt.Sprintf("\t\tPublic key: %x\n", nc.Details.PublicKey)
+	s += fmt.Sprintf("\t\tCurve: %s\n", nc.Details.Curve)
 	s += "\t}\n"
 	fp, err := nc.Sha256Sum()
 	if err == nil {
@@ -392,6 +816,7 @@ func (nc *NebulaCertificate) getRawDetails() *RawNebulaCertificateDetails {
 		NotAfter:  nc.Details.NotAfter.Unix(),
 		PublicKey: make([]byte, len(nc.Details.PublicKey)),
 		IsCA:      nc.Details.IsCA,
+		Curve:     nc.Details.Curve,
 	}
 
 	for _, ipNet := range nc.Details.Ips {
@@ -440,6 +865,25 @@ func (nc *NebulaCertificate) Sha256Sum() (string, error) {
 	return hex.EncodeToString(sum[:]), nil
 }
 
+// NOTE: This uses an internal cache that will not be invalidated automatically
+// if you manually change any fields in the NebulaCertificate.
+func (nc *NebulaCertificate) sha256SumWithCache(useCache bool) (string, error) {
+	if !useCache {
+		return nc.Sha256Sum()
+	}
+
+	if s := nc.sha256sum.Load(); s != nil {
+		return *s, nil
+	}
+	s, err := nc.Sha256Sum()
+	if err != nil {
+		return s, err
+	}
+
+	nc.sha256sum.Store(&s)
+	return s, nil
+}
+
 func (nc *NebulaCertificate) MarshalJSON() ([]byte, error) {
 	toString := func(ips []*net.IPNet) []string {
 		s := []string{}
@@ -461,6 +905,7 @@ func (nc *NebulaCertificate) MarshalJSON() ([]byte, error) {
 			"publicKey": fmt.Sprintf("%x", nc.Details.PublicKey),
 			"isCa":      nc.Details.IsCA,
 			"issuer":    nc.Details.Issuer,
+			"curve":     nc.Details.Curve.String(),
 		},
 		"fingerprint": fp,
 		"signature":   fmt.Sprintf("%x", nc.Signature),
@@ -468,6 +913,63 @@ func (nc *NebulaCertificate) MarshalJSON() ([]byte, error) {
 	return json.Marshal(jc)
 }
 
+//func (nc *NebulaCertificate) Copy() *NebulaCertificate {
+//	r, err := nc.Marshal()
+//	if err != nil {
+//		//TODO
+//		return nil
+//	}
+//
+//	c, err := UnmarshalNebulaCertificate(r)
+//	return c
+//}
+
+func (nc *NebulaCertificate) Copy() *NebulaCertificate {
+	c := &NebulaCertificate{
+		Details: NebulaCertificateDetails{
+			Name:           nc.Details.Name,
+			Groups:         make([]string, len(nc.Details.Groups)),
+			Ips:            make([]*net.IPNet, len(nc.Details.Ips)),
+			Subnets:        make([]*net.IPNet, len(nc.Details.Subnets)),
+			NotBefore:      nc.Details.NotBefore,
+			NotAfter:       nc.Details.NotAfter,
+			PublicKey:      make([]byte, len(nc.Details.PublicKey)),
+			IsCA:           nc.Details.IsCA,
+			Issuer:         nc.Details.Issuer,
+			InvertedGroups: make(map[string]struct{}, len(nc.Details.InvertedGroups)),
+		},
+		Signature: make([]byte, len(nc.Signature)),
+	}
+
+	copy(c.Signature, nc.Signature)
+	copy(c.Details.Groups, nc.Details.Groups)
+	copy(c.Details.PublicKey, nc.Details.PublicKey)
+
+	for i, p := range nc.Details.Ips {
+		c.Details.Ips[i] = &net.IPNet{
+			IP:   make(net.IP, len(p.IP)),
+			Mask: make(net.IPMask, len(p.Mask)),
+		}
+		copy(c.Details.Ips[i].IP, p.IP)
+		copy(c.Details.Ips[i].Mask, p.Mask)
+	}
+
+	for i, p := range nc.Details.Subnets {
+		c.Details.Subnets[i] = &net.IPNet{
+			IP:   make(net.IP, len(p.IP)),
+			Mask: make(net.IPMask, len(p.Mask)),
+		}
+		copy(c.Details.Subnets[i].IP, p.IP)
+		copy(c.Details.Subnets[i].Mask, p.Mask)
+	}
+
+	for g := range nc.Details.InvertedGroups {
+		c.Details.InvertedGroups[g] = struct{}{}
+	}
+
+	return c
+}
+
 func netMatch(certIp *net.IPNet, rootIps []*net.IPNet) bool {
 	for _, net := range rootIps {
 		if net.Contains(certIp.IP) && maskContains(net.Mask, certIp.Mask) {

cert/cert.pb.go

@@ -1,202 +1,620 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
+// versions:
+// 	protoc-gen-go v1.30.0
+// 	protoc        v3.21.5
 // source: cert.proto
 
 package cert
 
 import (
-	fmt "fmt"
-	proto "github.com/golang/protobuf/proto"
-	math "math"
+	protoreflect "google.golang.org/protobuf/reflect/protoreflect"
+	protoimpl "google.golang.org/protobuf/runtime/protoimpl"
+	reflect "reflect"
+	sync "sync"
 )
 
-// Reference imports to suppress errors if they are not otherwise used.
-var _ = proto.Marshal
-var _ = fmt.Errorf
-var _ = math.Inf
+const (
+	// Verify that this generated code is sufficiently up-to-date.
+	_ = protoimpl.EnforceVersion(20 - protoimpl.MinVersion)
+	// Verify that runtime/protoimpl is sufficiently up-to-date.
+	_ = protoimpl.EnforceVersion(protoimpl.MaxVersion - 20)
+)
 
-// This is a compile-time assertion to ensure that this generated file
-// is compatible with the proto package it is being compiled against.
-// A compilation error at this line likely means your copy of the
-// proto package needs to be updated.
-const _ = proto.ProtoPackageIsVersion3 // please upgrade the proto package
+type Curve int32
+
+const (
+	Curve_CURVE25519 Curve = 0
+	Curve_P256       Curve = 1
+)
+
+// Enum value maps for Curve.
+var (
+	Curve_name = map[int32]string{
+		0: "CURVE25519",
+		1: "P256",
+	}
+	Curve_value = map[string]int32{
+		"CURVE25519": 0,
+		"P256":       1,
+	}
+)
+
+func (x Curve) Enum() *Curve {
+	p := new(Curve)
+	*p = x
+	return p
+}
+
+func (x Curve) String() string {
+	return protoimpl.X.EnumStringOf(x.Descriptor(), protoreflect.EnumNumber(x))
+}
+
+func (Curve) Descriptor() protoreflect.EnumDescriptor {
+	return file_cert_proto_enumTypes[0].Descriptor()
+}
+
+func (Curve) Type() protoreflect.EnumType {
+	return &file_cert_proto_enumTypes[0]
+}
+
+func (x Curve) Number() protoreflect.EnumNumber {
+	return protoreflect.EnumNumber(x)
+}
+
+// Deprecated: Use Curve.Descriptor instead.
+func (Curve) EnumDescriptor() ([]byte, []int) {
+	return file_cert_proto_rawDescGZIP(), []int{0}
+}
 
 type RawNebulaCertificate struct {
-	Details              *RawNebulaCertificateDetails `protobuf:"bytes,1,opt,name=Details,json=details,proto3" json:"Details,omitempty"`
-	Signature            []byte                       `protobuf:"bytes,2,opt,name=Signature,json=signature,proto3" json:"Signature,omitempty"`
-	XXX_NoUnkeyedLiteral struct{}                     `json:"-"`
-	XXX_unrecognized     []byte                       `json:"-"`
-	XXX_sizecache        int32                        `json:"-"`
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	Details   *RawNebulaCertificateDetails `protobuf:"bytes,1,opt,name=Details,proto3" json:"Details,omitempty"`
+	Signature []byte                       `protobuf:"bytes,2,opt,name=Signature,proto3" json:"Signature,omitempty"`
 }
 
-func (m *RawNebulaCertificate) Reset()         { *m = RawNebulaCertificate{} }
-func (m *RawNebulaCertificate) String() string { return proto.CompactTextString(m) }
-func (*RawNebulaCertificate) ProtoMessage()    {}
+func (x *RawNebulaCertificate) Reset() {
+	*x = RawNebulaCertificate{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_cert_proto_msgTypes[0]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *RawNebulaCertificate) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*RawNebulaCertificate) ProtoMessage() {}
+
+func (x *RawNebulaCertificate) ProtoReflect() protoreflect.Message {
+	mi := &file_cert_proto_msgTypes[0]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use RawNebulaCertificate.ProtoReflect.Descriptor instead.
 func (*RawNebulaCertificate) Descriptor() ([]byte, []int) {
-	return fileDescriptor_a142e29cbef9b1cf, []int{0}
+	return file_cert_proto_rawDescGZIP(), []int{0}
 }
 
-func (m *RawNebulaCertificate) XXX_Unmarshal(b []byte) error {
-	return xxx_messageInfo_RawNebulaCertificate.Unmarshal(m, b)
-}
-func (m *RawNebulaCertificate) XXX_Marshal(b []byte, deterministic bool) ([]byte, error) {
-	return xxx_messageInfo_RawNebulaCertificate.Marshal(b, m, deterministic)
-}
-func (m *RawNebulaCertificate) XXX_Merge(src proto.Message) {
-	xxx_messageInfo_RawNebulaCertificate.Merge(m, src)
-}
-func (m *RawNebulaCertificate) XXX_Size() int {
-	return xxx_messageInfo_RawNebulaCertificate.Size(m)
-}
-func (m *RawNebulaCertificate) XXX_DiscardUnknown() {
-	xxx_messageInfo_RawNebulaCertificate.DiscardUnknown(m)
-}
-
-var xxx_messageInfo_RawNebulaCertificate proto.InternalMessageInfo
-
-func (m *RawNebulaCertificate) GetDetails() *RawNebulaCertificateDetails {
-	if m != nil {
-		return m.Details
+func (x *RawNebulaCertificate) GetDetails() *RawNebulaCertificateDetails {
+	if x != nil {
+		return x.Details
 	}
 	return nil
 }
 
-func (m *RawNebulaCertificate) GetSignature() []byte {
-	if m != nil {
-		return m.Signature
+func (x *RawNebulaCertificate) GetSignature() []byte {
+	if x != nil {
+		return x.Signature
 	}
 	return nil
 }
 
 type RawNebulaCertificateDetails struct {
-	Name string `protobuf:"bytes,1,opt,name=Name,json=name,proto3" json:"Name,omitempty"`
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	Name string `protobuf:"bytes,1,opt,name=Name,proto3" json:"Name,omitempty"`
 	// Ips and Subnets are in big endian 32 bit pairs, 1st the ip, 2nd the mask
-	Ips       []uint32 `protobuf:"varint,2,rep,packed,name=Ips,json=ips,proto3" json:"Ips,omitempty"`
-	Subnets   []uint32 `protobuf:"varint,3,rep,packed,name=Subnets,json=subnets,proto3" json:"Subnets,omitempty"`
-	Groups    []string `protobuf:"bytes,4,rep,name=Groups,json=groups,proto3" json:"Groups,omitempty"`
-	NotBefore int64    `protobuf:"varint,5,opt,name=NotBefore,json=notBefore,proto3" json:"NotBefore,omitempty"`
-	NotAfter  int64    `protobuf:"varint,6,opt,name=NotAfter,json=notAfter,proto3" json:"NotAfter,omitempty"`
-	PublicKey []byte   `protobuf:"bytes,7,opt,name=PublicKey,json=publicKey,proto3" json:"PublicKey,omitempty"`
-	IsCA      bool     `protobuf:"varint,8,opt,name=IsCA,json=isCA,proto3" json:"IsCA,omitempty"`
+	Ips       []uint32 `protobuf:"varint,2,rep,packed,name=Ips,proto3" json:"Ips,omitempty"`
+	Subnets   []uint32 `protobuf:"varint,3,rep,packed,name=Subnets,proto3" json:"Subnets,omitempty"`
+	Groups    []string `protobuf:"bytes,4,rep,name=Groups,proto3" json:"Groups,omitempty"`
+	NotBefore int64    `protobuf:"varint,5,opt,name=NotBefore,proto3" json:"NotBefore,omitempty"`
+	NotAfter  int64    `protobuf:"varint,6,opt,name=NotAfter,proto3" json:"NotAfter,omitempty"`
+	PublicKey []byte   `protobuf:"bytes,7,opt,name=PublicKey,proto3" json:"PublicKey,omitempty"`
+	IsCA      bool     `protobuf:"varint,8,opt,name=IsCA,proto3" json:"IsCA,omitempty"`
 	// sha-256 of the issuer certificate, if this field is blank the cert is self-signed
-	Issuer               []byte   `protobuf:"bytes,9,opt,name=Issuer,json=issuer,proto3" json:"Issuer,omitempty"`
-	XXX_NoUnkeyedLiteral struct{} `json:"-"`
-	XXX_unrecognized     []byte   `json:"-"`
-	XXX_sizecache        int32    `json:"-"`
+	Issuer []byte `protobuf:"bytes,9,opt,name=Issuer,proto3" json:"Issuer,omitempty"`
+	Curve  Curve  `protobuf:"varint,100,opt,name=curve,proto3,enum=cert.Curve" json:"curve,omitempty"`
 }
 
-func (m *RawNebulaCertificateDetails) Reset()         { *m = RawNebulaCertificateDetails{} }
-func (m *RawNebulaCertificateDetails) String() string { return proto.CompactTextString(m) }
-func (*RawNebulaCertificateDetails) ProtoMessage()    {}
+func (x *RawNebulaCertificateDetails) Reset() {
+	*x = RawNebulaCertificateDetails{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_cert_proto_msgTypes[1]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *RawNebulaCertificateDetails) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*RawNebulaCertificateDetails) ProtoMessage() {}
+
+func (x *RawNebulaCertificateDetails) ProtoReflect() protoreflect.Message {
+	mi := &file_cert_proto_msgTypes[1]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use RawNebulaCertificateDetails.ProtoReflect.Descriptor instead.
 func (*RawNebulaCertificateDetails) Descriptor() ([]byte, []int) {
-	return fileDescriptor_a142e29cbef9b1cf, []int{1}
+	return file_cert_proto_rawDescGZIP(), []int{1}
 }
 
-func (m *RawNebulaCertificateDetails) XXX_Unmarshal(b []byte) error {
-	return xxx_messageInfo_RawNebulaCertificateDetails.Unmarshal(m, b)
-}
-func (m *RawNebulaCertificateDetails) XXX_Marshal(b []byte, deterministic bool) ([]byte, error) {
-	return xxx_messageInfo_RawNebulaCertificateDetails.Marshal(b, m, deterministic)
-}
-func (m *RawNebulaCertificateDetails) XXX_Merge(src proto.Message) {
-	xxx_messageInfo_RawNebulaCertificateDetails.Merge(m, src)
-}
-func (m *RawNebulaCertificateDetails) XXX_Size() int {
-	return xxx_messageInfo_RawNebulaCertificateDetails.Size(m)
-}
-func (m *RawNebulaCertificateDetails) XXX_DiscardUnknown() {
-	xxx_messageInfo_RawNebulaCertificateDetails.DiscardUnknown(m)
-}
-
-var xxx_messageInfo_RawNebulaCertificateDetails proto.InternalMessageInfo
-
-func (m *RawNebulaCertificateDetails) GetName() string {
-	if m != nil {
-		return m.Name
+func (x *RawNebulaCertificateDetails) GetName() string {
+	if x != nil {
+		return x.Name
 	}
 	return ""
 }
 
-func (m *RawNebulaCertificateDetails) GetIps() []uint32 {
-	if m != nil {
-		return m.Ips
+func (x *RawNebulaCertificateDetails) GetIps() []uint32 {
+	if x != nil {
+		return x.Ips
 	}
 	return nil
 }
 
-func (m *RawNebulaCertificateDetails) GetSubnets() []uint32 {
-	if m != nil {
-		return m.Subnets
+func (x *RawNebulaCertificateDetails) GetSubnets() []uint32 {
+	if x != nil {
+		return x.Subnets
 	}
 	return nil
 }
 
-func (m *RawNebulaCertificateDetails) GetGroups() []string {
-	if m != nil {
-		return m.Groups
+func (x *RawNebulaCertificateDetails) GetGroups() []string {
+	if x != nil {
+		return x.Groups
 	}
 	return nil
 }
 
-func (m *RawNebulaCertificateDetails) GetNotBefore() int64 {
-	if m != nil {
-		return m.NotBefore
+func (x *RawNebulaCertificateDetails) GetNotBefore() int64 {
+	if x != nil {
+		return x.NotBefore
 	}
 	return 0
 }
 
-func (m *RawNebulaCertificateDetails) GetNotAfter() int64 {
-	if m != nil {
-		return m.NotAfter
+func (x *RawNebulaCertificateDetails) GetNotAfter() int64 {
+	if x != nil {
+		return x.NotAfter
 	}
 	return 0
 }
 
-func (m *RawNebulaCertificateDetails) GetPublicKey() []byte {
-	if m != nil {
-		return m.PublicKey
+func (x *RawNebulaCertificateDetails) GetPublicKey() []byte {
+	if x != nil {
+		return x.PublicKey
 	}
 	return nil
 }
 
-func (m *RawNebulaCertificateDetails) GetIsCA() bool {
-	if m != nil {
-		return m.IsCA
+func (x *RawNebulaCertificateDetails) GetIsCA() bool {
+	if x != nil {
+		return x.IsCA
 	}
 	return false
 }
 
-func (m *RawNebulaCertificateDetails) GetIssuer() []byte {
-	if m != nil {
-		return m.Issuer
+func (x *RawNebulaCertificateDetails) GetIssuer() []byte {
+	if x != nil {
+		return x.Issuer
 	}
 	return nil
 }
 
-func init() {
-	proto.RegisterType((*RawNebulaCertificate)(nil), "cert.RawNebulaCertificate")
-	proto.RegisterType((*RawNebulaCertificateDetails)(nil), "cert.RawNebulaCertificateDetails")
+func (x *RawNebulaCertificateDetails) GetCurve() Curve {
+	if x != nil {
+		return x.Curve
+	}
+	return Curve_CURVE25519
 }
 
-func init() { proto.RegisterFile("cert.proto", fileDescriptor_a142e29cbef9b1cf) }
+type RawNebulaEncryptedData struct {
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
 
-var fileDescriptor_a142e29cbef9b1cf = []byte{
-	// 279 bytes of a gzipped FileDescriptorProto
-	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x7c, 0x90, 0xcf, 0x4a, 0xf4, 0x30,
-	0x14, 0xc5, 0xc9, 0xa4, 0x5f, 0xdb, 0xe4, 0x53, 0x90, 0x20, 0x12, 0xd4, 0x45, 0x9c, 0x55, 0x56,
-	0xb3, 0xd0, 0xa5, 0xab, 0x71, 0x04, 0x29, 0x42, 0x91, 0xcc, 0x13, 0xa4, 0xf5, 0x76, 0x08, 0x74,
-	0x9a, 0x9a, 0x3f, 0x88, 0x8f, 0xee, 0x4e, 0x9a, 0x4e, 0x77, 0xe2, 0xee, 0x9e, 0x5f, 0xce, 0x49,
-	0x4e, 0x2e, 0xa5, 0x2d, 0xb8, 0xb0, 0x19, 0x9d, 0x0d, 0x96, 0x65, 0xd3, 0xbc, 0xfe, 0xa0, 0x97,
-	0x4a, 0x7f, 0xd6, 0xd0, 0xc4, 0x5e, 0xef, 0xc0, 0x05, 0xd3, 0x99, 0x56, 0x07, 0x60, 0x8f, 0xb4,
-	0x78, 0x86, 0xa0, 0x4d, 0xef, 0x39, 0x12, 0x48, 0xfe, 0xbf, 0xbf, 0xdb, 0xa4, 0xec, 0x6f, 0xe6,
-	0x93, 0x51, 0x15, 0xef, 0xf3, 0xc0, 0x6e, 0x29, 0xd9, 0x9b, 0xc3, 0xa0, 0x43, 0x74, 0xc0, 0x57,
-	0x02, 0xc9, 0x33, 0x45, 0xfc, 0x02, 0xd6, 0xdf, 0x88, 0xde, 0xfc, 0x71, 0x0d, 0x63, 0x34, 0xab,
-	0xf5, 0x11, 0xd2, 0xbb, 0x44, 0x65, 0x83, 0x3e, 0x02, 0xbb, 0xa0, 0xb8, 0x1a, 0x3d, 0x5f, 0x09,
-	0x2c, 0xcf, 0x15, 0x36, 0xa3, 0x67, 0x9c, 0x16, 0xfb, 0xd8, 0x0c, 0x10, 0x3c, 0xc7, 0x89, 0x16,
-	0x7e, 0x96, 0xec, 0x8a, 0xe6, 0x2f, 0xce, 0xc6, 0xd1, 0xf3, 0x4c, 0x60, 0x49, 0x54, 0x7e, 0x48,
-	0x6a, 0x6a, 0x55, 0xdb, 0xf0, 0x04, 0x9d, 0x75, 0xc0, 0xff, 0x09, 0x24, 0xb1, 0x22, 0xc3, 0x02,
-	0xd8, 0x35, 0x2d, 0x6b, 0x1b, 0xb6, 0x5d, 0x00, 0xc7, 0xf3, 0x74, 0x58, 0x0e, 0x27, 0x3d, 0x25,
-	0xdf, 0x62, 0xd3, 0x9b, 0xf6, 0x15, 0xbe, 0x78, 0x31, 0xff, 0x67, 0x5c, 0xc0, 0xd4, 0xb7, 0xf2,
-	0xbb, 0x2d, 0x2f, 0x05, 0x92, 0xa5, 0xca, 0x8c, 0xdf, 0x6d, 0xa7, 0x0e, 0x95, 0xf7, 0x11, 0x1c,
-	0x27, 0xc9, 0x9e, 0x9b, 0xa4, 0x9a, 0x3c, 0xed, 0xfe, 0xe1, 0x27, 0x00, 0x00, 0xff, 0xff, 0x2c,
-	0xe3, 0x08, 0x37, 0x89, 0x01, 0x00, 0x00,
+	EncryptionMetadata *RawNebulaEncryptionMetadata `protobuf:"bytes,1,opt,name=EncryptionMetadata,proto3" json:"EncryptionMetadata,omitempty"`
+	Ciphertext         []byte                       `protobuf:"bytes,2,opt,name=Ciphertext,proto3" json:"Ciphertext,omitempty"`
+}
+
+func (x *RawNebulaEncryptedData) Reset() {
+	*x = RawNebulaEncryptedData{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_cert_proto_msgTypes[2]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *RawNebulaEncryptedData) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*RawNebulaEncryptedData) ProtoMessage() {}
+
+func (x *RawNebulaEncryptedData) ProtoReflect() protoreflect.Message {
+	mi := &file_cert_proto_msgTypes[2]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use RawNebulaEncryptedData.ProtoReflect.Descriptor instead.
+func (*RawNebulaEncryptedData) Descriptor() ([]byte, []int) {
+	return file_cert_proto_rawDescGZIP(), []int{2}
+}
+
+func (x *RawNebulaEncryptedData) GetEncryptionMetadata() *RawNebulaEncryptionMetadata {
+	if x != nil {
+		return x.EncryptionMetadata
+	}
+	return nil
+}
+
+func (x *RawNebulaEncryptedData) GetCiphertext() []byte {
+	if x != nil {
+		return x.Ciphertext
+	}
+	return nil
+}
+
+type RawNebulaEncryptionMetadata struct {
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	EncryptionAlgorithm string                     `protobuf:"bytes,1,opt,name=EncryptionAlgorithm,proto3" json:"EncryptionAlgorithm,omitempty"`
+	Argon2Parameters    *RawNebulaArgon2Parameters `protobuf:"bytes,2,opt,name=Argon2Parameters,proto3" json:"Argon2Parameters,omitempty"`
+}
+
+func (x *RawNebulaEncryptionMetadata) Reset() {
+	*x = RawNebulaEncryptionMetadata{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_cert_proto_msgTypes[3]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *RawNebulaEncryptionMetadata) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*RawNebulaEncryptionMetadata) ProtoMessage() {}
+
+func (x *RawNebulaEncryptionMetadata) ProtoReflect() protoreflect.Message {
+	mi := &file_cert_proto_msgTypes[3]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use RawNebulaEncryptionMetadata.ProtoReflect.Descriptor instead.
+func (*RawNebulaEncryptionMetadata) Descriptor() ([]byte, []int) {
+	return file_cert_proto_rawDescGZIP(), []int{3}
+}
+
+func (x *RawNebulaEncryptionMetadata) GetEncryptionAlgorithm() string {
+	if x != nil {
+		return x.EncryptionAlgorithm
+	}
+	return ""
+}
+
+func (x *RawNebulaEncryptionMetadata) GetArgon2Parameters() *RawNebulaArgon2Parameters {
+	if x != nil {
+		return x.Argon2Parameters
+	}
+	return nil
+}
+
+type RawNebulaArgon2Parameters struct {
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	Version     int32  `protobuf:"varint,1,opt,name=version,proto3" json:"version,omitempty"` // rune in Go
+	Memory      uint32 `protobuf:"varint,2,opt,name=memory,proto3" json:"memory,omitempty"`
+	Parallelism uint32 `protobuf:"varint,4,opt,name=parallelism,proto3" json:"parallelism,omitempty"` // uint8 in Go
+	Iterations  uint32 `protobuf:"varint,3,opt,name=iterations,proto3" json:"iterations,omitempty"`
+	Salt        []byte `protobuf:"bytes,5,opt,name=salt,proto3" json:"salt,omitempty"`
+}
+
+func (x *RawNebulaArgon2Parameters) Reset() {
+	*x = RawNebulaArgon2Parameters{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_cert_proto_msgTypes[4]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *RawNebulaArgon2Parameters) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*RawNebulaArgon2Parameters) ProtoMessage() {}
+
+func (x *RawNebulaArgon2Parameters) ProtoReflect() protoreflect.Message {
+	mi := &file_cert_proto_msgTypes[4]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use RawNebulaArgon2Parameters.ProtoReflect.Descriptor instead.
+func (*RawNebulaArgon2Parameters) Descriptor() ([]byte, []int) {
+	return file_cert_proto_rawDescGZIP(), []int{4}
+}
+
+func (x *RawNebulaArgon2Parameters) GetVersion() int32 {
+	if x != nil {
+		return x.Version
+	}
+	return 0
+}
+
+func (x *RawNebulaArgon2Parameters) GetMemory() uint32 {
+	if x != nil {
+		return x.Memory
+	}
+	return 0
+}
+
+func (x *RawNebulaArgon2Parameters) GetParallelism() uint32 {
+	if x != nil {
+		return x.Parallelism
+	}
+	return 0
+}
+
+func (x *RawNebulaArgon2Parameters) GetIterations() uint32 {
+	if x != nil {
+		return x.Iterations
+	}
+	return 0
+}
+
+func (x *RawNebulaArgon2Parameters) GetSalt() []byte {
+	if x != nil {
+		return x.Salt
+	}
+	return nil
+}
+
+var File_cert_proto protoreflect.FileDescriptor
+
+var file_cert_proto_rawDesc = []byte{
+	0x0a, 0x0a, 0x63, 0x65, 0x72, 0x74, 0x2e, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x12, 0x04, 0x63, 0x65,
+	0x72, 0x74, 0x22, 0x71, 0x0a, 0x14, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x43,
+	0x65, 0x72, 0x74, 0x69, 0x66, 0x69, 0x63, 0x61, 0x74, 0x65, 0x12, 0x3b, 0x0a, 0x07, 0x44, 0x65,
+	0x74, 0x61, 0x69, 0x6c, 0x73, 0x18, 0x01, 0x20, 0x01, 0x28, 0x0b, 0x32, 0x21, 0x2e, 0x63, 0x65,
+	0x72, 0x74, 0x2e, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x43, 0x65, 0x72, 0x74,
+	0x69, 0x66, 0x69, 0x63, 0x61, 0x74, 0x65, 0x44, 0x65, 0x74, 0x61, 0x69, 0x6c, 0x73, 0x52, 0x07,
+	0x44, 0x65, 0x74, 0x61, 0x69, 0x6c, 0x73, 0x12, 0x1c, 0x0a, 0x09, 0x53, 0x69, 0x67, 0x6e, 0x61,
+	0x74, 0x75, 0x72, 0x65, 0x18, 0x02, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x09, 0x53, 0x69, 0x67, 0x6e,
+	0x61, 0x74, 0x75, 0x72, 0x65, 0x22, 0x9c, 0x02, 0x0a, 0x1b, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62,
+	0x75, 0x6c, 0x61, 0x43, 0x65, 0x72, 0x74, 0x69, 0x66, 0x69, 0x63, 0x61, 0x74, 0x65, 0x44, 0x65,
+	0x74, 0x61, 0x69, 0x6c, 0x73, 0x12, 0x12, 0x0a, 0x04, 0x4e, 0x61, 0x6d, 0x65, 0x18, 0x01, 0x20,
+	0x01, 0x28, 0x09, 0x52, 0x04, 0x4e, 0x61, 0x6d, 0x65, 0x12, 0x10, 0x0a, 0x03, 0x49, 0x70, 0x73,
+	0x18, 0x02, 0x20, 0x03, 0x28, 0x0d, 0x52, 0x03, 0x49, 0x70, 0x73, 0x12, 0x18, 0x0a, 0x07, 0x53,
+	0x75, 0x62, 0x6e, 0x65, 0x74, 0x73, 0x18, 0x03, 0x20, 0x03, 0x28, 0x0d, 0x52, 0x07, 0x53, 0x75,
+	0x62, 0x6e, 0x65, 0x74, 0x73, 0x12, 0x16, 0x0a, 0x06, 0x47, 0x72, 0x6f, 0x75, 0x70, 0x73, 0x18,
+	0x04, 0x20, 0x03, 0x28, 0x09, 0x52, 0x06, 0x47, 0x72, 0x6f, 0x75, 0x70, 0x73, 0x12, 0x1c, 0x0a,
+	0x09, 0x4e, 0x6f, 0x74, 0x42, 0x65, 0x66, 0x6f, 0x72, 0x65, 0x18, 0x05, 0x20, 0x01, 0x28, 0x03,
+	0x52, 0x09, 0x4e, 0x6f, 0x74, 0x42, 0x65, 0x66, 0x6f, 0x72, 0x65, 0x12, 0x1a, 0x0a, 0x08, 0x4e,
+	0x6f, 0x74, 0x41, 0x66, 0x74, 0x65, 0x72, 0x18, 0x06, 0x20, 0x01, 0x28, 0x03, 0x52, 0x08, 0x4e,
+	0x6f, 0x74, 0x41, 0x66, 0x74, 0x65, 0x72, 0x12, 0x1c, 0x0a, 0x09, 0x50, 0x75, 0x62, 0x6c, 0x69,
+	0x63, 0x4b, 0x65, 0x79, 0x18, 0x07, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x09, 0x50, 0x75, 0x62, 0x6c,
+	0x69, 0x63, 0x4b, 0x65, 0x79, 0x12, 0x12, 0x0a, 0x04, 0x49, 0x73, 0x43, 0x41, 0x18, 0x08, 0x20,
+	0x01, 0x28, 0x08, 0x52, 0x04, 0x49, 0x73, 0x43, 0x41, 0x12, 0x16, 0x0a, 0x06, 0x49, 0x73, 0x73,
+	0x75, 0x65, 0x72, 0x18, 0x09, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x06, 0x49, 0x73, 0x73, 0x75, 0x65,
+	0x72, 0x12, 0x21, 0x0a, 0x05, 0x63, 0x75, 0x72, 0x76, 0x65, 0x18, 0x64, 0x20, 0x01, 0x28, 0x0e,
+	0x32, 0x0b, 0x2e, 0x63, 0x65, 0x72, 0x74, 0x2e, 0x43, 0x75, 0x72, 0x76, 0x65, 0x52, 0x05, 0x63,
+	0x75, 0x72, 0x76, 0x65, 0x22, 0x8b, 0x01, 0x0a, 0x16, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75,
+	0x6c, 0x61, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x65, 0x64, 0x44, 0x61, 0x74, 0x61, 0x12,
+	0x51, 0x0a, 0x12, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e, 0x4d, 0x65, 0x74,
+	0x61, 0x64, 0x61, 0x74, 0x61, 0x18, 0x01, 0x20, 0x01, 0x28, 0x0b, 0x32, 0x21, 0x2e, 0x63, 0x65,
+	0x72, 0x74, 0x2e, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x45, 0x6e, 0x63, 0x72,
+	0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e, 0x4d, 0x65, 0x74, 0x61, 0x64, 0x61, 0x74, 0x61, 0x52, 0x12,
+	0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e, 0x4d, 0x65, 0x74, 0x61, 0x64, 0x61,
+	0x74, 0x61, 0x12, 0x1e, 0x0a, 0x0a, 0x43, 0x69, 0x70, 0x68, 0x65, 0x72, 0x74, 0x65, 0x78, 0x74,
+	0x18, 0x02, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x0a, 0x43, 0x69, 0x70, 0x68, 0x65, 0x72, 0x74, 0x65,
+	0x78, 0x74, 0x22, 0x9c, 0x01, 0x0a, 0x1b, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61,
+	0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e, 0x4d, 0x65, 0x74, 0x61, 0x64, 0x61,
+	0x74, 0x61, 0x12, 0x30, 0x0a, 0x13, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e,
+	0x41, 0x6c, 0x67, 0x6f, 0x72, 0x69, 0x74, 0x68, 0x6d, 0x18, 0x01, 0x20, 0x01, 0x28, 0x09, 0x52,
+	0x13, 0x45, 0x6e, 0x63, 0x72, 0x79, 0x70, 0x74, 0x69, 0x6f, 0x6e, 0x41, 0x6c, 0x67, 0x6f, 0x72,
+	0x69, 0x74, 0x68, 0x6d, 0x12, 0x4b, 0x0a, 0x10, 0x41, 0x72, 0x67, 0x6f, 0x6e, 0x32, 0x50, 0x61,
+	0x72, 0x61, 0x6d, 0x65, 0x74, 0x65, 0x72, 0x73, 0x18, 0x02, 0x20, 0x01, 0x28, 0x0b, 0x32, 0x1f,
+	0x2e, 0x63, 0x65, 0x72, 0x74, 0x2e, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x41,
+	0x72, 0x67, 0x6f, 0x6e, 0x32, 0x50, 0x61, 0x72, 0x61, 0x6d, 0x65, 0x74, 0x65, 0x72, 0x73, 0x52,
+	0x10, 0x41, 0x72, 0x67, 0x6f, 0x6e, 0x32, 0x50, 0x61, 0x72, 0x61, 0x6d, 0x65, 0x74, 0x65, 0x72,
+	0x73, 0x22, 0xa3, 0x01, 0x0a, 0x19, 0x52, 0x61, 0x77, 0x4e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x41,
+	0x72, 0x67, 0x6f, 0x6e, 0x32, 0x50, 0x61, 0x72, 0x61, 0x6d, 0x65, 0x74, 0x65, 0x72, 0x73, 0x12,
+	0x18, 0x0a, 0x07, 0x76, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x18, 0x01, 0x20, 0x01, 0x28, 0x05,
+	0x52, 0x07, 0x76, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x12, 0x16, 0x0a, 0x06, 0x6d, 0x65, 0x6d,
+	0x6f, 0x72, 0x79, 0x18, 0x02, 0x20, 0x01, 0x28, 0x0d, 0x52, 0x06, 0x6d, 0x65, 0x6d, 0x6f, 0x72,
+	0x79, 0x12, 0x20, 0x0a, 0x0b, 0x70, 0x61, 0x72, 0x61, 0x6c, 0x6c, 0x65, 0x6c, 0x69, 0x73, 0x6d,
+	0x18, 0x04, 0x20, 0x01, 0x28, 0x0d, 0x52, 0x0b, 0x70, 0x61, 0x72, 0x61, 0x6c, 0x6c, 0x65, 0x6c,
+	0x69, 0x73, 0x6d, 0x12, 0x1e, 0x0a, 0x0a, 0x69, 0x74, 0x65, 0x72, 0x61, 0x74, 0x69, 0x6f, 0x6e,
+	0x73, 0x18, 0x03, 0x20, 0x01, 0x28, 0x0d, 0x52, 0x0a, 0x69, 0x74, 0x65, 0x72, 0x61, 0x74, 0x69,
+	0x6f, 0x6e, 0x73, 0x12, 0x12, 0x0a, 0x04, 0x73, 0x61, 0x6c, 0x74, 0x18, 0x05, 0x20, 0x01, 0x28,
+	0x0c, 0x52, 0x04, 0x73, 0x61, 0x6c, 0x74, 0x2a, 0x21, 0x0a, 0x05, 0x43, 0x75, 0x72, 0x76, 0x65,
+	0x12, 0x0e, 0x0a, 0x0a, 0x43, 0x55, 0x52, 0x56, 0x45, 0x32, 0x35, 0x35, 0x31, 0x39, 0x10, 0x00,
+	0x12, 0x08, 0x0a, 0x04, 0x50, 0x32, 0x35, 0x36, 0x10, 0x01, 0x42, 0x20, 0x5a, 0x1e, 0x67, 0x69,
+	0x74, 0x68, 0x75, 0x62, 0x2e, 0x63, 0x6f, 0x6d, 0x2f, 0x73, 0x6c, 0x61, 0x63, 0x6b, 0x68, 0x71,
+	0x2f, 0x6e, 0x65, 0x62, 0x75, 0x6c, 0x61, 0x2f, 0x63, 0x65, 0x72, 0x74, 0x62, 0x06, 0x70, 0x72,
+	0x6f, 0x74, 0x6f, 0x33,
+}
+
+var (
+	file_cert_proto_rawDescOnce sync.Once
+	file_cert_proto_rawDescData = file_cert_proto_rawDesc
+)
+
+func file_cert_proto_rawDescGZIP() []byte {
+	file_cert_proto_rawDescOnce.Do(func() {
+		file_cert_proto_rawDescData = protoimpl.X.CompressGZIP(file_cert_proto_rawDescData)
+	})
+	return file_cert_proto_rawDescData
+}
+
+var file_cert_proto_enumTypes = make([]protoimpl.EnumInfo, 1)
+var file_cert_proto_msgTypes = make([]protoimpl.MessageInfo, 5)
+var file_cert_proto_goTypes = []interface{}{
+	(Curve)(0),                          // 0: cert.Curve
+	(*RawNebulaCertificate)(nil),        // 1: cert.RawNebulaCertificate
+	(*RawNebulaCertificateDetails)(nil), // 2: cert.RawNebulaCertificateDetails
+	(*RawNebulaEncryptedData)(nil),      // 3: cert.RawNebulaEncryptedData
+	(*RawNebulaEncryptionMetadata)(nil), // 4: cert.RawNebulaEncryptionMetadata
+	(*RawNebulaArgon2Parameters)(nil),   // 5: cert.RawNebulaArgon2Parameters
+}
+var file_cert_proto_depIdxs = []int32{
+	2, // 0: cert.RawNebulaCertificate.Details:type_name -> cert.RawNebulaCertificateDetails
+	0, // 1: cert.RawNebulaCertificateDetails.curve:type_name -> cert.Curve
+	4, // 2: cert.RawNebulaEncryptedData.EncryptionMetadata:type_name -> cert.RawNebulaEncryptionMetadata
+	5, // 3: cert.RawNebulaEncryptionMetadata.Argon2Parameters:type_name -> cert.RawNebulaArgon2Parameters
+	4, // [4:4] is the sub-list for method output_type
+	4, // [4:4] is the sub-list for method input_type
+	4, // [4:4] is the sub-list for extension type_name
+	4, // [4:4] is the sub-list for extension extendee
+	0, // [0:4] is the sub-list for field type_name
+}
+
+func init() { file_cert_proto_init() }
+func file_cert_proto_init() {
+	if File_cert_proto != nil {
+		return
+	}
+	if !protoimpl.UnsafeEnabled {
+		file_cert_proto_msgTypes[0].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*RawNebulaCertificate); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+		file_cert_proto_msgTypes[1].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*RawNebulaCertificateDetails); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+		file_cert_proto_msgTypes[2].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*RawNebulaEncryptedData); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+		file_cert_proto_msgTypes[3].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*RawNebulaEncryptionMetadata); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+		file_cert_proto_msgTypes[4].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*RawNebulaArgon2Parameters); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+	}
+	type x struct{}
+	out := protoimpl.TypeBuilder{
+		File: protoimpl.DescBuilder{
+			GoPackagePath: reflect.TypeOf(x{}).PkgPath(),
+			RawDescriptor: file_cert_proto_rawDesc,
+			NumEnums:      1,
+			NumMessages:   5,
+			NumExtensions: 0,
+			NumServices:   0,
+		},
+		GoTypes:           file_cert_proto_goTypes,
+		DependencyIndexes: file_cert_proto_depIdxs,
+		EnumInfos:         file_cert_proto_enumTypes,
+		MessageInfos:      file_cert_proto_msgTypes,
+	}.Build()
+	File_cert_proto = out.File
+	file_cert_proto_rawDesc = nil
+	file_cert_proto_goTypes = nil
+	file_cert_proto_depIdxs = nil
 }

cert/cert.proto

@@ -1,8 +1,15 @@
 syntax = "proto3";
 package cert;
 
+option go_package = "github.com/slackhq/nebula/cert";
+
 //import "google/protobuf/timestamp.proto";
 
+enum Curve {
+  CURVE25519 = 0;
+  P256 = 1;
+}
+
 message RawNebulaCertificate {
     RawNebulaCertificateDetails Details = 1;
     bytes Signature = 2;
@@ -24,4 +31,24 @@ message RawNebulaCertificateDetails {
 
     // sha-256 of the issuer certificate, if this field is blank the cert is self-signed
     bytes Issuer = 9;
-}
+
+    Curve curve = 100;
+}
+
+message RawNebulaEncryptedData {
+	RawNebulaEncryptionMetadata EncryptionMetadata = 1;
+	bytes Ciphertext = 2;
+}
+
+message RawNebulaEncryptionMetadata {
+	string EncryptionAlgorithm = 1;
+	RawNebulaArgon2Parameters Argon2Parameters = 2;
+}
+
+message RawNebulaArgon2Parameters {
+	int32 version = 1; // rune in Go
+	uint32 memory = 2;
+	uint32 parallelism = 4; // uint8 in Go
+	uint32 iterations = 3;
+	bytes salt = 5;
+}

cert/cert_test.go

@@ -1,17 +1,22 @@
 package cert
 
 import (
+	"crypto/ecdh"
+	"crypto/ecdsa"
+	"crypto/elliptic"
 	"crypto/rand"
+	"errors"
 	"fmt"
 	"io"
 	"net"
 	"testing"
 	"time"
 
-	"github.com/golang/protobuf/proto"
+	"github.com/slackhq/nebula/test"
 	"github.com/stretchr/testify/assert"
 	"golang.org/x/crypto/curve25519"
 	"golang.org/x/crypto/ed25519"
+	"google.golang.org/protobuf/proto"
 )
 
 func TestMarshalingNebulaCertificate(t *testing.T) {
@@ -100,7 +105,49 @@ func TestNebulaCertificate_Sign(t *testing.T) {
 	pub, priv, err := ed25519.GenerateKey(rand.Reader)
 	assert.Nil(t, err)
 	assert.False(t, nc.CheckSignature(pub))
-	assert.Nil(t, nc.Sign(priv))
+	assert.Nil(t, nc.Sign(Curve_CURVE25519, priv))
+	assert.True(t, nc.CheckSignature(pub))
+
+	_, err = nc.Marshal()
+	assert.Nil(t, err)
+	//t.Log("Cert size:", len(b))
+}
+
+func TestNebulaCertificate_SignP256(t *testing.T) {
+	before := time.Now().Add(time.Second * -60).Round(time.Second)
+	after := time.Now().Add(time.Second * 60).Round(time.Second)
+	pubKey := []byte("01234567890abcedfghij1234567890ab1234567890abcedfghij1234567890ab")
+
+	nc := NebulaCertificate{
+		Details: NebulaCertificateDetails{
+			Name: "testing",
+			Ips: []*net.IPNet{
+				{IP: net.ParseIP("10.1.1.1"), Mask: net.IPMask(net.ParseIP("255.255.255.0"))},
+				{IP: net.ParseIP("10.1.1.2"), Mask: net.IPMask(net.ParseIP("255.255.0.0"))},
+				{IP: net.ParseIP("10.1.1.3"), Mask: net.IPMask(net.ParseIP("255.0.255.0"))},
+			},
+			Subnets: []*net.IPNet{
+				{IP: net.ParseIP("9.1.1.1"), Mask: net.IPMask(net.ParseIP("255.0.255.0"))},
+				{IP: net.ParseIP("9.1.1.2"), Mask: net.IPMask(net.ParseIP("255.255.255.0"))},
+				{IP: net.ParseIP("9.1.1.3"), Mask: net.IPMask(net.ParseIP("255.255.0.0"))},
+			},
+			Groups:    []string{"test-group1", "test-group2", "test-group3"},
+			NotBefore: before,
+			NotAfter:  after,
+			PublicKey: pubKey,
+			IsCA:      false,
+			Curve:     Curve_P256,
+			Issuer:    "1234567890abcedfghij1234567890ab",
+		},
+	}
+
+	priv, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+	pub := elliptic.Marshal(elliptic.P256(), priv.PublicKey.X, priv.PublicKey.Y)
+	rawPriv := priv.D.FillBytes(make([]byte, 32))
+
+	assert.Nil(t, err)
+	assert.False(t, nc.CheckSignature(pub))
+	assert.Nil(t, nc.Sign(Curve_P256, rawPriv))
 	assert.True(t, nc.CheckSignature(pub))
 
 	_, err = nc.Marshal()
@@ -152,7 +199,7 @@ func TestNebulaCertificate_MarshalJSON(t *testing.T) {
 	assert.Nil(t, err)
 	assert.Equal(
 		t,
-		"{\"details\":{\"groups\":[\"test-group1\",\"test-group2\",\"test-group3\"],\"ips\":[\"10.1.1.1/24\",\"10.1.1.2/16\",\"10.1.1.3/ff00ff00\"],\"isCa\":false,\"issuer\":\"1234567890abcedfghij1234567890ab\",\"name\":\"testing\",\"notAfter\":\"0000-11-30T02:00:00Z\",\"notBefore\":\"0000-11-30T01:00:00Z\",\"publicKey\":\"313233343536373839306162636564666768696a313233343536373839306162\",\"subnets\":[\"9.1.1.1/ff00ff00\",\"9.1.1.2/24\",\"9.1.1.3/16\"]},\"fingerprint\":\"26cb1c30ad7872c804c166b5150fa372f437aa3856b04edb4334b4470ec728e4\",\"signature\":\"313233343536373839306162636564666768696a313233343536373839306162\"}",
+		"{\"details\":{\"curve\":\"CURVE25519\",\"groups\":[\"test-group1\",\"test-group2\",\"test-group3\"],\"ips\":[\"10.1.1.1/24\",\"10.1.1.2/16\",\"10.1.1.3/ff00ff00\"],\"isCa\":false,\"issuer\":\"1234567890abcedfghij1234567890ab\",\"name\":\"testing\",\"notAfter\":\"0000-11-30T02:00:00Z\",\"notBefore\":\"0000-11-30T01:00:00Z\",\"publicKey\":\"313233343536373839306162636564666768696a313233343536373839306162\",\"subnets\":[\"9.1.1.1/ff00ff00\",\"9.1.1.2/24\",\"9.1.1.3/16\"]},\"fingerprint\":\"26cb1c30ad7872c804c166b5150fa372f437aa3856b04edb4334b4470ec728e4\",\"signature\":\"313233343536373839306162636564666768696a313233343536373839306162\"}",
 		string(b),
 	)
 }
@@ -172,13 +219,13 @@ func TestNebulaCertificate_Verify(t *testing.T) {
 
 	f, err := c.Sha256Sum()
 	assert.Nil(t, err)
-	caPool.BlacklistFingerprint(f)
+	caPool.BlocklistFingerprint(f)
 
 	v, err := c.Verify(time.Now(), caPool)
 	assert.False(t, v)
-	assert.EqualError(t, err, "certificate has been blacklisted")
+	assert.EqualError(t, err, "certificate is in the block list")
 
-	caPool.ResetCertBlacklist()
+	caPool.ResetCertBlocklist()
 	v, err = c.Verify(time.Now(), caPool)
 	assert.True(t, v)
 	assert.Nil(t, err)
@@ -216,6 +263,65 @@ func TestNebulaCertificate_Verify(t *testing.T) {
 	assert.Nil(t, err)
 }
 
+func TestNebulaCertificate_VerifyP256(t *testing.T) {
+	ca, _, caKey, err := newTestCaCertP256(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+
+	c, _, _, err := newTestCert(ca, caKey, time.Now(), time.Now().Add(5*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+
+	h, err := ca.Sha256Sum()
+	assert.Nil(t, err)
+
+	caPool := NewCAPool()
+	caPool.CAs[h] = ca
+
+	f, err := c.Sha256Sum()
+	assert.Nil(t, err)
+	caPool.BlocklistFingerprint(f)
+
+	v, err := c.Verify(time.Now(), caPool)
+	assert.False(t, v)
+	assert.EqualError(t, err, "certificate is in the block list")
+
+	caPool.ResetCertBlocklist()
+	v, err = c.Verify(time.Now(), caPool)
+	assert.True(t, v)
+	assert.Nil(t, err)
+
+	v, err = c.Verify(time.Now().Add(time.Hour*1000), caPool)
+	assert.False(t, v)
+	assert.EqualError(t, err, "root certificate is expired")
+
+	c, _, _, err = newTestCert(ca, caKey, time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	v, err = c.Verify(time.Now().Add(time.Minute*6), caPool)
+	assert.False(t, v)
+	assert.EqualError(t, err, "certificate is expired")
+
+	// Test group assertion
+	ca, _, caKey, err = newTestCaCertP256(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{"test1", "test2"})
+	assert.Nil(t, err)
+
+	caPem, err := ca.MarshalToPEM()
+	assert.Nil(t, err)
+
+	caPool = NewCAPool()
+	caPool.AddCACertificate(caPem)
+
+	c, _, _, err = newTestCert(ca, caKey, time.Now(), time.Now().Add(5*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{"test1", "bad"})
+	assert.Nil(t, err)
+	v, err = c.Verify(time.Now(), caPool)
+	assert.False(t, v)
+	assert.EqualError(t, err, "certificate contained a group not present on the signing ca: bad")
+
+	c, _, _, err = newTestCert(ca, caKey, time.Now(), time.Now().Add(5*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{"test1"})
+	assert.Nil(t, err)
+	v, err = c.Verify(time.Now(), caPool)
+	assert.True(t, v)
+	assert.Nil(t, err)
+}
+
 func TestNebulaCertificate_Verify_IPs(t *testing.T) {
 	_, caIp1, _ := net.ParseCIDR("10.0.0.0/16")
 	_, caIp2, _ := net.ParseCIDR("192.168.0.0/24")
@@ -374,16 +480,43 @@ func TestNebulaCertificate_Verify_Subnets(t *testing.T) {
 	assert.Nil(t, err)
 }
 
-func TestNebulaVerifyPrivateKey(t *testing.T) {
+func TestNebulaCertificate_VerifyPrivateKey(t *testing.T) {
 	ca, _, caKey, err := newTestCaCert(time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
 	assert.Nil(t, err)
+	err = ca.VerifyPrivateKey(Curve_CURVE25519, caKey)
+	assert.Nil(t, err)
+
+	_, _, caKey2, err := newTestCaCert(time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	err = ca.VerifyPrivateKey(Curve_CURVE25519, caKey2)
+	assert.NotNil(t, err)
 
 	c, _, priv, err := newTestCert(ca, caKey, time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
-	err = c.VerifyPrivateKey(priv)
+	err = c.VerifyPrivateKey(Curve_CURVE25519, priv)
 	assert.Nil(t, err)
 
 	_, priv2 := x25519Keypair()
-	err = c.VerifyPrivateKey(priv2)
+	err = c.VerifyPrivateKey(Curve_CURVE25519, priv2)
+	assert.NotNil(t, err)
+}
+
+func TestNebulaCertificate_VerifyPrivateKeyP256(t *testing.T) {
+	ca, _, caKey, err := newTestCaCertP256(time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	err = ca.VerifyPrivateKey(Curve_P256, caKey)
+	assert.Nil(t, err)
+
+	_, _, caKey2, err := newTestCaCertP256(time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	err = ca.VerifyPrivateKey(Curve_P256, caKey2)
+	assert.NotNil(t, err)
+
+	c, _, priv, err := newTestCert(ca, caKey, time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	err = c.VerifyPrivateKey(Curve_P256, priv)
+	assert.Nil(t, err)
+
+	_, priv2 := p256Keypair()
+	err = c.VerifyPrivateKey(Curve_P256, priv2)
 	assert.NotNil(t, err)
 }
 
@@ -421,6 +554,32 @@ BVG+oJpAoqokUBbI4U0N8CSfpUABEkB/Pm5A2xyH/nc8mg/wvGUWG3pZ7nHzaDMf
 8/phAUt+FLzqTECzQKisYswKvE3pl9mbEYKbOdIHrxdIp95mo4sF
 -----END NEBULA CERTIFICATE-----
 
+`
+
+	expired := `
+# expired certificate
+-----BEGIN NEBULA CERTIFICATE-----
+CjkKB2V4cGlyZWQouPmWjQYwufmWjQY6ILCRaoCkJlqHgv5jfDN4lzLHBvDzaQm4
+vZxfu144hmgjQAESQG4qlnZi8DncvD/LDZnLgJHOaX1DWCHHEh59epVsC+BNgTie
+WH1M9n4O7cFtGlM6sJJOS+rCVVEJ3ABS7+MPdQs=
+-----END NEBULA CERTIFICATE-----
+`
+
+	p256 := `
+# p256 certificate
+-----BEGIN NEBULA CERTIFICATE-----
+CmYKEG5lYnVsYSBQMjU2IHRlc3Qo4s+7mgYw4tXrsAc6QQRkaW2jFmllYvN4+/k2
+6tctO9sPT3jOx8ES6M1nIqOhpTmZeabF/4rELDqPV4aH5jfJut798DUXql0FlF8H
+76gvQAGgBgESRzBFAiEAib0/te6eMiZOKD8gdDeloMTS0wGuX2t0C7TFdUhAQzgC
+IBNWYMep3ysx9zCgknfG5dKtwGTaqF++BWKDYdyl34KX
+-----END NEBULA CERTIFICATE-----
+`
+
+	v2 := `
+# valid PEM with the V2 header
+-----BEGIN NEBULA CERTIFICATE V2-----
+CmYKEG5lYnVsYSBQMjU2IHRlc3Qo4s+7mgYw4tXrsAc6QQRkaW2jFmllYvN4+/k2
+-----END NEBULA CERTIFICATE V2-----
 `
 
 	rootCA := NebulaCertificate{
@@ -435,15 +594,428 @@ BVG+oJpAoqokUBbI4U0N8CSfpUABEkB/Pm5A2xyH/nc8mg/wvGUWG3pZ7nHzaDMf
 		},
 	}
 
-	p, err := NewCAPoolFromBytes([]byte(noNewLines))
+	rootCAP256 := NebulaCertificate{
+		Details: NebulaCertificateDetails{
+			Name: "nebula P256 test",
+		},
+	}
+
+	p, warn, err := NewCAPoolFromBytes([]byte(noNewLines))
 	assert.Nil(t, err)
+	assert.Nil(t, warn)
 	assert.Equal(t, p.CAs[string("c9bfaf7ce8e84b2eeda2e27b469f4b9617bde192efd214b68891ecda6ed49522")].Details.Name, rootCA.Details.Name)
 	assert.Equal(t, p.CAs[string("5c9c3f23e7ee7fe97637cbd3a0a5b854154d1d9aaaf7b566a51f4a88f76b64cd")].Details.Name, rootCA01.Details.Name)
 
-	pp, err := NewCAPoolFromBytes([]byte(withNewLines))
+	pp, warn, err := NewCAPoolFromBytes([]byte(withNewLines))
 	assert.Nil(t, err)
+	assert.Nil(t, warn)
 	assert.Equal(t, pp.CAs[string("c9bfaf7ce8e84b2eeda2e27b469f4b9617bde192efd214b68891ecda6ed49522")].Details.Name, rootCA.Details.Name)
 	assert.Equal(t, pp.CAs[string("5c9c3f23e7ee7fe97637cbd3a0a5b854154d1d9aaaf7b566a51f4a88f76b64cd")].Details.Name, rootCA01.Details.Name)
+
+	// expired cert, no valid certs
+	ppp, warn, err := NewCAPoolFromBytes([]byte(expired))
+	assert.Error(t, err, "no valid CA certificates present")
+	assert.Len(t, warn, 1)
+	assert.Error(t, warn[0], ErrExpired)
+	assert.Nil(t, ppp)
+
+	// expired cert, with valid certs
+	pppp, warn, err := NewCAPoolFromBytes(append([]byte(expired), noNewLines...))
+	assert.Len(t, warn, 1)
+	assert.Nil(t, err)
+	assert.Error(t, warn[0], ErrExpired)
+	assert.Equal(t, pppp.CAs[string("c9bfaf7ce8e84b2eeda2e27b469f4b9617bde192efd214b68891ecda6ed49522")].Details.Name, rootCA.Details.Name)
+	assert.Equal(t, pppp.CAs[string("5c9c3f23e7ee7fe97637cbd3a0a5b854154d1d9aaaf7b566a51f4a88f76b64cd")].Details.Name, rootCA01.Details.Name)
+	assert.Equal(t, pppp.CAs[string("152070be6bb19bc9e3bde4c2f0e7d8f4ff5448b4c9856b8eccb314fade0229b0")].Details.Name, "expired")
+	assert.Equal(t, len(pppp.CAs), 3)
+
+	ppppp, warn, err := NewCAPoolFromBytes([]byte(p256))
+	assert.Nil(t, err)
+	assert.Nil(t, warn)
+	assert.Equal(t, ppppp.CAs[string("a7938893ec8c4ef769b06d7f425e5e46f7a7f5ffa49c3bcf4a86b608caba9159")].Details.Name, rootCAP256.Details.Name)
+	assert.Equal(t, len(ppppp.CAs), 1)
+
+	pppppp, warn, err := NewCAPoolFromBytes(append([]byte(p256), []byte(v2)...))
+	assert.Nil(t, err)
+	assert.True(t, errors.Is(warn[0], ErrInvalidPEMCertificateUnsupported))
+	assert.Equal(t, pppppp.CAs[string("a7938893ec8c4ef769b06d7f425e5e46f7a7f5ffa49c3bcf4a86b608caba9159")].Details.Name, rootCAP256.Details.Name)
+	assert.Equal(t, len(pppppp.CAs), 1)
+}
+
+func appendByteSlices(b ...[]byte) []byte {
+	retSlice := []byte{}
+	for _, v := range b {
+		retSlice = append(retSlice, v...)
+	}
+	return retSlice
+}
+
+func TestUnmrshalCertPEM(t *testing.T) {
+	goodCert := []byte(`
+# A good cert
+-----BEGIN NEBULA CERTIFICATE-----
+CkAKDm5lYnVsYSByb290IGNhKJfap9AFMJfg1+YGOiCUQGByMuNRhIlQBOyzXWbL
+vcKBwDhov900phEfJ5DN3kABEkDCq5R8qBiu8sl54yVfgRcQXEDt3cHr8UTSLszv
+bzBEr00kERQxxTzTsH8cpYEgRoipvmExvg8WP8NdAJEYJosB
+-----END NEBULA CERTIFICATE-----
+`)
+	badBanner := []byte(`# A bad banner
+-----BEGIN NOT A NEBULA CERTIFICATE-----
+CkAKDm5lYnVsYSByb290IGNhKJfap9AFMJfg1+YGOiCUQGByMuNRhIlQBOyzXWbL
+vcKBwDhov900phEfJ5DN3kABEkDCq5R8qBiu8sl54yVfgRcQXEDt3cHr8UTSLszv
+bzBEr00kERQxxTzTsH8cpYEgRoipvmExvg8WP8NdAJEYJosB
+-----END NOT A NEBULA CERTIFICATE-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA CERTIFICATE-----
+CkAKDm5lYnVsYSByb290IGNhKJfap9AFMJfg1+YGOiCUQGByMuNRhIlQBOyzXWbL
+vcKBwDhov900phEfJ5DN3kABEkDCq5R8qBiu8sl54yVfgRcQXEDt3cHr8UTSLszv
+bzBEr00kERQxxTzTsH8cpYEgRoipvmExvg8WP8NdAJEYJosB
+-END NEBULA CERTIFICATE----`)
+
+	certBundle := appendByteSlices(goodCert, badBanner, invalidPem)
+
+	// Success test case
+	cert, rest, err := UnmarshalNebulaCertificateFromPEM(certBundle)
+	assert.NotNil(t, cert)
+	assert.Equal(t, rest, append(badBanner, invalidPem...))
+	assert.Nil(t, err)
+
+	// Fail due to invalid banner.
+	cert, rest, err = UnmarshalNebulaCertificateFromPEM(rest)
+	assert.Nil(t, cert)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula certificate banner")
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	cert, rest, err = UnmarshalNebulaCertificateFromPEM(rest)
+	assert.Nil(t, cert)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
+}
+
+func TestUnmarshalSigningPrivateKey(t *testing.T) {
+	privKey := []byte(`# A good key
+-----BEGIN NEBULA ED25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-----END NEBULA ED25519 PRIVATE KEY-----
+`)
+	privP256Key := []byte(`# A good key
+-----BEGIN NEBULA ECDSA P256 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA ECDSA P256 PRIVATE KEY-----
+`)
+	shortKey := []byte(`# A short key
+-----BEGIN NEBULA ED25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+-----END NEBULA ED25519 PRIVATE KEY-----
+`)
+	invalidBanner := []byte(`# Invalid banner
+-----BEGIN NOT A NEBULA PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-----END NOT A NEBULA PRIVATE KEY-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA ED25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-END NEBULA ED25519 PRIVATE KEY-----`)
+
+	keyBundle := appendByteSlices(privKey, privP256Key, shortKey, invalidBanner, invalidPem)
+
+	// Success test case
+	k, rest, curve, err := UnmarshalSigningPrivateKey(keyBundle)
+	assert.Len(t, k, 64)
+	assert.Equal(t, rest, appendByteSlices(privP256Key, shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_CURVE25519, curve)
+	assert.Nil(t, err)
+
+	// Success test case
+	k, rest, curve, err = UnmarshalSigningPrivateKey(rest)
+	assert.Len(t, k, 32)
+	assert.Equal(t, rest, appendByteSlices(shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_P256, curve)
+	assert.Nil(t, err)
+
+	// Fail due to short key
+	k, rest, curve, err = UnmarshalSigningPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, appendByteSlices(invalidBanner, invalidPem))
+	assert.EqualError(t, err, "key was not 64 bytes, is invalid Ed25519 private key")
+
+	// Fail due to invalid banner
+	k, rest, curve, err = UnmarshalSigningPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula Ed25519/ECDSA private key banner")
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	k, rest, curve, err = UnmarshalSigningPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
+}
+
+func TestDecryptAndUnmarshalSigningPrivateKey(t *testing.T) {
+	passphrase := []byte("DO NOT USE THIS KEY")
+	privKey := []byte(`# A good key
+-----BEGIN NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+CjwKC0FFUy0yNTYtR0NNEi0IExCAgIABGAEgBCognnjujd67Vsv99p22wfAjQaDT
+oCMW1mdjkU3gACKNW4MSXOWR9Sts4C81yk1RUku2gvGKs3TB9LYoklLsIizSYOLl
++Vs//O1T0I1Xbml2XBAROsb/VSoDln/6LMqR4B6fn6B3GOsLBBqRI8daDl9lRMPB
+qrlJ69wer3ZUHFXA
+-----END NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+`)
+	shortKey := []byte(`# A key which, once decrypted, is too short
+-----BEGIN NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+CjwKC0FFUy0yNTYtR0NNEi0IExCAgIABGAEgBCoga5h8owMEBWRSMMJKzuUvWce7
+k0qlBkQmCxiuLh80MuASW70YcKt8jeEIS2axo2V6zAKA9TSMcCsJW1kDDXEtL/xe
+GLF5T7sDl5COp4LU3pGxpV+KoeQ/S3gQCAAcnaOtnJQX+aSDnbO3jCHyP7U9CHbs
+rQr3bdH3Oy/WiYU=
+-----END NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+`)
+	invalidBanner := []byte(`# Invalid banner (not encrypted)
+-----BEGIN NEBULA ED25519 PRIVATE KEY-----
+bWRp2CTVFhW9HD/qCd28ltDgK3w8VXSeaEYczDWos8sMUBqDb9jP3+NYwcS4lURG
+XgLvodMXZJuaFPssp+WwtA==
+-----END NEBULA ED25519 PRIVATE KEY-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+CjwKC0FFUy0yNTYtR0NNEi0IExCAgIABGAEgBCognnjujd67Vsv99p22wfAjQaDT
+oCMW1mdjkU3gACKNW4MSXOWR9Sts4C81yk1RUku2gvGKs3TB9LYoklLsIizSYOLl
++Vs//O1T0I1Xbml2XBAROsb/VSoDln/6LMqR4B6fn6B3GOsLBBqRI8daDl9lRMPB
+qrlJ69wer3ZUHFXA
+-END NEBULA ED25519 ENCRYPTED PRIVATE KEY-----
+`)
+
+	keyBundle := appendByteSlices(privKey, shortKey, invalidBanner, invalidPem)
+
+	// Success test case
+	curve, k, rest, err := DecryptAndUnmarshalSigningPrivateKey(passphrase, keyBundle)
+	assert.Nil(t, err)
+	assert.Equal(t, Curve_CURVE25519, curve)
+	assert.Len(t, k, 64)
+	assert.Equal(t, rest, appendByteSlices(shortKey, invalidBanner, invalidPem))
+
+	// Fail due to short key
+	curve, k, rest, err = DecryptAndUnmarshalSigningPrivateKey(passphrase, rest)
+	assert.EqualError(t, err, "key was not 64 bytes, is invalid ed25519 private key")
+	assert.Nil(t, k)
+	assert.Equal(t, rest, appendByteSlices(invalidBanner, invalidPem))
+
+	// Fail due to invalid banner
+	curve, k, rest, err = DecryptAndUnmarshalSigningPrivateKey(passphrase, rest)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula encrypted Ed25519/ECDSA private key banner")
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	curve, k, rest, err = DecryptAndUnmarshalSigningPrivateKey(passphrase, rest)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+
+	// Fail due to invalid passphrase
+	curve, k, rest, err = DecryptAndUnmarshalSigningPrivateKey([]byte("invalid passphrase"), privKey)
+	assert.EqualError(t, err, "invalid passphrase or corrupt private key")
+	assert.Nil(t, k)
+	assert.Equal(t, rest, []byte{})
+}
+
+func TestEncryptAndMarshalSigningPrivateKey(t *testing.T) {
+	// Having proved that decryption works correctly above, we can test the
+	// encryption function produces a value which can be decrypted
+	passphrase := []byte("passphrase")
+	bytes := []byte("AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA")
+	kdfParams := NewArgon2Parameters(64*1024, 4, 3)
+	key, err := EncryptAndMarshalSigningPrivateKey(Curve_CURVE25519, bytes, passphrase, kdfParams)
+	assert.Nil(t, err)
+
+	// Verify the "key" can be decrypted successfully
+	curve, k, rest, err := DecryptAndUnmarshalSigningPrivateKey(passphrase, key)
+	assert.Len(t, k, 64)
+	assert.Equal(t, Curve_CURVE25519, curve)
+	assert.Equal(t, rest, []byte{})
+	assert.Nil(t, err)
+
+	// EncryptAndMarshalEd25519PrivateKey does not create any errors itself
+}
+
+func TestUnmarshalPrivateKey(t *testing.T) {
+	privKey := []byte(`# A good key
+-----BEGIN NEBULA X25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA X25519 PRIVATE KEY-----
+`)
+	privP256Key := []byte(`# A good key
+-----BEGIN NEBULA P256 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA P256 PRIVATE KEY-----
+`)
+	shortKey := []byte(`# A short key
+-----BEGIN NEBULA X25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-----END NEBULA X25519 PRIVATE KEY-----
+`)
+	invalidBanner := []byte(`# Invalid banner
+-----BEGIN NOT A NEBULA PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NOT A NEBULA PRIVATE KEY-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA X25519 PRIVATE KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-END NEBULA X25519 PRIVATE KEY-----`)
+
+	keyBundle := appendByteSlices(privKey, privP256Key, shortKey, invalidBanner, invalidPem)
+
+	// Success test case
+	k, rest, curve, err := UnmarshalPrivateKey(keyBundle)
+	assert.Len(t, k, 32)
+	assert.Equal(t, rest, appendByteSlices(privP256Key, shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_CURVE25519, curve)
+	assert.Nil(t, err)
+
+	// Success test case
+	k, rest, curve, err = UnmarshalPrivateKey(rest)
+	assert.Len(t, k, 32)
+	assert.Equal(t, rest, appendByteSlices(shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_P256, curve)
+	assert.Nil(t, err)
+
+	// Fail due to short key
+	k, rest, curve, err = UnmarshalPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, appendByteSlices(invalidBanner, invalidPem))
+	assert.EqualError(t, err, "key was not 32 bytes, is invalid CURVE25519 private key")
+
+	// Fail due to invalid banner
+	k, rest, curve, err = UnmarshalPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula private key banner")
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	k, rest, curve, err = UnmarshalPrivateKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
+}
+
+func TestUnmarshalEd25519PublicKey(t *testing.T) {
+	pubKey := []byte(`# A good key
+-----BEGIN NEBULA ED25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA ED25519 PUBLIC KEY-----
+`)
+	shortKey := []byte(`# A short key
+-----BEGIN NEBULA ED25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-----END NEBULA ED25519 PUBLIC KEY-----
+`)
+	invalidBanner := []byte(`# Invalid banner
+-----BEGIN NOT A NEBULA PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NOT A NEBULA PUBLIC KEY-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA ED25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-END NEBULA ED25519 PUBLIC KEY-----`)
+
+	keyBundle := appendByteSlices(pubKey, shortKey, invalidBanner, invalidPem)
+
+	// Success test case
+	k, rest, err := UnmarshalEd25519PublicKey(keyBundle)
+	assert.Equal(t, len(k), 32)
+	assert.Nil(t, err)
+	assert.Equal(t, rest, appendByteSlices(shortKey, invalidBanner, invalidPem))
+
+	// Fail due to short key
+	k, rest, err = UnmarshalEd25519PublicKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, appendByteSlices(invalidBanner, invalidPem))
+	assert.EqualError(t, err, "key was not 32 bytes, is invalid ed25519 public key")
+
+	// Fail due to invalid banner
+	k, rest, err = UnmarshalEd25519PublicKey(rest)
+	assert.Nil(t, k)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula Ed25519 public key banner")
+	assert.Equal(t, rest, invalidPem)
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	k, rest, err = UnmarshalEd25519PublicKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
+}
+
+func TestUnmarshalX25519PublicKey(t *testing.T) {
+	pubKey := []byte(`# A good key
+-----BEGIN NEBULA X25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA X25519 PUBLIC KEY-----
+`)
+	pubP256Key := []byte(`# A good key
+-----BEGIN NEBULA P256 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAA=
+-----END NEBULA P256 PUBLIC KEY-----
+`)
+	shortKey := []byte(`# A short key
+-----BEGIN NEBULA X25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==
+-----END NEBULA X25519 PUBLIC KEY-----
+`)
+	invalidBanner := []byte(`# Invalid banner
+-----BEGIN NOT A NEBULA PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-----END NOT A NEBULA PUBLIC KEY-----
+`)
+	invalidPem := []byte(`# Not a valid PEM format
+-BEGIN NEBULA X25519 PUBLIC KEY-----
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
+-END NEBULA X25519 PUBLIC KEY-----`)
+
+	keyBundle := appendByteSlices(pubKey, pubP256Key, shortKey, invalidBanner, invalidPem)
+
+	// Success test case
+	k, rest, curve, err := UnmarshalPublicKey(keyBundle)
+	assert.Equal(t, len(k), 32)
+	assert.Nil(t, err)
+	assert.Equal(t, rest, appendByteSlices(pubP256Key, shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_CURVE25519, curve)
+
+	// Success test case
+	k, rest, curve, err = UnmarshalPublicKey(rest)
+	assert.Equal(t, len(k), 65)
+	assert.Nil(t, err)
+	assert.Equal(t, rest, appendByteSlices(shortKey, invalidBanner, invalidPem))
+	assert.Equal(t, Curve_P256, curve)
+
+	// Fail due to short key
+	k, rest, curve, err = UnmarshalPublicKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, appendByteSlices(invalidBanner, invalidPem))
+	assert.EqualError(t, err, "key was not 32 bytes, is invalid CURVE25519 public key")
+
+	// Fail due to invalid banner
+	k, rest, curve, err = UnmarshalPublicKey(rest)
+	assert.Nil(t, k)
+	assert.EqualError(t, err, "bytes did not contain a proper nebula public key banner")
+	assert.Equal(t, rest, invalidPem)
+
+	// Fail due to ivalid PEM format, because
+	// it's missing the requisite pre-encapsulation boundary.
+	k, rest, curve, err = UnmarshalPublicKey(rest)
+	assert.Nil(t, k)
+	assert.Equal(t, rest, invalidPem)
+	assert.EqualError(t, err, "input did not contain a valid PEM encoded block")
 }
 
 // Ensure that upgrading the protobuf library does not change how certificates
@@ -487,6 +1059,24 @@ func TestMarshalingNebulaCertificateConsistency(t *testing.T) {
 	assert.Equal(t, "0a0774657374696e67121b8182845080feffff0f828284508080fcff0f8382845080fe83f80f1a1b8182844880fe83f80f8282844880feffff0f838284488080fcff0f220b746573742d67726f757031220b746573742d67726f757032220b746573742d67726f75703328f0e0e7d70430a08681c4053a20313233343536373839306162636564666768696a3132333435363738393061624a081234567890abcedf", fmt.Sprintf("%x", b))
 }
 
+func TestNebulaCertificate_Copy(t *testing.T) {
+	ca, _, caKey, err := newTestCaCert(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+
+	c, _, _, err := newTestCert(ca, caKey, time.Now(), time.Now().Add(5*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	cc := c.Copy()
+
+	test.AssertDeepCopyEqual(t, c, cc)
+}
+
+func TestUnmarshalNebulaCertificate(t *testing.T) {
+	// Test that we don't panic with an invalid certificate (#332)
+	data := []byte("\x98\x00\x00")
+	_, err := UnmarshalNebulaCertificate(data)
+	assert.EqualError(t, err, "encoded Details was nil")
+}
+
 func newTestCaCert(before, after time.Time, ips, subnets []*net.IPNet, groups []string) (*NebulaCertificate, []byte, []byte, error) {
 	pub, priv, err := ed25519.GenerateKey(rand.Reader)
 	if before.IsZero() {
@@ -498,11 +1088,12 @@ func newTestCaCert(before, after time.Time, ips, subnets []*net.IPNet, groups []
 
 	nc := &NebulaCertificate{
 		Details: NebulaCertificateDetails{
-			Name:      "test ca",
-			NotBefore: before,
-			NotAfter:  after,
-			PublicKey: pub,
-			IsCA:      true,
+			Name:           "test ca",
+			NotBefore:      time.Unix(before.Unix(), 0),
+			NotAfter:       time.Unix(after.Unix(), 0),
+			PublicKey:      pub,
+			IsCA:           true,
+			InvertedGroups: make(map[string]struct{}),
 		},
 	}
 
@@ -518,13 +1109,56 @@ func newTestCaCert(before, after time.Time, ips, subnets []*net.IPNet, groups []
 		nc.Details.Groups = groups
 	}
 
-	err = nc.Sign(priv)
+	err = nc.Sign(Curve_CURVE25519, priv)
 	if err != nil {
 		return nil, nil, nil, err
 	}
 	return nc, pub, priv, nil
 }
 
+func newTestCaCertP256(before, after time.Time, ips, subnets []*net.IPNet, groups []string) (*NebulaCertificate, []byte, []byte, error) {
+	priv, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+	pub := elliptic.Marshal(elliptic.P256(), priv.PublicKey.X, priv.PublicKey.Y)
+	rawPriv := priv.D.FillBytes(make([]byte, 32))
+
+	if before.IsZero() {
+		before = time.Now().Add(time.Second * -60).Round(time.Second)
+	}
+	if after.IsZero() {
+		after = time.Now().Add(time.Second * 60).Round(time.Second)
+	}
+
+	nc := &NebulaCertificate{
+		Details: NebulaCertificateDetails{
+			Name:           "test ca",
+			NotBefore:      time.Unix(before.Unix(), 0),
+			NotAfter:       time.Unix(after.Unix(), 0),
+			PublicKey:      pub,
+			IsCA:           true,
+			Curve:          Curve_P256,
+			InvertedGroups: make(map[string]struct{}),
+		},
+	}
+
+	if len(ips) > 0 {
+		nc.Details.Ips = ips
+	}
+
+	if len(subnets) > 0 {
+		nc.Details.Subnets = subnets
+	}
+
+	if len(groups) > 0 {
+		nc.Details.Groups = groups
+	}
+
+	err = nc.Sign(Curve_P256, rawPriv)
+	if err != nil {
+		return nil, nil, nil, err
+	}
+	return nc, pub, rawPriv, nil
+}
+
 func newTestCert(ca *NebulaCertificate, key []byte, before, after time.Time, ips, subnets []*net.IPNet, groups []string) (*NebulaCertificate, []byte, []byte, error) {
 	issuer, err := ca.Sha256Sum()
 	if err != nil {
@@ -544,37 +1178,48 @@ func newTestCert(ca *NebulaCertificate, key []byte, before, after time.Time, ips
 
 	if len(ips) == 0 {
 		ips = []*net.IPNet{
-			{IP: net.ParseIP("10.1.1.1"), Mask: net.IPMask(net.ParseIP("255.255.255.0"))},
-			{IP: net.ParseIP("10.1.1.2"), Mask: net.IPMask(net.ParseIP("255.255.0.0"))},
-			{IP: net.ParseIP("10.1.1.3"), Mask: net.IPMask(net.ParseIP("255.0.255.0"))},
+			{IP: net.ParseIP("10.1.1.1").To4(), Mask: net.IPMask(net.ParseIP("255.255.255.0").To4())},
+			{IP: net.ParseIP("10.1.1.2").To4(), Mask: net.IPMask(net.ParseIP("255.255.0.0").To4())},
+			{IP: net.ParseIP("10.1.1.3").To4(), Mask: net.IPMask(net.ParseIP("255.0.255.0").To4())},
 		}
 	}
 
 	if len(subnets) == 0 {
 		subnets = []*net.IPNet{
-			{IP: net.ParseIP("9.1.1.1"), Mask: net.IPMask(net.ParseIP("255.0.255.0"))},
-			{IP: net.ParseIP("9.1.1.2"), Mask: net.IPMask(net.ParseIP("255.255.255.0"))},
-			{IP: net.ParseIP("9.1.1.3"), Mask: net.IPMask(net.ParseIP("255.255.0.0"))},
+			{IP: net.ParseIP("9.1.1.1").To4(), Mask: net.IPMask(net.ParseIP("255.0.255.0").To4())},
+			{IP: net.ParseIP("9.1.1.2").To4(), Mask: net.IPMask(net.ParseIP("255.255.255.0").To4())},
+			{IP: net.ParseIP("9.1.1.3").To4(), Mask: net.IPMask(net.ParseIP("255.255.0.0").To4())},
 		}
 	}
 
-	pub, rawPriv := x25519Keypair()
+	var pub, rawPriv []byte
+
+	switch ca.Details.Curve {
+	case Curve_CURVE25519:
+		pub, rawPriv = x25519Keypair()
+	case Curve_P256:
+		pub, rawPriv = p256Keypair()
+	default:
+		return nil, nil, nil, fmt.Errorf("unknown curve: %v", ca.Details.Curve)
+	}
 
 	nc := &NebulaCertificate{
 		Details: NebulaCertificateDetails{
-			Name:      "testing",
-			Ips:       ips,
-			Subnets:   subnets,
-			Groups:    groups,
-			NotBefore: before,
-			NotAfter:  after,
-			PublicKey: pub,
-			IsCA:      false,
-			Issuer:    issuer,
+			Name:           "testing",
+			Ips:            ips,
+			Subnets:        subnets,
+			Groups:         groups,
+			NotBefore:      time.Unix(before.Unix(), 0),
+			NotAfter:       time.Unix(after.Unix(), 0),
+			PublicKey:      pub,
+			IsCA:           false,
+			Curve:          ca.Details.Curve,
+			Issuer:         issuer,
+			InvertedGroups: make(map[string]struct{}),
 		},
 	}
 
-	err = nc.Sign(key)
+	err = nc.Sign(ca.Details.Curve, key)
 	if err != nil {
 		return nil, nil, nil, err
 	}
@@ -583,10 +1228,24 @@ func newTestCert(ca *NebulaCertificate, key []byte, before, after time.Time, ips
 }
 
 func x25519Keypair() ([]byte, []byte) {
-	var pubkey, privkey [32]byte
-	if _, err := io.ReadFull(rand.Reader, privkey[:]); err != nil {
+	privkey := make([]byte, 32)
+	if _, err := io.ReadFull(rand.Reader, privkey); err != nil {
 		panic(err)
 	}
-	curve25519.ScalarBaseMult(&pubkey, &privkey)
-	return pubkey[:], privkey[:]
+
+	pubkey, err := curve25519.X25519(privkey, curve25519.Basepoint)
+	if err != nil {
+		panic(err)
+	}
+
+	return pubkey, privkey
+}
+
+func p256Keypair() ([]byte, []byte) {
+	privkey, err := ecdh.P256().GenerateKey(rand.Reader)
+	if err != nil {
+		panic(err)
+	}
+	pubkey := privkey.PublicKey()
+	return pubkey.Bytes(), privkey.Bytes()
 }

cert/crypto.go

@@ -0,0 +1,143 @@
+package cert
+
+import (
+	"crypto/aes"
+	"crypto/cipher"
+	"crypto/rand"
+	"fmt"
+	"io"
+
+	"golang.org/x/crypto/argon2"
+)
+
+// KDF factors
+type Argon2Parameters struct {
+	version     rune
+	Memory      uint32 // KiB
+	Parallelism uint8
+	Iterations  uint32
+	salt        []byte
+}
+
+// Returns a new Argon2Parameters object with current version set
+func NewArgon2Parameters(memory uint32, parallelism uint8, iterations uint32) *Argon2Parameters {
+	return &Argon2Parameters{
+		version:     argon2.Version,
+		Memory:      memory, // KiB
+		Parallelism: parallelism,
+		Iterations:  iterations,
+	}
+}
+
+// Encrypts data using AES-256-GCM and the Argon2id key derivation function
+func aes256Encrypt(passphrase []byte, kdfParams *Argon2Parameters, data []byte) ([]byte, error) {
+	key, err := aes256DeriveKey(passphrase, kdfParams)
+	if err != nil {
+		return nil, err
+	}
+
+	// this should never happen, but since this dictates how our calls into the
+	// aes package behave and could be catastraphic, let's sanity check this
+	if len(key) != 32 {
+		return nil, fmt.Errorf("invalid AES-256 key length (%d) - cowardly refusing to encrypt", len(key))
+	}
+
+	block, err := aes.NewCipher(key)
+	if err != nil {
+		return nil, err
+	}
+
+	gcm, err := cipher.NewGCM(block)
+	if err != nil {
+		return nil, err
+	}
+
+	nonce := make([]byte, gcm.NonceSize())
+	if _, err := io.ReadFull(rand.Reader, nonce); err != nil {
+		return nil, err
+	}
+
+	ciphertext := gcm.Seal(nil, nonce, data, nil)
+	blob := joinNonceCiphertext(nonce, ciphertext)
+
+	return blob, nil
+}
+
+// Decrypts data using AES-256-GCM and the Argon2id key derivation function
+// Expects the data to include an Argon2id parameter string before the encrypted data
+func aes256Decrypt(passphrase []byte, kdfParams *Argon2Parameters, data []byte) ([]byte, error) {
+	key, err := aes256DeriveKey(passphrase, kdfParams)
+	if err != nil {
+		return nil, err
+	}
+
+	block, err := aes.NewCipher(key)
+	if err != nil {
+		return nil, err
+	}
+
+	gcm, err := cipher.NewGCM(block)
+	if err != nil {
+		return nil, err
+	}
+
+	nonce, ciphertext, err := splitNonceCiphertext(data, gcm.NonceSize())
+	if err != nil {
+		return nil, err
+	}
+
+	plaintext, err := gcm.Open(nil, nonce, ciphertext, nil)
+	if err != nil {
+		return nil, fmt.Errorf("invalid passphrase or corrupt private key")
+	}
+
+	return plaintext, nil
+}
+
+func aes256DeriveKey(passphrase []byte, params *Argon2Parameters) ([]byte, error) {
+	if params.salt == nil {
+		params.salt = make([]byte, 32)
+		if _, err := rand.Read(params.salt); err != nil {
+			return nil, err
+		}
+	}
+
+	// keySize of 32 bytes will result in AES-256 encryption
+	key, err := deriveKey(passphrase, 32, params)
+	if err != nil {
+		return nil, err
+	}
+
+	return key, nil
+}
+
+// Derives a key from a passphrase using Argon2id
+func deriveKey(passphrase []byte, keySize uint32, params *Argon2Parameters) ([]byte, error) {
+	if params.version != argon2.Version {
+		return nil, fmt.Errorf("incompatible Argon2 version: %d", params.version)
+	}
+
+	if params.salt == nil {
+		return nil, fmt.Errorf("salt must be set in argon2Parameters")
+	} else if len(params.salt) < 16 {
+		return nil, fmt.Errorf("salt must be at least 128  bits")
+	}
+
+	key := argon2.IDKey(passphrase, params.salt, params.Iterations, params.Memory, params.Parallelism, keySize)
+
+	return key, nil
+}
+
+// Prepends nonce to ciphertext
+func joinNonceCiphertext(nonce []byte, ciphertext []byte) []byte {
+	return append(nonce, ciphertext...)
+}
+
+// Splits nonce from ciphertext
+func splitNonceCiphertext(blob []byte, nonceSize int) ([]byte, []byte, error) {
+	if len(blob) <= nonceSize {
+		return nil, nil, fmt.Errorf("invalid ciphertext blob - blob shorter than nonce length")
+	}
+
+	return blob[:nonceSize], blob[nonceSize:], nil
+}

cert/crypto_test.go

@@ -0,0 +1,25 @@
+package cert
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"golang.org/x/crypto/argon2"
+)
+
+func TestNewArgon2Parameters(t *testing.T) {
+	p := NewArgon2Parameters(64*1024, 4, 3)
+	assert.EqualValues(t, &Argon2Parameters{
+		version:     argon2.Version,
+		Memory:      64 * 1024,
+		Parallelism: 4,
+		Iterations:  3,
+	}, p)
+	p = NewArgon2Parameters(2*1024*1024, 2, 1)
+	assert.EqualValues(t, &Argon2Parameters{
+		version:     argon2.Version,
+		Memory:      2 * 1024 * 1024,
+		Parallelism: 2,
+		Iterations:  1,
+	}, p)
+}

cert/errors.go

@@ -0,0 +1,15 @@
+package cert
+
+import (
+	"errors"
+)
+
+var (
+	ErrRootExpired                      = errors.New("root certificate is expired")
+	ErrExpired                          = errors.New("certificate is expired")
+	ErrNotCA                            = errors.New("certificate is not a CA")
+	ErrNotSelfSigned                    = errors.New("certificate is not self-signed")
+	ErrBlockListed                      = errors.New("certificate is in the block list")
+	ErrSignatureMismatch                = errors.New("certificate signature did not match")
+	ErrInvalidPEMCertificateUnsupported = errors.New("bytes contain an unsupported certificate format")
+)

cidr_radix.go

@@ -1,170 +0,0 @@
-package nebula
-
-import (
-	"encoding/binary"
-	"fmt"
-	"net"
-)
-
-type CIDRNode struct {
-	left   *CIDRNode
-	right  *CIDRNode
-	parent *CIDRNode
-	value  interface{}
-}
-
-type CIDRTree struct {
-	root *CIDRNode
-}
-
-const (
-	startbit = uint32(0x80000000)
-)
-
-func NewCIDRTree() *CIDRTree {
-	tree := new(CIDRTree)
-	tree.root = &CIDRNode{}
-	return tree
-}
-
-func (tree *CIDRTree) AddCIDR(cidr *net.IPNet, val interface{}) {
-	bit := startbit
-	node := tree.root
-	next := tree.root
-
-	ip := ip2int(cidr.IP)
-	mask := ip2int(cidr.Mask)
-
-	// Find our last ancestor in the tree
-	for bit&mask != 0 {
-		if ip&bit != 0 {
-			next = node.right
-		} else {
-			next = node.left
-		}
-
-		if next == nil {
-			break
-		}
-
-		bit = bit >> 1
-		node = next
-	}
-
-	// We already have this range so update the value
-	if next != nil {
-		node.value = val
-		return
-	}
-
-	// Build up the rest of the tree we don't already have
-	for bit&mask != 0 {
-		next = &CIDRNode{}
-		next.parent = node
-
-		if ip&bit != 0 {
-			node.right = next
-		} else {
-			node.left = next
-		}
-
-		bit >>= 1
-		node = next
-	}
-
-	// Final node marks our cidr, set the value
-	node.value = val
-}
-
-// Finds the first match, which way be the least specific
-func (tree *CIDRTree) Contains(ip uint32) (value interface{}) {
-	bit := startbit
-	node := tree.root
-
-	for node != nil {
-		if node.value != nil {
-			return node.value
-		}
-
-		if ip&bit != 0 {
-			node = node.right
-		} else {
-			node = node.left
-		}
-
-		bit >>= 1
-
-	}
-
-	return value
-}
-
-// Finds the most specific match
-func (tree *CIDRTree) MostSpecificContains(ip uint32) (value interface{}) {
-	bit := startbit
-	node := tree.root
-
-	for node != nil {
-		if node.value != nil {
-			value = node.value
-		}
-
-		if ip&bit != 0 {
-			node = node.right
-		} else {
-			node = node.left
-		}
-
-		bit >>= 1
-
-	}
-
-	return value
-}
-
-// Finds the most specific match
-func (tree *CIDRTree) Match(ip uint32) (value interface{}) {
-	bit := startbit
-	node := tree.root
-	lastNode := node
-
-	for node != nil {
-		lastNode = node
-		if ip&bit != 0 {
-			node = node.right
-		} else {
-			node = node.left
-		}
-
-		bit >>= 1
-	}
-
-	if bit == 0 && lastNode != nil {
-		value = lastNode.value
-	}
-	return value
-}
-
-// A helper type to avoid converting to IP when logging
-type IntIp uint32
-
-func (ip IntIp) String() string {
-	return fmt.Sprintf("%v", int2ip(uint32(ip)))
-}
-
-func (ip IntIp) MarshalJSON() ([]byte, error) {
-	return []byte(fmt.Sprintf("\"%s\"", int2ip(uint32(ip)).String())), nil
-}
-
-func ip2int(ip []byte) uint32 {
-	if len(ip) == 16 {
-		return binary.BigEndian.Uint32(ip[12:16])
-	}
-	return binary.BigEndian.Uint32(ip)
-}
-
-func int2ip(nn uint32) net.IP {
-	ip := make(net.IP, 4)
-	binary.BigEndian.PutUint32(ip, nn)
-	return ip
-}

cidr_radix_test.go

@@ -1,157 +0,0 @@
-package nebula
-
-import (
-	"net"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-)
-
-func TestCIDRTree_Contains(t *testing.T) {
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.0.0.0/8"), "1")
-	tree.AddCIDR(getCIDR("2.1.0.0/16"), "2")
-	tree.AddCIDR(getCIDR("3.1.1.0/24"), "3")
-	tree.AddCIDR(getCIDR("4.1.1.0/24"), "4a")
-	tree.AddCIDR(getCIDR("4.1.1.1/32"), "4b")
-	tree.AddCIDR(getCIDR("4.1.2.1/32"), "4c")
-	tree.AddCIDR(getCIDR("254.0.0.0/4"), "5")
-
-	tests := []struct {
-		Result interface{}
-		IP     string
-	}{
-		{"1", "1.0.0.0"},
-		{"1", "1.255.255.255"},
-		{"2", "2.1.0.0"},
-		{"2", "2.1.255.255"},
-		{"3", "3.1.1.0"},
-		{"3", "3.1.1.255"},
-		{"4a", "4.1.1.255"},
-		{"4a", "4.1.1.1"},
-		{"5", "240.0.0.0"},
-		{"5", "255.255.255.255"},
-		{nil, "239.0.0.0"},
-		{nil, "4.1.2.2"},
-	}
-
-	for _, tt := range tests {
-		assert.Equal(t, tt.Result, tree.Contains(ip2int(net.ParseIP(tt.IP))))
-	}
-
-	tree = NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.1.1.1/0"), "cool")
-	assert.Equal(t, "cool", tree.Contains(ip2int(net.ParseIP("0.0.0.0"))))
-	assert.Equal(t, "cool", tree.Contains(ip2int(net.ParseIP("255.255.255.255"))))
-}
-
-func TestCIDRTree_MostSpecificContains(t *testing.T) {
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.0.0.0/8"), "1")
-	tree.AddCIDR(getCIDR("2.1.0.0/16"), "2")
-	tree.AddCIDR(getCIDR("3.1.1.0/24"), "3")
-	tree.AddCIDR(getCIDR("4.1.1.0/24"), "4a")
-	tree.AddCIDR(getCIDR("4.1.1.0/30"), "4b")
-	tree.AddCIDR(getCIDR("4.1.1.1/32"), "4c")
-	tree.AddCIDR(getCIDR("254.0.0.0/4"), "5")
-
-	tests := []struct {
-		Result interface{}
-		IP     string
-	}{
-		{"1", "1.0.0.0"},
-		{"1", "1.255.255.255"},
-		{"2", "2.1.0.0"},
-		{"2", "2.1.255.255"},
-		{"3", "3.1.1.0"},
-		{"3", "3.1.1.255"},
-		{"4a", "4.1.1.255"},
-		{"4b", "4.1.1.2"},
-		{"4c", "4.1.1.1"},
-		{"5", "240.0.0.0"},
-		{"5", "255.255.255.255"},
-		{nil, "239.0.0.0"},
-		{nil, "4.1.2.2"},
-	}
-
-	for _, tt := range tests {
-		assert.Equal(t, tt.Result, tree.MostSpecificContains(ip2int(net.ParseIP(tt.IP))))
-	}
-
-	tree = NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.1.1.1/0"), "cool")
-	assert.Equal(t, "cool", tree.MostSpecificContains(ip2int(net.ParseIP("0.0.0.0"))))
-	assert.Equal(t, "cool", tree.MostSpecificContains(ip2int(net.ParseIP("255.255.255.255"))))
-}
-
-func TestCIDRTree_Match(t *testing.T) {
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("4.1.1.0/32"), "1a")
-	tree.AddCIDR(getCIDR("4.1.1.1/32"), "1b")
-
-	tests := []struct {
-		Result interface{}
-		IP     string
-	}{
-		{"1a", "4.1.1.0"},
-		{"1b", "4.1.1.1"},
-	}
-
-	for _, tt := range tests {
-		assert.Equal(t, tt.Result, tree.Match(ip2int(net.ParseIP(tt.IP))))
-	}
-
-	tree = NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.1.1.1/0"), "cool")
-	assert.Equal(t, "cool", tree.Contains(ip2int(net.ParseIP("0.0.0.0"))))
-	assert.Equal(t, "cool", tree.Contains(ip2int(net.ParseIP("255.255.255.255"))))
-}
-
-func BenchmarkCIDRTree_Contains(b *testing.B) {
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.1.0.0/16"), "1")
-	tree.AddCIDR(getCIDR("1.2.1.1/32"), "1")
-	tree.AddCIDR(getCIDR("192.2.1.1/32"), "1")
-	tree.AddCIDR(getCIDR("172.2.1.1/32"), "1")
-
-	ip := ip2int(net.ParseIP("1.2.1.1"))
-	b.Run("found", func(b *testing.B) {
-		for i := 0; i < b.N; i++ {
-			tree.Contains(ip)
-		}
-	})
-
-	ip = ip2int(net.ParseIP("1.2.1.255"))
-	b.Run("not found", func(b *testing.B) {
-		for i := 0; i < b.N; i++ {
-			tree.Contains(ip)
-		}
-	})
-}
-
-func BenchmarkCIDRTree_Match(b *testing.B) {
-	tree := NewCIDRTree()
-	tree.AddCIDR(getCIDR("1.1.0.0/16"), "1")
-	tree.AddCIDR(getCIDR("1.2.1.1/32"), "1")
-	tree.AddCIDR(getCIDR("192.2.1.1/32"), "1")
-	tree.AddCIDR(getCIDR("172.2.1.1/32"), "1")
-
-	ip := ip2int(net.ParseIP("1.2.1.1"))
-	b.Run("found", func(b *testing.B) {
-		for i := 0; i < b.N; i++ {
-			tree.Match(ip)
-		}
-	})
-
-	ip = ip2int(net.ParseIP("1.2.1.255"))
-	b.Run("not found", func(b *testing.B) {
-		for i := 0; i < b.N; i++ {
-			tree.Match(ip)
-		}
-	})
-}
-
-func getCIDR(s string) *net.IPNet {
-	_, c, _ := net.ParseCIDR(s)
-	return c
-}

cmd/nebula-cert/ca.go

@@ -1,29 +1,39 @@
 package main
 
 import (
+	"crypto/ecdsa"
+	"crypto/elliptic"
 	"crypto/rand"
 	"flag"
 	"fmt"
 	"io"
-	"io/ioutil"
+	"math"
 	"net"
 	"os"
 	"strings"
 	"time"
 
+	"github.com/skip2/go-qrcode"
 	"github.com/slackhq/nebula/cert"
 	"golang.org/x/crypto/ed25519"
 )
 
 type caFlags struct {
-	set         *flag.FlagSet
-	name        *string
-	duration    *time.Duration
-	outKeyPath  *string
-	outCertPath *string
-	groups      *string
-	ips         *string
-	subnets     *string
+	set              *flag.FlagSet
+	name             *string
+	duration         *time.Duration
+	outKeyPath       *string
+	outCertPath      *string
+	outQRPath        *string
+	groups           *string
+	ips              *string
+	subnets          *string
+	argonMemory      *uint
+	argonIterations  *uint
+	argonParallelism *uint
+	encryption       *bool
+
+	curve *string
 }
 
 func newCaFlags() *caFlags {
@@ -33,13 +43,33 @@ func newCaFlags() *caFlags {
 	cf.duration = cf.set.Duration("duration", time.Duration(time.Hour*8760), "Optional: amount of time the certificate should be valid for. Valid time units are seconds: \"s\", minutes: \"m\", hours: \"h\"")
 	cf.outKeyPath = cf.set.String("out-key", "ca.key", "Optional: path to write the private key to")
 	cf.outCertPath = cf.set.String("out-crt", "ca.crt", "Optional: path to write the certificate to")
+	cf.outQRPath = cf.set.String("out-qr", "", "Optional: output a qr code image (png) of the certificate")
 	cf.groups = cf.set.String("groups", "", "Optional: comma separated list of groups. This will limit which groups subordinate certs can use")
-	cf.ips = cf.set.String("ips", "", "Optional: comma separated list of ip and network in CIDR notation. This will limit which ip addresses and networks subordinate certs can use")
-	cf.subnets = cf.set.String("subnets", "", "Optional: comma separated list of ip and network in CIDR notation. This will limit which subnet addresses and networks subordinate certs can use")
+	cf.ips = cf.set.String("ips", "", "Optional: comma separated list of ipv4 address and network in CIDR notation. This will limit which ipv4 addresses and networks subordinate certs can use for ip addresses")
+	cf.subnets = cf.set.String("subnets", "", "Optional: comma separated list of ipv4 address and network in CIDR notation. This will limit which ipv4 addresses and networks subordinate certs can use in subnets")
+	cf.argonMemory = cf.set.Uint("argon-memory", 2*1024*1024, "Optional: Argon2 memory parameter (in KiB) used for encrypted private key passphrase")
+	cf.argonParallelism = cf.set.Uint("argon-parallelism", 4, "Optional: Argon2 parallelism parameter used for encrypted private key passphrase")
+	cf.argonIterations = cf.set.Uint("argon-iterations", 1, "Optional: Argon2 iterations parameter used for encrypted private key passphrase")
+	cf.encryption = cf.set.Bool("encrypt", false, "Optional: prompt for passphrase and write out-key in an encrypted format")
+	cf.curve = cf.set.String("curve", "25519", "EdDSA/ECDSA Curve (25519, P256)")
 	return &cf
 }
 
-func ca(args []string, out io.Writer, errOut io.Writer) error {
+func parseArgonParameters(memory uint, parallelism uint, iterations uint) (*cert.Argon2Parameters, error) {
+	if memory <= 0 || memory > math.MaxUint32 {
+		return nil, newHelpErrorf("-argon-memory must be be greater than 0 and no more than %d KiB", uint32(math.MaxUint32))
+	}
+	if parallelism <= 0 || parallelism > math.MaxUint8 {
+		return nil, newHelpErrorf("-argon-parallelism must be be greater than 0 and no more than %d", math.MaxUint8)
+	}
+	if iterations <= 0 || iterations > math.MaxUint32 {
+		return nil, newHelpErrorf("-argon-iterations must be be greater than 0 and no more than %d", uint32(math.MaxUint32))
+	}
+
+	return cert.NewArgon2Parameters(uint32(memory), uint8(parallelism), uint32(iterations)), nil
+}
+
+func ca(args []string, out io.Writer, errOut io.Writer, pr PasswordReader) error {
 	cf := newCaFlags()
 	err := cf.set.Parse(args)
 	if err != nil {
@@ -55,6 +85,12 @@ func ca(args []string, out io.Writer, errOut io.Writer) error {
 	if err := mustFlagString("out-crt", cf.outCertPath); err != nil {
 		return err
 	}
+	var kdfParams *cert.Argon2Parameters
+	if *cf.encryption {
+		if kdfParams, err = parseArgonParameters(*cf.argonMemory, *cf.argonParallelism, *cf.argonIterations); err != nil {
+			return err
+		}
+	}
 
 	if *cf.duration <= 0 {
 		return &helpError{"-duration must be greater than 0"}
@@ -79,6 +115,9 @@ func ca(args []string, out io.Writer, errOut io.Writer) error {
 				if err != nil {
 					return newHelpErrorf("invalid ip definition: %s", err)
 				}
+				if ip.To4() == nil {
+					return newHelpErrorf("invalid ip definition: can only be ipv4, have %s", rs)
+				}
 
 				ipNet.IP = ip
 				ips = append(ips, ipNet)
@@ -95,14 +134,61 @@ func ca(args []string, out io.Writer, errOut io.Writer) error {
 				if err != nil {
 					return newHelpErrorf("invalid subnet definition: %s", err)
 				}
+				if s.IP.To4() == nil {
+					return newHelpErrorf("invalid subnet definition: can only be ipv4, have %s", rs)
+				}
 				subnets = append(subnets, s)
 			}
 		}
 	}
 
-	pub, rawPriv, err := ed25519.GenerateKey(rand.Reader)
-	if err != nil {
-		return fmt.Errorf("error while generating ed25519 keys: %s", err)
+	var passphrase []byte
+	if *cf.encryption {
+		for i := 0; i < 5; i++ {
+			out.Write([]byte("Enter passphrase: "))
+			passphrase, err = pr.ReadPassword()
+
+			if err == ErrNoTerminal {
+				return fmt.Errorf("out-key must be encrypted interactively")
+			} else if err != nil {
+				return fmt.Errorf("error reading passphrase: %s", err)
+			}
+
+			if len(passphrase) > 0 {
+				break
+			}
+		}
+
+		if len(passphrase) == 0 {
+			return fmt.Errorf("no passphrase specified, remove -encrypt flag to write out-key in plaintext")
+		}
+	}
+
+	var curve cert.Curve
+	var pub, rawPriv []byte
+	switch *cf.curve {
+	case "25519", "X25519", "Curve25519", "CURVE25519":
+		curve = cert.Curve_CURVE25519
+		pub, rawPriv, err = ed25519.GenerateKey(rand.Reader)
+		if err != nil {
+			return fmt.Errorf("error while generating ed25519 keys: %s", err)
+		}
+	case "P256":
+		var key *ecdsa.PrivateKey
+		curve = cert.Curve_P256
+		key, err = ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+		if err != nil {
+			return fmt.Errorf("error while generating ecdsa keys: %s", err)
+		}
+
+		// ecdh.PrivateKey lets us get at the encoded bytes, even though
+		// we aren't using ECDH here.
+		eKey, err := key.ECDH()
+		if err != nil {
+			return fmt.Errorf("error while converting ecdsa key: %s", err)
+		}
+		rawPriv = eKey.Bytes()
+		pub = eKey.PublicKey().Bytes()
 	}
 
 	nc := cert.NebulaCertificate{
@@ -115,6 +201,7 @@ func ca(args []string, out io.Writer, errOut io.Writer) error {
 			NotAfter:  time.Now().Add(*cf.duration),
 			PublicKey: pub,
 			IsCA:      true,
+			Curve:     curve,
 		},
 	}
 
@@ -126,26 +213,48 @@ func ca(args []string, out io.Writer, errOut io.Writer) error {
 		return fmt.Errorf("refusing to overwrite existing CA cert: %s", *cf.outCertPath)
 	}
 
-	err = nc.Sign(rawPriv)
+	err = nc.Sign(curve, rawPriv)
 	if err != nil {
 		return fmt.Errorf("error while signing: %s", err)
 	}
 
-	err = ioutil.WriteFile(*cf.outKeyPath, cert.MarshalEd25519PrivateKey(rawPriv), 0600)
+	var b []byte
+	if *cf.encryption {
+		b, err = cert.EncryptAndMarshalSigningPrivateKey(curve, rawPriv, passphrase, kdfParams)
+		if err != nil {
+			return fmt.Errorf("error while encrypting out-key: %s", err)
+		}
+	} else {
+		b = cert.MarshalSigningPrivateKey(curve, rawPriv)
+	}
+
+	err = os.WriteFile(*cf.outKeyPath, b, 0600)
 	if err != nil {
 		return fmt.Errorf("error while writing out-key: %s", err)
 	}
 
-	b, err := nc.MarshalToPEM()
+	b, err = nc.MarshalToPEM()
 	if err != nil {
 		return fmt.Errorf("error while marshalling certificate: %s", err)
 	}
 
-	err = ioutil.WriteFile(*cf.outCertPath, b, 0600)
+	err = os.WriteFile(*cf.outCertPath, b, 0600)
 	if err != nil {
 		return fmt.Errorf("error while writing out-crt: %s", err)
 	}
 
+	if *cf.outQRPath != "" {
+		b, err = qrcode.Encode(string(b), qrcode.Medium, -5)
+		if err != nil {
+			return fmt.Errorf("error while generating qr code: %s", err)
+		}
+
+		err = os.WriteFile(*cf.outQRPath, b, 0600)
+		if err != nil {
+			return fmt.Errorf("error while writing out-qr: %s", err)
+		}
+	}
+
 	return nil
 }
 

cmd/nebula-cert/ca_test.go

@@ -1,11 +1,14 @@
+//go:build !windows
 // +build !windows
 
 package main
 
 import (
 	"bytes"
-	"io/ioutil"
+	"encoding/pem"
+	"errors"
 	"os"
+	"strings"
 	"testing"
 	"time"
 
@@ -25,20 +28,32 @@ func Test_caHelp(t *testing.T) {
 	assert.Equal(
 		t,
 		"Usage of "+os.Args[0]+" ca <flags>: create a self signed certificate authority\n"+
+			"  -argon-iterations uint\n"+
+			"    \tOptional: Argon2 iterations parameter used for encrypted private key passphrase (default 1)\n"+
+			"  -argon-memory uint\n"+
+			"    \tOptional: Argon2 memory parameter (in KiB) used for encrypted private key passphrase (default 2097152)\n"+
+			"  -argon-parallelism uint\n"+
+			"    \tOptional: Argon2 parallelism parameter used for encrypted private key passphrase (default 4)\n"+
+			"  -curve string\n"+
+			"    \tEdDSA/ECDSA Curve (25519, P256) (default \"25519\")\n"+
 			"  -duration duration\n"+
 			"    \tOptional: amount of time the certificate should be valid for. Valid time units are seconds: \"s\", minutes: \"m\", hours: \"h\" (default 8760h0m0s)\n"+
+			"  -encrypt\n"+
+			"    \tOptional: prompt for passphrase and write out-key in an encrypted format\n"+
 			"  -groups string\n"+
 			"    \tOptional: comma separated list of groups. This will limit which groups subordinate certs can use\n"+
 			"  -ips string\n"+
-			"    \tOptional: comma separated list of ip and network in CIDR notation. This will limit which ip addresses and networks subordinate certs can use\n"+
+			"    \tOptional: comma separated list of ipv4 address and network in CIDR notation. This will limit which ipv4 addresses and networks subordinate certs can use for ip addresses\n"+
 			"  -name string\n"+
 			"    \tRequired: name of the certificate authority\n"+
 			"  -out-crt string\n"+
 			"    \tOptional: path to write the certificate to (default \"ca.crt\")\n"+
 			"  -out-key string\n"+
 			"    \tOptional: path to write the private key to (default \"ca.key\")\n"+
+			"  -out-qr string\n"+
+			"    \tOptional: output a qr code image (png) of the certificate\n"+
 			"  -subnets string\n"+
-			"    \tOptional: comma separated list of ip and network in CIDR notation. This will limit which subnet addresses and networks subordinate certs can use\n",
+			"    \tOptional: comma separated list of ipv4 address and network in CIDR notation. This will limit which ipv4 addresses and networks subordinate certs can use in subnets\n",
 		ob.String(),
 	)
 }
@@ -47,8 +62,38 @@ func Test_ca(t *testing.T) {
 	ob := &bytes.Buffer{}
 	eb := &bytes.Buffer{}
 
+	nopw := &StubPasswordReader{
+		password: []byte(""),
+		err:      nil,
+	}
+
+	errpw := &StubPasswordReader{
+		password: []byte(""),
+		err:      errors.New("stub error"),
+	}
+
+	passphrase := []byte("DO NOT USE THIS KEY")
+	testpw := &StubPasswordReader{
+		password: passphrase,
+		err:      nil,
+	}
+
+	pwPromptOb := "Enter passphrase: "
+
 	// required args
-	assertHelpError(t, ca([]string{"-out-key", "nope", "-out-crt", "nope", "duration", "100m"}, ob, eb), "-name is required")
+	assertHelpError(t, ca(
+		[]string{"-out-key", "nope", "-out-crt", "nope", "duration", "100m"}, ob, eb, nopw,
+	), "-name is required")
+	assert.Equal(t, "", ob.String())
+	assert.Equal(t, "", eb.String())
+
+	// ipv4 only ips
+	assertHelpError(t, ca([]string{"-name", "ipv6", "-ips", "100::100/100"}, ob, eb, nopw), "invalid ip definition: can only be ipv4, have 100::100/100")
+	assert.Equal(t, "", ob.String())
+	assert.Equal(t, "", eb.String())
+
+	// ipv4 only subnets
+	assertHelpError(t, ca([]string{"-name", "ipv6", "-subnets", "100::100/100"}, ob, eb, nopw), "invalid subnet definition: can only be ipv4, have 100::100/100")
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 
@@ -56,12 +101,12 @@ func Test_ca(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args := []string{"-name", "test", "-duration", "100m", "-out-crt", "/do/not/write/pleasecrt", "-out-key", "/do/not/write/pleasekey"}
-	assert.EqualError(t, ca(args, ob, eb), "error while writing out-key: open /do/not/write/pleasekey: "+NoSuchDirError)
+	assert.EqualError(t, ca(args, ob, eb, nopw), "error while writing out-key: open /do/not/write/pleasekey: "+NoSuchDirError)
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 
 	// create temp key file
-	keyF, err := ioutil.TempFile("", "test.key")
+	keyF, err := os.CreateTemp("", "test.key")
 	assert.Nil(t, err)
 	os.Remove(keyF.Name())
 
@@ -69,12 +114,12 @@ func Test_ca(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-name", "test", "-duration", "100m", "-out-crt", "/do/not/write/pleasecrt", "-out-key", keyF.Name()}
-	assert.EqualError(t, ca(args, ob, eb), "error while writing out-crt: open /do/not/write/pleasecrt: "+NoSuchDirError)
+	assert.EqualError(t, ca(args, ob, eb, nopw), "error while writing out-crt: open /do/not/write/pleasecrt: "+NoSuchDirError)
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 
 	// create temp cert file
-	crtF, err := ioutil.TempFile("", "test.crt")
+	crtF, err := os.CreateTemp("", "test.crt")
 	assert.Nil(t, err)
 	os.Remove(crtF.Name())
 	os.Remove(keyF.Name())
@@ -83,18 +128,18 @@ func Test_ca(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-name", "test", "-duration", "100m", "-groups", "1,,   2    ,        ,,,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
-	assert.Nil(t, ca(args, ob, eb))
+	assert.Nil(t, ca(args, ob, eb, nopw))
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 
 	// read cert and key files
-	rb, _ := ioutil.ReadFile(keyF.Name())
+	rb, _ := os.ReadFile(keyF.Name())
 	lKey, b, err := cert.UnmarshalEd25519PrivateKey(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
 	assert.Len(t, lKey, 64)
 
-	rb, _ = ioutil.ReadFile(crtF.Name())
+	rb, _ = os.ReadFile(crtF.Name())
 	lCrt, b, err := cert.UnmarshalNebulaCertificateFromPEM(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
@@ -109,19 +154,67 @@ func Test_ca(t *testing.T) {
 	assert.Equal(t, "", lCrt.Details.Issuer)
 	assert.True(t, lCrt.CheckSignature(lCrt.Details.PublicKey))
 
+	// test encrypted key
+	os.Remove(keyF.Name())
+	os.Remove(crtF.Name())
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-encrypt", "-name", "test", "-duration", "100m", "-groups", "1,2,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
+	assert.Nil(t, ca(args, ob, eb, testpw))
+	assert.Equal(t, pwPromptOb, ob.String())
+	assert.Equal(t, "", eb.String())
+
+	// read encrypted key file and verify default params
+	rb, _ = os.ReadFile(keyF.Name())
+	k, _ := pem.Decode(rb)
+	ned, err := cert.UnmarshalNebulaEncryptedData(k.Bytes)
+	assert.Nil(t, err)
+	// we won't know salt in advance, so just check start of string
+	assert.Equal(t, uint32(2*1024*1024), ned.EncryptionMetadata.Argon2Parameters.Memory)
+	assert.Equal(t, uint8(4), ned.EncryptionMetadata.Argon2Parameters.Parallelism)
+	assert.Equal(t, uint32(1), ned.EncryptionMetadata.Argon2Parameters.Iterations)
+
+	// verify the key is valid and decrypt-able
+	var curve cert.Curve
+	curve, lKey, b, err = cert.DecryptAndUnmarshalSigningPrivateKey(passphrase, rb)
+	assert.Equal(t, cert.Curve_CURVE25519, curve)
+	assert.Nil(t, err)
+	assert.Len(t, b, 0)
+	assert.Len(t, lKey, 64)
+
+	// test when reading passsword results in an error
+	os.Remove(keyF.Name())
+	os.Remove(crtF.Name())
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-encrypt", "-name", "test", "-duration", "100m", "-groups", "1,2,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
+	assert.Error(t, ca(args, ob, eb, errpw))
+	assert.Equal(t, pwPromptOb, ob.String())
+	assert.Equal(t, "", eb.String())
+
+	// test when user fails to enter a password
+	os.Remove(keyF.Name())
+	os.Remove(crtF.Name())
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-encrypt", "-name", "test", "-duration", "100m", "-groups", "1,2,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
+	assert.EqualError(t, ca(args, ob, eb, nopw), "no passphrase specified, remove -encrypt flag to write out-key in plaintext")
+	assert.Equal(t, strings.Repeat(pwPromptOb, 5), ob.String()) // prompts 5 times before giving up
+	assert.Equal(t, "", eb.String())
+
 	// create valid cert/key for overwrite tests
 	os.Remove(keyF.Name())
 	os.Remove(crtF.Name())
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-name", "test", "-duration", "100m", "-groups", "1,,   2    ,        ,,,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
-	assert.Nil(t, ca(args, ob, eb))
+	assert.Nil(t, ca(args, ob, eb, nopw))
 
 	// test that we won't overwrite existing certificate file
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-name", "test", "-duration", "100m", "-groups", "1,,   2    ,        ,,,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
-	assert.EqualError(t, ca(args, ob, eb), "refusing to overwrite existing CA key: "+keyF.Name())
+	assert.EqualError(t, ca(args, ob, eb, nopw), "refusing to overwrite existing CA key: "+keyF.Name())
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 
@@ -130,7 +223,7 @@ func Test_ca(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-name", "test", "-duration", "100m", "-groups", "1,,   2    ,        ,,,3,4,5", "-out-crt", crtF.Name(), "-out-key", keyF.Name()}
-	assert.EqualError(t, ca(args, ob, eb), "refusing to overwrite existing CA cert: "+crtF.Name())
+	assert.EqualError(t, ca(args, ob, eb, nopw), "refusing to overwrite existing CA cert: "+crtF.Name())
 	assert.Equal(t, "", ob.String())
 	assert.Equal(t, "", eb.String())
 	os.Remove(keyF.Name())

cmd/nebula-cert/keygen.go

@@ -4,7 +4,6 @@ import (
 	"flag"
 	"fmt"
 	"io"
-	"io/ioutil"
 	"os"
 
 	"github.com/slackhq/nebula/cert"
@@ -14,6 +13,8 @@ type keygenFlags struct {
 	set        *flag.FlagSet
 	outKeyPath *string
 	outPubPath *string
+
+	curve *string
 }
 
 func newKeygenFlags() *keygenFlags {
@@ -21,6 +22,7 @@ func newKeygenFlags() *keygenFlags {
 	cf.set.Usage = func() {}
 	cf.outPubPath = cf.set.String("out-pub", "", "Required: path to write the public key to")
 	cf.outKeyPath = cf.set.String("out-key", "", "Required: path to write the private key to")
+	cf.curve = cf.set.String("curve", "25519", "ECDH Curve (25519, P256)")
 	return &cf
 }
 
@@ -38,14 +40,25 @@ func keygen(args []string, out io.Writer, errOut io.Writer) error {
 		return err
 	}
 
-	pub, rawPriv := x25519Keypair()
+	var pub, rawPriv []byte
+	var curve cert.Curve
+	switch *cf.curve {
+	case "25519", "X25519", "Curve25519", "CURVE25519":
+		pub, rawPriv = x25519Keypair()
+		curve = cert.Curve_CURVE25519
+	case "P256":
+		pub, rawPriv = p256Keypair()
+		curve = cert.Curve_P256
+	default:
+		return fmt.Errorf("invalid curve: %s", *cf.curve)
+	}
 
-	err = ioutil.WriteFile(*cf.outKeyPath, cert.MarshalX25519PrivateKey(rawPriv), 0600)
+	err = os.WriteFile(*cf.outKeyPath, cert.MarshalPrivateKey(curve, rawPriv), 0600)
 	if err != nil {
 		return fmt.Errorf("error while writing out-key: %s", err)
 	}
 
-	err = ioutil.WriteFile(*cf.outPubPath, cert.MarshalX25519PublicKey(pub), 0600)
+	err = os.WriteFile(*cf.outPubPath, cert.MarshalPublicKey(curve, pub), 0600)
 	if err != nil {
 		return fmt.Errorf("error while writing out-pub: %s", err)
 	}

cmd/nebula-cert/keygen_test.go

@@ -2,7 +2,6 @@ package main
 
 import (
 	"bytes"
-	"io/ioutil"
 	"os"
 	"testing"
 
@@ -22,6 +21,8 @@ func Test_keygenHelp(t *testing.T) {
 	assert.Equal(
 		t,
 		"Usage of "+os.Args[0]+" keygen <flags>: create a public/private key pair. the public key can be passed to `nebula-cert sign`\n"+
+			"  -curve string\n"+
+			"    \tECDH Curve (25519, P256) (default \"25519\")\n"+
 			"  -out-key string\n"+
 			"    \tRequired: path to write the private key to\n"+
 			"  -out-pub string\n"+
@@ -52,7 +53,7 @@ func Test_keygen(t *testing.T) {
 	assert.Equal(t, "", eb.String())
 
 	// create temp key file
-	keyF, err := ioutil.TempFile("", "test.key")
+	keyF, err := os.CreateTemp("", "test.key")
 	assert.Nil(t, err)
 	defer os.Remove(keyF.Name())
 
@@ -65,7 +66,7 @@ func Test_keygen(t *testing.T) {
 	assert.Equal(t, "", eb.String())
 
 	// create temp pub file
-	pubF, err := ioutil.TempFile("", "test.pub")
+	pubF, err := os.CreateTemp("", "test.pub")
 	assert.Nil(t, err)
 	defer os.Remove(pubF.Name())
 
@@ -78,13 +79,13 @@ func Test_keygen(t *testing.T) {
 	assert.Equal(t, "", eb.String())
 
 	// read cert and key files
-	rb, _ := ioutil.ReadFile(keyF.Name())
+	rb, _ := os.ReadFile(keyF.Name())
 	lKey, b, err := cert.UnmarshalX25519PrivateKey(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
 	assert.Len(t, lKey, 32)
 
-	rb, _ = ioutil.ReadFile(pubF.Name())
+	rb, _ = os.ReadFile(pubF.Name())
 	lPub, b, err := cert.UnmarshalX25519PublicKey(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)

cmd/nebula-cert/main.go

@@ -62,11 +62,11 @@ func main() {
 
 	switch args[0] {
 	case "ca":
-		err = ca(args[1:], os.Stdout, os.Stderr)
+		err = ca(args[1:], os.Stdout, os.Stderr, StdinPasswordReader{})
 	case "keygen":
 		err = keygen(args[1:], os.Stdout, os.Stderr)
 	case "sign":
-		err = signCert(args[1:], os.Stdout, os.Stderr)
+		err = signCert(args[1:], os.Stdout, os.Stderr, StdinPasswordReader{})
 	case "print":
 		err = printCert(args[1:], os.Stdout, os.Stderr)
 	case "verify":
@@ -127,6 +127,8 @@ func help(err string, out io.Writer) {
 	fmt.Fprintln(out, "    "+signSummary())
 	fmt.Fprintln(out, "    "+printSummary())
 	fmt.Fprintln(out, "    "+verifySummary())
+	fmt.Fprintln(out, "")
+	fmt.Fprintf(out, "  To see usage for a given mode, use %s <mode> -h\n", os.Args[0])
 }
 
 func mustFlagString(name string, val *string) error {

cmd/nebula-cert/main_test.go

@@ -3,10 +3,11 @@ package main
 import (
 	"bytes"
 	"errors"
-	"github.com/stretchr/testify/assert"
 	"io"
 	"os"
 	"testing"
+
+	"github.com/stretchr/testify/assert"
 )
 
 //TODO: all flag parsing continueOnError will print to stderr on its own currently
@@ -21,7 +22,9 @@ func Test_help(t *testing.T) {
 		"    " + keygenSummary() + "\n" +
 		"    " + signSummary() + "\n" +
 		"    " + printSummary() + "\n" +
-		"    " + verifySummary() + "\n"
+		"    " + verifySummary() + "\n" +
+		"\n" +
+		"  To see usage for a given mode, use " + os.Args[0] + " <mode> -h\n"
 
 	ob := &bytes.Buffer{}
 

cmd/nebula-cert/passwords.go

@@ -0,0 +1,28 @@
+package main
+
+import (
+	"errors"
+	"fmt"
+	"os"
+
+	"golang.org/x/term"
+)
+
+var ErrNoTerminal = errors.New("cannot read password from nonexistent terminal")
+
+type PasswordReader interface {
+	ReadPassword() ([]byte, error)
+}
+
+type StdinPasswordReader struct{}
+
+func (pr StdinPasswordReader) ReadPassword() ([]byte, error) {
+	if !term.IsTerminal(int(os.Stdin.Fd())) {
+		return nil, ErrNoTerminal
+	}
+
+	password, err := term.ReadPassword(int(os.Stdin.Fd()))
+	fmt.Println()
+
+	return password, err
+}

cmd/nebula-cert/passwords_test.go

@@ -0,0 +1,10 @@
+package main
+
+type StubPasswordReader struct {
+	password []byte
+	err      error
+}
+
+func (pr *StubPasswordReader) ReadPassword() ([]byte, error) {
+	return pr.password, pr.err
+}

cmd/nebula-cert/print.go

@@ -4,23 +4,26 @@ import (
 	"encoding/json"
 	"flag"
 	"fmt"
-	"github.com/slackhq/nebula/cert"
 	"io"
-	"io/ioutil"
 	"os"
 	"strings"
+
+	"github.com/skip2/go-qrcode"
+	"github.com/slackhq/nebula/cert"
 )
 
 type printFlags struct {
-	set  *flag.FlagSet
-	json *bool
-	path *string
+	set       *flag.FlagSet
+	json      *bool
+	outQRPath *string
+	path      *string
 }
 
 func newPrintFlags() *printFlags {
 	pf := printFlags{set: flag.NewFlagSet("print", flag.ContinueOnError)}
 	pf.set.Usage = func() {}
 	pf.json = pf.set.Bool("json", false, "Optional: outputs certificates in json format")
+	pf.outQRPath = pf.set.String("out-qr", "", "Optional: output a qr code image (png) of the certificate")
 	pf.path = pf.set.String("path", "", "Required: path to the certificate")
 
 	return &pf
@@ -37,12 +40,14 @@ func printCert(args []string, out io.Writer, errOut io.Writer) error {
 		return err
 	}
 
-	rawCert, err := ioutil.ReadFile(*pf.path)
+	rawCert, err := os.ReadFile(*pf.path)
 	if err != nil {
 		return fmt.Errorf("unable to read cert; %s", err)
 	}
 
 	var c *cert.NebulaCertificate
+	var qrBytes []byte
+	part := 0
 
 	for {
 		c, rawCert, err = cert.UnmarshalNebulaCertificateFromPEM(rawCert)
@@ -60,9 +65,31 @@ func printCert(args []string, out io.Writer, errOut io.Writer) error {
 			out.Write([]byte("\n"))
 		}
 
+		if *pf.outQRPath != "" {
+			b, err := c.MarshalToPEM()
+			if err != nil {
+				return fmt.Errorf("error while marshalling cert to PEM: %s", err)
+			}
+			qrBytes = append(qrBytes, b...)
+		}
+
 		if rawCert == nil || len(rawCert) == 0 || strings.TrimSpace(string(rawCert)) == "" {
 			break
 		}
+
+		part++
+	}
+
+	if *pf.outQRPath != "" {
+		b, err := qrcode.Encode(string(qrBytes), qrcode.Medium, -5)
+		if err != nil {
+			return fmt.Errorf("error while generating qr code: %s", err)
+		}
+
+		err = os.WriteFile(*pf.outQRPath, b, 0600)
+		if err != nil {
+			return fmt.Errorf("error while writing out-qr: %s", err)
+		}
 	}
 
 	return nil

cmd/nebula-cert/print_test.go

@@ -2,12 +2,12 @@ package main
 
 import (
 	"bytes"
-	"github.com/slackhq/nebula/cert"
-	"github.com/stretchr/testify/assert"
-	"io/ioutil"
 	"os"
 	"testing"
 	"time"
+
+	"github.com/slackhq/nebula/cert"
+	"github.com/stretchr/testify/assert"
 )
 
 func Test_printSummary(t *testing.T) {
@@ -22,6 +22,8 @@ func Test_printHelp(t *testing.T) {
 		"Usage of "+os.Args[0]+" print <flags>: prints details about a certificate\n"+
 			"  -json\n"+
 			"    \tOptional: outputs certificates in json format\n"+
+			"  -out-qr string\n"+
+			"    \tOptional: output a qr code image (png) of the certificate\n"+
 			"  -path string\n"+
 			"    \tRequired: path to the certificate\n",
 		ob.String(),
@@ -51,7 +53,7 @@ func Test_printCert(t *testing.T) {
 	// invalid cert at path
 	ob.Reset()
 	eb.Reset()
-	tf, err := ioutil.TempFile("", "print-cert")
+	tf, err := os.CreateTemp("", "print-cert")
 	assert.Nil(t, err)
 	defer os.Remove(tf.Name())
 
@@ -84,7 +86,7 @@ func Test_printCert(t *testing.T) {
 	assert.Nil(t, err)
 	assert.Equal(
 		t,
-		"NebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\nNebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\nNebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\n",
+		"NebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t\tCurve: CURVE25519\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\nNebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t\tCurve: CURVE25519\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\nNebulaCertificate {\n\tDetails {\n\t\tName: test\n\t\tIps: []\n\t\tSubnets: []\n\t\tGroups: [\n\t\t\t\"hi\"\n\t\t]\n\t\tNot before: 0001-01-01 00:00:00 +0000 UTC\n\t\tNot After: 0001-01-01 00:00:00 +0000 UTC\n\t\tIs CA: false\n\t\tIssuer: \n\t\tPublic key: 0102030405060708090001020304050607080900010203040506070809000102\n\t\tCurve: CURVE25519\n\t}\n\tFingerprint: cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\n\tSignature: 0102030405060708090001020304050607080900010203040506070809000102\n}\n",
 		ob.String(),
 	)
 	assert.Equal(t, "", eb.String())
@@ -112,7 +114,7 @@ func Test_printCert(t *testing.T) {
 	assert.Nil(t, err)
 	assert.Equal(
 		t,
-		"{\"details\":{\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n{\"details\":{\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n{\"details\":{\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n",
+		"{\"details\":{\"curve\":\"CURVE25519\",\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n{\"details\":{\"curve\":\"CURVE25519\",\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n{\"details\":{\"curve\":\"CURVE25519\",\"groups\":[\"hi\"],\"ips\":[],\"isCa\":false,\"issuer\":\"\",\"name\":\"test\",\"notAfter\":\"0001-01-01T00:00:00Z\",\"notBefore\":\"0001-01-01T00:00:00Z\",\"publicKey\":\"0102030405060708090001020304050607080900010203040506070809000102\",\"subnets\":[]},\"fingerprint\":\"cc3492c0e9c48f17547f5987ea807462ebb3451e622590a10bb3763c344c82bd\",\"signature\":\"0102030405060708090001020304050607080900010203040506070809000102\"}\n",
 		ob.String(),
 	)
 	assert.Equal(t, "", eb.String())

cmd/nebula-cert/sign.go

@@ -1,16 +1,17 @@
 package main
 
 import (
+	"crypto/ecdh"
 	"crypto/rand"
 	"flag"
 	"fmt"
 	"io"
-	"io/ioutil"
 	"net"
 	"os"
 	"strings"
 	"time"
 
+	"github.com/skip2/go-qrcode"
 	"github.com/slackhq/nebula/cert"
 	"golang.org/x/crypto/curve25519"
 )
@@ -25,6 +26,7 @@ type signFlags struct {
 	inPubPath   *string
 	outKeyPath  *string
 	outCertPath *string
+	outQRPath   *string
 	groups      *string
 	subnets     *string
 }
@@ -35,18 +37,19 @@ func newSignFlags() *signFlags {
 	sf.caKeyPath = sf.set.String("ca-key", "ca.key", "Optional: path to the signing CA key")
 	sf.caCertPath = sf.set.String("ca-crt", "ca.crt", "Optional: path to the signing CA cert")
 	sf.name = sf.set.String("name", "", "Required: name of the cert, usually a hostname")
-	sf.ip = sf.set.String("ip", "", "Required: ip and network in CIDR notation to assign the cert")
+	sf.ip = sf.set.String("ip", "", "Required: ipv4 address and network in CIDR notation to assign the cert")
 	sf.duration = sf.set.Duration("duration", 0, "Optional: how long the cert should be valid for. The default is 1 second before the signing cert expires. Valid time units are seconds: \"s\", minutes: \"m\", hours: \"h\"")
 	sf.inPubPath = sf.set.String("in-pub", "", "Optional (if out-key not set): path to read a previously generated public key")
 	sf.outKeyPath = sf.set.String("out-key", "", "Optional (if in-pub not set): path to write the private key to")
 	sf.outCertPath = sf.set.String("out-crt", "", "Optional: path to write the certificate to")
+	sf.outQRPath = sf.set.String("out-qr", "", "Optional: output a qr code image (png) of the certificate")
 	sf.groups = sf.set.String("groups", "", "Optional: comma separated list of groups")
-	sf.subnets = sf.set.String("subnets", "", "Optional: comma seperated list of subnet this cert can serve for")
+	sf.subnets = sf.set.String("subnets", "", "Optional: comma separated list of ipv4 address and network in CIDR notation. Subnets this cert can serve for")
 	return &sf
 
 }
 
-func signCert(args []string, out io.Writer, errOut io.Writer) error {
+func signCert(args []string, out io.Writer, errOut io.Writer, pr PasswordReader) error {
 	sf := newSignFlags()
 	err := sf.set.Parse(args)
 	if err != nil {
@@ -69,17 +72,46 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 		return newHelpErrorf("cannot set both -in-pub and -out-key")
 	}
 
-	rawCAKey, err := ioutil.ReadFile(*sf.caKeyPath)
+	rawCAKey, err := os.ReadFile(*sf.caKeyPath)
 	if err != nil {
 		return fmt.Errorf("error while reading ca-key: %s", err)
 	}
 
-	caKey, _, err := cert.UnmarshalEd25519PrivateKey(rawCAKey)
-	if err != nil {
+	var curve cert.Curve
+	var caKey []byte
+
+	// naively attempt to decode the private key as though it is not encrypted
+	caKey, _, curve, err = cert.UnmarshalSigningPrivateKey(rawCAKey)
+	if err == cert.ErrPrivateKeyEncrypted {
+		// ask for a passphrase until we get one
+		var passphrase []byte
+		for i := 0; i < 5; i++ {
+			out.Write([]byte("Enter passphrase: "))
+			passphrase, err = pr.ReadPassword()
+
+			if err == ErrNoTerminal {
+				return fmt.Errorf("ca-key is encrypted and must be decrypted interactively")
+			} else if err != nil {
+				return fmt.Errorf("error reading password: %s", err)
+			}
+
+			if len(passphrase) > 0 {
+				break
+			}
+		}
+		if len(passphrase) == 0 {
+			return fmt.Errorf("cannot open encrypted ca-key without passphrase")
+		}
+
+		curve, caKey, _, err = cert.DecryptAndUnmarshalSigningPrivateKey(passphrase, rawCAKey)
+		if err != nil {
+			return fmt.Errorf("error while parsing encrypted ca-key: %s", err)
+		}
+	} else if err != nil {
 		return fmt.Errorf("error while parsing ca-key: %s", err)
 	}
 
-	rawCACert, err := ioutil.ReadFile(*sf.caCertPath)
+	rawCACert, err := os.ReadFile(*sf.caCertPath)
 	if err != nil {
 		return fmt.Errorf("error while reading ca-crt: %s", err)
 	}
@@ -89,6 +121,10 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 		return fmt.Errorf("error while parsing ca-crt: %s", err)
 	}
 
+	if err := caCert.VerifyPrivateKey(curve, caKey); err != nil {
+		return fmt.Errorf("refusing to sign, root certificate does not match private key")
+	}
+
 	issuer, err := caCert.Sha256Sum()
 	if err != nil {
 		return fmt.Errorf("error while getting -ca-crt fingerprint: %s", err)
@@ -107,6 +143,9 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 	if err != nil {
 		return newHelpErrorf("invalid ip definition: %s", err)
 	}
+	if ip.To4() == nil {
+		return newHelpErrorf("invalid ip definition: can only be ipv4, have %s", *sf.ip)
+	}
 	ipNet.IP = ip
 
 	groups := []string{}
@@ -128,6 +167,9 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 				if err != nil {
 					return newHelpErrorf("invalid subnet definition: %s", err)
 				}
+				if s.IP.To4() == nil {
+					return newHelpErrorf("invalid subnet definition: can only be ipv4, have %s", rs)
+				}
 				subnets = append(subnets, s)
 			}
 		}
@@ -135,16 +177,20 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 
 	var pub, rawPriv []byte
 	if *sf.inPubPath != "" {
-		rawPub, err := ioutil.ReadFile(*sf.inPubPath)
+		rawPub, err := os.ReadFile(*sf.inPubPath)
 		if err != nil {
 			return fmt.Errorf("error while reading in-pub: %s", err)
 		}
-		pub, _, err = cert.UnmarshalX25519PublicKey(rawPub)
+		var pubCurve cert.Curve
+		pub, _, pubCurve, err = cert.UnmarshalPublicKey(rawPub)
 		if err != nil {
 			return fmt.Errorf("error while parsing in-pub: %s", err)
 		}
+		if pubCurve != curve {
+			return fmt.Errorf("curve of in-pub does not match ca")
+		}
 	} else {
-		pub, rawPriv = x25519Keypair()
+		pub, rawPriv = newKeypair(curve)
 	}
 
 	nc := cert.NebulaCertificate{
@@ -158,6 +204,7 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 			PublicKey: pub,
 			IsCA:      false,
 			Issuer:    issuer,
+			Curve:     curve,
 		},
 	}
 
@@ -177,7 +224,7 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 		return fmt.Errorf("refusing to overwrite existing cert: %s", *sf.outCertPath)
 	}
 
-	err = nc.Sign(caKey)
+	err = nc.Sign(curve, caKey)
 	if err != nil {
 		return fmt.Errorf("error while signing: %s", err)
 	}
@@ -187,7 +234,7 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 			return fmt.Errorf("refusing to overwrite existing key: %s", *sf.outKeyPath)
 		}
 
-		err = ioutil.WriteFile(*sf.outKeyPath, cert.MarshalX25519PrivateKey(rawPriv), 0600)
+		err = os.WriteFile(*sf.outKeyPath, cert.MarshalPrivateKey(curve, rawPriv), 0600)
 		if err != nil {
 			return fmt.Errorf("error while writing out-key: %s", err)
 		}
@@ -198,21 +245,58 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 		return fmt.Errorf("error while marshalling certificate: %s", err)
 	}
 
-	err = ioutil.WriteFile(*sf.outCertPath, b, 0600)
+	err = os.WriteFile(*sf.outCertPath, b, 0600)
 	if err != nil {
 		return fmt.Errorf("error while writing out-crt: %s", err)
 	}
 
+	if *sf.outQRPath != "" {
+		b, err = qrcode.Encode(string(b), qrcode.Medium, -5)
+		if err != nil {
+			return fmt.Errorf("error while generating qr code: %s", err)
+		}
+
+		err = os.WriteFile(*sf.outQRPath, b, 0600)
+		if err != nil {
+			return fmt.Errorf("error while writing out-qr: %s", err)
+		}
+	}
+
 	return nil
 }
 
+func newKeypair(curve cert.Curve) ([]byte, []byte) {
+	switch curve {
+	case cert.Curve_CURVE25519:
+		return x25519Keypair()
+	case cert.Curve_P256:
+		return p256Keypair()
+	default:
+		return nil, nil
+	}
+}
+
 func x25519Keypair() ([]byte, []byte) {
-	var pubkey, privkey [32]byte
-	if _, err := io.ReadFull(rand.Reader, privkey[:]); err != nil {
+	privkey := make([]byte, 32)
+	if _, err := io.ReadFull(rand.Reader, privkey); err != nil {
 		panic(err)
 	}
-	curve25519.ScalarBaseMult(&pubkey, &privkey)
-	return pubkey[:], privkey[:]
+
+	pubkey, err := curve25519.X25519(privkey, curve25519.Basepoint)
+	if err != nil {
+		panic(err)
+	}
+
+	return pubkey, privkey
+}
+
+func p256Keypair() ([]byte, []byte) {
+	privkey, err := ecdh.P256().GenerateKey(rand.Reader)
+	if err != nil {
+		panic(err)
+	}
+	pubkey := privkey.PublicKey()
+	return pubkey.Bytes(), privkey.Bytes()
 }
 
 func signSummary() string {

cmd/nebula-cert/sign_test.go

@@ -1,3 +1,4 @@
+//go:build !windows
 // +build !windows
 
 package main
@@ -5,7 +6,7 @@ package main
 import (
 	"bytes"
 	"crypto/rand"
-	"io/ioutil"
+	"errors"
 	"os"
 	"testing"
 	"time"
@@ -38,15 +39,17 @@ func Test_signHelp(t *testing.T) {
 			"  -in-pub string\n"+
 			"    \tOptional (if out-key not set): path to read a previously generated public key\n"+
 			"  -ip string\n"+
-			"    \tRequired: ip and network in CIDR notation to assign the cert\n"+
+			"    \tRequired: ipv4 address and network in CIDR notation to assign the cert\n"+
 			"  -name string\n"+
 			"    \tRequired: name of the cert, usually a hostname\n"+
 			"  -out-crt string\n"+
 			"    \tOptional: path to write the certificate to\n"+
 			"  -out-key string\n"+
 			"    \tOptional (if in-pub not set): path to write the private key to\n"+
+			"  -out-qr string\n"+
+			"    \tOptional: output a qr code image (png) of the certificate\n"+
 			"  -subnets string\n"+
-			"    \tOptional: comma seperated list of subnet this cert can serve for\n",
+			"    \tOptional: comma separated list of ipv4 address and network in CIDR notation. Subnets this cert can serve for\n",
 		ob.String(),
 	)
 }
@@ -55,18 +58,39 @@ func Test_signCert(t *testing.T) {
 	ob := &bytes.Buffer{}
 	eb := &bytes.Buffer{}
 
-	// required args
+	nopw := &StubPasswordReader{
+		password: []byte(""),
+		err:      nil,
+	}
 
-	assertHelpError(t, signCert([]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-ip", "1.1.1.1/24", "-out-key", "nope", "-out-crt", "nope"}, ob, eb), "-name is required")
+	errpw := &StubPasswordReader{
+		password: []byte(""),
+		err:      errors.New("stub error"),
+	}
+
+	passphrase := []byte("DO NOT USE THIS KEY")
+	testpw := &StubPasswordReader{
+		password: passphrase,
+		err:      nil,
+	}
+
+	// required args
+	assertHelpError(t, signCert(
+		[]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-ip", "1.1.1.1/24", "-out-key", "nope", "-out-crt", "nope"}, ob, eb, nopw,
+	), "-name is required")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
-	assertHelpError(t, signCert([]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-name", "test", "-out-key", "nope", "-out-crt", "nope"}, ob, eb), "-ip is required")
+	assertHelpError(t, signCert(
+		[]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-name", "test", "-out-key", "nope", "-out-crt", "nope"}, ob, eb, nopw,
+	), "-ip is required")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// cannot set -in-pub and -out-key
-	assertHelpError(t, signCert([]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-name", "test", "-in-pub", "nope", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope"}, ob, eb), "cannot set both -in-pub and -out-key")
+	assertHelpError(t, signCert(
+		[]string{"-ca-crt", "./nope", "-ca-key", "./nope", "-name", "test", "-in-pub", "nope", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope"}, ob, eb, nopw,
+	), "cannot set both -in-pub and -out-key")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -74,17 +98,17 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args := []string{"-ca-crt", "./nope", "-ca-key", "./nope", "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while reading ca-key: open ./nope: "+NoSuchFileError)
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while reading ca-key: open ./nope: "+NoSuchFileError)
 
 	// failed to unmarshal key
 	ob.Reset()
 	eb.Reset()
-	caKeyF, err := ioutil.TempFile("", "sign-cert.key")
+	caKeyF, err := os.CreateTemp("", "sign-cert.key")
 	assert.Nil(t, err)
 	defer os.Remove(caKeyF.Name())
 
 	args = []string{"-ca-crt", "./nope", "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while parsing ca-key: input did not contain a valid PEM encoded block")
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while parsing ca-key: input did not contain a valid PEM encoded block")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -96,19 +120,19 @@ func Test_signCert(t *testing.T) {
 
 	// failed to read cert
 	args = []string{"-ca-crt", "./nope", "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while reading ca-crt: open ./nope: "+NoSuchFileError)
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while reading ca-crt: open ./nope: "+NoSuchFileError)
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// failed to unmarshal cert
 	ob.Reset()
 	eb.Reset()
-	caCrtF, err := ioutil.TempFile("", "sign-cert.crt")
+	caCrtF, err := os.CreateTemp("", "sign-cert.crt")
 	assert.Nil(t, err)
 	defer os.Remove(caCrtF.Name())
 
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while parsing ca-crt: input did not contain a valid PEM encoded block")
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while parsing ca-crt: input did not contain a valid PEM encoded block")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -127,19 +151,19 @@ func Test_signCert(t *testing.T) {
 
 	// failed to read pub
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-in-pub", "./nope", "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while reading in-pub: open ./nope: "+NoSuchFileError)
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while reading in-pub: open ./nope: "+NoSuchFileError)
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// failed to unmarshal pub
 	ob.Reset()
 	eb.Reset()
-	inPubF, err := ioutil.TempFile("", "in.pub")
+	inPubF, err := os.CreateTemp("", "in.pub")
 	assert.Nil(t, err)
 	defer os.Remove(inPubF.Name())
 
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-in-pub", inPubF.Name(), "-duration", "100m"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while parsing in-pub: input did not contain a valid PEM encoded block")
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while parsing in-pub: input did not contain a valid PEM encoded block")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -153,7 +177,14 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "a1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
-	assertHelpError(t, signCert(args, ob, eb), "invalid ip definition: invalid CIDR address: a1.1.1.1/24")
+	assertHelpError(t, signCert(args, ob, eb, nopw), "invalid ip definition: invalid CIDR address: a1.1.1.1/24")
+	assert.Empty(t, ob.String())
+	assert.Empty(t, eb.String())
+
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "100::100/100", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m"}
+	assertHelpError(t, signCert(args, ob, eb, nopw), "invalid ip definition: can only be ipv4, have 100::100/100")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -161,7 +192,28 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m", "-subnets", "a"}
-	assertHelpError(t, signCert(args, ob, eb), "invalid subnet definition: invalid CIDR address: a")
+	assertHelpError(t, signCert(args, ob, eb, nopw), "invalid subnet definition: invalid CIDR address: a")
+	assert.Empty(t, ob.String())
+	assert.Empty(t, eb.String())
+
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m", "-subnets", "100::100/100"}
+	assertHelpError(t, signCert(args, ob, eb, nopw), "invalid subnet definition: can only be ipv4, have 100::100/100")
+	assert.Empty(t, ob.String())
+	assert.Empty(t, eb.String())
+
+	// mismatched ca key
+	_, caPriv2, _ := ed25519.GenerateKey(rand.Reader)
+	caKeyF2, err := os.CreateTemp("", "sign-cert-2.key")
+	assert.Nil(t, err)
+	defer os.Remove(caKeyF2.Name())
+	caKeyF2.Write(cert.MarshalEd25519PrivateKey(caPriv2))
+
+	ob.Reset()
+	eb.Reset()
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF2.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "nope", "-out-key", "nope", "-duration", "100m", "-subnets", "a"}
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "refusing to sign, root certificate does not match private key")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -169,12 +221,12 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "/do/not/write/pleasecrt", "-out-key", "/do/not/write/pleasekey", "-duration", "100m", "-subnets", "10.1.1.1/32"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while writing out-key: open /do/not/write/pleasekey: "+NoSuchDirError)
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while writing out-key: open /do/not/write/pleasekey: "+NoSuchDirError)
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// create temp key file
-	keyF, err := ioutil.TempFile("", "test.key")
+	keyF, err := os.CreateTemp("", "test.key")
 	assert.Nil(t, err)
 	os.Remove(keyF.Name())
 
@@ -182,13 +234,13 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", "/do/not/write/pleasecrt", "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32"}
-	assert.EqualError(t, signCert(args, ob, eb), "error while writing out-crt: open /do/not/write/pleasecrt: "+NoSuchDirError)
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "error while writing out-crt: open /do/not/write/pleasecrt: "+NoSuchDirError)
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 	os.Remove(keyF.Name())
 
 	// create temp cert file
-	crtF, err := ioutil.TempFile("", "test.crt")
+	crtF, err := os.CreateTemp("", "test.crt")
 	assert.Nil(t, err)
 	os.Remove(crtF.Name())
 
@@ -196,18 +248,18 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.Nil(t, signCert(args, ob, eb))
+	assert.Nil(t, signCert(args, ob, eb, nopw))
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// read cert and key files
-	rb, _ := ioutil.ReadFile(keyF.Name())
+	rb, _ := os.ReadFile(keyF.Name())
 	lKey, b, err := cert.UnmarshalX25519PrivateKey(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
 	assert.Len(t, lKey, 32)
 
-	rb, _ = ioutil.ReadFile(crtF.Name())
+	rb, _ = os.ReadFile(crtF.Name())
 	lCrt, b, err := cert.UnmarshalNebulaCertificateFromPEM(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
@@ -238,12 +290,12 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-in-pub", inPubF.Name(), "-duration", "100m", "-groups", "1"}
-	assert.Nil(t, signCert(args, ob, eb))
+	assert.Nil(t, signCert(args, ob, eb, nopw))
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
 	// read cert file and check pub key matches in-pub
-	rb, _ = ioutil.ReadFile(crtF.Name())
+	rb, _ = os.ReadFile(crtF.Name())
 	lCrt, b, err = cert.UnmarshalNebulaCertificateFromPEM(rb)
 	assert.Len(t, b, 0)
 	assert.Nil(t, err)
@@ -253,7 +305,7 @@ func Test_signCert(t *testing.T) {
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "1000m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.EqualError(t, signCert(args, ob, eb), "refusing to sign, root certificate constraints violated: certificate expires after signing certificate")
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "refusing to sign, root certificate constraints violated: certificate expires after signing certificate")
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -261,14 +313,14 @@ func Test_signCert(t *testing.T) {
 	os.Remove(keyF.Name())
 	os.Remove(crtF.Name())
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.Nil(t, signCert(args, ob, eb))
+	assert.Nil(t, signCert(args, ob, eb, nopw))
 
 	// test that we won't overwrite existing key file
 	os.Remove(crtF.Name())
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.EqualError(t, signCert(args, ob, eb), "refusing to overwrite existing key: "+keyF.Name())
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "refusing to overwrite existing key: "+keyF.Name())
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
@@ -276,15 +328,83 @@ func Test_signCert(t *testing.T) {
 	os.Remove(keyF.Name())
 	os.Remove(crtF.Name())
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.Nil(t, signCert(args, ob, eb))
+	assert.Nil(t, signCert(args, ob, eb, nopw))
 
 	// test that we won't overwrite existing certificate file
 	os.Remove(keyF.Name())
 	ob.Reset()
 	eb.Reset()
 	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
-	assert.EqualError(t, signCert(args, ob, eb), "refusing to overwrite existing cert: "+crtF.Name())
+	assert.EqualError(t, signCert(args, ob, eb, nopw), "refusing to overwrite existing cert: "+crtF.Name())
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
+	// create valid cert/key using encrypted CA key
+	os.Remove(caKeyF.Name())
+	os.Remove(caCrtF.Name())
+	os.Remove(keyF.Name())
+	os.Remove(crtF.Name())
+	ob.Reset()
+	eb.Reset()
+
+	caKeyF, err = os.CreateTemp("", "sign-cert.key")
+	assert.Nil(t, err)
+	defer os.Remove(caKeyF.Name())
+
+	caCrtF, err = os.CreateTemp("", "sign-cert.crt")
+	assert.Nil(t, err)
+	defer os.Remove(caCrtF.Name())
+
+	// generate the encrypted key
+	caPub, caPriv, _ = ed25519.GenerateKey(rand.Reader)
+	kdfParams := cert.NewArgon2Parameters(64*1024, 4, 3)
+	b, _ = cert.EncryptAndMarshalSigningPrivateKey(cert.Curve_CURVE25519, caPriv, passphrase, kdfParams)
+	caKeyF.Write(b)
+
+	ca = cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:      "ca",
+			NotBefore: time.Now(),
+			NotAfter:  time.Now().Add(time.Minute * 200),
+			PublicKey: caPub,
+			IsCA:      true,
+		},
+	}
+	b, _ = ca.MarshalToPEM()
+	caCrtF.Write(b)
+
+	// test with the proper password
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
+	assert.Nil(t, signCert(args, ob, eb, testpw))
+	assert.Equal(t, "Enter passphrase: ", ob.String())
+	assert.Empty(t, eb.String())
+
+	// test with the wrong password
+	ob.Reset()
+	eb.Reset()
+
+	testpw.password = []byte("invalid password")
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
+	assert.Error(t, signCert(args, ob, eb, testpw))
+	assert.Equal(t, "Enter passphrase: ", ob.String())
+	assert.Empty(t, eb.String())
+
+	// test with the user not entering a password
+	ob.Reset()
+	eb.Reset()
+
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
+	assert.Error(t, signCert(args, ob, eb, nopw))
+	// normally the user hitting enter on the prompt would add newlines between these
+	assert.Equal(t, "Enter passphrase: Enter passphrase: Enter passphrase: Enter passphrase: Enter passphrase: ", ob.String())
+	assert.Empty(t, eb.String())
+
+	// test an error condition
+	ob.Reset()
+	eb.Reset()
+
+	args = []string{"-ca-crt", caCrtF.Name(), "-ca-key", caKeyF.Name(), "-name", "test", "-ip", "1.1.1.1/24", "-out-crt", crtF.Name(), "-out-key", keyF.Name(), "-duration", "100m", "-subnets", "10.1.1.1/32, ,   10.2.2.2/32   ,   ,  ,, 10.5.5.5/32", "-groups", "1,,   2    ,        ,,,3,4,5"}
+	assert.Error(t, signCert(args, ob, eb, errpw))
+	assert.Equal(t, "Enter passphrase: ", ob.String())
+	assert.Empty(t, eb.String())
 }

cmd/nebula-cert/verify.go

@@ -3,12 +3,12 @@ package main
 import (
 	"flag"
 	"fmt"
-	"github.com/slackhq/nebula/cert"
 	"io"
-	"io/ioutil"
 	"os"
 	"strings"
 	"time"
+
+	"github.com/slackhq/nebula/cert"
 )
 
 type verifyFlags struct {
@@ -39,7 +39,7 @@ func verify(args []string, out io.Writer, errOut io.Writer) error {
 		return err
 	}
 
-	rawCACert, err := ioutil.ReadFile(*vf.caPath)
+	rawCACert, err := os.ReadFile(*vf.caPath)
 	if err != nil {
 		return fmt.Errorf("error while reading ca: %s", err)
 	}
@@ -56,7 +56,7 @@ func verify(args []string, out io.Writer, errOut io.Writer) error {
 		}
 	}
 
-	rawCert, err := ioutil.ReadFile(*vf.certPath)
+	rawCert, err := os.ReadFile(*vf.certPath)
 	if err != nil {
 		return fmt.Errorf("unable to read crt; %s", err)
 	}

cmd/nebula-cert/verify_test.go

@@ -3,13 +3,13 @@ package main
 import (
 	"bytes"
 	"crypto/rand"
-	"github.com/slackhq/nebula/cert"
-	"github.com/stretchr/testify/assert"
-	"golang.org/x/crypto/ed25519"
-	"io/ioutil"
 	"os"
 	"testing"
 	"time"
+
+	"github.com/slackhq/nebula/cert"
+	"github.com/stretchr/testify/assert"
+	"golang.org/x/crypto/ed25519"
 )
 
 func Test_verifySummary(t *testing.T) {
@@ -55,7 +55,7 @@ func Test_verify(t *testing.T) {
 	// invalid ca at path
 	ob.Reset()
 	eb.Reset()
-	caFile, err := ioutil.TempFile("", "verify-ca")
+	caFile, err := os.CreateTemp("", "verify-ca")
 	assert.Nil(t, err)
 	defer os.Remove(caFile.Name())
 
@@ -71,12 +71,12 @@ func Test_verify(t *testing.T) {
 		Details: cert.NebulaCertificateDetails{
 			Name:      "test-ca",
 			NotBefore: time.Now().Add(time.Hour * -1),
-			NotAfter:  time.Now().Add(time.Hour),
+			NotAfter:  time.Now().Add(time.Hour * 2),
 			PublicKey: caPub,
 			IsCA:      true,
 		},
 	}
-	ca.Sign(caPriv)
+	ca.Sign(cert.Curve_CURVE25519, caPriv)
 	b, _ := ca.MarshalToPEM()
 	caFile.Truncate(0)
 	caFile.Seek(0, 0)
@@ -91,7 +91,7 @@ func Test_verify(t *testing.T) {
 	// invalid crt at path
 	ob.Reset()
 	eb.Reset()
-	certFile, err := ioutil.TempFile("", "verify-cert")
+	certFile, err := os.CreateTemp("", "verify-cert")
 	assert.Nil(t, err)
 	defer os.Remove(certFile.Name())
 
@@ -116,7 +116,7 @@ func Test_verify(t *testing.T) {
 		},
 	}
 
-	crt.Sign(badPriv)
+	crt.Sign(cert.Curve_CURVE25519, badPriv)
 	b, _ = crt.MarshalToPEM()
 	certFile.Truncate(0)
 	certFile.Seek(0, 0)
@@ -128,7 +128,7 @@ func Test_verify(t *testing.T) {
 	assert.EqualError(t, err, "certificate signature did not match")
 
 	// verified cert at path
-	crt.Sign(caPriv)
+	crt.Sign(cert.Curve_CURVE25519, caPriv)
 	b, _ = crt.MarshalToPEM()
 	certFile.Truncate(0)
 	certFile.Seek(0, 0)

cmd/nebula-service/logs_generic.go

@@ -0,0 +1,10 @@
+//go:build !windows
+// +build !windows
+
+package main
+
+import "github.com/sirupsen/logrus"
+
+func HookLogger(l *logrus.Logger) {
+	// Do nothing, let the logs flow to stdout/stderr
+}

cmd/nebula-service/logs_windows.go

@@ -0,0 +1,54 @@
+package main
+
+import (
+	"fmt"
+	"io/ioutil"
+	"os"
+
+	"github.com/kardianos/service"
+	"github.com/sirupsen/logrus"
+)
+
+// HookLogger routes the logrus logs through the service logger so that they end up in the Windows Event Viewer
+// logrus output will be discarded
+func HookLogger(l *logrus.Logger) {
+	l.AddHook(newLogHook(logger))
+	l.SetOutput(ioutil.Discard)
+}
+
+type logHook struct {
+	sl service.Logger
+}
+
+func newLogHook(sl service.Logger) *logHook {
+	return &logHook{sl: sl}
+}
+
+func (h *logHook) Fire(entry *logrus.Entry) error {
+	line, err := entry.String()
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Unable to read entry, %v", err)
+		return err
+	}
+
+	switch entry.Level {
+	case logrus.PanicLevel:
+		return h.sl.Error(line)
+	case logrus.FatalLevel:
+		return h.sl.Error(line)
+	case logrus.ErrorLevel:
+		return h.sl.Error(line)
+	case logrus.WarnLevel:
+		return h.sl.Warning(line)
+	case logrus.InfoLevel:
+		return h.sl.Info(line)
+	case logrus.DebugLevel:
+		return h.sl.Info(line)
+	default:
+		return nil
+	}
+}
+
+func (h *logHook) Levels() []logrus.Level {
+	return logrus.AllLevels
+}

cmd/nebula-service/main.go

@@ -5,12 +5,15 @@ import (
 	"fmt"
 	"os"
 
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/util"
 )
 
 // A version string that can be set with
 //
-//     -ldflags "-X main.Build=SOMEVERSION"
+//	-ldflags "-X main.Build=SOMEVERSION"
 //
 // at compile-time.
 var Build string
@@ -45,5 +48,26 @@ func main() {
 		os.Exit(1)
 	}
 
-	nebula.Main(*configPath, *configTest, Build)
+	l := logrus.New()
+	l.Out = os.Stdout
+
+	c := config.NewC(l)
+	err := c.Load(*configPath)
+	if err != nil {
+		fmt.Printf("failed to load config: %s", err)
+		os.Exit(1)
+	}
+
+	ctrl, err := nebula.Main(c, *configTest, Build, l, nil)
+	if err != nil {
+		util.LogWithContextIfNeeded("Failed to start", err, l)
+		os.Exit(1)
+	}
+
+	if !*configTest {
+		ctrl.Start()
+		ctrl.ShutdownBlock()
+	}
+
+	os.Exit(0)
 }

cmd/nebula-service/service.go

@@ -1,50 +1,72 @@
 package main
 
 import (
+	"fmt"
 	"log"
 	"os"
 	"path/filepath"
 
 	"github.com/kardianos/service"
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
 )
 
 var logger service.Logger
 
 type program struct {
-	exit       chan struct{}
 	configPath *string
 	configTest *bool
 	build      string
+	control    *nebula.Control
 }
 
 func (p *program) Start(s service.Service) error {
-	logger.Info("Nebula service starting.")
-	p.exit = make(chan struct{})
 	// Start should not block.
-	go p.run()
-	return nil
-}
+	logger.Info("Nebula service starting.")
 
-func (p *program) run() error {
-	nebula.Main(*p.configPath, *p.configTest, Build)
+	l := logrus.New()
+	HookLogger(l)
+
+	c := config.NewC(l)
+	err := c.Load(*p.configPath)
+	if err != nil {
+		return fmt.Errorf("failed to load config: %s", err)
+	}
+
+	p.control, err = nebula.Main(c, *p.configTest, Build, l, nil)
+	if err != nil {
+		return err
+	}
+
+	p.control.Start()
 	return nil
 }
 
 func (p *program) Stop(s service.Service) error {
 	logger.Info("Nebula service stopping.")
-	close(p.exit)
+	p.control.Stop()
 	return nil
 }
 
-func doService(configPath *string, configTest *bool, build string, serviceFlag *string) {
+func fileExists(filename string) bool {
+	_, err := os.Stat(filename)
+	if os.IsNotExist(err) {
+		return false
+	}
+	return true
+}
 
+func doService(configPath *string, configTest *bool, build string, serviceFlag *string) {
 	if *configPath == "" {
 		ex, err := os.Executable()
 		if err != nil {
 			panic(err)
 		}
 		*configPath = filepath.Dir(ex) + "/config.yaml"
+		if !fileExists(*configPath) {
+			*configPath = filepath.Dir(ex) + "/config.yml"
+		}
 	}
 
 	svcConfig := &service.Config{
@@ -60,6 +82,10 @@ func doService(configPath *string, configTest *bool, build string, serviceFlag *
 		build:      build,
 	}
 
+	// Here are what the different loggers are doing:
+	// - `log` is the standard go log utility, meant to be used while the process is still attached to stdout/stderr
+	// - `logger` is the service log utility that may be attached to a special place depending on OS (Windows will have it attached to the event log)
+	// - above, in `Run` we create a `logrus.Logger` which is what nebula expects to use
 	s, err := service.New(prg, svcConfig)
 	if err != nil {
 		log.Fatal(err)
@@ -75,6 +101,7 @@ func doService(configPath *string, configTest *bool, build string, serviceFlag *
 		for {
 			err := <-errs
 			if err != nil {
+				// Route any errors from the system logger to stdout as a best effort to notice issues there
 				log.Print(err)
 			}
 		}
@@ -84,6 +111,7 @@ func doService(configPath *string, configTest *bool, build string, serviceFlag *
 	case "run":
 		err = s.Run()
 		if err != nil {
+			// Route any errors to the system logger
 			logger.Error(err)
 		}
 	default:

cmd/nebula/main.go

@@ -5,12 +5,15 @@ import (
 	"fmt"
 	"os"
 
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/util"
 )
 
 // A version string that can be set with
 //
-//     -ldflags "-X main.Build=SOMEVERSION"
+//	-ldflags "-X main.Build=SOMEVERSION"
 //
 // at compile-time.
 var Build string
@@ -39,5 +42,27 @@ func main() {
 		os.Exit(1)
 	}
 
-	nebula.Main(*configPath, *configTest, Build)
+	l := logrus.New()
+	l.Out = os.Stdout
+
+	c := config.NewC(l)
+	err := c.Load(*configPath)
+	if err != nil {
+		fmt.Printf("failed to load config: %s", err)
+		os.Exit(1)
+	}
+
+	ctrl, err := nebula.Main(c, *configTest, Build, l, nil)
+	if err != nil {
+		util.LogWithContextIfNeeded("Failed to start", err, l)
+		os.Exit(1)
+	}
+
+	if !*configTest {
+		ctrl.Start()
+		notifyReady(l)
+		ctrl.ShutdownBlock()
+	}
+
+	os.Exit(0)
 }

cmd/nebula/notify_linux.go

@@ -0,0 +1,42 @@
+package main
+
+import (
+	"net"
+	"os"
+	"time"
+
+	"github.com/sirupsen/logrus"
+)
+
+// SdNotifyReady tells systemd the service is ready and dependent services can now be started
+// https://www.freedesktop.org/software/systemd/man/sd_notify.html
+// https://www.freedesktop.org/software/systemd/man/systemd.service.html
+const SdNotifyReady = "READY=1"
+
+func notifyReady(l *logrus.Logger) {
+	sockName := os.Getenv("NOTIFY_SOCKET")
+	if sockName == "" {
+		l.Debugln("NOTIFY_SOCKET systemd env var not set, not sending ready signal")
+		return
+	}
+
+	conn, err := net.DialTimeout("unixgram", sockName, time.Second)
+	if err != nil {
+		l.WithError(err).Error("failed to connect to systemd notification socket")
+		return
+	}
+	defer conn.Close()
+
+	err = conn.SetWriteDeadline(time.Now().Add(time.Second))
+	if err != nil {
+		l.WithError(err).Error("failed to set the write deadline for the systemd notification socket")
+		return
+	}
+
+	if _, err = conn.Write([]byte(SdNotifyReady)); err != nil {
+		l.WithError(err).Error("failed to signal the systemd notification socket")
+		return
+	}
+
+	l.Debugln("notified systemd the service is ready")
+}

cmd/nebula/notify_notlinux.go

@@ -0,0 +1,10 @@
+//go:build !linux
+// +build !linux
+
+package main
+
+import "github.com/sirupsen/logrus"
+
+func notifyReady(_ *logrus.Logger) {
+	// No init service to notify
+}

config/config.go

@@ -1,39 +1,44 @@
-package nebula
+package config
 
 import (
+	"context"
+	"errors"
 	"fmt"
-	"github.com/imdario/mergo"
-	"github.com/sirupsen/logrus"
-	"gopkg.in/yaml.v2"
-	"io/ioutil"
-	"net"
+	"math"
 	"os"
 	"os/signal"
 	"path/filepath"
-	"regexp"
 	"sort"
 	"strconv"
 	"strings"
+	"sync"
 	"syscall"
 	"time"
+
+	"dario.cat/mergo"
+	"github.com/sirupsen/logrus"
+	"gopkg.in/yaml.v2"
 )
 
-type Config struct {
+type C struct {
 	path        string
 	files       []string
 	Settings    map[interface{}]interface{}
 	oldSettings map[interface{}]interface{}
-	callbacks   []func(*Config)
+	callbacks   []func(*C)
+	l           *logrus.Logger
+	reloadLock  sync.Mutex
 }
 
-func NewConfig() *Config {
-	return &Config{
+func NewC(l *logrus.Logger) *C {
+	return &C{
 		Settings: make(map[interface{}]interface{}),
+		l:        l,
 	}
 }
 
 // Load will find all yaml files within path and load them in lexical order
-func (c *Config) Load(path string) error {
+func (c *C) Load(path string) error {
 	c.path = path
 	c.files = make([]string, 0)
 
@@ -56,20 +61,32 @@ func (c *Config) Load(path string) error {
 	return nil
 }
 
+func (c *C) LoadString(raw string) error {
+	if raw == "" {
+		return errors.New("Empty configuration")
+	}
+	return c.parseRaw([]byte(raw))
+}
+
 // RegisterReloadCallback stores a function to be called when a config reload is triggered. The functions registered
 // here should decide if they need to make a change to the current process before making the change. HasChanged can be
 // used to help decide if a change is necessary.
 // These functions should return quickly or spawn their own go routine if they will take a while
-func (c *Config) RegisterReloadCallback(f func(*Config)) {
+func (c *C) RegisterReloadCallback(f func(*C)) {
 	c.callbacks = append(c.callbacks, f)
 }
 
+// InitialLoad returns true if this is the first load of the config, and ReloadConfig has not been called yet.
+func (c *C) InitialLoad() bool {
+	return c.oldSettings == nil
+}
+
 // HasChanged checks if the underlying structure of the provided key has changed after a config reload. The value of
 // k in both the old and new settings will be serialized, the result of the string comparison is returned.
 // If k is an empty string the entire config is tested.
 // It's important to note that this is very rudimentary and susceptible to configuration ordering issues indicating
 // there is change when there actually wasn't any.
-func (c *Config) HasChanged(k string) bool {
+func (c *C) HasChanged(k string) bool {
 	if c.oldSettings == nil {
 		return false
 	}
@@ -90,12 +107,12 @@ func (c *Config) HasChanged(k string) bool {
 
 	newVals, err := yaml.Marshal(nv)
 	if err != nil {
-		l.WithField("config_path", k).WithError(err).Error("Error while marshaling new config")
+		c.l.WithField("config_path", k).WithError(err).Error("Error while marshaling new config")
 	}
 
 	oldVals, err := yaml.Marshal(ov)
 	if err != nil {
-		l.WithField("config_path", k).WithError(err).Error("Error while marshaling old config")
+		c.l.WithField("config_path", k).WithError(err).Error("Error while marshaling old config")
 	}
 
 	return string(newVals) != string(oldVals)
@@ -103,19 +120,33 @@ func (c *Config) HasChanged(k string) bool {
 
 // CatchHUP will listen for the HUP signal in a go routine and reload all configs found in the
 // original path provided to Load. The old settings are shallow copied for change detection after the reload.
-func (c *Config) CatchHUP() {
+func (c *C) CatchHUP(ctx context.Context) {
+	if c.path == "" {
+		return
+	}
+
 	ch := make(chan os.Signal, 1)
 	signal.Notify(ch, syscall.SIGHUP)
 
 	go func() {
-		for range ch {
-			l.Info("Caught HUP, reloading config")
-			c.ReloadConfig()
+		for {
+			select {
+			case <-ctx.Done():
+				signal.Stop(ch)
+				close(ch)
+				return
+			case <-ch:
+				c.l.Info("Caught HUP, reloading config")
+				c.ReloadConfig()
+			}
 		}
 	}()
 }
 
-func (c *Config) ReloadConfig() {
+func (c *C) ReloadConfig() {
+	c.reloadLock.Lock()
+	defer c.reloadLock.Unlock()
+
 	c.oldSettings = make(map[interface{}]interface{})
 	for k, v := range c.Settings {
 		c.oldSettings[k] = v
@@ -123,7 +154,7 @@ func (c *Config) ReloadConfig() {
 
 	err := c.Load(c.path)
 	if err != nil {
-		l.WithField("config_path", c.path).WithError(err).Error("Error occurred while reloading config")
+		c.l.WithField("config_path", c.path).WithError(err).Error("Error occurred while reloading config")
 		return
 	}
 
@@ -132,8 +163,29 @@ func (c *Config) ReloadConfig() {
 	}
 }
 
+func (c *C) ReloadConfigString(raw string) error {
+	c.reloadLock.Lock()
+	defer c.reloadLock.Unlock()
+
+	c.oldSettings = make(map[interface{}]interface{})
+	for k, v := range c.Settings {
+		c.oldSettings[k] = v
+	}
+
+	err := c.LoadString(raw)
+	if err != nil {
+		return err
+	}
+
+	for _, v := range c.callbacks {
+		v(c)
+	}
+
+	return nil
+}
+
 // GetString will get the string for k or return the default d if not found or invalid
-func (c *Config) GetString(k, d string) string {
+func (c *C) GetString(k, d string) string {
 	r := c.Get(k)
 	if r == nil {
 		return d
@@ -143,7 +195,7 @@ func (c *Config) GetString(k, d string) string {
 }
 
 // GetStringSlice will get the slice of strings for k or return the default d if not found or invalid
-func (c *Config) GetStringSlice(k string, d []string) []string {
+func (c *C) GetStringSlice(k string, d []string) []string {
 	r := c.Get(k)
 	if r == nil {
 		return d
@@ -163,7 +215,7 @@ func (c *Config) GetStringSlice(k string, d []string) []string {
 }
 
 // GetMap will get the map for k or return the default d if not found or invalid
-func (c *Config) GetMap(k string, d map[interface{}]interface{}) map[interface{}]interface{} {
+func (c *C) GetMap(k string, d map[interface{}]interface{}) map[interface{}]interface{} {
 	r := c.Get(k)
 	if r == nil {
 		return d
@@ -178,7 +230,7 @@ func (c *Config) GetMap(k string, d map[interface{}]interface{}) map[interface{}
 }
 
 // GetInt will get the int for k or return the default d if not found or invalid
-func (c *Config) GetInt(k string, d int) int {
+func (c *C) GetInt(k string, d int) int {
 	r := c.GetString(k, strconv.Itoa(d))
 	v, err := strconv.Atoi(r)
 	if err != nil {
@@ -188,8 +240,17 @@ func (c *Config) GetInt(k string, d int) int {
 	return v
 }
 
+// GetUint32 will get the uint32 for k or return the default d if not found or invalid
+func (c *C) GetUint32(k string, d uint32) uint32 {
+	r := c.GetInt(k, int(d))
+	if uint64(r) > uint64(math.MaxUint32) {
+		return d
+	}
+	return uint32(r)
+}
+
 // GetBool will get the bool for k or return the default d if not found or invalid
-func (c *Config) GetBool(k string, d bool) bool {
+func (c *C) GetBool(k string, d bool) bool {
 	r := strings.ToLower(c.GetString(k, fmt.Sprintf("%v", d)))
 	v, err := strconv.ParseBool(r)
 	if err != nil {
@@ -206,7 +267,7 @@ func (c *Config) GetBool(k string, d bool) bool {
 }
 
 // GetDuration will get the duration for k or return the default d if not found or invalid
-func (c *Config) GetDuration(k string, d time.Duration) time.Duration {
+func (c *C) GetDuration(k string, d time.Duration) time.Duration {
 	r := c.GetString(k, "")
 	v, err := time.ParseDuration(r)
 	if err != nil {
@@ -215,138 +276,15 @@ func (c *Config) GetDuration(k string, d time.Duration) time.Duration {
 	return v
 }
 
-func (c *Config) GetAllowList(k string, allowInterfaces bool) (*AllowList, error) {
-	r := c.Get(k)
-	if r == nil {
-		return nil, nil
-	}
-
-	rawMap, ok := r.(map[interface{}]interface{})
-	if !ok {
-		return nil, fmt.Errorf("config `%s` has invalid type: %T", k, r)
-	}
-
-	tree := NewCIDRTree()
-	var nameRules []AllowListNameRule
-
-	firstValue := true
-	allValuesMatch := true
-	defaultSet := false
-	var allValues bool
-
-	for rawKey, rawValue := range rawMap {
-		rawCIDR, ok := rawKey.(string)
-		if !ok {
-			return nil, fmt.Errorf("config `%s` has invalid key (type %T): %v", k, rawKey, rawKey)
-		}
-
-		// Special rule for interface names
-		if rawCIDR == "interfaces" {
-			if !allowInterfaces {
-				return nil, fmt.Errorf("config `%s` does not support `interfaces`", k)
-			}
-			var err error
-			nameRules, err = c.getAllowListInterfaces(k, rawValue)
-			if err != nil {
-				return nil, err
-			}
-
-			continue
-		}
-
-		value, ok := rawValue.(bool)
-		if !ok {
-			return nil, fmt.Errorf("config `%s` has invalid value (type %T): %v", k, rawValue, rawValue)
-		}
-
-		_, cidr, err := net.ParseCIDR(rawCIDR)
-		if err != nil {
-			return nil, fmt.Errorf("config `%s` has invalid CIDR: %s", k, rawCIDR)
-		}
-
-		// TODO: should we error on duplicate CIDRs in the config?
-		tree.AddCIDR(cidr, value)
-
-		if firstValue {
-			allValues = value
-			firstValue = false
-		} else {
-			if value != allValues {
-				allValuesMatch = false
-			}
-		}
-
-		// Check if this is 0.0.0.0/0
-		bits, size := cidr.Mask.Size()
-		if bits == 0 && size == 32 {
-			defaultSet = true
-		}
-	}
-
-	if !defaultSet {
-		if allValuesMatch {
-			_, zeroCIDR, _ := net.ParseCIDR("0.0.0.0/0")
-			tree.AddCIDR(zeroCIDR, !allValues)
-		} else {
-			return nil, fmt.Errorf("config `%s` contains both true and false rules, but no default set for 0.0.0.0/0", k)
-		}
-	}
-
-	return &AllowList{cidrTree: tree, nameRules: nameRules}, nil
-}
-
-func (c *Config) getAllowListInterfaces(k string, v interface{}) ([]AllowListNameRule, error) {
-	var nameRules []AllowListNameRule
-
-	rawRules, ok := v.(map[interface{}]interface{})
-	if !ok {
-		return nil, fmt.Errorf("config `%s.interfaces` is invalid (type %T): %v", k, v, v)
-	}
-
-	firstEntry := true
-	var allValues bool
-	for rawName, rawAllow := range rawRules {
-		name, ok := rawName.(string)
-		if !ok {
-			return nil, fmt.Errorf("config `%s.interfaces` has invalid key (type %T): %v", k, rawName, rawName)
-		}
-		allow, ok := rawAllow.(bool)
-		if !ok {
-			return nil, fmt.Errorf("config `%s.interfaces` has invalid value (type %T): %v", k, rawAllow, rawAllow)
-		}
-
-		nameRE, err := regexp.Compile("^" + name + "$")
-		if err != nil {
-			return nil, fmt.Errorf("config `%s.interfaces` has invalid key: %s: %v", k, name, err)
-		}
-
-		nameRules = append(nameRules, AllowListNameRule{
-			Name:  nameRE,
-			Allow: allow,
-		})
-
-		if firstEntry {
-			allValues = allow
-			firstEntry = false
-		} else {
-			if allow != allValues {
-				return nil, fmt.Errorf("config `%s.interfaces` values must all be the same true/false value", k)
-			}
-		}
-	}
-
-	return nameRules, nil
-}
-
-func (c *Config) Get(k string) interface{} {
+func (c *C) Get(k string) interface{} {
 	return c.get(k, c.Settings)
 }
 
-func (c *Config) IsSet(k string) bool {
+func (c *C) IsSet(k string) bool {
 	return c.get(k, c.Settings) != nil
 }
 
-func (c *Config) get(k string, v interface{}) interface{} {
+func (c *C) get(k string, v interface{}) interface{} {
 	parts := strings.Split(k, ".")
 	for _, p := range parts {
 		m, ok := v.(map[interface{}]interface{})
@@ -365,7 +303,7 @@ func (c *Config) get(k string, v interface{}) interface{} {
 
 // direct signifies if this is the config path directly specified by the user,
 // versus a file/dir found by recursing into that path
-func (c *Config) resolve(path string, direct bool) error {
+func (c *C) resolve(path string, direct bool) error {
 	i, err := os.Stat(path)
 	if err != nil {
 		return nil
@@ -391,7 +329,7 @@ func (c *Config) resolve(path string, direct bool) error {
 	return nil
 }
 
-func (c *Config) addFile(path string, direct bool) error {
+func (c *C) addFile(path string, direct bool) error {
 	ext := filepath.Ext(path)
 
 	if !direct && ext != ".yaml" && ext != ".yml" {
@@ -407,11 +345,23 @@ func (c *Config) addFile(path string, direct bool) error {
 	return nil
 }
 
-func (c *Config) parse() error {
+func (c *C) parseRaw(b []byte) error {
+	var m map[interface{}]interface{}
+
+	err := yaml.Unmarshal(b, &m)
+	if err != nil {
+		return err
+	}
+
+	c.Settings = m
+	return nil
+}
+
+func (c *C) parse() error {
 	var m map[interface{}]interface{}
 
 	for _, path := range c.files {
-		b, err := ioutil.ReadFile(path)
+		b, err := os.ReadFile(path)
 		if err != nil {
 			return err
 		}
@@ -450,35 +400,3 @@ func readDirNames(path string) ([]string, error) {
 	sort.Strings(paths)
 	return paths, nil
 }
-
-func configLogger(c *Config) error {
-	// set up our logging level
-	logLevel, err := logrus.ParseLevel(strings.ToLower(c.GetString("logging.level", "info")))
-	if err != nil {
-		return fmt.Errorf("%s; possible levels: %s", err, logrus.AllLevels)
-	}
-	l.SetLevel(logLevel)
-
-	timestampFormat := c.GetString("logging.timestamp_format", "")
-	fullTimestamp := (timestampFormat != "")
-	if timestampFormat == "" {
-		timestampFormat = time.RFC3339
-	}
-
-	logFormat := strings.ToLower(c.GetString("logging.format", "text"))
-	switch logFormat {
-	case "text":
-		l.Formatter = &logrus.TextFormatter{
-			TimestampFormat: timestampFormat,
-			FullTimestamp:   fullTimestamp,
-		}
-	case "json":
-		l.Formatter = &logrus.JSONFormatter{
-			TimestampFormat: timestampFormat,
-		}
-	default:
-		return fmt.Errorf("unknown log format `%s`. possible formats: %s", logFormat, []string{"text", "json"})
-	}
-
-	return nil
-}

config/config_test.go

@@ -1,30 +1,35 @@
-package nebula
+package config
 
 import (
-	"github.com/stretchr/testify/assert"
-	"io/ioutil"
 	"os"
 	"path/filepath"
 	"testing"
 	"time"
+
+	"dario.cat/mergo"
+	"github.com/slackhq/nebula/test"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"gopkg.in/yaml.v2"
 )
 
 func TestConfig_Load(t *testing.T) {
-	dir, err := ioutil.TempDir("", "config-test")
+	l := test.NewLogger()
+	dir, err := os.MkdirTemp("", "config-test")
 	// invalid yaml
-	c := NewConfig()
-	ioutil.WriteFile(filepath.Join(dir, "01.yaml"), []byte(" invalid yaml"), 0644)
+	c := NewC(l)
+	os.WriteFile(filepath.Join(dir, "01.yaml"), []byte(" invalid yaml"), 0644)
 	assert.EqualError(t, c.Load(dir), "yaml: unmarshal errors:\n  line 1: cannot unmarshal !!str `invalid...` into map[interface {}]interface {}")
 
 	// simple multi config merge
-	c = NewConfig()
+	c = NewC(l)
 	os.RemoveAll(dir)
 	os.Mkdir(dir, 0755)
 
 	assert.Nil(t, err)
 
-	ioutil.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: hi"), 0644)
-	ioutil.WriteFile(filepath.Join(dir, "02.yml"), []byte("outer:\n  inner: override\nnew: hi"), 0644)
+	os.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: hi"), 0644)
+	os.WriteFile(filepath.Join(dir, "02.yml"), []byte("outer:\n  inner: override\nnew: hi"), 0644)
 	assert.Nil(t, c.Load(dir))
 	expected := map[interface{}]interface{}{
 		"outer": map[interface{}]interface{}{
@@ -39,8 +44,9 @@ func TestConfig_Load(t *testing.T) {
 }
 
 func TestConfig_Get(t *testing.T) {
+	l := test.NewLogger()
 	// test simple type
-	c := NewConfig()
+	c := NewC(l)
 	c.Settings["firewall"] = map[interface{}]interface{}{"outbound": "hi"}
 	assert.Equal(t, "hi", c.Get("firewall.outbound"))
 
@@ -54,13 +60,15 @@ func TestConfig_Get(t *testing.T) {
 }
 
 func TestConfig_GetStringSlice(t *testing.T) {
-	c := NewConfig()
+	l := test.NewLogger()
+	c := NewC(l)
 	c.Settings["slice"] = []interface{}{"one", "two"}
 	assert.Equal(t, []string{"one", "two"}, c.GetStringSlice("slice", []string{}))
 }
 
 func TestConfig_GetBool(t *testing.T) {
-	c := NewConfig()
+	l := test.NewLogger()
+	c := NewC(l)
 	c.Settings["bool"] = true
 	assert.Equal(t, true, c.GetBool("bool", false))
 
@@ -86,91 +94,22 @@ func TestConfig_GetBool(t *testing.T) {
 	assert.Equal(t, false, c.GetBool("bool", true))
 }
 
-func TestConfig_GetAllowList(t *testing.T) {
-	c := NewConfig()
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"192.168.0.0": true,
-	}
-	r, err := c.GetAllowList("allowlist", false)
-	assert.EqualError(t, err, "config `allowlist` has invalid CIDR: 192.168.0.0")
-	assert.Nil(t, r)
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"192.168.0.0/16": "abc",
-	}
-	r, err = c.GetAllowList("allowlist", false)
-	assert.EqualError(t, err, "config `allowlist` has invalid value (type string): abc")
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"192.168.0.0/16": true,
-		"10.0.0.0/8":     false,
-	}
-	r, err = c.GetAllowList("allowlist", false)
-	assert.EqualError(t, err, "config `allowlist` contains both true and false rules, but no default set for 0.0.0.0/0")
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"0.0.0.0/0":     true,
-		"10.0.0.0/8":    false,
-		"10.42.42.0/24": true,
-	}
-	r, err = c.GetAllowList("allowlist", false)
-	if assert.NoError(t, err) {
-		assert.NotNil(t, r)
-	}
-
-	// Test interface names
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"interfaces": map[interface{}]interface{}{
-			`docker.*`: false,
-		},
-	}
-	r, err = c.GetAllowList("allowlist", false)
-	assert.EqualError(t, err, "config `allowlist` does not support `interfaces`")
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"interfaces": map[interface{}]interface{}{
-			`docker.*`: "foo",
-		},
-	}
-	r, err = c.GetAllowList("allowlist", true)
-	assert.EqualError(t, err, "config `allowlist.interfaces` has invalid value (type string): foo")
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"interfaces": map[interface{}]interface{}{
-			`docker.*`: false,
-			`eth.*`:    true,
-		},
-	}
-	r, err = c.GetAllowList("allowlist", true)
-	assert.EqualError(t, err, "config `allowlist.interfaces` values must all be the same true/false value")
-
-	c.Settings["allowlist"] = map[interface{}]interface{}{
-		"interfaces": map[interface{}]interface{}{
-			`docker.*`: false,
-		},
-	}
-	r, err = c.GetAllowList("allowlist", true)
-	if assert.NoError(t, err) {
-		assert.NotNil(t, r)
-	}
-}
-
 func TestConfig_HasChanged(t *testing.T) {
+	l := test.NewLogger()
 	// No reload has occurred, return false
-	c := NewConfig()
+	c := NewC(l)
 	c.Settings["test"] = "hi"
 	assert.False(t, c.HasChanged(""))
 
 	// Test key change
-	c = NewConfig()
+	c = NewC(l)
 	c.Settings["test"] = "hi"
 	c.oldSettings = map[interface{}]interface{}{"test": "no"}
 	assert.True(t, c.HasChanged("test"))
 	assert.True(t, c.HasChanged(""))
 
 	// No key change
-	c = NewConfig()
+	c = NewC(l)
 	c.Settings["test"] = "hi"
 	c.oldSettings = map[interface{}]interface{}{"test": "hi"}
 	assert.False(t, c.HasChanged("test"))
@@ -178,21 +117,22 @@ func TestConfig_HasChanged(t *testing.T) {
 }
 
 func TestConfig_ReloadConfig(t *testing.T) {
+	l := test.NewLogger()
 	done := make(chan bool, 1)
-	dir, err := ioutil.TempDir("", "config-test")
+	dir, err := os.MkdirTemp("", "config-test")
 	assert.Nil(t, err)
-	ioutil.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: hi"), 0644)
+	os.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: hi"), 0644)
 
-	c := NewConfig()
+	c := NewC(l)
 	assert.Nil(t, c.Load(dir))
 
 	assert.False(t, c.HasChanged("outer.inner"))
 	assert.False(t, c.HasChanged("outer"))
 	assert.False(t, c.HasChanged(""))
 
-	ioutil.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: ho"), 0644)
+	os.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: ho"), 0644)
 
-	c.RegisterReloadCallback(func(c *Config) {
+	c.RegisterReloadCallback(func(c *C) {
 		done <- true
 	})
 
@@ -209,3 +149,77 @@ func TestConfig_ReloadConfig(t *testing.T) {
 	}
 
 }
+
+// Ensure mergo merges are done the way we expect.
+// This is needed to test for potential regressions, like:
+// - https://github.com/imdario/mergo/issues/187
+func TestConfig_MergoMerge(t *testing.T) {
+	configs := [][]byte{
+		[]byte(`
+listen:
+  port: 1234
+`),
+		[]byte(`
+firewall:
+  inbound:
+    - port: 443
+      proto: tcp
+      groups:
+        - server
+    - port: 443
+      proto: tcp
+      groups:
+        - webapp
+`),
+		[]byte(`
+listen:
+  host: 0.0.0.0
+  port: 4242
+firewall:
+  outbound:
+    - port: any
+      proto: any
+      host: any
+  inbound:
+    - port: any
+      proto: icmp
+      host: any
+`),
+	}
+
+	var m map[any]any
+
+	// merge the same way config.parse() merges
+	for _, b := range configs {
+		var nm map[any]any
+		err := yaml.Unmarshal(b, &nm)
+		require.NoError(t, err)
+
+		// We need to use WithAppendSlice so that firewall rules in separate
+		// files are appended together
+		err = mergo.Merge(&nm, m, mergo.WithAppendSlice)
+		m = nm
+		require.NoError(t, err)
+	}
+
+	t.Logf("Merged Config: %#v", m)
+	mYaml, err := yaml.Marshal(m)
+	require.NoError(t, err)
+	t.Logf("Merged Config as YAML:\n%s", mYaml)
+
+	// If a bug is present, some items might be replaced instead of merged like we expect
+	expected := map[any]any{
+		"firewall": map[any]any{
+			"inbound": []any{
+				map[any]any{"host": "any", "port": "any", "proto": "icmp"},
+				map[any]any{"groups": []any{"server"}, "port": 443, "proto": "tcp"},
+				map[any]any{"groups": []any{"webapp"}, "port": 443, "proto": "tcp"}},
+			"outbound": []any{
+				map[any]any{"host": "any", "port": "any", "proto": "any"}}},
+		"listen": map[any]any{
+			"host": "0.0.0.0",
+			"port": 4242,
+		},
+	}
+	assert.Equal(t, expected, m)
+}

connection_manager.go

@@ -1,254 +1,519 @@
 package nebula
 
 import (
+	"bytes"
+	"context"
+	"encoding/binary"
+	"fmt"
+	"net/netip"
 	"sync"
+	"sync/atomic"
 	"time"
 
+	"github.com/rcrowley/go-metrics"
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/header"
 )
 
-// TODO: incount and outcount are intended as a shortcut to locking the mutexes for every single packet
-// and something like every 10 packets we could lock, send 10, then unlock for a moment
+type trafficDecision int
+
+const (
+	doNothing      trafficDecision = 0
+	deleteTunnel   trafficDecision = 1 // delete the hostinfo on our side, do not notify the remote
+	closeTunnel    trafficDecision = 2 // delete the hostinfo and notify the remote
+	swapPrimary    trafficDecision = 3
+	migrateRelays  trafficDecision = 4
+	tryRehandshake trafficDecision = 5
+	sendTestPacket trafficDecision = 6
+)
 
 type connectionManager struct {
+	// relayUsed holds which relay localIndexs are in use
+	relayUsed     map[uint32]struct{}
+	relayUsedLock *sync.RWMutex
+
 	hostMap      *HostMap
-	in           map[uint32]struct{}
-	inLock       *sync.RWMutex
-	inCount      int
-	out          map[uint32]struct{}
-	outLock      *sync.RWMutex
-	outCount     int
-	TrafficTimer *SystemTimerWheel
+	trafficTimer *LockingTimerWheel[uint32]
 	intf         *Interface
+	punchy       *Punchy
 
-	pendingDeletion      map[uint32]int
-	pendingDeletionLock  *sync.RWMutex
-	pendingDeletionTimer *SystemTimerWheel
+	// Configuration settings
+	checkInterval           time.Duration
+	pendingDeletionInterval time.Duration
+	inactivityTimeout       atomic.Int64
+	dropInactive            atomic.Bool
 
-	checkInterval           int
-	pendingDeletionInterval int
+	metricsTxPunchy metrics.Counter
 
-	// I wanted to call one matLock
+	l *logrus.Logger
 }
 
-func newConnectionManager(intf *Interface, checkInterval, pendingDeletionInterval int) *connectionManager {
-	nc := &connectionManager{
-		hostMap:                 intf.hostMap,
-		in:                      make(map[uint32]struct{}),
-		inLock:                  &sync.RWMutex{},
-		inCount:                 0,
-		out:                     make(map[uint32]struct{}),
-		outLock:                 &sync.RWMutex{},
-		outCount:                0,
-		TrafficTimer:            NewSystemTimerWheel(time.Millisecond*500, time.Second*60),
-		intf:                    intf,
-		pendingDeletion:         make(map[uint32]int),
-		pendingDeletionLock:     &sync.RWMutex{},
-		pendingDeletionTimer:    NewSystemTimerWheel(time.Millisecond*500, time.Second*60),
-		checkInterval:           checkInterval,
-		pendingDeletionInterval: pendingDeletionInterval,
+func newConnectionManagerFromConfig(l *logrus.Logger, c *config.C, hm *HostMap, p *Punchy) *connectionManager {
+	cm := &connectionManager{
+		hostMap:         hm,
+		l:               l,
+		punchy:          p,
+		relayUsed:       make(map[uint32]struct{}),
+		relayUsedLock:   &sync.RWMutex{},
+		metricsTxPunchy: metrics.GetOrRegisterCounter("messages.tx.punchy", nil),
 	}
-	nc.Start()
-	return nc
+
+	cm.reload(c, true)
+	c.RegisterReloadCallback(func(c *config.C) {
+		cm.reload(c, false)
+	})
+
+	return cm
 }
 
-func (n *connectionManager) In(ip uint32) {
-	n.inLock.RLock()
-	// If this already exists, return
-	if _, ok := n.in[ip]; ok {
-		n.inLock.RUnlock()
-		return
+func (cm *connectionManager) reload(c *config.C, initial bool) {
+	if initial {
+		cm.checkInterval = time.Duration(c.GetInt("timers.connection_alive_interval", 5)) * time.Second
+		cm.pendingDeletionInterval = time.Duration(c.GetInt("timers.pending_deletion_interval", 10)) * time.Second
+
+		// We want at least a minimum resolution of 500ms per tick so that we can hit these intervals
+		// pretty close to their configured duration.
+		// The inactivity duration is checked each time a hostinfo ticks through so we don't need the wheel to contain it.
+		minDuration := min(time.Millisecond*500, cm.checkInterval, cm.pendingDeletionInterval)
+		maxDuration := max(cm.checkInterval, cm.pendingDeletionInterval)
+		cm.trafficTimer = NewLockingTimerWheel[uint32](minDuration, maxDuration)
 	}
-	n.inLock.RUnlock()
-	n.inLock.Lock()
-	n.in[ip] = struct{}{}
-	n.inLock.Unlock()
-}
 
-func (n *connectionManager) Out(ip uint32) {
-	n.outLock.RLock()
-	// If this already exists, return
-	if _, ok := n.out[ip]; ok {
-		n.outLock.RUnlock()
-		return
-	}
-	n.outLock.RUnlock()
-	n.outLock.Lock()
-	// double check since we dropped the lock temporarily
-	if _, ok := n.out[ip]; ok {
-		n.outLock.Unlock()
-		return
-	}
-	n.out[ip] = struct{}{}
-	n.AddTrafficWatch(ip, n.checkInterval)
-	n.outLock.Unlock()
-}
-
-func (n *connectionManager) CheckIn(vpnIP uint32) bool {
-	n.inLock.RLock()
-	if _, ok := n.in[vpnIP]; ok {
-		n.inLock.RUnlock()
-		return true
-	}
-	n.inLock.RUnlock()
-	return false
-}
-
-func (n *connectionManager) ClearIP(ip uint32) {
-	n.inLock.Lock()
-	n.outLock.Lock()
-	delete(n.in, ip)
-	delete(n.out, ip)
-	n.inLock.Unlock()
-	n.outLock.Unlock()
-}
-
-func (n *connectionManager) ClearPendingDeletion(ip uint32) {
-	n.pendingDeletionLock.Lock()
-	delete(n.pendingDeletion, ip)
-	n.pendingDeletionLock.Unlock()
-}
-
-func (n *connectionManager) AddPendingDeletion(ip uint32) {
-	n.pendingDeletionLock.Lock()
-	if _, ok := n.pendingDeletion[ip]; ok {
-		n.pendingDeletion[ip] += 1
-	} else {
-		n.pendingDeletion[ip] = 0
-	}
-	n.pendingDeletionTimer.Add(ip, time.Second*time.Duration(n.pendingDeletionInterval))
-	n.pendingDeletionLock.Unlock()
-}
-
-func (n *connectionManager) checkPendingDeletion(ip uint32) bool {
-	n.pendingDeletionLock.RLock()
-	if _, ok := n.pendingDeletion[ip]; ok {
-
-		n.pendingDeletionLock.RUnlock()
-		return true
-	}
-	n.pendingDeletionLock.RUnlock()
-	return false
-}
-
-func (n *connectionManager) AddTrafficWatch(vpnIP uint32, seconds int) {
-	n.TrafficTimer.Add(vpnIP, time.Second*time.Duration(seconds))
-}
-
-func (n *connectionManager) Start() {
-	go n.Run()
-}
-
-func (n *connectionManager) Run() {
-	clockSource := time.Tick(500 * time.Millisecond)
-
-	for now := range clockSource {
-		n.HandleMonitorTick(now)
-		n.HandleDeletionTick(now)
-	}
-}
-
-func (n *connectionManager) HandleMonitorTick(now time.Time) {
-	n.TrafficTimer.advance(now)
-	for {
-		ep := n.TrafficTimer.Purge()
-		if ep == nil {
-			break
+	if initial || c.HasChanged("tunnels.inactivity_timeout") {
+		old := cm.getInactivityTimeout()
+		cm.inactivityTimeout.Store((int64)(c.GetDuration("tunnels.inactivity_timeout", 10*time.Minute)))
+		if !initial {
+			cm.l.WithField("oldDuration", old).
+				WithField("newDuration", cm.getInactivityTimeout()).
+				Info("Inactivity timeout has changed")
 		}
+	}
 
-		vpnIP := ep.(uint32)
+	if initial || c.HasChanged("tunnels.drop_inactive") {
+		old := cm.dropInactive.Load()
+		cm.dropInactive.Store(c.GetBool("tunnels.drop_inactive", false))
+		if !initial {
+			cm.l.WithField("oldBool", old).
+				WithField("newBool", cm.dropInactive.Load()).
+				Info("Drop inactive setting has changed")
+		}
+	}
+}
 
-		// Check for traffic coming back in from this host.
-		traf := n.CheckIn(vpnIP)
+func (cm *connectionManager) getInactivityTimeout() time.Duration {
+	return (time.Duration)(cm.inactivityTimeout.Load())
+}
 
-		// If we saw incoming packets from this ip, just return
-		if traf {
-			if l.Level >= logrus.DebugLevel {
-				l.WithField("vpnIp", IntIp(vpnIP)).
-					WithField("tunnelCheck", m{"state": "alive", "method": "passive"}).
-					Debug("Tunnel status")
+func (cm *connectionManager) In(h *HostInfo) {
+	h.in.Store(true)
+}
+
+func (cm *connectionManager) Out(h *HostInfo) {
+	h.out.Store(true)
+}
+
+func (cm *connectionManager) RelayUsed(localIndex uint32) {
+	cm.relayUsedLock.RLock()
+	// If this already exists, return
+	if _, ok := cm.relayUsed[localIndex]; ok {
+		cm.relayUsedLock.RUnlock()
+		return
+	}
+	cm.relayUsedLock.RUnlock()
+	cm.relayUsedLock.Lock()
+	cm.relayUsed[localIndex] = struct{}{}
+	cm.relayUsedLock.Unlock()
+}
+
+// getAndResetTrafficCheck returns if there was any inbound or outbound traffic within the last tick and
+// resets the state for this local index
+func (cm *connectionManager) getAndResetTrafficCheck(h *HostInfo, now time.Time) (bool, bool) {
+	in := h.in.Swap(false)
+	out := h.out.Swap(false)
+	if in || out {
+		h.lastUsed = now
+	}
+	return in, out
+}
+
+// AddTrafficWatch must be called for every new HostInfo.
+// We will continue to monitor the HostInfo until the tunnel is dropped.
+func (cm *connectionManager) AddTrafficWatch(h *HostInfo) {
+	if h.out.Swap(true) == false {
+		cm.trafficTimer.Add(h.localIndexId, cm.checkInterval)
+	}
+}
+
+func (cm *connectionManager) Start(ctx context.Context) {
+	clockSource := time.NewTicker(cm.trafficTimer.t.tickDuration)
+	defer clockSource.Stop()
+
+	p := []byte("")
+	nb := make([]byte, 12, 12)
+	out := make([]byte, mtu)
+
+	for {
+		select {
+		case <-ctx.Done():
+			return
+
+		case now := <-clockSource.C:
+			cm.trafficTimer.Advance(now)
+			for {
+				localIndex, has := cm.trafficTimer.Purge()
+				if !has {
+					break
+				}
+
+				cm.doTrafficCheck(localIndex, p, nb, out, now)
 			}
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			continue
 		}
-
-		// If we didn't we may need to probe or destroy the conn
-		hostinfo, err := n.hostMap.QueryVpnIP(vpnIP)
-		if err != nil {
-			l.Debugf("Not found in hostmap: %s", IntIp(vpnIP))
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			continue
-		}
-
-		hostinfo.logger().
-			WithField("tunnelCheck", m{"state": "testing", "method": "active"}).
-			Debug("Tunnel status")
-
-		if hostinfo != nil && hostinfo.ConnectionState != nil {
-			// Send a test packet to trigger an authenticated tunnel test, this should suss out any lingering tunnel issues
-			n.intf.SendMessageToVpnIp(test, testRequest, vpnIP, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
-
-		} else {
-			hostinfo.logger().Debugf("Hostinfo sadness: %s", IntIp(vpnIP))
-		}
-		n.AddPendingDeletion(vpnIP)
 	}
-
 }
 
-func (n *connectionManager) HandleDeletionTick(now time.Time) {
-	n.pendingDeletionTimer.advance(now)
-	for {
-		ep := n.pendingDeletionTimer.Purge()
-		if ep == nil {
-			break
+func (cm *connectionManager) doTrafficCheck(localIndex uint32, p, nb, out []byte, now time.Time) {
+	decision, hostinfo, primary := cm.makeTrafficDecision(localIndex, now)
+
+	switch decision {
+	case deleteTunnel:
+		if cm.hostMap.DeleteHostInfo(hostinfo) {
+			// Only clearing the lighthouse cache if this is the last hostinfo for this vpn ip in the hostmap
+			cm.intf.lightHouse.DeleteVpnIp(hostinfo.vpnIp)
 		}
 
-		vpnIP := ep.(uint32)
+	case closeTunnel:
+		cm.intf.sendCloseTunnel(hostinfo)
+		cm.intf.closeTunnel(hostinfo)
 
-		// If we saw incoming packets from this ip, just return
-		traf := n.CheckIn(vpnIP)
-		if traf {
-			l.WithField("vpnIp", IntIp(vpnIP)).
-				WithField("tunnelCheck", m{"state": "alive", "method": "active"}).
+	case swapPrimary:
+		cm.swapPrimary(hostinfo, primary)
+
+	case migrateRelays:
+		cm.migrateRelayUsed(hostinfo, primary)
+
+	case tryRehandshake:
+		cm.tryRehandshake(hostinfo)
+
+	case sendTestPacket:
+		cm.intf.SendMessageToHostInfo(header.Test, header.TestRequest, hostinfo, p, nb, out)
+	}
+
+	cm.resetRelayTrafficCheck(hostinfo)
+}
+
+func (cm *connectionManager) resetRelayTrafficCheck(hostinfo *HostInfo) {
+	if hostinfo != nil {
+		cm.relayUsedLock.Lock()
+		defer cm.relayUsedLock.Unlock()
+		// No need to migrate any relays, delete usage info now.
+		for _, idx := range hostinfo.relayState.CopyRelayForIdxs() {
+			delete(cm.relayUsed, idx)
+		}
+	}
+}
+
+func (cm *connectionManager) migrateRelayUsed(oldhostinfo, newhostinfo *HostInfo) {
+	relayFor := oldhostinfo.relayState.CopyAllRelayFor()
+
+	for _, r := range relayFor {
+		existing, ok := newhostinfo.relayState.QueryRelayForByIp(r.PeerIp)
+
+		var index uint32
+		var relayFrom netip.Addr
+		var relayTo netip.Addr
+		switch {
+		case ok:
+			switch existing.State {
+			case Established, PeerRequested, Disestablished:
+				// This relay already exists in newhostinfo, then do nothing.
+				continue
+			case Requested:
+				// The relayed connection exists in a Requested state; re-send the request
+				index = existing.LocalIndex
+				switch r.Type {
+				case TerminalType:
+					relayFrom = cm.intf.myVpnNet.Addr()
+					relayTo = existing.PeerIp
+				case ForwardingType:
+					relayFrom = existing.PeerIp
+					relayTo = newhostinfo.vpnIp
+				default:
+					// should never happen
+					panic(fmt.Sprintf("Migrating unknown relay type: %v", r.Type))
+				}
+			}
+		case !ok:
+			cm.relayUsedLock.RLock()
+			if _, relayUsed := cm.relayUsed[r.LocalIndex]; !relayUsed {
+				// The relay hasn't been used; don't migrate it.
+				cm.relayUsedLock.RUnlock()
+				continue
+			}
+			cm.relayUsedLock.RUnlock()
+			// The relay doesn't exist at all; create some relay state and send the request.
+			var err error
+			index, err = AddRelay(cm.l, newhostinfo, cm.hostMap, r.PeerIp, nil, r.Type, Requested)
+			if err != nil {
+				cm.l.WithError(err).Error("failed to migrate relay to new hostinfo")
+				continue
+			}
+			switch r.Type {
+			case TerminalType:
+				relayFrom = cm.intf.myVpnNet.Addr()
+				relayTo = r.PeerIp
+			case ForwardingType:
+				relayFrom = r.PeerIp
+				relayTo = newhostinfo.vpnIp
+			default:
+				// should never happen
+				panic(fmt.Sprintf("Migrating unknown relay type: %v", r.Type))
+			}
+		}
+
+		//TODO: IPV6-WORK
+		relayFromB := relayFrom.As4()
+		relayToB := relayTo.As4()
+
+		// Send a CreateRelayRequest to the peer.
+		req := NebulaControl{
+			Type:                NebulaControl_CreateRelayRequest,
+			InitiatorRelayIndex: index,
+			RelayFromIp:         binary.BigEndian.Uint32(relayFromB[:]),
+			RelayToIp:           binary.BigEndian.Uint32(relayToB[:]),
+		}
+		msg, err := req.Marshal()
+		if err != nil {
+			cm.l.WithError(err).Error("failed to marshal Control message to migrate relay")
+		} else {
+			cm.intf.SendMessageToHostInfo(header.Control, 0, newhostinfo, msg, make([]byte, 12), make([]byte, mtu))
+			cm.l.WithFields(logrus.Fields{
+				"relayFrom":           req.RelayFromIp,
+				"relayTo":             req.RelayToIp,
+				"initiatorRelayIndex": req.InitiatorRelayIndex,
+				"responderRelayIndex": req.ResponderRelayIndex,
+				"vpnIp":               newhostinfo.vpnIp}).
+				Info("send CreateRelayRequest")
+		}
+	}
+}
+
+func (cm *connectionManager) makeTrafficDecision(localIndex uint32, now time.Time) (trafficDecision, *HostInfo, *HostInfo) {
+	// Read lock the main hostmap to order decisions based on tunnels being the primary tunnel
+	cm.hostMap.RLock()
+	defer cm.hostMap.RUnlock()
+
+	hostinfo := cm.hostMap.Indexes[localIndex]
+	if hostinfo == nil {
+		cm.l.WithField("localIndex", localIndex).Debugln("Not found in hostmap")
+		return doNothing, nil, nil
+	}
+
+	if cm.isInvalidCertificate(now, hostinfo) {
+		return closeTunnel, hostinfo, nil
+	}
+
+	primary := cm.hostMap.Hosts[hostinfo.vpnIp]
+	mainHostInfo := true
+	if primary != nil && primary != hostinfo {
+		mainHostInfo = false
+	}
+
+	// Check for traffic on this hostinfo
+	inTraffic, outTraffic := cm.getAndResetTrafficCheck(hostinfo, now)
+
+	// A hostinfo is determined alive if there is incoming traffic
+	if inTraffic {
+		decision := doNothing
+		if cm.l.Level >= logrus.DebugLevel {
+			hostinfo.logger(cm.l).
+				WithField("tunnelCheck", m{"state": "alive", "method": "passive"}).
 				Debug("Tunnel status")
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			continue
 		}
+		hostinfo.pendingDeletion.Store(false)
 
-		hostinfo, err := n.hostMap.QueryVpnIP(vpnIP)
-		if err != nil {
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			l.Debugf("Not found in hostmap: %s", IntIp(vpnIP))
-			continue
-		}
+		if mainHostInfo {
+			decision = tryRehandshake
 
-		// If it comes around on deletion wheel and hasn't resolved itself, delete
-		if n.checkPendingDeletion(vpnIP) {
-			cn := ""
-			if hostinfo.ConnectionState != nil && hostinfo.ConnectionState.peerCert != nil {
-				cn = hostinfo.ConnectionState.peerCert.Details.Name
-			}
-			hostinfo.logger().
-				WithField("tunnelCheck", m{"state": "dead", "method": "active"}).
-				WithField("certName", cn).
-				Info("Tunnel status")
-
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			// TODO: This is only here to let tests work. Should do proper mocking
-			if n.intf.lightHouse != nil {
-				n.intf.lightHouse.DeleteVpnIP(vpnIP)
-			}
-			n.hostMap.DeleteVpnIP(vpnIP)
-			n.hostMap.DeleteIndex(hostinfo.localIndexId)
 		} else {
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
+			if cm.shouldSwapPrimary(hostinfo, primary) {
+				decision = swapPrimary
+			} else {
+				// migrate the relays to the primary, if in use.
+				decision = migrateRelays
+			}
+		}
+
+		cm.trafficTimer.Add(hostinfo.localIndexId, cm.checkInterval)
+
+		if !outTraffic {
+			// Send a punch packet to keep the NAT state alive
+			cm.sendPunch(hostinfo)
+		}
+
+		return decision, hostinfo, primary
+	}
+
+	if hostinfo.pendingDeletion.Load() {
+		// We have already sent a test packet and nothing was returned, this hostinfo is dead
+		hostinfo.logger(cm.l).
+			WithField("tunnelCheck", m{"state": "dead", "method": "active"}).
+			Info("Tunnel status")
+
+		return deleteTunnel, hostinfo, nil
+	}
+
+	decision := doNothing
+	if hostinfo != nil && hostinfo.ConnectionState != nil && mainHostInfo {
+		if !outTraffic {
+			inactiveFor, isInactive := cm.isInactive(hostinfo, now)
+			if isInactive {
+				// Tunnel is inactive, tear it down
+				hostinfo.logger(cm.l).
+					WithField("inactiveDuration", inactiveFor).
+					WithField("primary", mainHostInfo).
+					Info("Dropping tunnel due to inactivity")
+
+				return closeTunnel, hostinfo, primary
+			}
+
+			// If we aren't sending or receiving traffic then its an unused tunnel and we don't to test the tunnel.
+			// Just maintain NAT state if configured to do so.
+			cm.sendPunch(hostinfo)
+			cm.trafficTimer.Add(hostinfo.localIndexId, cm.checkInterval)
+			return doNothing, nil, nil
+		}
+
+		if cm.punchy.GetTargetEverything() {
+			// This is similar to the old punchy behavior with a slight optimization.
+			// We aren't receiving traffic but we are sending it, punch on all known
+			// ips in case we need to re-prime NAT state
+			cm.sendPunch(hostinfo)
+		}
+
+		if cm.l.Level >= logrus.DebugLevel {
+			hostinfo.logger(cm.l).
+				WithField("tunnelCheck", m{"state": "testing", "method": "active"}).
+				Debug("Tunnel status")
+		}
+
+		// Send a test packet to trigger an authenticated tunnel test, this should suss out any lingering tunnel issues
+		decision = sendTestPacket
+
+	} else {
+		if cm.l.Level >= logrus.DebugLevel {
+			hostinfo.logger(cm.l).Debugf("Hostinfo sadness")
 		}
 	}
+
+	hostinfo.pendingDeletion.Store(true)
+	cm.trafficTimer.Add(hostinfo.localIndexId, cm.pendingDeletionInterval)
+	return decision, hostinfo, nil
+}
+
+func (cm *connectionManager) isInactive(hostinfo *HostInfo, now time.Time) (time.Duration, bool) {
+	if cm.dropInactive.Load() == false {
+		// We aren't configured to drop inactive tunnels
+		return 0, false
+	}
+
+	inactiveDuration := now.Sub(hostinfo.lastUsed)
+	if inactiveDuration < cm.getInactivityTimeout() {
+		// It's not considered inactive
+		return inactiveDuration, false
+	}
+
+	// The tunnel is inactive
+	return inactiveDuration, true
+}
+
+func (cm *connectionManager) shouldSwapPrimary(current, primary *HostInfo) bool {
+	// The primary tunnel is the most recent handshake to complete locally and should work entirely fine.
+	// If we are here then we have multiple tunnels for a host pair and neither side believes the same tunnel is primary.
+	// Let's sort this out.
+
+	if current.vpnIp.Compare(cm.intf.myVpnNet.Addr()) < 0 {
+		// Only one side should flip primary because if both flip then we may never resolve to a single tunnel.
+		// vpn ip is static across all tunnels for this host pair so lets use that to determine who is flipping.
+		// The remotes vpn ip is lower than mine. I will not flip.
+		return false
+	}
+
+	certState := cm.intf.pki.GetCertState()
+	return bytes.Equal(current.ConnectionState.myCert.Signature, certState.Certificate.Signature)
+}
+
+func (cm *connectionManager) swapPrimary(current, primary *HostInfo) {
+	cm.hostMap.Lock()
+	// Make sure the primary is still the same after the write lock. This avoids a race with a rehandshake.
+	if cm.hostMap.Hosts[current.vpnIp] == primary {
+		cm.hostMap.unlockedMakePrimary(current)
+	}
+	cm.hostMap.Unlock()
+}
+
+// isInvalidCertificate will check if we should destroy a tunnel if pki.disconnect_invalid is true and
+// the certificate is no longer valid. Block listed certificates will skip the pki.disconnect_invalid
+// check and return true.
+func (cm *connectionManager) isInvalidCertificate(now time.Time, hostinfo *HostInfo) bool {
+	remoteCert := hostinfo.GetCert()
+	if remoteCert == nil {
+		return false
+	}
+
+	valid, err := remoteCert.VerifyWithCache(now, cm.intf.pki.GetCAPool())
+	if valid {
+		return false
+	}
+
+	if !cm.intf.disconnectInvalid.Load() && err != cert.ErrBlockListed {
+		// Block listed certificates should always be disconnected
+		return false
+	}
+
+	fingerprint, _ := remoteCert.Sha256Sum()
+	hostinfo.logger(cm.l).WithError(err).
+		WithField("fingerprint", fingerprint).
+		Info("Remote certificate is no longer valid, tearing down the tunnel")
+
+	return true
+}
+
+func (cm *connectionManager) sendPunch(hostinfo *HostInfo) {
+	if !cm.punchy.GetPunch() {
+		// Punching is disabled
+		return
+	}
+
+	if cm.intf.lightHouse.IsLighthouseIP(hostinfo.vpnIp) {
+		// Do not punch to lighthouses, we assume our lighthouse update interval is good enough.
+		// In the event the update interval is not sufficient to maintain NAT state then a publicly available lighthouse
+		// would lose the ability to notify us and punchy.respond would become unreliable.
+		return
+	}
+
+	if cm.punchy.GetTargetEverything() {
+		hostinfo.remotes.ForEach(cm.hostMap.GetPreferredRanges(), func(addr netip.AddrPort, preferred bool) {
+			cm.metricsTxPunchy.Inc(1)
+			cm.intf.outside.WriteTo([]byte{1}, addr)
+		})
+
+	} else if hostinfo.remote.IsValid() {
+		cm.metricsTxPunchy.Inc(1)
+		cm.intf.outside.WriteTo([]byte{1}, hostinfo.remote)
+	}
+}
+
+func (cm *connectionManager) tryRehandshake(hostinfo *HostInfo) {
+	certState := cm.intf.pki.GetCertState()
+	if bytes.Equal(hostinfo.ConnectionState.myCert.Signature, certState.Certificate.Signature) {
+		return
+	}
+
+	cm.l.WithField("vpnIp", hostinfo.vpnIp).
+		WithField("reason", "local certificate is not current").
+		Info("Re-handshaking with remote")
+
+	cm.intf.handshakeManager.StartHandshake(hostinfo.vpnIp, nil)
 }

connection_manager_test.go

@@ -1,142 +1,402 @@
 package nebula
 
 import (
+	"crypto/ed25519"
+	"crypto/rand"
 	"net"
+	"net/netip"
 	"testing"
 	"time"
 
 	"github.com/flynn/noise"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/test"
+	"github.com/slackhq/nebula/udp"
 	"github.com/stretchr/testify/assert"
 )
 
-var vpnIP uint32
+func newTestLighthouse() *LightHouse {
+	lh := &LightHouse{
+		l:         test.NewLogger(),
+		addrMap:   map[netip.Addr]*RemoteList{},
+		queryChan: make(chan netip.Addr, 10),
+	}
+	lighthouses := map[netip.Addr]struct{}{}
+	staticList := map[netip.Addr]struct{}{}
+
+	lh.lighthouses.Store(&lighthouses)
+	lh.staticList.Store(&staticList)
+
+	return lh
+}
 
 func Test_NewConnectionManagerTest(t *testing.T) {
+	l := test.NewLogger()
 	//_, tuncidr, _ := net.ParseCIDR("1.1.1.1/24")
-	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	vpnIP = ip2int(net.ParseIP("172.1.1.2"))
-	preferredRanges := []*net.IPNet{localrange}
+	vpncidr := netip.MustParsePrefix("172.1.1.1/24")
+	localrange := netip.MustParsePrefix("10.1.1.1/24")
+	vpnIp := netip.MustParseAddr("172.1.1.2")
+	preferredRanges := []netip.Prefix{localrange}
 
 	// Very incomplete mock objects
-	hostMap := NewHostMap("test", vpncidr, preferredRanges)
+	hostMap := newHostMap(l, vpncidr)
+	hostMap.preferredRanges.Store(&preferredRanges)
+
 	cs := &CertState{
-		rawCertificate:      []byte{},
-		privateKey:          []byte{},
-		certificate:         &cert.NebulaCertificate{},
-		rawCertificateNoKey: []byte{},
+		RawCertificate:      []byte{},
+		PrivateKey:          []byte{},
+		Certificate:         &cert.NebulaCertificate{},
+		RawCertificateNoKey: []byte{},
 	}
 
-	lh := NewLightHouse(false, 0, []uint32{}, 1000, 0, &udpConn{}, false, 1)
+	lh := newTestLighthouse()
 	ifce := &Interface{
 		hostMap:          hostMap,
-		inside:           &Tun{},
-		outside:          &udpConn{},
-		certState:        cs,
+		inside:           &test.NoopTun{},
+		outside:          &udp.NoopConn{},
 		firewall:         &Firewall{},
 		lightHouse:       lh,
-		handshakeManager: NewHandshakeManager(vpncidr, preferredRanges, hostMap, lh, &udpConn{}, defaultHandshakeConfig),
+		pki:              &PKI{},
+		handshakeManager: NewHandshakeManager(l, hostMap, lh, &udp.NoopConn{}, defaultHandshakeConfig),
+		l:                l,
 	}
-	now := time.Now()
+	ifce.pki.cs.Store(cs)
 
 	// Create manager
-	nc := newConnectionManager(ifce, 5, 10)
-	nc.HandleMonitorTick(now)
+	conf := config.NewC(l)
+	punchy := NewPunchyFromConfig(l, conf)
+	nc := newConnectionManagerFromConfig(l, conf, hostMap, punchy)
+	nc.intf = ifce
+	p := []byte("")
+	nb := make([]byte, 12, 12)
+	out := make([]byte, mtu)
+
 	// Add an ip we have established a connection w/ to hostmap
-	hostinfo := nc.hostMap.AddVpnIP(vpnIP)
-	hostinfo.ConnectionState = &ConnectionState{
-		certState:      cs,
-		H:              &noise.HandshakeState{},
-		messageCounter: new(uint64),
+	hostinfo := &HostInfo{
+		vpnIp:         vpnIp,
+		localIndexId:  1099,
+		remoteIndexId: 9901,
 	}
+	hostinfo.ConnectionState = &ConnectionState{
+		myCert: &cert.NebulaCertificate{},
+		H:      &noise.HandshakeState{},
+	}
+	nc.hostMap.unlockedAddHostInfo(hostinfo, ifce)
 
-	// We saw traffic out to vpnIP
-	nc.Out(vpnIP)
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
-	// Move ahead 5s. Nothing should happen
-	next_tick := now.Add(5 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// Move ahead 6s. We haven't heard back
-	next_tick = now.Add(6 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// This host should now be up for deletion
-	assert.Contains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
-	// Move ahead some more
-	next_tick = now.Add(45 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// The host should be evicted
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.NotContains(t, nc.hostMap.Hosts, vpnIP)
+	// We saw traffic out to vpnIp
+	nc.Out(hostinfo)
+	nc.In(hostinfo)
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.True(t, hostinfo.out.Load())
+	assert.True(t, hostinfo.in.Load())
 
+	// Do a traffic check tick, should not be pending deletion but should not have any in/out packets recorded
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+
+	// Do another traffic check tick, this host should be pending deletion now
+	nc.Out(hostinfo)
+	assert.True(t, hostinfo.out.Load())
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.True(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+
+	// Do a final traffic check tick, the host should now be removed
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.NotContains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+	assert.NotContains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
 }
 
 func Test_NewConnectionManagerTest2(t *testing.T) {
+	l := test.NewLogger()
 	//_, tuncidr, _ := net.ParseCIDR("1.1.1.1/24")
-	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	preferredRanges := []*net.IPNet{localrange}
+	vpncidr := netip.MustParsePrefix("172.1.1.1/24")
+	localrange := netip.MustParsePrefix("10.1.1.1/24")
+	vpnIp := netip.MustParseAddr("172.1.1.2")
+	preferredRanges := []netip.Prefix{localrange}
 
 	// Very incomplete mock objects
-	hostMap := NewHostMap("test", vpncidr, preferredRanges)
+	hostMap := newHostMap(l, vpncidr)
+	hostMap.preferredRanges.Store(&preferredRanges)
+
 	cs := &CertState{
-		rawCertificate:      []byte{},
-		privateKey:          []byte{},
-		certificate:         &cert.NebulaCertificate{},
-		rawCertificateNoKey: []byte{},
+		RawCertificate:      []byte{},
+		PrivateKey:          []byte{},
+		Certificate:         &cert.NebulaCertificate{},
+		RawCertificateNoKey: []byte{},
 	}
 
-	lh := NewLightHouse(false, 0, []uint32{}, 1000, 0, &udpConn{}, false, 1)
+	lh := newTestLighthouse()
 	ifce := &Interface{
 		hostMap:          hostMap,
-		inside:           &Tun{},
-		outside:          &udpConn{},
-		certState:        cs,
+		inside:           &test.NoopTun{},
+		outside:          &udp.NoopConn{},
 		firewall:         &Firewall{},
 		lightHouse:       lh,
-		handshakeManager: NewHandshakeManager(vpncidr, preferredRanges, hostMap, lh, &udpConn{}, defaultHandshakeConfig),
+		pki:              &PKI{},
+		handshakeManager: NewHandshakeManager(l, hostMap, lh, &udp.NoopConn{}, defaultHandshakeConfig),
+		l:                l,
 	}
-	now := time.Now()
+	ifce.pki.cs.Store(cs)
 
 	// Create manager
-	nc := newConnectionManager(ifce, 5, 10)
-	nc.HandleMonitorTick(now)
+	conf := config.NewC(l)
+	punchy := NewPunchyFromConfig(l, conf)
+	nc := newConnectionManagerFromConfig(l, conf, hostMap, punchy)
+	nc.intf = ifce
+	p := []byte("")
+	nb := make([]byte, 12, 12)
+	out := make([]byte, mtu)
+
 	// Add an ip we have established a connection w/ to hostmap
-	hostinfo := nc.hostMap.AddVpnIP(vpnIP)
+	hostinfo := &HostInfo{
+		vpnIp:         vpnIp,
+		localIndexId:  1099,
+		remoteIndexId: 9901,
+	}
 	hostinfo.ConnectionState = &ConnectionState{
-		certState:      cs,
-		H:              &noise.HandshakeState{},
-		messageCounter: new(uint64),
+		myCert: &cert.NebulaCertificate{},
+		H:      &noise.HandshakeState{},
+	}
+	nc.hostMap.unlockedAddHostInfo(hostinfo, ifce)
+
+	// We saw traffic out to vpnIp
+	nc.Out(hostinfo)
+	nc.In(hostinfo)
+	assert.True(t, hostinfo.in.Load())
+	assert.True(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+
+	// Do a traffic check tick, should not be pending deletion but should not have any in/out packets recorded
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+
+	// Do another traffic check tick, this host should be pending deletion now
+	nc.Out(hostinfo)
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.True(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+
+	// We saw traffic, should no longer be pending deletion
+	nc.In(hostinfo)
+	nc.doTrafficCheck(hostinfo.localIndexId, p, nb, out, time.Now())
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+}
+
+func Test_NewConnectionManager_DisconnectInactive(t *testing.T) {
+	l := test.NewLogger()
+	vpncidr := netip.MustParsePrefix("172.1.1.1/24")
+	localrange := netip.MustParsePrefix("10.1.1.1/24")
+	vpnIp := netip.MustParseAddr("172.1.1.2")
+	preferredRanges := []netip.Prefix{localrange}
+
+	// Very incomplete mock objects
+	hostMap := newHostMap(l, vpncidr)
+	hostMap.preferredRanges.Store(&preferredRanges)
+
+	cs := &CertState{
+		RawCertificate:      []byte{},
+		PrivateKey:          []byte{},
+		Certificate:         &cert.NebulaCertificate{},
+		RawCertificateNoKey: []byte{},
 	}
 
-	// We saw traffic out to vpnIP
-	nc.Out(vpnIP)
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
-	// Move ahead 5s. Nothing should happen
-	next_tick := now.Add(5 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// Move ahead 6s. We haven't heard back
-	next_tick = now.Add(6 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// This host should now be up for deletion
-	assert.Contains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
-	// We heard back this time
-	nc.In(vpnIP)
-	// Move ahead some more
-	next_tick = now.Add(45 * time.Second)
-	nc.HandleMonitorTick(next_tick)
-	nc.HandleDeletionTick(next_tick)
-	// The host should be evicted
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
+	lh := newTestLighthouse()
+	ifce := &Interface{
+		hostMap:          hostMap,
+		inside:           &test.NoopTun{},
+		outside:          &udp.NoopConn{},
+		firewall:         &Firewall{},
+		lightHouse:       lh,
+		pki:              &PKI{},
+		handshakeManager: NewHandshakeManager(l, hostMap, lh, &udp.NoopConn{}, defaultHandshakeConfig),
+		l:                l,
+	}
+	ifce.pki.cs.Store(cs)
 
+	// Create manager
+	conf := config.NewC(l)
+	conf.Settings["tunnels"] = map[interface{}]interface{}{
+		"drop_inactive": true,
+	}
+	punchy := NewPunchyFromConfig(l, conf)
+	nc := newConnectionManagerFromConfig(l, conf, hostMap, punchy)
+	assert.True(t, nc.dropInactive.Load())
+	nc.intf = ifce
+
+	// Add an ip we have established a connection w/ to hostmap
+	hostinfo := &HostInfo{
+		vpnIp:         vpnIp,
+		localIndexId:  1099,
+		remoteIndexId: 9901,
+	}
+	hostinfo.ConnectionState = &ConnectionState{
+		myCert: &cert.NebulaCertificate{},
+		H:      &noise.HandshakeState{},
+	}
+	nc.hostMap.unlockedAddHostInfo(hostinfo, ifce)
+
+	// Do a traffic check tick, in and out should be cleared but should not be pending deletion
+	nc.Out(hostinfo)
+	nc.In(hostinfo)
+	assert.True(t, hostinfo.out.Load())
+	assert.True(t, hostinfo.in.Load())
+
+	now := time.Now()
+	decision, _, _ := nc.makeTrafficDecision(hostinfo.localIndexId, now)
+	assert.Equal(t, tryRehandshake, decision)
+	assert.Equal(t, now, hostinfo.lastUsed)
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+
+	decision, _, _ = nc.makeTrafficDecision(hostinfo.localIndexId, now.Add(time.Second*5))
+	assert.Equal(t, doNothing, decision)
+	assert.Equal(t, now, hostinfo.lastUsed)
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+
+	// Do another traffic check tick, should still not be pending deletion
+	decision, _, _ = nc.makeTrafficDecision(hostinfo.localIndexId, now.Add(time.Second*10))
+	assert.Equal(t, doNothing, decision)
+	assert.Equal(t, now, hostinfo.lastUsed)
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+
+	// Finally advance beyond the inactivity timeout
+	decision, _, _ = nc.makeTrafficDecision(hostinfo.localIndexId, now.Add(time.Minute*10))
+	assert.Equal(t, closeTunnel, decision)
+	assert.Equal(t, now, hostinfo.lastUsed)
+	assert.False(t, hostinfo.pendingDeletion.Load())
+	assert.False(t, hostinfo.out.Load())
+	assert.False(t, hostinfo.in.Load())
+	assert.Contains(t, nc.hostMap.Indexes, hostinfo.localIndexId)
+	assert.Contains(t, nc.hostMap.Hosts, hostinfo.vpnIp)
+}
+
+// Check if we can disconnect the peer.
+// Validate if the peer's certificate is invalid (expired, etc.)
+// Disconnect only if disconnectInvalid: true is set.
+func Test_NewConnectionManagerTest_DisconnectInvalid(t *testing.T) {
+	now := time.Now()
+	l := test.NewLogger()
+	ipNet := net.IPNet{
+		IP:   net.IPv4(172, 1, 1, 2),
+		Mask: net.IPMask{255, 255, 255, 0},
+	}
+	vpncidr := netip.MustParsePrefix("172.1.1.1/24")
+	localrange := netip.MustParsePrefix("10.1.1.1/24")
+	vpnIp := netip.MustParseAddr("172.1.1.2")
+	preferredRanges := []netip.Prefix{localrange}
+	hostMap := newHostMap(l, vpncidr)
+	hostMap.preferredRanges.Store(&preferredRanges)
+
+	// Generate keys for CA and peer's cert.
+	pubCA, privCA, _ := ed25519.GenerateKey(rand.Reader)
+	caCert := cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:      "ca",
+			NotBefore: now,
+			NotAfter:  now.Add(1 * time.Hour),
+			IsCA:      true,
+			PublicKey: pubCA,
+		},
+	}
+
+	assert.NoError(t, caCert.Sign(cert.Curve_CURVE25519, privCA))
+	ncp := &cert.NebulaCAPool{
+		CAs: cert.NewCAPool().CAs,
+	}
+	ncp.CAs["ca"] = &caCert
+
+	pubCrt, _, _ := ed25519.GenerateKey(rand.Reader)
+	peerCert := cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:      "host",
+			Ips:       []*net.IPNet{&ipNet},
+			Subnets:   []*net.IPNet{},
+			NotBefore: now,
+			NotAfter:  now.Add(60 * time.Second),
+			PublicKey: pubCrt,
+			IsCA:      false,
+			Issuer:    "ca",
+		},
+	}
+	assert.NoError(t, peerCert.Sign(cert.Curve_CURVE25519, privCA))
+
+	cs := &CertState{
+		RawCertificate:      []byte{},
+		PrivateKey:          []byte{},
+		Certificate:         &cert.NebulaCertificate{},
+		RawCertificateNoKey: []byte{},
+	}
+
+	lh := newTestLighthouse()
+	ifce := &Interface{
+		hostMap:          hostMap,
+		inside:           &test.NoopTun{},
+		outside:          &udp.NoopConn{},
+		firewall:         &Firewall{},
+		lightHouse:       lh,
+		handshakeManager: NewHandshakeManager(l, hostMap, lh, &udp.NoopConn{}, defaultHandshakeConfig),
+		l:                l,
+		pki:              &PKI{},
+	}
+	ifce.pki.cs.Store(cs)
+	ifce.pki.caPool.Store(ncp)
+	ifce.disconnectInvalid.Store(true)
+
+	// Create manager
+	conf := config.NewC(l)
+	punchy := NewPunchyFromConfig(l, conf)
+	nc := newConnectionManagerFromConfig(l, conf, hostMap, punchy)
+	nc.intf = ifce
+	ifce.connectionManager = nc
+
+	hostinfo := &HostInfo{
+		vpnIp: vpnIp,
+		ConnectionState: &ConnectionState{
+			myCert:   &cert.NebulaCertificate{},
+			peerCert: &peerCert,
+			H:        &noise.HandshakeState{},
+		},
+	}
+	nc.hostMap.unlockedAddHostInfo(hostinfo, ifce)
+
+	// Move ahead 45s.
+	// Check if to disconnect with invalid certificate.
+	// Should be alive.
+	nextTick := now.Add(45 * time.Second)
+	invalid := nc.isInvalidCertificate(nextTick, hostinfo)
+	assert.False(t, invalid)
+
+	// Move ahead 61s.
+	// Check if to disconnect with invalid certificate.
+	// Should be disconnected.
+	nextTick = now.Add(61 * time.Second)
+	invalid = nc.isInvalidCertificate(nextTick, hostinfo)
+	assert.True(t, invalid)
 }

connection_state.go

@@ -4,9 +4,12 @@ import (
 	"crypto/rand"
 	"encoding/json"
 	"sync"
+	"sync/atomic"
 
 	"github.com/flynn/noise"
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/noiseutil"
 )
 
 const ReplayWindow = 1024
@@ -15,28 +18,38 @@ type ConnectionState struct {
 	eKey           *NebulaCipherState
 	dKey           *NebulaCipherState
 	H              *noise.HandshakeState
-	certState      *CertState
+	myCert         *cert.NebulaCertificate
 	peerCert       *cert.NebulaCertificate
 	initiator      bool
-	messageCounter *uint64
+	messageCounter atomic.Uint64
 	window         *Bits
-	queueLock      sync.Mutex
 	writeLock      sync.Mutex
-	ready          bool
 }
 
-func (f *Interface) newConnectionState(initiator bool, pattern noise.HandshakePattern, psk []byte, pskStage int) *ConnectionState {
-	cs := noise.NewCipherSuite(noise.DH25519, noise.CipherAESGCM, noise.HashSHA256)
-	if f.cipher == "chachapoly" {
-		cs = noise.NewCipherSuite(noise.DH25519, noise.CipherChaChaPoly, noise.HashSHA256)
+func NewConnectionState(l *logrus.Logger, cipher string, certState *CertState, initiator bool, pattern noise.HandshakePattern, psk []byte, pskStage int) *ConnectionState {
+	var dhFunc noise.DHFunc
+	switch certState.Certificate.Details.Curve {
+	case cert.Curve_CURVE25519:
+		dhFunc = noise.DH25519
+	case cert.Curve_P256:
+		dhFunc = noiseutil.DHP256
+	default:
+		l.Errorf("invalid curve: %s", certState.Certificate.Details.Curve)
+		return nil
 	}
 
-	curCertState := f.certState
-	static := noise.DHKey{Private: curCertState.privateKey, Public: curCertState.publicKey}
+	var cs noise.CipherSuite
+	if cipher == "chachapoly" {
+		cs = noise.NewCipherSuite(dhFunc, noise.CipherChaChaPoly, noise.HashSHA256)
+	} else {
+		cs = noise.NewCipherSuite(dhFunc, noiseutil.CipherAESGCM, noise.HashSHA256)
+	}
+
+	static := noise.DHKey{Private: certState.PrivateKey, Public: certState.PublicKey}
 
 	b := NewBits(ReplayWindow)
 	// Clear out bit 0, we never transmit it and we don't want it showing as packet loss
-	b.Update(0)
+	b.Update(l, 0)
 
 	hs, err := noise.NewHandshakeState(noise.Config{
 		CipherSuite:           cs,
@@ -54,13 +67,13 @@ func (f *Interface) newConnectionState(initiator bool, pattern noise.HandshakePa
 	// The queue and ready params prevent a counter race that would happen when
 	// sending stored packets and simultaneously accepting new traffic.
 	ci := &ConnectionState{
-		H:              hs,
-		initiator:      initiator,
-		window:         b,
-		ready:          false,
-		certState:      curCertState,
-		messageCounter: new(uint64),
+		H:         hs,
+		initiator: initiator,
+		window:    b,
+		myCert:    certState.Certificate,
 	}
+	// always start the counter from 2, as packet 1 and packet 2 are handshake packets.
+	ci.messageCounter.Add(2)
 
 	return ci
 }
@@ -69,7 +82,6 @@ func (cs *ConnectionState) MarshalJSON() ([]byte, error) {
 	return json.Marshal(m{
 		"certificate":     cs.peerCert,
 		"initiator":       cs.initiator,
-		"message_counter": cs.messageCounter,
-		"ready":           cs.ready,
+		"message_counter": cs.messageCounter.Load(),
 	})
 }

control.go

@@ -0,0 +1,319 @@
+package nebula
+
+import (
+	"context"
+	"net/netip"
+	"os"
+	"os/signal"
+	"syscall"
+
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/overlay"
+)
+
+// Every interaction here needs to take extra care to copy memory and not return or use arguments "as is" when touching
+// core. This means copying IP objects, slices, de-referencing pointers and taking the actual value, etc
+
+type controlEach func(h *HostInfo)
+
+type controlHostLister interface {
+	QueryVpnIp(vpnIp netip.Addr) *HostInfo
+	ForEachIndex(each controlEach)
+	ForEachVpnIp(each controlEach)
+	GetPreferredRanges() []netip.Prefix
+}
+
+type Control struct {
+	f                      *Interface
+	l                      *logrus.Logger
+	ctx                    context.Context
+	cancel                 context.CancelFunc
+	sshStart               func()
+	statsStart             func()
+	dnsStart               func()
+	lighthouseStart        func()
+	connectionManagerStart func(context.Context)
+}
+
+type ControlHostInfo struct {
+	VpnIp                  netip.Addr              `json:"vpnIp"`
+	LocalIndex             uint32                  `json:"localIndex"`
+	RemoteIndex            uint32                  `json:"remoteIndex"`
+	RemoteAddrs            []netip.AddrPort        `json:"remoteAddrs"`
+	Cert                   *cert.NebulaCertificate `json:"cert"`
+	MessageCounter         uint64                  `json:"messageCounter"`
+	CurrentRemote          netip.AddrPort          `json:"currentRemote"`
+	CurrentRelaysToMe      []netip.Addr            `json:"currentRelaysToMe"`
+	CurrentRelaysThroughMe []netip.Addr            `json:"currentRelaysThroughMe"`
+}
+
+// Start actually runs nebula, this is a nonblocking call. To block use Control.ShutdownBlock()
+func (c *Control) Start() {
+	// Activate the interface
+	c.f.activate()
+
+	// Call all the delayed funcs that waited patiently for the interface to be created.
+	if c.sshStart != nil {
+		go c.sshStart()
+	}
+	if c.statsStart != nil {
+		go c.statsStart()
+	}
+	if c.dnsStart != nil {
+		go c.dnsStart()
+	}
+	if c.connectionManagerStart != nil {
+		go c.connectionManagerStart(c.ctx)
+	}
+	if c.lighthouseStart != nil {
+		c.lighthouseStart()
+	}
+
+	// Start reading packets.
+	c.f.run()
+}
+
+func (c *Control) Context() context.Context {
+	return c.ctx
+}
+
+// Stop signals nebula to shutdown and close all tunnels, returns after the shutdown is complete
+func (c *Control) Stop() {
+	// Stop the handshakeManager (and other services), to prevent new tunnels from
+	// being created while we're shutting them all down.
+	c.cancel()
+
+	c.CloseAllTunnels(false)
+	if err := c.f.Close(); err != nil {
+		c.l.WithError(err).Error("Close interface failed")
+	}
+	c.l.Info("Goodbye")
+}
+
+// ShutdownBlock will listen for and block on term and interrupt signals, calling Control.Stop() once signalled
+func (c *Control) ShutdownBlock() {
+	sigChan := make(chan os.Signal, 1)
+	signal.Notify(sigChan, syscall.SIGTERM)
+	signal.Notify(sigChan, syscall.SIGINT)
+
+	rawSig := <-sigChan
+	sig := rawSig.String()
+	c.l.WithField("signal", sig).Info("Caught signal, shutting down")
+	c.Stop()
+}
+
+// RebindUDPServer asks the UDP listener to rebind it's listener. Mainly used on mobile clients when interfaces change
+func (c *Control) RebindUDPServer() {
+	_ = c.f.outside.Rebind()
+
+	// Trigger a lighthouse update, useful for mobile clients that should have an update interval of 0
+	c.f.lightHouse.SendUpdate()
+
+	// Let the main interface know that we rebound so that underlying tunnels know to trigger punches from their remotes
+	c.f.rebindCount++
+}
+
+// ListHostmapHosts returns details about the actual or pending (handshaking) hostmap by vpn ip
+func (c *Control) ListHostmapHosts(pendingMap bool) []ControlHostInfo {
+	if pendingMap {
+		return listHostMapHosts(c.f.handshakeManager)
+	} else {
+		return listHostMapHosts(c.f.hostMap)
+	}
+}
+
+// ListHostmapIndexes returns details about the actual or pending (handshaking) hostmap by local index id
+func (c *Control) ListHostmapIndexes(pendingMap bool) []ControlHostInfo {
+	if pendingMap {
+		return listHostMapIndexes(c.f.handshakeManager)
+	} else {
+		return listHostMapIndexes(c.f.hostMap)
+	}
+}
+
+// GetCertByVpnIp returns the authenticated certificate of the given vpn IP, or nil if not found
+func (c *Control) GetCertByVpnIp(vpnIp netip.Addr) *cert.NebulaCertificate {
+	if c.f.myVpnNet.Addr() == vpnIp {
+		return c.f.pki.GetCertState().Certificate
+	}
+	hi := c.f.hostMap.QueryVpnIp(vpnIp)
+	if hi == nil {
+		return nil
+	}
+	return hi.GetCert()
+}
+
+// CreateTunnel creates a new tunnel to the given vpn ip.
+func (c *Control) CreateTunnel(vpnIp netip.Addr) {
+	c.f.handshakeManager.StartHandshake(vpnIp, nil)
+}
+
+// PrintTunnel creates a new tunnel to the given vpn ip.
+func (c *Control) PrintTunnel(vpnIp netip.Addr) *ControlHostInfo {
+	hi := c.f.hostMap.QueryVpnIp(vpnIp)
+	if hi == nil {
+		return nil
+	}
+	chi := copyHostInfo(hi, c.f.hostMap.GetPreferredRanges())
+	return &chi
+}
+
+// QueryLighthouse queries the lighthouse.
+func (c *Control) QueryLighthouse(vpnIp netip.Addr) *CacheMap {
+	hi := c.f.lightHouse.Query(vpnIp)
+	if hi == nil {
+		return nil
+	}
+	return hi.CopyCache()
+}
+
+// GetHostInfoByVpnIp returns a single tunnels hostInfo, or nil if not found
+// Caller should take care to Unmap() any 4in6 addresses prior to calling.
+func (c *Control) GetHostInfoByVpnIp(vpnIp netip.Addr, pending bool) *ControlHostInfo {
+	var hl controlHostLister
+	if pending {
+		hl = c.f.handshakeManager
+	} else {
+		hl = c.f.hostMap
+	}
+
+	h := hl.QueryVpnIp(vpnIp)
+	if h == nil {
+		return nil
+	}
+
+	ch := copyHostInfo(h, c.f.hostMap.GetPreferredRanges())
+	return &ch
+}
+
+// SetRemoteForTunnel forces a tunnel to use a specific remote
+// Caller should take care to Unmap() any 4in6 addresses prior to calling.
+func (c *Control) SetRemoteForTunnel(vpnIp netip.Addr, addr netip.AddrPort) *ControlHostInfo {
+	hostInfo := c.f.hostMap.QueryVpnIp(vpnIp)
+	if hostInfo == nil {
+		return nil
+	}
+
+	hostInfo.SetRemote(addr)
+	ch := copyHostInfo(hostInfo, c.f.hostMap.GetPreferredRanges())
+	return &ch
+}
+
+// CloseTunnel closes a fully established tunnel. If localOnly is false it will notify the remote end as well.
+// Caller should take care to Unmap() any 4in6 addresses prior to calling.
+func (c *Control) CloseTunnel(vpnIp netip.Addr, localOnly bool) bool {
+	hostInfo := c.f.hostMap.QueryVpnIp(vpnIp)
+	if hostInfo == nil {
+		return false
+	}
+
+	if !localOnly {
+		c.f.send(
+			header.CloseTunnel,
+			0,
+			hostInfo.ConnectionState,
+			hostInfo,
+			[]byte{},
+			make([]byte, 12, 12),
+			make([]byte, mtu),
+		)
+	}
+
+	c.f.closeTunnel(hostInfo)
+	return true
+}
+
+// CloseAllTunnels is just like CloseTunnel except it goes through and shuts them all down, optionally you can avoid shutting down lighthouse tunnels
+// the int returned is a count of tunnels closed
+func (c *Control) CloseAllTunnels(excludeLighthouses bool) (closed int) {
+	//TODO: this is probably better as a function in ConnectionManager or HostMap directly
+	lighthouses := c.f.lightHouse.GetLighthouses()
+
+	shutdown := func(h *HostInfo) {
+		if excludeLighthouses {
+			if _, ok := lighthouses[h.vpnIp]; ok {
+				return
+			}
+		}
+		c.f.send(header.CloseTunnel, 0, h.ConnectionState, h, []byte{}, make([]byte, 12, 12), make([]byte, mtu))
+		c.f.closeTunnel(h)
+
+		c.l.WithField("vpnIp", h.vpnIp).WithField("udpAddr", h.remote).
+			Debug("Sending close tunnel message")
+		closed++
+	}
+
+	// Learn which hosts are being used as relays, so we can shut them down last.
+	relayingHosts := map[netip.Addr]*HostInfo{}
+	// Grab the hostMap lock to access the Relays map
+	c.f.hostMap.Lock()
+	for _, relayingHost := range c.f.hostMap.Relays {
+		relayingHosts[relayingHost.vpnIp] = relayingHost
+	}
+	c.f.hostMap.Unlock()
+
+	hostInfos := []*HostInfo{}
+	// Grab the hostMap lock to access the Hosts map
+	c.f.hostMap.Lock()
+	for _, relayHost := range c.f.hostMap.Indexes {
+		if _, ok := relayingHosts[relayHost.vpnIp]; !ok {
+			hostInfos = append(hostInfos, relayHost)
+		}
+	}
+	c.f.hostMap.Unlock()
+
+	for _, h := range hostInfos {
+		shutdown(h)
+	}
+	for _, h := range relayingHosts {
+		shutdown(h)
+	}
+	return
+}
+
+func (c *Control) Device() overlay.Device {
+	return c.f.inside
+}
+
+func copyHostInfo(h *HostInfo, preferredRanges []netip.Prefix) ControlHostInfo {
+
+	chi := ControlHostInfo{
+		VpnIp:                  h.vpnIp,
+		LocalIndex:             h.localIndexId,
+		RemoteIndex:            h.remoteIndexId,
+		RemoteAddrs:            h.remotes.CopyAddrs(preferredRanges),
+		CurrentRelaysToMe:      h.relayState.CopyRelayIps(),
+		CurrentRelaysThroughMe: h.relayState.CopyRelayForIps(),
+		CurrentRemote:          h.remote,
+	}
+
+	if h.ConnectionState != nil {
+		chi.MessageCounter = h.ConnectionState.messageCounter.Load()
+	}
+
+	if c := h.GetCert(); c != nil {
+		chi.Cert = c.Copy()
+	}
+
+	return chi
+}
+
+func listHostMapHosts(hl controlHostLister) []ControlHostInfo {
+	hosts := make([]ControlHostInfo, 0)
+	pr := hl.GetPreferredRanges()
+	hl.ForEachVpnIp(func(hostinfo *HostInfo) {
+		hosts = append(hosts, copyHostInfo(hostinfo, pr))
+	})
+	return hosts
+}
+
+func listHostMapIndexes(hl controlHostLister) []ControlHostInfo {
+	hosts := make([]ControlHostInfo, 0)
+	pr := hl.GetPreferredRanges()
+	hl.ForEachIndex(func(hostinfo *HostInfo) {
+		hosts = append(hosts, copyHostInfo(hostinfo, pr))
+	})
+	return hosts
+}

control_test.go

@@ -0,0 +1,135 @@
+package nebula
+
+import (
+	"net"
+	"net/netip"
+	"reflect"
+	"testing"
+	"time"
+
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/test"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestControl_GetHostInfoByVpnIp(t *testing.T) {
+	l := test.NewLogger()
+	// Special care must be taken to re-use all objects provided to the hostmap and certificate in the expectedInfo object
+	// To properly ensure we are not exposing core memory to the caller
+	hm := newHostMap(l, netip.Prefix{})
+	hm.preferredRanges.Store(&[]netip.Prefix{})
+
+	remote1 := netip.MustParseAddrPort("0.0.0.100:4444")
+	remote2 := netip.MustParseAddrPort("[1:2:3:4:5:6:7:8]:4444")
+
+	ipNet := net.IPNet{
+		IP:   remote1.Addr().AsSlice(),
+		Mask: net.IPMask{255, 255, 255, 0},
+	}
+
+	ipNet2 := net.IPNet{
+		IP:   remote2.Addr().AsSlice(),
+		Mask: net.IPMask{255, 255, 255, 0},
+	}
+
+	crt := &cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:           "test",
+			Ips:            []*net.IPNet{&ipNet},
+			Subnets:        []*net.IPNet{},
+			Groups:         []string{"default-group"},
+			NotBefore:      time.Unix(1, 0),
+			NotAfter:       time.Unix(2, 0),
+			PublicKey:      []byte{5, 6, 7, 8},
+			IsCA:           false,
+			Issuer:         "the-issuer",
+			InvertedGroups: map[string]struct{}{"default-group": {}},
+		},
+		Signature: []byte{1, 2, 1, 2, 1, 3},
+	}
+
+	remotes := NewRemoteList(nil)
+	remotes.unlockedPrependV4(netip.IPv4Unspecified(), NewIp4AndPortFromNetIP(remote1.Addr(), remote1.Port()))
+	remotes.unlockedPrependV6(netip.IPv4Unspecified(), NewIp6AndPortFromNetIP(remote2.Addr(), remote2.Port()))
+
+	vpnIp, ok := netip.AddrFromSlice(ipNet.IP)
+	assert.True(t, ok)
+
+	hm.unlockedAddHostInfo(&HostInfo{
+		remote:  remote1,
+		remotes: remotes,
+		ConnectionState: &ConnectionState{
+			peerCert: crt,
+		},
+		remoteIndexId: 200,
+		localIndexId:  201,
+		vpnIp:         vpnIp,
+		relayState: RelayState{
+			relays:        nil,
+			relayForByIp:  map[netip.Addr]*Relay{},
+			relayForByIdx: map[uint32]*Relay{},
+		},
+	}, &Interface{})
+
+	vpnIp2, ok := netip.AddrFromSlice(ipNet2.IP)
+	assert.True(t, ok)
+
+	hm.unlockedAddHostInfo(&HostInfo{
+		remote:  remote1,
+		remotes: remotes,
+		ConnectionState: &ConnectionState{
+			peerCert: nil,
+		},
+		remoteIndexId: 200,
+		localIndexId:  201,
+		vpnIp:         vpnIp2,
+		relayState: RelayState{
+			relays:        nil,
+			relayForByIp:  map[netip.Addr]*Relay{},
+			relayForByIdx: map[uint32]*Relay{},
+		},
+	}, &Interface{})
+
+	c := Control{
+		f: &Interface{
+			hostMap: hm,
+		},
+		l: logrus.New(),
+	}
+
+	thi := c.GetHostInfoByVpnIp(vpnIp, false)
+
+	expectedInfo := ControlHostInfo{
+		VpnIp:                  vpnIp,
+		LocalIndex:             201,
+		RemoteIndex:            200,
+		RemoteAddrs:            []netip.AddrPort{remote2, remote1},
+		Cert:                   crt.Copy(),
+		MessageCounter:         0,
+		CurrentRemote:          remote1,
+		CurrentRelaysToMe:      []netip.Addr{},
+		CurrentRelaysThroughMe: []netip.Addr{},
+	}
+
+	// Make sure we don't have any unexpected fields
+	assertFields(t, []string{"VpnIp", "LocalIndex", "RemoteIndex", "RemoteAddrs", "Cert", "MessageCounter", "CurrentRemote", "CurrentRelaysToMe", "CurrentRelaysThroughMe"}, thi)
+	assert.EqualValues(t, &expectedInfo, thi)
+	//TODO: netip.Addr reuses global memory for zone identifiers which breaks our "no reused memory check" here
+	//test.AssertDeepCopyEqual(t, &expectedInfo, thi)
+
+	// Make sure we don't panic if the host info doesn't have a cert yet
+	assert.NotPanics(t, func() {
+		thi = c.GetHostInfoByVpnIp(vpnIp2, false)
+	})
+}
+
+func assertFields(t *testing.T, expected []string, actualStruct interface{}) {
+	val := reflect.ValueOf(actualStruct).Elem()
+	fields := make([]string, val.NumField())
+	for i := 0; i < val.NumField(); i++ {
+		fields[i] = val.Type().Field(i).Name
+	}
+
+	assert.Equal(t, expected, fields)
+}

control_tester.go

@@ -0,0 +1,162 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package nebula
+
+import (
+	"net/netip"
+
+	"github.com/slackhq/nebula/cert"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/overlay"
+	"github.com/slackhq/nebula/udp"
+)
+
+// WaitForType will pipe all messages from this control device into the pipeTo control device
+// returning after a message matching the criteria has been piped
+func (c *Control) WaitForType(msgType header.MessageType, subType header.MessageSubType, pipeTo *Control) {
+	h := &header.H{}
+	for {
+		p := c.f.outside.(*udp.TesterConn).Get(true)
+		if err := h.Parse(p.Data); err != nil {
+			panic(err)
+		}
+		pipeTo.InjectUDPPacket(p)
+		if h.Type == msgType && h.Subtype == subType {
+			return
+		}
+	}
+}
+
+// WaitForTypeByIndex is similar to WaitForType except it adds an index check
+// Useful if you have many nodes communicating and want to wait to find a specific nodes packet
+func (c *Control) WaitForTypeByIndex(toIndex uint32, msgType header.MessageType, subType header.MessageSubType, pipeTo *Control) {
+	h := &header.H{}
+	for {
+		p := c.f.outside.(*udp.TesterConn).Get(true)
+		if err := h.Parse(p.Data); err != nil {
+			panic(err)
+		}
+		pipeTo.InjectUDPPacket(p)
+		if h.RemoteIndex == toIndex && h.Type == msgType && h.Subtype == subType {
+			return
+		}
+	}
+}
+
+// InjectLightHouseAddr will push toAddr into the local lighthouse cache for the vpnIp
+// This is necessary if you did not configure static hosts or are not running a lighthouse
+func (c *Control) InjectLightHouseAddr(vpnIp netip.Addr, toAddr netip.AddrPort) {
+	c.f.lightHouse.Lock()
+	remoteList := c.f.lightHouse.unlockedGetRemoteList(vpnIp)
+	remoteList.Lock()
+	defer remoteList.Unlock()
+	c.f.lightHouse.Unlock()
+
+	if toAddr.Addr().Is4() {
+		remoteList.unlockedPrependV4(vpnIp, NewIp4AndPortFromNetIP(toAddr.Addr(), toAddr.Port()))
+	} else {
+		remoteList.unlockedPrependV6(vpnIp, NewIp6AndPortFromNetIP(toAddr.Addr(), toAddr.Port()))
+	}
+}
+
+// InjectRelays will push relayVpnIps into the local lighthouse cache for the vpnIp
+// This is necessary to inform an initiator of possible relays for communicating with a responder
+func (c *Control) InjectRelays(vpnIp netip.Addr, relayVpnIps []netip.Addr) {
+	c.f.lightHouse.Lock()
+	remoteList := c.f.lightHouse.unlockedGetRemoteList(vpnIp)
+	remoteList.Lock()
+	defer remoteList.Unlock()
+	c.f.lightHouse.Unlock()
+
+	remoteList.unlockedSetRelay(vpnIp, vpnIp, relayVpnIps)
+}
+
+// GetFromTun will pull a packet off the tun side of nebula
+func (c *Control) GetFromTun(block bool) []byte {
+	return c.f.inside.(*overlay.TestTun).Get(block)
+}
+
+// GetFromUDP will pull a udp packet off the udp side of nebula
+func (c *Control) GetFromUDP(block bool) *udp.Packet {
+	return c.f.outside.(*udp.TesterConn).Get(block)
+}
+
+func (c *Control) GetUDPTxChan() <-chan *udp.Packet {
+	return c.f.outside.(*udp.TesterConn).TxPackets
+}
+
+func (c *Control) GetTunTxChan() <-chan []byte {
+	return c.f.inside.(*overlay.TestTun).TxPackets
+}
+
+// InjectUDPPacket will inject a packet into the udp side of nebula
+func (c *Control) InjectUDPPacket(p *udp.Packet) {
+	c.f.outside.(*udp.TesterConn).Send(p)
+}
+
+// InjectTunUDPPacket puts a udp packet on the tun interface. Using UDP here because it's a simpler protocol
+func (c *Control) InjectTunUDPPacket(toIp netip.Addr, toPort uint16, fromPort uint16, data []byte) {
+	//TODO: IPV6-WORK
+	ip := layers.IPv4{
+		Version:  4,
+		TTL:      64,
+		Protocol: layers.IPProtocolUDP,
+		SrcIP:    c.f.inside.Cidr().Addr().Unmap().AsSlice(),
+		DstIP:    toIp.Unmap().AsSlice(),
+	}
+
+	udp := layers.UDP{
+		SrcPort: layers.UDPPort(fromPort),
+		DstPort: layers.UDPPort(toPort),
+	}
+	err := udp.SetNetworkLayerForChecksum(&ip)
+	if err != nil {
+		panic(err)
+	}
+
+	buffer := gopacket.NewSerializeBuffer()
+	opt := gopacket.SerializeOptions{
+		ComputeChecksums: true,
+		FixLengths:       true,
+	}
+	err = gopacket.SerializeLayers(buffer, opt, &ip, &udp, gopacket.Payload(data))
+	if err != nil {
+		panic(err)
+	}
+
+	c.f.inside.(*overlay.TestTun).Send(buffer.Bytes())
+}
+
+func (c *Control) GetVpnIp() netip.Addr {
+	return c.f.myVpnNet.Addr()
+}
+
+func (c *Control) GetUDPAddr() netip.AddrPort {
+	return c.f.outside.(*udp.TesterConn).Addr
+}
+
+func (c *Control) KillPendingTunnel(vpnIp netip.Addr) bool {
+	hostinfo := c.f.handshakeManager.QueryVpnIp(vpnIp)
+	if hostinfo == nil {
+		return false
+	}
+
+	c.f.handshakeManager.DeleteHostInfo(hostinfo)
+	return true
+}
+
+func (c *Control) GetHostmap() *HostMap {
+	return c.f.hostMap
+}
+
+func (c *Control) GetCert() *cert.NebulaCertificate {
+	return c.f.pki.GetCertState().Certificate
+}
+
+func (c *Control) ReHandshake(vpnIp netip.Addr) {
+	c.f.handshakeManager.StartHandshake(vpnIp, nil)
+}

dist/arch/nebula.service

@@ -1,15 +0,0 @@
-[Unit]
-Description=nebula
-Wants=basic.target
-After=basic.target network.target
-
-[Service]
-SyslogIdentifier=nebula
-StandardOutput=syslog
-StandardError=syslog
-ExecReload=/bin/kill -HUP $MAINPID
-ExecStart=/usr/bin/nebula -config /etc/nebula/config.yml
-Restart=always
-
-[Install]
-WantedBy=multi-user.target

dist/windows/wintun/LICENSE.txt

@@ -0,0 +1,84 @@
+Prebuilt Binaries License
+-------------------------
+
+1. DEFINITIONS. "Software" means the precise contents of the "wintun.dll"
+   files that are included in the .zip file that contains this document as
+   downloaded from wintun.net/builds.
+
+2. LICENSE GRANT. WireGuard LLC grants to you a non-exclusive and
+   non-transferable right to use Software for lawful purposes under certain
+   obligations and limited rights as set forth in this agreement.
+
+3. RESTRICTIONS. Software is owned and copyrighted by WireGuard LLC. It is
+   licensed, not sold. Title to Software and all associated intellectual
+   property rights are retained by WireGuard. You must not:
+   a. reverse engineer, decompile, disassemble, extract from, or otherwise
+      modify the Software;
+   b. modify or create derivative work based upon Software in whole or in
+      parts, except insofar as only the API interfaces of the "wintun.h" file
+      distributed alongside the Software (the "Permitted API") are used;
+   c. remove any proprietary notices, labels, or copyrights from the Software;
+   d. resell, redistribute, lease, rent, transfer, sublicense, or otherwise
+      transfer rights of the Software without the prior written consent of
+      WireGuard LLC, except insofar as the Software is distributed alongside
+      other software that uses the Software only via the Permitted API;
+   e. use the name of WireGuard LLC, the WireGuard project, the Wintun
+      project, or the names of its contributors to endorse or promote products
+      derived from the Software without specific prior written consent.
+
+4. LIMITED WARRANTY. THE SOFTWARE IS PROVIDED "AS IS" AND WITHOUT WARRANTY OF
+   ANY KIND. WIREGUARD LLC HEREBY EXCLUDES AND DISCLAIMS ALL IMPLIED OR
+   STATUTORY WARRANTIES, INCLUDING ANY WARRANTIES OF MERCHANTABILITY, FITNESS
+   FOR A PARTICULAR PURPOSE, QUALITY, NON-INFRINGEMENT, TITLE, RESULTS,
+   EFFORTS, OR QUIET ENJOYMENT. THERE IS NO WARRANTY THAT THE PRODUCT WILL BE
+   ERROR-FREE OR WILL FUNCTION WITHOUT INTERRUPTION. YOU ASSUME THE ENTIRE
+   RISK FOR THE RESULTS OBTAINED USING THE PRODUCT. TO THE EXTENT THAT
+   WIREGUARD LLC MAY NOT DISCLAIM ANY WARRANTY AS A MATTER OF APPLICABLE LAW,
+   THE SCOPE AND DURATION OF SUCH WARRANTY WILL BE THE MINIMUM PERMITTED UNDER
+   SUCH LAW. ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND
+   WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR
+   A PARTICULAR PURPOSE OR NON-INFRINGEMENT ARE DISCLAIMED, EXCEPT TO THE
+   EXTENT THAT THESE DISCLAIMERS ARE HELD TO BE LEGALLY INVALID.
+
+5. LIMITATION OF LIABILITY. To the extent not prohibited by law, in no event
+   WireGuard LLC or any third-party-developer will be liable for any lost
+   revenue, profit or data or for special, indirect, consequential, incidental
+   or punitive damages, however caused regardless of the theory of liability,
+   arising out of or related to the use of or inability to use Software, even
+   if WireGuard LLC has been advised of the possibility of such damages.
+   Solely you are responsible for determining the appropriateness of using
+   Software and accept full responsibility for all risks associated with its
+   exercise of rights under this agreement, including but not limited to the
+   risks and costs of program errors, compliance with applicable laws, damage
+   to or loss of data, programs or equipment, and unavailability or
+   interruption of operations. The foregoing limitations will apply even if
+   the above stated warranty fails of its essential purpose. You acknowledge,
+   that it is in the nature of software that software is complex and not
+   completely free of errors. In no event shall WireGuard LLC or any
+   third-party-developer be liable to you under any theory for any damages
+   suffered by you or any user of Software or for any special, incidental,
+   indirect, consequential or similar damages (including without limitation
+   damages for loss of business profits, business interruption, loss of
+   business information or any other pecuniary loss) arising out of the use or
+   inability to use Software, even if WireGuard LLC has been advised of the
+   possibility of such damages and regardless of the legal or quitable theory
+   (contract, tort, or otherwise) upon which the claim is based.
+
+6. TERMINATION. This agreement is affected until terminated. You may
+   terminate this agreement at any time. This agreement will terminate
+   immediately without notice from WireGuard LLC if you fail to comply with
+   the terms and conditions of this agreement. Upon termination, you must
+   delete Software and all copies of Software and cease all forms of
+   distribution of Software.
+
+7. SEVERABILITY. If any provision of this agreement is held to be
+   unenforceable, this agreement will remain in effect with the provision
+   omitted, unless omission would frustrate the intent of the parties, in
+   which case this agreement will immediately terminate.
+
+8. RESERVATION OF RIGHTS. All rights not expressly granted in this agreement
+   are reserved by WireGuard LLC. For example, WireGuard LLC reserves the
+   right at any time to cease development of Software, to alter distribution
+   details, features, specifications, capabilities, functions, licensing
+   terms, release dates, APIs, ABIs, general availability, or other
+   characteristics of the Software.

dist/windows/wintun/README.md

@@ -0,0 +1,339 @@
+# [Wintun Network Adapter](https://www.wintun.net/)
+### TUN Device Driver for Windows
+
+This is a layer 3 TUN driver for Windows 7, 8, 8.1, and 10. Originally created for [WireGuard](https://www.wireguard.com/), it is intended to be useful to a wide variety of projects that require layer 3 tunneling devices with implementations primarily in userspace.
+
+## Installation
+
+Wintun is deployed as a platform-specific `wintun.dll` file. Install the `wintun.dll` file side-by-side with your application. Download the dll from [wintun.net](https://www.wintun.net/), alongside the header file for your application described below.
+
+## Usage
+
+Include the [`wintun.h` file](https://git.zx2c4.com/wintun/tree/api/wintun.h) in your project simply by copying it there and dynamically load the `wintun.dll` using [`LoadLibraryEx()`](https://docs.microsoft.com/en-us/windows/win32/api/libloaderapi/nf-libloaderapi-loadlibraryexa) and [`GetProcAddress()`](https://docs.microsoft.com/en-us/windows/win32/api/libloaderapi/nf-libloaderapi-getprocaddress) to resolve each function, using the typedefs provided in the header file. The [`InitializeWintun` function in the example.c code](https://git.zx2c4.com/wintun/tree/example/example.c) provides this in a function that you can simply copy and paste.
+
+With the library setup, Wintun can then be used by first creating an adapter, configuring it, and then setting its status to "up". Adapters have names (e.g. "OfficeNet") and types (e.g. "Wintun").
+
+```C
+WINTUN_ADAPTER_HANDLE Adapter1 = WintunCreateAdapter(L"OfficeNet", L"Wintun", &SomeFixedGUID1);
+WINTUN_ADAPTER_HANDLE Adapter2 = WintunCreateAdapter(L"HomeNet", L"Wintun", &SomeFixedGUID2);
+WINTUN_ADAPTER_HANDLE Adapter3 = WintunCreateAdapter(L"Data Center", L"Wintun", &SomeFixedGUID3);
+```
+
+After creating an adapter, we can use it by starting a session:
+
+```C
+WINTUN_SESSION_HANDLE Session = WintunStartSession(Adapter2, 0x400000);
+```
+
+Then, the `WintunAllocateSendPacket` and `WintunSendPacket` functions can be used for sending packets ([used by `SendPackets` in the example.c code](https://git.zx2c4.com/wintun/tree/example/example.c)):
+
+```C
+BYTE *OutgoingPacket = WintunAllocateSendPacket(Session, PacketDataSize);
+if (OutgoingPacket)
+{
+    memcpy(OutgoingPacket, PacketData, PacketDataSize);
+    WintunSendPacket(Session, OutgoingPacket);
+}
+else if (GetLastError() != ERROR_BUFFER_OVERFLOW) // Silently drop packets if the ring is full
+    Log(L"Packet write failed");
+```
+
+And the `WintunReceivePacket` and `WintunReleaseReceivePacket` functions can be used for receiving packets ([used by `ReceivePackets` in the example.c code](https://git.zx2c4.com/wintun/tree/example/example.c)):
+
+```C
+for (;;)
+{
+    DWORD IncomingPacketSize;
+    BYTE *IncomingPacket = WintunReceivePacket(Session, &IncomingPacketSize);
+    if (IncomingPacket)
+    {
+        DoSomethingWithPacket(IncomingPacket, IncomingPacketSize);
+        WintunReleaseReceivePacket(Session, IncomingPacket);
+    }
+    else if (GetLastError() == ERROR_NO_MORE_ITEMS)
+        WaitForSingleObject(WintunGetReadWaitEvent(Session), INFINITE);
+    else
+    {
+        Log(L"Packet read failed");
+        break;
+    }
+}
+```
+
+Some high performance use cases may want to spin on `WintunReceivePackets` for a number of cycles before falling back to waiting on the read-wait event.
+
+You are **highly encouraged** to read the [**example.c short example**](https://git.zx2c4.com/wintun/tree/example/example.c) to see how to put together a simple userspace network tunnel.
+
+The various functions and definitions are [documented in the reference below](#Reference).
+
+## Reference
+
+### Macro Definitions
+
+#### WINTUN\_MAX\_POOL
+
+`#define WINTUN_MAX_POOL   256`
+
+Maximum pool name length including zero terminator
+
+#### WINTUN\_MIN\_RING\_CAPACITY
+
+`#define WINTUN_MIN_RING_CAPACITY   0x20000 /* 128kiB */`
+
+Minimum ring capacity.
+
+#### WINTUN\_MAX\_RING\_CAPACITY
+
+`#define WINTUN_MAX_RING_CAPACITY   0x4000000 /* 64MiB */`
+
+Maximum ring capacity.
+
+#### WINTUN\_MAX\_IP\_PACKET\_SIZE
+
+`#define WINTUN_MAX_IP_PACKET_SIZE   0xFFFF`
+
+Maximum IP packet size
+
+### Typedefs
+
+#### WINTUN\_ADAPTER\_HANDLE
+
+`typedef void* WINTUN_ADAPTER_HANDLE`
+
+A handle representing Wintun adapter
+
+#### WINTUN\_ENUM\_CALLBACK
+
+`typedef BOOL(* WINTUN_ENUM_CALLBACK) (WINTUN_ADAPTER_HANDLE Adapter, LPARAM Param)`
+
+Called by WintunEnumAdapters for each adapter in the pool.
+
+**Parameters**
+
+- *Adapter*: Adapter handle, which will be freed when this function returns.
+- *Param*: An application-defined value passed to the WintunEnumAdapters.
+
+**Returns**
+
+Non-zero to continue iterating adapters; zero to stop.
+
+#### WINTUN\_LOGGER\_CALLBACK
+
+`typedef void(* WINTUN_LOGGER_CALLBACK) (WINTUN_LOGGER_LEVEL Level, DWORD64 Timestamp, const WCHAR *Message)`
+
+Called by internal logger to report diagnostic messages
+
+**Parameters**
+
+- *Level*: Message level.
+- *Timestamp*: Message timestamp in in 100ns intervals since 1601-01-01 UTC.
+- *Message*: Message text.
+
+#### WINTUN\_SESSION\_HANDLE
+
+`typedef void* WINTUN_SESSION_HANDLE`
+
+A handle representing Wintun session
+
+### Enumeration Types
+
+#### WINTUN\_LOGGER\_LEVEL
+
+`enum WINTUN_LOGGER_LEVEL`
+
+Determines the level of logging, passed to WINTUN\_LOGGER\_CALLBACK.
+
+- *WINTUN\_LOG\_INFO*: Informational
+- *WINTUN\_LOG\_WARN*: Warning
+- *WINTUN\_LOG\_ERR*: Error
+
+Enumerator
+
+### Functions
+
+#### WintunCreateAdapter()
+
+`WINTUN_ADAPTER_HANDLE WintunCreateAdapter (const WCHAR * Name, const WCHAR * TunnelType, const GUID * RequestedGUID)`
+
+Creates a new Wintun adapter.
+
+**Parameters**
+
+- *Name*: The requested name of the adapter. Zero-terminated string of up to MAX\_ADAPTER\_NAME-1 characters.
+- *Name*: Name of the adapter tunnel type. Zero-terminated string of up to MAX\_ADAPTER\_NAME-1 characters.
+- *RequestedGUID*: The GUID of the created network adapter, which then influences NLA generation deterministically. If it is set to NULL, the GUID is chosen by the system at random, and hence a new NLA entry is created for each new adapter. It is called "requested" GUID because the API it uses is completely undocumented, and so there could be minor interesting complications with its usage.
+
+**Returns**
+
+If the function succeeds, the return value is the adapter handle. Must be released with WintunCloseAdapter. If the function fails, the return value is NULL. To get extended error information, call GetLastError.
+
+#### WintunOpenAdapter()
+
+`WINTUN_ADAPTER_HANDLE WintunOpenAdapter (const WCHAR * Name)`
+
+Opens an existing Wintun adapter.
+
+**Parameters**
+
+- *Name*: The requested name of the adapter. Zero-terminated string of up to MAX\_ADAPTER\_NAME-1 characters.
+
+**Returns**
+
+If the function succeeds, the return value is adapter handle. Must be released with WintunCloseAdapter. If the function fails, the return value is NULL. To get extended error information, call GetLastError.
+
+#### WintunCloseAdapter()
+
+`void WintunCloseAdapter (WINTUN_ADAPTER_HANDLE Adapter)`
+
+Releases Wintun adapter resources and, if adapter was created with WintunCreateAdapter, removes adapter.
+
+**Parameters**
+
+- *Adapter*: Adapter handle obtained with WintunCreateAdapter or WintunOpenAdapter.
+
+#### WintunDeleteDriver()
+
+`BOOL WintunDeleteDriver ()`
+
+Deletes the Wintun driver if there are no more adapters in use.
+
+**Returns**
+
+If the function succeeds, the return value is nonzero. If the function fails, the return value is zero. To get extended error information, call GetLastError.
+
+#### WintunGetAdapterLuid()
+
+`void WintunGetAdapterLuid (WINTUN_ADAPTER_HANDLE Adapter, NET_LUID * Luid)`
+
+Returns the LUID of the adapter.
+
+**Parameters**
+
+- *Adapter*: Adapter handle obtained with WintunOpenAdapter or WintunCreateAdapter
+- *Luid*: Pointer to LUID to receive adapter LUID.
+
+#### WintunGetRunningDriverVersion()
+
+`DWORD WintunGetRunningDriverVersion (void )`
+
+Determines the version of the Wintun driver currently loaded.
+
+**Returns**
+
+If the function succeeds, the return value is the version number. If the function fails, the return value is zero. To get extended error information, call GetLastError. Possible errors include the following: ERROR\_FILE\_NOT\_FOUND Wintun not loaded
+
+#### WintunSetLogger()
+
+`void WintunSetLogger (WINTUN_LOGGER_CALLBACK NewLogger)`
+
+Sets logger callback function.
+
+**Parameters**
+
+- *NewLogger*: Pointer to callback function to use as a new global logger. NewLogger may be called from various threads concurrently. Should the logging require serialization, you must handle serialization in NewLogger. Set to NULL to disable.
+
+#### WintunStartSession()
+
+`WINTUN_SESSION_HANDLE WintunStartSession (WINTUN_ADAPTER_HANDLE Adapter, DWORD Capacity)`
+
+Starts Wintun session.
+
+**Parameters**
+
+- *Adapter*: Adapter handle obtained with WintunOpenAdapter or WintunCreateAdapter
+- *Capacity*: Rings capacity. Must be between WINTUN\_MIN\_RING\_CAPACITY and WINTUN\_MAX\_RING\_CAPACITY (incl.) Must be a power of two.
+
+**Returns**
+
+Wintun session handle. Must be released with WintunEndSession. If the function fails, the return value is NULL. To get extended error information, call GetLastError.
+
+#### WintunEndSession()
+
+`void WintunEndSession (WINTUN_SESSION_HANDLE Session)`
+
+Ends Wintun session.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+
+#### WintunGetReadWaitEvent()
+
+`HANDLE WintunGetReadWaitEvent (WINTUN_SESSION_HANDLE Session)`
+
+Gets Wintun session's read-wait event handle.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+
+**Returns**
+
+Pointer to receive event handle to wait for available data when reading. Should WintunReceivePackets return ERROR\_NO\_MORE\_ITEMS (after spinning on it for a while under heavy load), wait for this event to become signaled before retrying WintunReceivePackets. Do not call CloseHandle on this event - it is managed by the session.
+
+#### WintunReceivePacket()
+
+`BYTE* WintunReceivePacket (WINTUN_SESSION_HANDLE Session, DWORD * PacketSize)`
+
+Retrieves one or packet. After the packet content is consumed, call WintunReleaseReceivePacket with Packet returned from this function to release internal buffer. This function is thread-safe.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+- *PacketSize*: Pointer to receive packet size.
+
+**Returns**
+
+Pointer to layer 3 IPv4 or IPv6 packet. Client may modify its content at will. If the function fails, the return value is NULL. To get extended error information, call GetLastError. Possible errors include the following: ERROR\_HANDLE\_EOF Wintun adapter is terminating; ERROR\_NO\_MORE\_ITEMS Wintun buffer is exhausted; ERROR\_INVALID\_DATA Wintun buffer is corrupt
+
+#### WintunReleaseReceivePacket()
+
+`void WintunReleaseReceivePacket (WINTUN_SESSION_HANDLE Session, const BYTE * Packet)`
+
+Releases internal buffer after the received packet has been processed by the client. This function is thread-safe.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+- *Packet*: Packet obtained with WintunReceivePacket
+
+#### WintunAllocateSendPacket()
+
+`BYTE* WintunAllocateSendPacket (WINTUN_SESSION_HANDLE Session, DWORD PacketSize)`
+
+Allocates memory for a packet to send. After the memory is filled with packet data, call WintunSendPacket to send and release internal buffer. WintunAllocateSendPacket is thread-safe and the WintunAllocateSendPacket order of calls define the packet sending order.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+- *PacketSize*: Exact packet size. Must be less or equal to WINTUN\_MAX\_IP\_PACKET\_SIZE.
+
+**Returns**
+
+Returns pointer to memory where to prepare layer 3 IPv4 or IPv6 packet for sending. If the function fails, the return value is NULL. To get extended error information, call GetLastError. Possible errors include the following: ERROR\_HANDLE\_EOF Wintun adapter is terminating; ERROR\_BUFFER\_OVERFLOW Wintun buffer is full;
+
+#### WintunSendPacket()
+
+`void WintunSendPacket (WINTUN_SESSION_HANDLE Session, const BYTE * Packet)`
+
+Sends the packet and releases internal buffer. WintunSendPacket is thread-safe, but the WintunAllocateSendPacket order of calls define the packet sending order. This means the packet is not guaranteed to be sent in the WintunSendPacket yet.
+
+**Parameters**
+
+- *Session*: Wintun session handle obtained with WintunStartSession
+- *Packet*: Packet obtained with WintunAllocateSendPacket
+
+## Building
+
+**Do not distribute drivers or files named "Wintun", as they will most certainly clash with official deployments. Instead distribute [`wintun.dll` as downloaded from wintun.net](https://www.wintun.net).**
+
+General requirements:
+
+- [Visual Studio 2019](https://visualstudio.microsoft.com/downloads/) with Windows SDK
+- [Windows Driver Kit](https://docs.microsoft.com/en-us/windows-hardware/drivers/download-the-wdk)
+
+`wintun.sln` may be opened in Visual Studio for development and building. Be sure to run `bcdedit /set testsigning on` and then reboot before to enable unsigned driver loading. The default run sequence (F5) in Visual Studio will build the example project and its dependencies.
+
+## License
+
+The entire contents of [the repository](https://git.zx2c4.com/wintun/), including all documentation and example code, is "Copyright © 2018-2021 WireGuard LLC. All Rights Reserved." Source code is licensed under the [GPLv2](COPYING). Prebuilt binaries from [wintun.net](https://www.wintun.net/) are released under a more permissive license suitable for more forms of software contained inside of the .zip files distributed there.

dist/windows/wintun/include/wintun.h

@@ -0,0 +1,270 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT
+ *
+ * Copyright (C) 2018-2021 WireGuard LLC. All Rights Reserved.
+ */
+
+#pragma once
+
+#include <winsock2.h>
+#include <windows.h>
+#include <ipexport.h>
+#include <ifdef.h>
+#include <ws2ipdef.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifndef ALIGNED
+#    if defined(_MSC_VER)
+#        define ALIGNED(n) __declspec(align(n))
+#    elif defined(__GNUC__)
+#        define ALIGNED(n) __attribute__((aligned(n)))
+#    else
+#        error "Unable to define ALIGNED"
+#    endif
+#endif
+
+/* MinGW is missing this one, unfortunately. */
+#ifndef _Post_maybenull_
+#    define _Post_maybenull_
+#endif
+
+#pragma warning(push)
+#pragma warning(disable : 4324) /* structure was padded due to alignment specifier */
+
+/**
+ * A handle representing Wintun adapter
+ */
+typedef struct _WINTUN_ADAPTER *WINTUN_ADAPTER_HANDLE;
+
+/**
+ * Creates a new Wintun adapter.
+ *
+ * @param Name          The requested name of the adapter. Zero-terminated string of up to MAX_ADAPTER_NAME-1
+ *                      characters.
+ *
+ * @param TunnelType    Name of the adapter tunnel type. Zero-terminated string of up to MAX_ADAPTER_NAME-1
+ *                      characters.
+ *
+ * @param RequestedGUID The GUID of the created network adapter, which then influences NLA generation deterministically.
+ *                      If it is set to NULL, the GUID is chosen by the system at random, and hence a new NLA entry is
+ *                      created for each new adapter. It is called "requested" GUID because the API it uses is
+ *                      completely undocumented, and so there could be minor interesting complications with its usage.
+ *
+ * @return If the function succeeds, the return value is the adapter handle. Must be released with
+ * WintunCloseAdapter. If the function fails, the return value is NULL. To get extended error information, call
+ * GetLastError.
+ */
+typedef _Must_inspect_result_
+_Return_type_success_(return != NULL)
+_Post_maybenull_
+WINTUN_ADAPTER_HANDLE(WINAPI WINTUN_CREATE_ADAPTER_FUNC)
+(_In_z_ LPCWSTR Name, _In_z_ LPCWSTR TunnelType, _In_opt_ const GUID *RequestedGUID);
+
+/**
+ * Opens an existing Wintun adapter.
+ *
+ * @param Name          The requested name of the adapter. Zero-terminated string of up to MAX_ADAPTER_NAME-1
+ *                      characters.
+ *
+ * @return If the function succeeds, the return value is the adapter handle. Must be released with
+ * WintunCloseAdapter. If the function fails, the return value is NULL. To get extended error information, call
+ * GetLastError.
+ */
+typedef _Must_inspect_result_
+_Return_type_success_(return != NULL)
+_Post_maybenull_
+WINTUN_ADAPTER_HANDLE(WINAPI WINTUN_OPEN_ADAPTER_FUNC)(_In_z_ LPCWSTR Name);
+
+/**
+ * Releases Wintun adapter resources and, if adapter was created with WintunCreateAdapter, removes adapter.
+ *
+ * @param Adapter       Adapter handle obtained with WintunCreateAdapter or WintunOpenAdapter.
+ */
+typedef VOID(WINAPI WINTUN_CLOSE_ADAPTER_FUNC)(_In_opt_ WINTUN_ADAPTER_HANDLE Adapter);
+
+/**
+ * Deletes the Wintun driver if there are no more adapters in use.
+ *
+ * @return If the function succeeds, the return value is nonzero. If the function fails, the return value is zero. To
+ *         get extended error information, call GetLastError.
+ */
+typedef _Return_type_success_(return != FALSE)
+BOOL(WINAPI WINTUN_DELETE_DRIVER_FUNC)(VOID);
+
+/**
+ * Returns the LUID of the adapter.
+ *
+ * @param Adapter       Adapter handle obtained with WintunCreateAdapter or WintunOpenAdapter
+ *
+ * @param Luid          Pointer to LUID to receive adapter LUID.
+ */
+typedef VOID(WINAPI WINTUN_GET_ADAPTER_LUID_FUNC)(_In_ WINTUN_ADAPTER_HANDLE Adapter, _Out_ NET_LUID *Luid);
+
+/**
+ * Determines the version of the Wintun driver currently loaded.
+ *
+ * @return If the function succeeds, the return value is the version number. If the function fails, the return value is
+ *         zero. To get extended error information, call GetLastError. Possible errors include the following:
+ *         ERROR_FILE_NOT_FOUND  Wintun not loaded
+ */
+typedef _Return_type_success_(return != 0)
+DWORD(WINAPI WINTUN_GET_RUNNING_DRIVER_VERSION_FUNC)(VOID);
+
+/**
+ * Determines the level of logging, passed to WINTUN_LOGGER_CALLBACK.
+ */
+typedef enum
+{
+    WINTUN_LOG_INFO, /**< Informational */
+    WINTUN_LOG_WARN, /**< Warning */
+    WINTUN_LOG_ERR   /**< Error */
+} WINTUN_LOGGER_LEVEL;
+
+/**
+ * Called by internal logger to report diagnostic messages
+ *
+ * @param Level         Message level.
+ *
+ * @param Timestamp     Message timestamp in in 100ns intervals since 1601-01-01 UTC.
+ *
+ * @param Message       Message text.
+ */
+typedef VOID(CALLBACK *WINTUN_LOGGER_CALLBACK)(
+    _In_ WINTUN_LOGGER_LEVEL Level,
+    _In_ DWORD64 Timestamp,
+    _In_z_ LPCWSTR Message);
+
+/**
+ * Sets logger callback function.
+ *
+ * @param NewLogger     Pointer to callback function to use as a new global logger. NewLogger may be called from various
+ *                      threads concurrently. Should the logging require serialization, you must handle serialization in
+ *                      NewLogger. Set to NULL to disable.
+ */
+typedef VOID(WINAPI WINTUN_SET_LOGGER_FUNC)(_In_ WINTUN_LOGGER_CALLBACK NewLogger);
+
+/**
+ * Minimum ring capacity.
+ */
+#define WINTUN_MIN_RING_CAPACITY 0x20000 /* 128kiB */
+
+/**
+ * Maximum ring capacity.
+ */
+#define WINTUN_MAX_RING_CAPACITY 0x4000000 /* 64MiB */
+
+/**
+ * A handle representing Wintun session
+ */
+typedef struct _TUN_SESSION *WINTUN_SESSION_HANDLE;
+
+/**
+ * Starts Wintun session.
+ *
+ * @param Adapter       Adapter handle obtained with WintunOpenAdapter or WintunCreateAdapter
+ *
+ * @param Capacity      Rings capacity. Must be between WINTUN_MIN_RING_CAPACITY and WINTUN_MAX_RING_CAPACITY (incl.)
+ *                      Must be a power of two.
+ *
+ * @return Wintun session handle. Must be released with WintunEndSession. If the function fails, the return value is
+ *         NULL. To get extended error information, call GetLastError.
+ */
+typedef _Must_inspect_result_
+_Return_type_success_(return != NULL)
+_Post_maybenull_
+WINTUN_SESSION_HANDLE(WINAPI WINTUN_START_SESSION_FUNC)(_In_ WINTUN_ADAPTER_HANDLE Adapter, _In_ DWORD Capacity);
+
+/**
+ * Ends Wintun session.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ */
+typedef VOID(WINAPI WINTUN_END_SESSION_FUNC)(_In_ WINTUN_SESSION_HANDLE Session);
+
+/**
+ * Gets Wintun session's read-wait event handle.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ *
+ * @return Pointer to receive event handle to wait for available data when reading. Should
+ *         WintunReceivePackets return ERROR_NO_MORE_ITEMS (after spinning on it for a while under heavy
+ *         load), wait for this event to become signaled before retrying WintunReceivePackets. Do not call
+ *         CloseHandle on this event - it is managed by the session.
+ */
+typedef HANDLE(WINAPI WINTUN_GET_READ_WAIT_EVENT_FUNC)(_In_ WINTUN_SESSION_HANDLE Session);
+
+/**
+ * Maximum IP packet size
+ */
+#define WINTUN_MAX_IP_PACKET_SIZE 0xFFFF
+
+/**
+ * Retrieves one or packet. After the packet content is consumed, call WintunReleaseReceivePacket with Packet returned
+ * from this function to release internal buffer. This function is thread-safe.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ *
+ * @param PacketSize    Pointer to receive packet size.
+ *
+ * @return Pointer to layer 3 IPv4 or IPv6 packet. Client may modify its content at will. If the function fails, the
+ *         return value is NULL. To get extended error information, call GetLastError. Possible errors include the
+ *         following:
+ *         ERROR_HANDLE_EOF     Wintun adapter is terminating;
+ *         ERROR_NO_MORE_ITEMS  Wintun buffer is exhausted;
+ *         ERROR_INVALID_DATA   Wintun buffer is corrupt
+ */
+typedef _Must_inspect_result_
+_Return_type_success_(return != NULL)
+_Post_maybenull_
+_Post_writable_byte_size_(*PacketSize)
+BYTE *(WINAPI WINTUN_RECEIVE_PACKET_FUNC)(_In_ WINTUN_SESSION_HANDLE Session, _Out_ DWORD *PacketSize);
+
+/**
+ * Releases internal buffer after the received packet has been processed by the client. This function is thread-safe.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ *
+ * @param Packet        Packet obtained with WintunReceivePacket
+ */
+typedef VOID(
+    WINAPI WINTUN_RELEASE_RECEIVE_PACKET_FUNC)(_In_ WINTUN_SESSION_HANDLE Session, _In_ const BYTE *Packet);
+
+/**
+ * Allocates memory for a packet to send. After the memory is filled with packet data, call WintunSendPacket to send
+ * and release internal buffer. WintunAllocateSendPacket is thread-safe and the WintunAllocateSendPacket order of
+ * calls define the packet sending order.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ *
+ * @param PacketSize    Exact packet size. Must be less or equal to WINTUN_MAX_IP_PACKET_SIZE.
+ *
+ * @return Returns pointer to memory where to prepare layer 3 IPv4 or IPv6 packet for sending. If the function fails,
+ *         the return value is NULL. To get extended error information, call GetLastError. Possible errors include the
+ *         following:
+ *         ERROR_HANDLE_EOF       Wintun adapter is terminating;
+ *         ERROR_BUFFER_OVERFLOW  Wintun buffer is full;
+ */
+typedef _Must_inspect_result_
+_Return_type_success_(return != NULL)
+_Post_maybenull_
+_Post_writable_byte_size_(PacketSize)
+BYTE *(WINAPI WINTUN_ALLOCATE_SEND_PACKET_FUNC)(_In_ WINTUN_SESSION_HANDLE Session, _In_ DWORD PacketSize);
+
+/**
+ * Sends the packet and releases internal buffer. WintunSendPacket is thread-safe, but the WintunAllocateSendPacket
+ * order of calls define the packet sending order. This means the packet is not guaranteed to be sent in the
+ * WintunSendPacket yet.
+ *
+ * @param Session       Wintun session handle obtained with WintunStartSession
+ *
+ * @param Packet        Packet obtained with WintunAllocateSendPacket
+ */
+typedef VOID(WINAPI WINTUN_SEND_PACKET_FUNC)(_In_ WINTUN_SESSION_HANDLE Session, _In_ const BYTE *Packet);
+
+#pragma warning(pop)
+
+#ifdef __cplusplus
+}
+#endif

dns_server.go

@@ -3,10 +3,14 @@ package nebula
 import (
 	"fmt"
 	"net"
+	"net/netip"
 	"strconv"
+	"strings"
 	"sync"
 
 	"github.com/miekg/dns"
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
 )
 
 // This whole thing should be rewritten to use context
@@ -30,40 +34,41 @@ func newDnsRecords(hostMap *HostMap) *dnsRecords {
 
 func (d *dnsRecords) Query(data string) string {
 	d.RLock()
-	if r, ok := d.dnsMap[data]; ok {
-		d.RUnlock()
+	defer d.RUnlock()
+	if r, ok := d.dnsMap[strings.ToLower(data)]; ok {
 		return r
 	}
-	d.RUnlock()
 	return ""
 }
 
 func (d *dnsRecords) QueryCert(data string) string {
-	ip := net.ParseIP(data[:len(data)-1])
-	if ip == nil {
-		return ""
-	}
-	iip := ip2int(ip)
-	hostinfo, err := d.hostMap.QueryVpnIP(iip)
+	ip, err := netip.ParseAddr(data[:len(data)-1])
 	if err != nil {
 		return ""
 	}
+
+	hostinfo := d.hostMap.QueryVpnIp(ip)
+	if hostinfo == nil {
+		return ""
+	}
+
 	q := hostinfo.GetCert()
 	if q == nil {
 		return ""
 	}
+
 	cert := q.Details
-	c := fmt.Sprintf("\"Name: %s\" \"Ips: %s\" \"Subnets %s\" \"Groups %s\" \"NotBefore %s\" \"NotAFter %s\" \"PublicKey %x\" \"IsCA %t\" \"Issuer %s\"", cert.Name, cert.Ips, cert.Subnets, cert.Groups, cert.NotBefore, cert.NotAfter, cert.PublicKey, cert.IsCA, cert.Issuer)
+	c := fmt.Sprintf("\"Name: %s\" \"Ips: %s\" \"Subnets %s\" \"Groups %s\" \"NotBefore %s\" \"NotAfter %s\" \"PublicKey %x\" \"IsCA %t\" \"Issuer %s\"", cert.Name, cert.Ips, cert.Subnets, cert.Groups, cert.NotBefore, cert.NotAfter, cert.PublicKey, cert.IsCA, cert.Issuer)
 	return c
 }
 
 func (d *dnsRecords) Add(host, data string) {
 	d.Lock()
-	d.dnsMap[host] = data
-	d.Unlock()
+	defer d.Unlock()
+	d.dnsMap[strings.ToLower(host)] = data
 }
 
-func parseQuery(m *dns.Msg, w dns.ResponseWriter) {
+func parseQuery(l *logrus.Logger, m *dns.Msg, w dns.ResponseWriter) {
 	for _, q := range m.Question {
 		switch q.Qtype {
 		case dns.TypeA:
@@ -77,7 +82,11 @@ func parseQuery(m *dns.Msg, w dns.ResponseWriter) {
 			}
 		case dns.TypeTXT:
 			a, _, _ := net.SplitHostPort(w.RemoteAddr().String())
-			b := net.ParseIP(a)
+			b, err := netip.ParseAddr(a)
+			if err != nil {
+				return
+			}
+
 			// We don't answer these queries from non nebula nodes or localhost
 			//l.Debugf("Does %s contain %s", b, dnsR.hostMap.vpnCIDR)
 			if !dnsR.hostMap.vpnCIDR.Contains(b) && a != "127.0.0.1" {
@@ -93,39 +102,55 @@ func parseQuery(m *dns.Msg, w dns.ResponseWriter) {
 			}
 		}
 	}
+
+	if len(m.Answer) == 0 {
+		m.Rcode = dns.RcodeNameError
+	}
 }
 
-func handleDnsRequest(w dns.ResponseWriter, r *dns.Msg) {
+func handleDnsRequest(l *logrus.Logger, w dns.ResponseWriter, r *dns.Msg) {
 	m := new(dns.Msg)
 	m.SetReply(r)
 	m.Compress = false
 
 	switch r.Opcode {
 	case dns.OpcodeQuery:
-		parseQuery(m, w)
+		parseQuery(l, m, w)
 	}
 
 	w.WriteMsg(m)
 }
 
-func dnsMain(hostMap *HostMap, c *Config) {
+func dnsMain(l *logrus.Logger, hostMap *HostMap, c *config.C) func() {
 	dnsR = newDnsRecords(hostMap)
 
 	// attach request handler func
-	dns.HandleFunc(".", handleDnsRequest)
+	dns.HandleFunc(".", func(w dns.ResponseWriter, r *dns.Msg) {
+		handleDnsRequest(l, w, r)
+	})
 
-	c.RegisterReloadCallback(reloadDns)
-	startDns(c)
+	c.RegisterReloadCallback(func(c *config.C) {
+		reloadDns(l, c)
+	})
+
+	return func() {
+		startDns(l, c)
+	}
 }
 
-func getDnsServerAddr(c *Config) string {
-	return c.GetString("lighthouse.dns.host", "") + ":" + strconv.Itoa(c.GetInt("lighthouse.dns.port", 53))
+func getDnsServerAddr(c *config.C) string {
+	dnsHost := strings.TrimSpace(c.GetString("lighthouse.dns.host", ""))
+	// Old guidance was to provide the literal `[::]` in `lighthouse.dns.host` but that won't resolve.
+	if dnsHost == "[::]" {
+		dnsHost = "::"
+	}
+	return net.JoinHostPort(dnsHost, strconv.Itoa(c.GetInt("lighthouse.dns.port", 53)))
 }
 
-func startDns(c *Config) {
+func startDns(l *logrus.Logger, c *config.C) {
 	dnsAddr = getDnsServerAddr(c)
 	dnsServer = &dns.Server{Addr: dnsAddr, Net: "udp"}
-	l.Debugf("Starting DNS responder at %s\n", dnsAddr)
+	l.WithField("dnsListener", dnsAddr).Info("Starting DNS responder")
 	err := dnsServer.ListenAndServe()
 	defer dnsServer.Shutdown()
 	if err != nil {
@@ -133,7 +158,7 @@ func startDns(c *Config) {
 	}
 }
 
-func reloadDns(c *Config) {
+func reloadDns(l *logrus.Logger, c *config.C) {
 	if dnsAddr == getDnsServerAddr(c) {
 		l.Debug("No DNS server config change detected")
 		return
@@ -141,5 +166,5 @@ func reloadDns(c *Config) {
 
 	l.Debug("Restarting DNS server")
 	dnsServer.Shutdown()
-	go startDns(c)
+	go startDns(l, c)
 }

dns_server_test.go

@@ -4,6 +4,8 @@ import (
 	"testing"
 
 	"github.com/miekg/dns"
+	"github.com/slackhq/nebula/config"
+	"github.com/stretchr/testify/assert"
 )
 
 func TestParsequery(t *testing.T) {
@@ -17,3 +19,40 @@ func TestParsequery(t *testing.T) {
 
 	//parseQuery(m)
 }
+
+func Test_getDnsServerAddr(t *testing.T) {
+	c := config.NewC(nil)
+
+	c.Settings["lighthouse"] = map[interface{}]interface{}{
+		"dns": map[interface{}]interface{}{
+			"host": "0.0.0.0",
+			"port": "1",
+		},
+	}
+	assert.Equal(t, "0.0.0.0:1", getDnsServerAddr(c))
+
+	c.Settings["lighthouse"] = map[interface{}]interface{}{
+		"dns": map[interface{}]interface{}{
+			"host": "::",
+			"port": "1",
+		},
+	}
+	assert.Equal(t, "[::]:1", getDnsServerAddr(c))
+
+	c.Settings["lighthouse"] = map[interface{}]interface{}{
+		"dns": map[interface{}]interface{}{
+			"host": "[::]",
+			"port": "1",
+		},
+	}
+	assert.Equal(t, "[::]:1", getDnsServerAddr(c))
+
+	// Make sure whitespace doesn't mess us up
+	c.Settings["lighthouse"] = map[interface{}]interface{}{
+		"dns": map[interface{}]interface{}{
+			"host": "[::] ",
+			"port": "1",
+		},
+	}
+	assert.Equal(t, "[::]:1", getDnsServerAddr(c))
+}

docker/Dockerfile

@@ -0,0 +1,11 @@
+FROM gcr.io/distroless/static:latest
+
+ARG TARGETOS TARGETARCH
+COPY build/$TARGETOS-$TARGETARCH/nebula /nebula
+COPY build/$TARGETOS-$TARGETARCH/nebula-cert /nebula-cert
+
+VOLUME ["/config"]
+
+ENTRYPOINT ["/nebula"]
+# Allow users to override the args passed to nebula
+CMD ["-config", "/config/config.yml"]

docker/README.md

@@ -0,0 +1,24 @@
+# NebulaOSS/nebula Docker Image
+
+## Building
+
+From the root of the repository, run `make docker`.
+
+## Running
+
+To run the built image, use the following command:
+
+```
+docker run \
+    --name nebula \
+    --network host \
+    --cap-add NET_ADMIN \
+    --volume ./config:/config \
+    --rm \
+    nebulaoss/nebula
+```
+
+A few notes:
+
+- The `NET_ADMIN` capability is necessary to create the tun adapter on the host (this is unnecessary if the tun device is disabled.)
+- `--volume ./config:/config` should point to a directory that contains your `config.yml` and any other necessary files.

e2e/doc.go

@@ -0,0 +1,3 @@
+package e2e
+
+// This file exists to allow `go fmt` to traverse here on its own. The build tags were keeping it out before

e2e/helpers.go

@@ -0,0 +1,125 @@
+package e2e
+
+import (
+	"crypto/rand"
+	"io"
+	"net"
+	"net/netip"
+	"time"
+
+	"github.com/slackhq/nebula/cert"
+	"golang.org/x/crypto/curve25519"
+	"golang.org/x/crypto/ed25519"
+)
+
+// NewTestCaCert will generate a CA cert
+func NewTestCaCert(before, after time.Time, ips, subnets []netip.Prefix, groups []string) (*cert.NebulaCertificate, []byte, []byte, []byte) {
+	pub, priv, err := ed25519.GenerateKey(rand.Reader)
+	if before.IsZero() {
+		before = time.Now().Add(time.Second * -60).Round(time.Second)
+	}
+	if after.IsZero() {
+		after = time.Now().Add(time.Second * 60).Round(time.Second)
+	}
+
+	nc := &cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:           "test ca",
+			NotBefore:      time.Unix(before.Unix(), 0),
+			NotAfter:       time.Unix(after.Unix(), 0),
+			PublicKey:      pub,
+			IsCA:           true,
+			InvertedGroups: make(map[string]struct{}),
+		},
+	}
+
+	if len(ips) > 0 {
+		nc.Details.Ips = make([]*net.IPNet, len(ips))
+		for i, ip := range ips {
+			nc.Details.Ips[i] = &net.IPNet{IP: ip.Addr().AsSlice(), Mask: net.CIDRMask(ip.Bits(), ip.Addr().BitLen())}
+		}
+	}
+
+	if len(subnets) > 0 {
+		nc.Details.Subnets = make([]*net.IPNet, len(subnets))
+		for i, ip := range subnets {
+			nc.Details.Ips[i] = &net.IPNet{IP: ip.Addr().AsSlice(), Mask: net.CIDRMask(ip.Bits(), ip.Addr().BitLen())}
+		}
+	}
+
+	if len(groups) > 0 {
+		nc.Details.Groups = groups
+	}
+
+	err = nc.Sign(cert.Curve_CURVE25519, priv)
+	if err != nil {
+		panic(err)
+	}
+
+	pem, err := nc.MarshalToPEM()
+	if err != nil {
+		panic(err)
+	}
+
+	return nc, pub, priv, pem
+}
+
+// NewTestCert will generate a signed certificate with the provided details.
+// Expiry times are defaulted if you do not pass them in
+func NewTestCert(ca *cert.NebulaCertificate, key []byte, name string, before, after time.Time, ip netip.Prefix, subnets []netip.Prefix, groups []string) (*cert.NebulaCertificate, []byte, []byte, []byte) {
+	issuer, err := ca.Sha256Sum()
+	if err != nil {
+		panic(err)
+	}
+
+	if before.IsZero() {
+		before = time.Now().Add(time.Second * -60).Round(time.Second)
+	}
+
+	if after.IsZero() {
+		after = time.Now().Add(time.Second * 60).Round(time.Second)
+	}
+
+	pub, rawPriv := x25519Keypair()
+	ipb := ip.Addr().AsSlice()
+	nc := &cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name: name,
+			Ips:  []*net.IPNet{{IP: ipb[:], Mask: net.CIDRMask(ip.Bits(), ip.Addr().BitLen())}},
+			//Subnets:        subnets,
+			Groups:         groups,
+			NotBefore:      time.Unix(before.Unix(), 0),
+			NotAfter:       time.Unix(after.Unix(), 0),
+			PublicKey:      pub,
+			IsCA:           false,
+			Issuer:         issuer,
+			InvertedGroups: make(map[string]struct{}),
+		},
+	}
+
+	err = nc.Sign(ca.Details.Curve, key)
+	if err != nil {
+		panic(err)
+	}
+
+	pem, err := nc.MarshalToPEM()
+	if err != nil {
+		panic(err)
+	}
+
+	return nc, pub, cert.MarshalX25519PrivateKey(rawPriv), pem
+}
+
+func x25519Keypair() ([]byte, []byte) {
+	privkey := make([]byte, 32)
+	if _, err := io.ReadFull(rand.Reader, privkey); err != nil {
+		panic(err)
+	}
+
+	pubkey, err := curve25519.X25519(privkey, curve25519.Basepoint)
+	if err != nil {
+		panic(err)
+	}
+
+	return pubkey, privkey
+}

e2e/helpers_test.go

@@ -0,0 +1,220 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package e2e
+
+import (
+	"fmt"
+	"io"
+	"net/netip"
+	"os"
+	"testing"
+	"time"
+
+	"dario.cat/mergo"
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/e2e/router"
+	"github.com/stretchr/testify/assert"
+	"gopkg.in/yaml.v2"
+)
+
+type m map[string]interface{}
+
+// newSimpleServer creates a nebula instance with many assumptions
+func newSimpleServer(caCrt *cert.NebulaCertificate, caKey []byte, name string, sVpnIpNet string, overrides m) (*nebula.Control, netip.Prefix, netip.AddrPort, *config.C) {
+	l := NewTestLogger()
+
+	vpnIpNet, err := netip.ParsePrefix(sVpnIpNet)
+	if err != nil {
+		panic(err)
+	}
+
+	var udpAddr netip.AddrPort
+	if vpnIpNet.Addr().Is4() {
+		budpIp := vpnIpNet.Addr().As4()
+		budpIp[1] -= 128
+		udpAddr = netip.AddrPortFrom(netip.AddrFrom4(budpIp), 4242)
+	} else {
+		budpIp := vpnIpNet.Addr().As16()
+		budpIp[13] -= 128
+		udpAddr = netip.AddrPortFrom(netip.AddrFrom16(budpIp), 4242)
+	}
+	_, _, myPrivKey, myPEM := NewTestCert(caCrt, caKey, name, time.Now(), time.Now().Add(5*time.Minute), vpnIpNet, nil, []string{})
+
+	caB, err := caCrt.MarshalToPEM()
+	if err != nil {
+		panic(err)
+	}
+
+	mc := m{
+		"pki": m{
+			"ca":   string(caB),
+			"cert": string(myPEM),
+			"key":  string(myPrivKey),
+		},
+		//"tun": m{"disabled": true},
+		"firewall": m{
+			"outbound": []m{{
+				"proto": "any",
+				"port":  "any",
+				"host":  "any",
+			}},
+			"inbound": []m{{
+				"proto": "any",
+				"port":  "any",
+				"host":  "any",
+			}},
+		},
+		//"handshakes": m{
+		//	"try_interval": "1s",
+		//},
+		"listen": m{
+			"host": udpAddr.Addr().String(),
+			"port": udpAddr.Port(),
+		},
+		"logging": m{
+			"timestamp_format": fmt.Sprintf("%v 15:04:05.000000", name),
+			"level":            l.Level.String(),
+		},
+		"timers": m{
+			"pending_deletion_interval": 2,
+			"connection_alive_interval": 2,
+		},
+	}
+
+	if overrides != nil {
+		err = mergo.Merge(&overrides, mc, mergo.WithAppendSlice)
+		if err != nil {
+			panic(err)
+		}
+		mc = overrides
+	}
+
+	cb, err := yaml.Marshal(mc)
+	if err != nil {
+		panic(err)
+	}
+
+	c := config.NewC(l)
+	c.LoadString(string(cb))
+
+	control, err := nebula.Main(c, false, "e2e-test", l, nil)
+
+	if err != nil {
+		panic(err)
+	}
+
+	return control, vpnIpNet, udpAddr, c
+}
+
+type doneCb func()
+
+func deadline(t *testing.T, seconds time.Duration) doneCb {
+	timeout := time.After(seconds * time.Second)
+	done := make(chan bool)
+	go func() {
+		select {
+		case <-timeout:
+			t.Fatal("Test did not finish in time")
+		case <-done:
+		}
+	}()
+
+	return func() {
+		done <- true
+	}
+}
+
+func assertTunnel(t *testing.T, vpnIpA, vpnIpB netip.Addr, controlA, controlB *nebula.Control, r *router.R) {
+	// Send a packet from them to me
+	controlB.InjectTunUDPPacket(vpnIpA, 80, 90, []byte("Hi from B"))
+	bPacket := r.RouteForAllUntilTxTun(controlA)
+	assertUdpPacket(t, []byte("Hi from B"), bPacket, vpnIpB, vpnIpA, 90, 80)
+
+	// And once more from me to them
+	controlA.InjectTunUDPPacket(vpnIpB, 80, 90, []byte("Hello from A"))
+	aPacket := r.RouteForAllUntilTxTun(controlB)
+	assertUdpPacket(t, []byte("Hello from A"), aPacket, vpnIpA, vpnIpB, 90, 80)
+}
+
+func assertHostInfoPair(t *testing.T, addrA, addrB netip.AddrPort, vpnIpA, vpnIpB netip.Addr, controlA, controlB *nebula.Control) {
+	// Get both host infos
+	hBinA := controlA.GetHostInfoByVpnIp(vpnIpB, false)
+	assert.NotNil(t, hBinA, "Host B was not found by vpnIp in controlA")
+
+	hAinB := controlB.GetHostInfoByVpnIp(vpnIpA, false)
+	assert.NotNil(t, hAinB, "Host A was not found by vpnIp in controlB")
+
+	// Check that both vpn and real addr are correct
+	assert.Equal(t, vpnIpB, hBinA.VpnIp, "Host B VpnIp is wrong in control A")
+	assert.Equal(t, vpnIpA, hAinB.VpnIp, "Host A VpnIp is wrong in control B")
+
+	assert.Equal(t, addrB, hBinA.CurrentRemote, "Host B remote is wrong in control A")
+	assert.Equal(t, addrA, hAinB.CurrentRemote, "Host A remote is wrong in control B")
+
+	// Check that our indexes match
+	assert.Equal(t, hBinA.LocalIndex, hAinB.RemoteIndex, "Host B local index does not match host A remote index")
+	assert.Equal(t, hBinA.RemoteIndex, hAinB.LocalIndex, "Host B remote index does not match host A local index")
+
+	//TODO: Would be nice to assert this memory
+	//checkIndexes := func(name string, hm *HostMap, hi *HostInfo) {
+	//	hBbyIndex := hmA.Indexes[hBinA.localIndexId]
+	//	assert.NotNil(t, hBbyIndex, "Could not host info by local index in %s", name)
+	//	assert.Equal(t, &hBbyIndex, &hBinA, "%s Indexes map did not point to the right host info", name)
+	//
+	//	//TODO: remote indexes are susceptible to collision
+	//	hBbyRemoteIndex := hmA.RemoteIndexes[hBinA.remoteIndexId]
+	//	assert.NotNil(t, hBbyIndex, "Could not host info by remote index in %s", name)
+	//	assert.Equal(t, &hBbyRemoteIndex, &hBinA, "%s RemoteIndexes did not point to the right host info", name)
+	//}
+	//
+	//// Check hostmap indexes too
+	//checkIndexes("hmA", hmA, hBinA)
+	//checkIndexes("hmB", hmB, hAinB)
+}
+
+func assertUdpPacket(t *testing.T, expected, b []byte, fromIp, toIp netip.Addr, fromPort, toPort uint16) {
+	packet := gopacket.NewPacket(b, layers.LayerTypeIPv4, gopacket.Lazy)
+	v4 := packet.Layer(layers.LayerTypeIPv4).(*layers.IPv4)
+	assert.NotNil(t, v4, "No ipv4 data found")
+
+	assert.Equal(t, fromIp.AsSlice(), []byte(v4.SrcIP), "Source ip was incorrect")
+	assert.Equal(t, toIp.AsSlice(), []byte(v4.DstIP), "Dest ip was incorrect")
+
+	udp := packet.Layer(layers.LayerTypeUDP).(*layers.UDP)
+	assert.NotNil(t, udp, "No udp data found")
+
+	assert.Equal(t, fromPort, uint16(udp.SrcPort), "Source port was incorrect")
+	assert.Equal(t, toPort, uint16(udp.DstPort), "Dest port was incorrect")
+
+	data := packet.ApplicationLayer()
+	assert.NotNil(t, data)
+	assert.Equal(t, expected, data.Payload(), "Data was incorrect")
+}
+
+func NewTestLogger() *logrus.Logger {
+	l := logrus.New()
+
+	v := os.Getenv("TEST_LOGS")
+	if v == "" {
+		l.SetOutput(io.Discard)
+		l.SetLevel(logrus.PanicLevel)
+		return l
+	}
+
+	switch v {
+	case "2":
+		l.SetLevel(logrus.DebugLevel)
+	case "3":
+		l.SetLevel(logrus.TraceLevel)
+	default:
+		l.SetLevel(logrus.InfoLevel)
+	}
+
+	return l
+}

e2e/router/doc.go

@@ -0,0 +1,3 @@
+package router
+
+// This file exists to allow `go fmt` to traverse here on its own. The build tags were keeping it out before

e2e/router/hostmap.go

@@ -0,0 +1,145 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package router
+
+import (
+	"fmt"
+	"net/netip"
+	"sort"
+	"strings"
+
+	"github.com/slackhq/nebula"
+)
+
+type edge struct {
+	from string
+	to   string
+	dual bool
+}
+
+func renderHostmaps(controls ...*nebula.Control) string {
+	var lines []*edge
+	r := "graph TB\n"
+	for _, c := range controls {
+		sr, se := renderHostmap(c)
+		r += sr
+		for _, e := range se {
+			add := true
+
+			// Collapse duplicate edges into a bi-directionally connected edge
+			for _, ge := range lines {
+				if e.to == ge.from && e.from == ge.to {
+					add = false
+					ge.dual = true
+					break
+				}
+			}
+
+			if add {
+				lines = append(lines, e)
+			}
+		}
+	}
+
+	for _, line := range lines {
+		if line.dual {
+			r += fmt.Sprintf("\t%v <--> %v\n", line.from, line.to)
+		} else {
+			r += fmt.Sprintf("\t%v --> %v\n", line.from, line.to)
+		}
+
+	}
+
+	return r
+}
+
+func renderHostmap(c *nebula.Control) (string, []*edge) {
+	var lines []string
+	var globalLines []*edge
+
+	clusterName := strings.Trim(c.GetCert().Details.Name, " ")
+	clusterVpnIp := c.GetCert().Details.Ips[0].IP
+	r := fmt.Sprintf("\tsubgraph %s[\"%s (%s)\"]\n", clusterName, clusterName, clusterVpnIp)
+
+	hm := c.GetHostmap()
+	hm.RLock()
+	defer hm.RUnlock()
+
+	// Draw the vpn to index nodes
+	r += fmt.Sprintf("\t\tsubgraph %s.hosts[\"Hosts (vpn ip to index)\"]\n", clusterName)
+	hosts := sortedHosts(hm.Hosts)
+	for _, vpnIp := range hosts {
+		hi := hm.Hosts[vpnIp]
+		r += fmt.Sprintf("\t\t\t%v.%v[\"%v\"]\n", clusterName, vpnIp, vpnIp)
+		lines = append(lines, fmt.Sprintf("%v.%v --> %v.%v", clusterName, vpnIp, clusterName, hi.GetLocalIndex()))
+
+		rs := hi.GetRelayState()
+		for _, relayIp := range rs.CopyRelayIps() {
+			lines = append(lines, fmt.Sprintf("%v.%v --> %v.%v", clusterName, vpnIp, clusterName, relayIp))
+		}
+
+		for _, relayIp := range rs.CopyRelayForIdxs() {
+			lines = append(lines, fmt.Sprintf("%v.%v --> %v.%v", clusterName, vpnIp, clusterName, relayIp))
+		}
+	}
+	r += "\t\tend\n"
+
+	// Draw the relay hostinfos
+	if len(hm.Relays) > 0 {
+		r += fmt.Sprintf("\t\tsubgraph %s.relays[\"Relays (relay index to hostinfo)\"]\n", clusterName)
+		for relayIndex, hi := range hm.Relays {
+			r += fmt.Sprintf("\t\t\t%v.%v[\"%v\"]\n", clusterName, relayIndex, relayIndex)
+			lines = append(lines, fmt.Sprintf("%v.%v --> %v.%v", clusterName, relayIndex, clusterName, hi.GetLocalIndex()))
+		}
+		r += "\t\tend\n"
+	}
+
+	// Draw the local index to relay or remote index nodes
+	r += fmt.Sprintf("\t\tsubgraph indexes.%s[\"Indexes (index to hostinfo)\"]\n", clusterName)
+	indexes := sortedIndexes(hm.Indexes)
+	for _, idx := range indexes {
+		hi, ok := hm.Indexes[idx]
+		if ok {
+			r += fmt.Sprintf("\t\t\t%v.%v[\"%v (%v)\"]\n", clusterName, idx, idx, hi.GetVpnIp())
+			remoteClusterName := strings.Trim(hi.GetCert().Details.Name, " ")
+			globalLines = append(globalLines, &edge{from: fmt.Sprintf("%v.%v", clusterName, idx), to: fmt.Sprintf("%v.%v", remoteClusterName, hi.GetRemoteIndex())})
+			_ = hi
+		}
+	}
+	r += "\t\tend\n"
+
+	// Add the edges inside this host
+	for _, line := range lines {
+		r += fmt.Sprintf("\t\t%v\n", line)
+	}
+
+	r += "\tend\n"
+	return r, globalLines
+}
+
+func sortedHosts(hosts map[netip.Addr]*nebula.HostInfo) []netip.Addr {
+	keys := make([]netip.Addr, 0, len(hosts))
+	for key := range hosts {
+		keys = append(keys, key)
+	}
+
+	sort.SliceStable(keys, func(i, j int) bool {
+		return keys[i].Compare(keys[j]) > 0
+	})
+
+	return keys
+}
+
+func sortedIndexes(indexes map[uint32]*nebula.HostInfo) []uint32 {
+	keys := make([]uint32, 0, len(indexes))
+	for key := range indexes {
+		keys = append(keys, key)
+	}
+
+	sort.SliceStable(keys, func(i, j int) bool {
+		return keys[i] > keys[j]
+	})
+
+	return keys
+}

e2e/router/router.go

@@ -0,0 +1,741 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package router
+
+import (
+	"context"
+	"fmt"
+	"net/netip"
+	"os"
+	"path/filepath"
+	"reflect"
+	"sort"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/udp"
+	"golang.org/x/exp/maps"
+)
+
+type R struct {
+	// Simple map of the ip:port registered on a control to the control
+	// Basically a router, right?
+	controls map[netip.AddrPort]*nebula.Control
+
+	// A map for inbound packets for a control that doesn't know about this address
+	inNat map[netip.AddrPort]*nebula.Control
+
+	// A last used map, if an inbound packet hit the inNat map then
+	// all return packets should use the same last used inbound address for the outbound sender
+	// map[from address + ":" + to address] => ip:port to rewrite in the udp packet to receiver
+	outNat map[string]netip.AddrPort
+
+	// A map of vpn ip to the nebula control it belongs to
+	vpnControls map[netip.Addr]*nebula.Control
+
+	ignoreFlows []ignoreFlow
+	flow        []flowEntry
+
+	// A set of additional mermaid graphs to draw in the flow log markdown file
+	// Currently consisting only of hostmap renders
+	additionalGraphs []mermaidGraph
+
+	// All interactions are locked to help serialize behavior
+	sync.Mutex
+
+	fn           string
+	cancelRender context.CancelFunc
+	t            testing.TB
+}
+
+type ignoreFlow struct {
+	tun         NullBool
+	messageType header.MessageType
+	subType     header.MessageSubType
+	//from
+	//to
+}
+
+type mermaidGraph struct {
+	title   string
+	content string
+}
+
+type NullBool struct {
+	HasValue bool
+	IsTrue   bool
+}
+
+type flowEntry struct {
+	note   string
+	packet *packet
+}
+
+type packet struct {
+	from   *nebula.Control
+	to     *nebula.Control
+	packet *udp.Packet
+	tun    bool // a packet pulled off a tun device
+	rx     bool // the packet was received by a udp device
+}
+
+func (p *packet) WasReceived() {
+	if p != nil {
+		p.rx = true
+	}
+}
+
+type ExitType int
+
+const (
+	// KeepRouting the function will get called again on the next packet
+	KeepRouting ExitType = 0
+	// ExitNow does not route this packet and exits immediately
+	ExitNow ExitType = 1
+	// RouteAndExit routes this packet and exits immediately afterwards
+	RouteAndExit ExitType = 2
+)
+
+type ExitFunc func(packet *udp.Packet, receiver *nebula.Control) ExitType
+
+// NewR creates a new router to pass packets in a controlled fashion between the provided controllers.
+// The packet flow will be recorded in a file within the mermaid directory under the same name as the test.
+// Renders will occur automatically, roughly every 100ms, until a call to RenderFlow() is made
+func NewR(t testing.TB, controls ...*nebula.Control) *R {
+	ctx, cancel := context.WithCancel(context.Background())
+
+	if err := os.MkdirAll("mermaid", 0755); err != nil {
+		panic(err)
+	}
+
+	r := &R{
+		controls:     make(map[netip.AddrPort]*nebula.Control),
+		vpnControls:  make(map[netip.Addr]*nebula.Control),
+		inNat:        make(map[netip.AddrPort]*nebula.Control),
+		outNat:       make(map[string]netip.AddrPort),
+		flow:         []flowEntry{},
+		ignoreFlows:  []ignoreFlow{},
+		fn:           filepath.Join("mermaid", fmt.Sprintf("%s.md", t.Name())),
+		t:            t,
+		cancelRender: cancel,
+	}
+
+	// Try to remove our render file
+	os.Remove(r.fn)
+
+	for _, c := range controls {
+		addr := c.GetUDPAddr()
+		if _, ok := r.controls[addr]; ok {
+			panic("Duplicate listen address: " + addr.String())
+		}
+
+		r.vpnControls[c.GetVpnIp()] = c
+		r.controls[addr] = c
+	}
+
+	// Spin the renderer in case we go nuts and the test never completes
+	go func() {
+		clockSource := time.NewTicker(time.Millisecond * 100)
+		defer clockSource.Stop()
+
+		for {
+			select {
+			case <-ctx.Done():
+				return
+			case <-clockSource.C:
+				r.renderHostmaps("clock tick")
+				r.renderFlow()
+			}
+		}
+	}()
+
+	return r
+}
+
+// AddRoute will place the nebula controller at the ip and port specified.
+// It does not look at the addr attached to the instance.
+// If a route is used, this will behave like a NAT for the return path.
+// Rewriting the source ip:port to what was last sent to from the origin
+func (r *R) AddRoute(ip netip.Addr, port uint16, c *nebula.Control) {
+	r.Lock()
+	defer r.Unlock()
+
+	inAddr := netip.AddrPortFrom(ip, port)
+	if _, ok := r.inNat[inAddr]; ok {
+		panic("Duplicate listen address inNat: " + inAddr.String())
+	}
+	r.inNat[inAddr] = c
+}
+
+// RenderFlow renders the packet flow seen up until now and stops further automatic renders from happening.
+func (r *R) RenderFlow() {
+	r.cancelRender()
+	r.renderFlow()
+}
+
+// CancelFlowLogs stops flow logs from being tracked and destroys any logs already collected
+func (r *R) CancelFlowLogs() {
+	r.cancelRender()
+	r.flow = nil
+}
+
+func (r *R) renderFlow() {
+	if r.flow == nil {
+		return
+	}
+
+	f, err := os.OpenFile(r.fn, os.O_CREATE|os.O_TRUNC|os.O_RDWR, 0644)
+	if err != nil {
+		panic(err)
+	}
+
+	var participants = map[netip.AddrPort]struct{}{}
+	var participantsVals []string
+
+	fmt.Fprintln(f, "```mermaid")
+	fmt.Fprintln(f, "sequenceDiagram")
+
+	// Assemble participants
+	for _, e := range r.flow {
+		if e.packet == nil {
+			continue
+		}
+
+		addr := e.packet.from.GetUDPAddr()
+		if _, ok := participants[addr]; ok {
+			continue
+		}
+		participants[addr] = struct{}{}
+		sanAddr := strings.Replace(addr.String(), ":", "-", 1)
+		participantsVals = append(participantsVals, sanAddr)
+		fmt.Fprintf(
+			f, "    participant %s as Nebula: %s<br/>UDP: %s\n",
+			sanAddr, e.packet.from.GetVpnIp(), sanAddr,
+		)
+	}
+
+	if len(participantsVals) > 2 {
+		// Get the first and last participantVals for notes
+		participantsVals = []string{participantsVals[0], participantsVals[len(participantsVals)-1]}
+	}
+
+	// Print packets
+	h := &header.H{}
+	for _, e := range r.flow {
+		if e.packet == nil {
+			//fmt.Fprintf(f, "    note over %s: %s\n", strings.Join(participantsVals, ", "), e.note)
+			continue
+		}
+
+		p := e.packet
+		if p.tun {
+			fmt.Fprintln(f, r.formatUdpPacket(p))
+
+		} else {
+			if err := h.Parse(p.packet.Data); err != nil {
+				panic(err)
+			}
+
+			line := "--x"
+			if p.rx {
+				line = "->>"
+			}
+
+			fmt.Fprintf(f,
+				"    %s%s%s: %s(%s), index %v, counter: %v\n",
+				strings.Replace(p.from.GetUDPAddr().String(), ":", "-", 1),
+				line,
+				strings.Replace(p.to.GetUDPAddr().String(), ":", "-", 1),
+				h.TypeName(), h.SubTypeName(), h.RemoteIndex, h.MessageCounter,
+			)
+		}
+	}
+	fmt.Fprintln(f, "```")
+
+	for _, g := range r.additionalGraphs {
+		fmt.Fprintf(f, "## %s\n", g.title)
+		fmt.Fprintln(f, "```mermaid")
+		fmt.Fprintln(f, g.content)
+		fmt.Fprintln(f, "```")
+	}
+}
+
+// IgnoreFlow tells the router to stop recording future flows that matches the provided criteria.
+// messageType and subType will target nebula underlay packets while tun will target nebula overlay packets
+// NOTE: This is a very broad system, if you set tun to true then no more tun traffic will be rendered
+func (r *R) IgnoreFlow(messageType header.MessageType, subType header.MessageSubType, tun NullBool) {
+	r.Lock()
+	defer r.Unlock()
+	r.ignoreFlows = append(r.ignoreFlows, ignoreFlow{
+		tun,
+		messageType,
+		subType,
+	})
+}
+
+func (r *R) RenderHostmaps(title string, controls ...*nebula.Control) {
+	r.Lock()
+	defer r.Unlock()
+
+	s := renderHostmaps(controls...)
+	if len(r.additionalGraphs) > 0 {
+		lastGraph := r.additionalGraphs[len(r.additionalGraphs)-1]
+		if lastGraph.content == s && lastGraph.title == title {
+			// Ignore this rendering if it matches the last rendering added
+			// This is useful if you want to track rendering changes
+			return
+		}
+	}
+
+	r.additionalGraphs = append(r.additionalGraphs, mermaidGraph{
+		title:   title,
+		content: s,
+	})
+}
+
+func (r *R) renderHostmaps(title string) {
+	c := maps.Values(r.controls)
+	sort.SliceStable(c, func(i, j int) bool {
+		return c[i].GetVpnIp().Compare(c[j].GetVpnIp()) > 0
+	})
+
+	s := renderHostmaps(c...)
+	if len(r.additionalGraphs) > 0 {
+		lastGraph := r.additionalGraphs[len(r.additionalGraphs)-1]
+		if lastGraph.content == s {
+			// Ignore this rendering if it matches the last rendering added
+			// This is useful if you want to track rendering changes
+			return
+		}
+	}
+
+	r.additionalGraphs = append(r.additionalGraphs, mermaidGraph{
+		title:   title,
+		content: s,
+	})
+}
+
+// InjectFlow can be used to record packet flow if the test is handling the routing on its own.
+// The packet is assumed to have been received
+func (r *R) InjectFlow(from, to *nebula.Control, p *udp.Packet) {
+	r.Lock()
+	defer r.Unlock()
+	r.unlockedInjectFlow(from, to, p, false)
+}
+
+func (r *R) Log(arg ...any) {
+	if r.flow == nil {
+		return
+	}
+
+	r.Lock()
+	r.flow = append(r.flow, flowEntry{note: fmt.Sprint(arg...)})
+	r.t.Log(arg...)
+	r.Unlock()
+}
+
+func (r *R) Logf(format string, arg ...any) {
+	if r.flow == nil {
+		return
+	}
+
+	r.Lock()
+	r.flow = append(r.flow, flowEntry{note: fmt.Sprintf(format, arg...)})
+	r.t.Logf(format, arg...)
+	r.Unlock()
+}
+
+// unlockedInjectFlow is used by the router to record a packet has been transmitted, the packet is returned and
+// should be marked as received AFTER it has been placed on the receivers channel.
+// If flow logs have been disabled this function will return nil
+func (r *R) unlockedInjectFlow(from, to *nebula.Control, p *udp.Packet, tun bool) *packet {
+	if r.flow == nil {
+		return nil
+	}
+
+	r.renderHostmaps(fmt.Sprintf("Packet %v", len(r.flow)))
+
+	if len(r.ignoreFlows) > 0 {
+		var h header.H
+		err := h.Parse(p.Data)
+		if err != nil {
+			panic(err)
+		}
+
+		for _, i := range r.ignoreFlows {
+			if !tun {
+				if i.messageType == h.Type && i.subType == h.Subtype {
+					return nil
+				}
+			} else if i.tun.HasValue && i.tun.IsTrue {
+				return nil
+			}
+		}
+	}
+
+	fp := &packet{
+		from:   from,
+		to:     to,
+		packet: p.Copy(),
+		tun:    tun,
+	}
+
+	r.flow = append(r.flow, flowEntry{packet: fp})
+	return fp
+}
+
+// OnceFrom will route a single packet from sender then return
+// If the router doesn't have the nebula controller for that address, we panic
+func (r *R) OnceFrom(sender *nebula.Control) {
+	r.RouteExitFunc(sender, func(*udp.Packet, *nebula.Control) ExitType {
+		return RouteAndExit
+	})
+}
+
+// RouteUntilTxTun will route for sender and return when a packet is seen on receivers tun
+// If the router doesn't have the nebula controller for that address, we panic
+func (r *R) RouteUntilTxTun(sender *nebula.Control, receiver *nebula.Control) []byte {
+	tunTx := receiver.GetTunTxChan()
+	udpTx := sender.GetUDPTxChan()
+
+	for {
+		select {
+		// Maybe we already have something on the tun for us
+		case b := <-tunTx:
+			r.Lock()
+			np := udp.Packet{Data: make([]byte, len(b))}
+			copy(np.Data, b)
+			r.unlockedInjectFlow(receiver, receiver, &np, true)
+			r.Unlock()
+			return b
+
+		// Nope, lets push the sender along
+		case p := <-udpTx:
+			r.Lock()
+			c := r.getControl(sender.GetUDPAddr(), p.To, p)
+			if c == nil {
+				r.Unlock()
+				panic("No control for udp tx")
+			}
+			fp := r.unlockedInjectFlow(sender, c, p, false)
+			c.InjectUDPPacket(p)
+			fp.WasReceived()
+			r.Unlock()
+		}
+	}
+}
+
+// RouteForAllUntilTxTun will route for everyone and return when a packet is seen on receivers tun
+// If the router doesn't have the nebula controller for that address, we panic
+func (r *R) RouteForAllUntilTxTun(receiver *nebula.Control) []byte {
+	sc := make([]reflect.SelectCase, len(r.controls)+1)
+	cm := make([]*nebula.Control, len(r.controls)+1)
+
+	i := 0
+	sc[i] = reflect.SelectCase{
+		Dir:  reflect.SelectRecv,
+		Chan: reflect.ValueOf(receiver.GetTunTxChan()),
+		Send: reflect.Value{},
+	}
+	cm[i] = receiver
+
+	i++
+	for _, c := range r.controls {
+		sc[i] = reflect.SelectCase{
+			Dir:  reflect.SelectRecv,
+			Chan: reflect.ValueOf(c.GetUDPTxChan()),
+			Send: reflect.Value{},
+		}
+
+		cm[i] = c
+		i++
+	}
+
+	for {
+		x, rx, _ := reflect.Select(sc)
+		r.Lock()
+
+		if x == 0 {
+			// we are the tun tx, we can exit
+			p := rx.Interface().([]byte)
+			np := udp.Packet{Data: make([]byte, len(p))}
+			copy(np.Data, p)
+
+			r.unlockedInjectFlow(cm[x], cm[x], &np, true)
+			r.Unlock()
+			return p
+
+		} else {
+			// we are a udp tx, route and continue
+			p := rx.Interface().(*udp.Packet)
+			c := r.getControl(cm[x].GetUDPAddr(), p.To, p)
+			if c == nil {
+				r.Unlock()
+				panic("No control for udp tx")
+			}
+			fp := r.unlockedInjectFlow(cm[x], c, p, false)
+			c.InjectUDPPacket(p)
+			fp.WasReceived()
+		}
+		r.Unlock()
+	}
+}
+
+// RouteExitFunc will call the whatDo func with each udp packet from sender.
+// whatDo can return:
+//   - exitNow: the packet will not be routed and this call will return immediately
+//   - routeAndExit: this call will return immediately after routing the last packet from sender
+//   - keepRouting: the packet will be routed and whatDo will be called again on the next packet from sender
+func (r *R) RouteExitFunc(sender *nebula.Control, whatDo ExitFunc) {
+	h := &header.H{}
+	for {
+		p := sender.GetFromUDP(true)
+		r.Lock()
+		if err := h.Parse(p.Data); err != nil {
+			panic(err)
+		}
+
+		receiver := r.getControl(sender.GetUDPAddr(), p.To, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't RouteExitFunc for host: " + p.To.String())
+		}
+
+		e := whatDo(p, receiver)
+		switch e {
+		case ExitNow:
+			r.Unlock()
+			return
+
+		case RouteAndExit:
+			fp := r.unlockedInjectFlow(sender, receiver, p, false)
+			receiver.InjectUDPPacket(p)
+			fp.WasReceived()
+			r.Unlock()
+			return
+
+		case KeepRouting:
+			fp := r.unlockedInjectFlow(sender, receiver, p, false)
+			receiver.InjectUDPPacket(p)
+			fp.WasReceived()
+
+		default:
+			panic(fmt.Sprintf("Unknown exitFunc return: %v", e))
+		}
+
+		r.Unlock()
+	}
+}
+
+// RouteUntilAfterMsgType will route for sender until a message type is seen and sent from sender
+// If the router doesn't have the nebula controller for that address, we panic
+func (r *R) RouteUntilAfterMsgType(sender *nebula.Control, msgType header.MessageType, subType header.MessageSubType) {
+	h := &header.H{}
+	r.RouteExitFunc(sender, func(p *udp.Packet, r *nebula.Control) ExitType {
+		if err := h.Parse(p.Data); err != nil {
+			panic(err)
+		}
+		if h.Type == msgType && h.Subtype == subType {
+			return RouteAndExit
+		}
+
+		return KeepRouting
+	})
+}
+
+func (r *R) RouteForAllUntilAfterMsgTypeTo(receiver *nebula.Control, msgType header.MessageType, subType header.MessageSubType) {
+	h := &header.H{}
+	r.RouteForAllExitFunc(func(p *udp.Packet, r *nebula.Control) ExitType {
+		if r != receiver {
+			return KeepRouting
+		}
+
+		if err := h.Parse(p.Data); err != nil {
+			panic(err)
+		}
+
+		if h.Type == msgType && h.Subtype == subType {
+			return RouteAndExit
+		}
+
+		return KeepRouting
+	})
+}
+
+func (r *R) InjectUDPPacket(sender, receiver *nebula.Control, packet *udp.Packet) {
+	r.Lock()
+	defer r.Unlock()
+
+	fp := r.unlockedInjectFlow(sender, receiver, packet, false)
+	receiver.InjectUDPPacket(packet)
+	fp.WasReceived()
+}
+
+// RouteForUntilAfterToAddr will route for sender and return only after it sees and sends a packet destined for toAddr
+// finish can be any of the exitType values except `keepRouting`, the default value is `routeAndExit`
+// If the router doesn't have the nebula controller for that address, we panic
+func (r *R) RouteForUntilAfterToAddr(sender *nebula.Control, toAddr netip.AddrPort, finish ExitType) {
+	if finish == KeepRouting {
+		finish = RouteAndExit
+	}
+
+	r.RouteExitFunc(sender, func(p *udp.Packet, r *nebula.Control) ExitType {
+		if p.To == toAddr {
+			return finish
+		}
+
+		return KeepRouting
+	})
+}
+
+// RouteForAllExitFunc will route for every registered controller and calls the whatDo func with each udp packet from
+// whatDo can return:
+//   - exitNow: the packet will not be routed and this call will return immediately
+//   - routeAndExit: this call will return immediately after routing the last packet from sender
+//   - keepRouting: the packet will be routed and whatDo will be called again on the next packet from sender
+func (r *R) RouteForAllExitFunc(whatDo ExitFunc) {
+	sc := make([]reflect.SelectCase, len(r.controls))
+	cm := make([]*nebula.Control, len(r.controls))
+
+	i := 0
+	for _, c := range r.controls {
+		sc[i] = reflect.SelectCase{
+			Dir:  reflect.SelectRecv,
+			Chan: reflect.ValueOf(c.GetUDPTxChan()),
+			Send: reflect.Value{},
+		}
+
+		cm[i] = c
+		i++
+	}
+
+	for {
+		x, rx, _ := reflect.Select(sc)
+		r.Lock()
+
+		p := rx.Interface().(*udp.Packet)
+		receiver := r.getControl(cm[x].GetUDPAddr(), p.To, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't RouteForAllExitFunc for host: " + p.To.String())
+		}
+
+		e := whatDo(p, receiver)
+		switch e {
+		case ExitNow:
+			r.Unlock()
+			return
+
+		case RouteAndExit:
+			fp := r.unlockedInjectFlow(cm[x], receiver, p, false)
+			receiver.InjectUDPPacket(p)
+			fp.WasReceived()
+			r.Unlock()
+			return
+
+		case KeepRouting:
+			fp := r.unlockedInjectFlow(cm[x], receiver, p, false)
+			receiver.InjectUDPPacket(p)
+			fp.WasReceived()
+
+		default:
+			panic(fmt.Sprintf("Unknown exitFunc return: %v", e))
+		}
+		r.Unlock()
+	}
+}
+
+// FlushAll will route for every registered controller, exiting once there are no packets left to route
+func (r *R) FlushAll() {
+	sc := make([]reflect.SelectCase, len(r.controls))
+	cm := make([]*nebula.Control, len(r.controls))
+
+	i := 0
+	for _, c := range r.controls {
+		sc[i] = reflect.SelectCase{
+			Dir:  reflect.SelectRecv,
+			Chan: reflect.ValueOf(c.GetUDPTxChan()),
+			Send: reflect.Value{},
+		}
+
+		cm[i] = c
+		i++
+	}
+
+	// Add a default case to exit when nothing is left to send
+	sc = append(sc, reflect.SelectCase{
+		Dir:  reflect.SelectDefault,
+		Chan: reflect.Value{},
+		Send: reflect.Value{},
+	})
+
+	for {
+		x, rx, ok := reflect.Select(sc)
+		if !ok {
+			return
+		}
+		r.Lock()
+
+		p := rx.Interface().(*udp.Packet)
+
+		receiver := r.getControl(cm[x].GetUDPAddr(), p.To, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't FlushAll for host: " + p.To.String())
+		}
+		receiver.InjectUDPPacket(p)
+		r.Unlock()
+	}
+}
+
+// getControl performs or seeds NAT translation and returns the control for toAddr, p from fields may change
+// This is an internal router function, the caller must hold the lock
+func (r *R) getControl(fromAddr, toAddr netip.AddrPort, p *udp.Packet) *nebula.Control {
+	if newAddr, ok := r.outNat[fromAddr.String()+":"+toAddr.String()]; ok {
+		p.From = newAddr
+	}
+
+	c, ok := r.inNat[toAddr]
+	if ok {
+		r.outNat[c.GetUDPAddr().String()+":"+fromAddr.String()] = toAddr
+		return c
+	}
+
+	return r.controls[toAddr]
+}
+
+func (r *R) formatUdpPacket(p *packet) string {
+	packet := gopacket.NewPacket(p.packet.Data, layers.LayerTypeIPv4, gopacket.Lazy)
+	v4 := packet.Layer(layers.LayerTypeIPv4).(*layers.IPv4)
+	if v4 == nil {
+		panic("not an ipv4 packet")
+	}
+
+	from := "unknown"
+	srcAddr, _ := netip.AddrFromSlice(v4.SrcIP)
+	if c, ok := r.vpnControls[srcAddr]; ok {
+		from = c.GetUDPAddr().String()
+	}
+
+	udp := packet.Layer(layers.LayerTypeUDP).(*layers.UDP)
+	if udp == nil {
+		panic("not a udp packet")
+	}
+
+	data := packet.ApplicationLayer()
+	return fmt.Sprintf(
+		"    %s-->>%s: src port: %v<br/>dest port: %v<br/>data: \"%v\"\n",
+		strings.Replace(from, ":", "-", 1),
+		strings.Replace(p.to.GetUDPAddr().String(), ":", "-", 1),
+		udp.SrcPort,
+		udp.DstPort,
+		string(data.Payload()),
+	)
+}

e2e/tunnels_test.go

@@ -0,0 +1,55 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package e2e
+
+import (
+	"testing"
+	"time"
+
+	"github.com/slackhq/nebula/e2e/router"
+)
+
+func TestDropInactiveTunnels(t *testing.T) {
+	// The goal of this test is to ensure the shortest inactivity timeout will close the tunnel on both sides
+	// under ideal conditions
+	ca, _, caKey, _ := NewTestCaCert(time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
+	myControl, myVpnIpNet, myUdpAddr, _ := newSimpleServer(ca, caKey, "me", "10.128.0.1/24", m{"tunnels": m{"drop_inactive": true, "inactivity_timeout": "5s"}})
+	theirControl, theirVpnIpNet, theirUdpAddr, _ := newSimpleServer(ca, caKey, "them", "10.128.0.2/24", m{"tunnels": m{"drop_inactive": true, "inactivity_timeout": "10m"}})
+
+	// Share our underlay information
+	myControl.InjectLightHouseAddr(theirVpnIpNet.Addr(), theirUdpAddr)
+	theirControl.InjectLightHouseAddr(myVpnIpNet.Addr(), myUdpAddr)
+
+	// Start the servers
+	myControl.Start()
+	theirControl.Start()
+
+	r := router.NewR(t, myControl, theirControl)
+
+	r.Log("Assert the tunnel between me and them works")
+	assertTunnel(t, myVpnIpNet.Addr(), theirVpnIpNet.Addr(), myControl, theirControl, r)
+
+	r.Log("Go inactive and wait for the tunnels to get dropped")
+	waitStart := time.Now()
+	for {
+		myIndexes := len(myControl.GetHostmap().Indexes)
+		theirIndexes := len(theirControl.GetHostmap().Indexes)
+		if myIndexes == 0 && theirIndexes == 0 {
+			break
+		}
+
+		since := time.Since(waitStart)
+		r.Logf("my tunnels: %v; their tunnels: %v; duration: %v", myIndexes, theirIndexes, since)
+		if since > time.Second*30 {
+			t.Fatal("Tunnel should have been declared inactive after 5 seconds and before 30 seconds")
+		}
+
+		time.Sleep(1 * time.Second)
+		r.FlushAll()
+	}
+
+	r.Logf("Inactive tunnels were dropped within %v", time.Since(waitStart))
+	myControl.Stop()
+	theirControl.Stop()
+}

examples/config.yml

@@ -7,9 +7,11 @@ pki:
   ca: /etc/nebula/ca.crt
   cert: /etc/nebula/host.crt
   key: /etc/nebula/host.key
-  #blacklist is a list of certificate fingerprints that we will refuse to talk to
-  #blacklist:
+  # blocklist is a list of certificate fingerprints that we will refuse to talk to
+  #blocklist:
   #  - c99d4e650533b92061b09918e838a5a0a6aaee21eed1d12fd937682865936c72
+  # disconnect_invalid is a toggle to force a client to be disconnected if the certificate is expired or invalid.
+  #disconnect_invalid: true
 
 # The static host map defines a set of hosts with fixed IP addresses on the internet (or any network).
 # A host can have multiple fixed IP addresses defined here, and nebula will try each when establishing a tunnel.
@@ -19,6 +21,19 @@ pki:
 static_host_map:
   "192.168.100.1": ["100.64.22.11:4242"]
 
+# The static_map config stanza can be used to configure how the static_host_map behaves.
+#static_map:
+  # cadence determines how frequently DNS is re-queried for updated IP addresses when a static_host_map entry contains
+  # a DNS name.
+  #cadence: 30s
+
+  # network determines the type of IP addresses to ask the DNS server for. The default is "ip4" because nodes typically
+  # do not know their public IPv4 address. Connecting to the Lighthouse via IPv4 allows the Lighthouse to detect the
+  # public address. Other valid options are "ip6" and "ip" (returns both.)
+  #network: ip4
+
+  # lookup_timeout is the DNS query timeout.
+  #lookup_timeout: 250ms
 
 lighthouse:
   # am_lighthouse is used to enable lighthouse functionality for a node. This should ONLY be true on nodes
@@ -45,8 +60,9 @@ lighthouse:
   # allowed. You can provide CIDRs here with `true` to allow and `false` to
   # deny. The most specific CIDR rule applies to each remote. If all rules are
   # "allow", the default will be "deny", and vice-versa. If both "allow" and
-  # "deny" rules are present, then you MUST set a rule for "0.0.0.0/0" as the
-  # default.
+  # "deny" IPv4 rules are present, then you MUST set a rule for "0.0.0.0/0" as
+  # the default. Similarly if both "allow" and "deny" IPv6 rules are present,
+  # then you MUST set a rule for "::/0" as the default.
   #remote_allow_list:
     # Example to block IPs from this subnet from being used for remote IPs.
     #"172.16.0.0/12": false
@@ -56,6 +72,14 @@ lighthouse:
     #"10.0.0.0/8": false
     #"10.42.42.0/24": true
 
+  # EXPERIMENTAL: This option may change or disappear in the future.
+  # Optionally allows the definition of remote_allow_list blocks
+  # specific to an inside VPN IP CIDR.
+  #remote_allow_ranges:
+    # This rule would only allow only private IPs for this VPN range
+    #"10.42.42.0/24":
+      #"192.168.0.0/16": true
+
   # local_allow_list allows you to filter which local IP addresses we advertise
   # to the lighthouses. This uses the same logic as `remote_allow_list`, but
   # additionally, you can specify an `interfaces` map of regular expressions
@@ -64,16 +88,39 @@ lighthouse:
   # the inverse). CIDR rules are matched after interface name rules.
   # Default is all local IP addresses.
   #local_allow_list:
-    # Example to blacklist tun0 and all docker interfaces.
+    # Example to block tun0 and all docker interfaces.
     #interfaces:
       #tun0: false
       #'docker.*': false
     # Example to only advertise this subnet to the lighthouse.
     #"10.0.0.0/8": true
 
+  # advertise_addrs are routable addresses that will be included along with discovered addresses to report to the
+  # lighthouse, the format is "ip:port". `port` can be `0`, in which case the actual listening port will be used in its
+  # place, useful if `listen.port` is set to 0.
+  # This option is mainly useful when there are static ip addresses the host can be reached at that nebula can not
+  # typically discover on its own. Examples being port forwarding or multiple paths to the internet.
+  #advertise_addrs:
+    #- "1.1.1.1:4242"
+    #- "1.2.3.4:0" # port will be replaced with the real listening port
+
+  # EXPERIMENTAL: This option may change or disappear in the future.
+  # This setting allows us to "guess" what the remote might be for a host
+  # while we wait for the lighthouse response.
+  #calculated_remotes:
+    # For any Nebula IPs in 10.0.10.0/24, this will apply the mask and add
+    # the calculated IP as an initial remote (while we wait for the response
+    # from the lighthouse). Both CIDRs must have the same mask size.
+    # For example, Nebula IP 10.0.10.123 will have a calculated remote of
+    # 192.168.1.123
+    #10.0.10.0/24:
+      #- mask: 192.168.1.0/24
+      #  port: 4242
+
 # Port Nebula will be listening on. The default here is 4242. For a lighthouse node, the port should be defined,
 # however using port 0 will dynamically assign a port and is recommended for roaming nodes.
 listen:
+  # To listen on both any ipv4 and ipv6 use "::"
   host: 0.0.0.0
   port: 4242
   # Sets the max number of packets to pull from the kernel for each syscall (under systems that support recvmmsg)
@@ -85,6 +132,19 @@ listen:
   # max, net.core.rmem_max and net.core.wmem_max
   #read_buffer: 10485760
   #write_buffer: 10485760
+  # By default, Nebula replies to packets it has no tunnel for with a "recv_error" packet. This packet helps speed up reconnection
+  # in the case that Nebula on either side did not shut down cleanly. This response can be abused as a way to discover if Nebula is running
+  # on a host though. This option lets you configure if you want to send "recv_error" packets always, never, or only to private network remotes.
+  # valid values: always, never, private
+  # This setting is reloadable.
+  #send_recv_error: always
+
+# Routines is the number of thread pairs to run that consume from the tun and UDP queues.
+# Currently, this defaults to 1 which means we have 1 tun queue reader and 1
+# UDP queue reader. Setting this above one will set IFF_MULTI_QUEUE on the tun
+# device and SO_REUSEPORT on the UDP socket to allow multiple queues.
+# This option is only supported on Linux.
+#routines: 1
 
 punchy:
   # Continues to punch inbound/outbound at a regular interval to avoid expiration of firewall nat mappings
@@ -95,18 +155,23 @@ punchy:
   # Default is false
   #respond: true
 
-  # delays a punch response for misbehaving NATs, default is 1 second, respond must be true to take effect
+  # delays a punch response for misbehaving NATs, default is 1 second.
   #delay: 1s
 
-# Cipher allows you to choose between the available ciphers for your network.
+  # set the delay before attempting punchy.respond. Default is 5 seconds. respond must be true to take effect.
+  #respond_delay: 5s
+
+# Cipher allows you to choose between the available ciphers for your network. Options are chachapoly or aes
 # IMPORTANT: this value must be identical on ALL NODES/LIGHTHOUSES. We do not/will not support use of different ciphers simultaneously!
-#cipher: chachapoly
+#cipher: aes
 
-# Local range is used to define a hint about the local network range, which speeds up discovering the fastest
+# Preferred ranges is used to define a hint about the local network ranges, which speeds up discovering the fastest
 # path to a network adjacent nebula node.
-#local_range: "172.16.0.0/24"
+# This setting is reloadable.
+#preferred_ranges: ["172.16.0.0/24"]
 
-# sshd can expose informational and administrative functions via ssh this is a
+# sshd can expose informational and administrative functions via ssh. This can expose informational and administrative
+# functions, and allows manual tweaking of various network settings when debugging or testing.
 #sshd:
   # Toggles the feature
   #enabled: true
@@ -115,16 +180,37 @@ punchy:
   # A file containing the ssh host private key to use
   # A decent way to generate one: ssh-keygen -t ed25519 -f ssh_host_ed25519_key -N "" < /dev/null
   #host_key: ./ssh_host_ed25519_key
-  # A file containing a list of authorized public keys
+  # Authorized users and their public keys
   #authorized_users:
     #- user: steeeeve
       # keys can be an array of strings or single string
       #keys:
         #- "ssh public key string"
+  # Trusted SSH CA public keys. These are the public keys of the CAs that are allowed to sign SSH keys for access.
+  #trusted_cas:
+    #- "ssh public key string"
+
+# EXPERIMENTAL: relay support for networks that can't establish direct connections.
+relay:
+  # Relays are a list of Nebula IP's that peers can use to relay packets to me.
+  # IPs in this list must have am_relay set to true in their configs, otherwise
+  # they will reject relay requests.
+  #relays:
+    #- 192.168.100.1
+    #- <other Nebula VPN IPs of hosts used as relays to access me>
+  # Set am_relay to true to permit other hosts to list my IP in their relays config. Default false.
+  am_relay: false
+  # Set use_relays to false to prevent this instance from attempting to establish connections through relays.
+  # default true
+  use_relays: true
 
 # Configure the private interface. Note: addr is baked into the nebula certificate
 tun:
-  # Name of the device
+  # When tun is disabled, a lighthouse can be started without a local tun interface (and therefore without root)
+  disabled: false
+  # Name of the device. If not set, a default will be chosen by the OS.
+  # For macOS: if set, must be in the form `utun[0-9]+`.
+  # For NetBSD: Required to be set, must be in the form `tun[0-9]+`
   dev: nebula1
   # Toggles forwarding of local broadcast packets, the address of which depends on the ip/mask encoded in pki.cert
   drop_local_broadcast: false
@@ -134,26 +220,42 @@ tun:
   tx_queue: 500
   # Default MTU for every packet, safe setting is (and the default) 1300 for internet based traffic
   mtu: 1300
+
   # Route based MTU overrides, you have known vpn ip paths that can support larger MTUs you can increase/decrease them here
   routes:
     #- mtu: 8800
     #  route: 10.0.0.0/16
+
   # Unsafe routes allows you to route traffic over nebula to non-nebula nodes
   # Unsafe routes should be avoided unless you have hosts/services that cannot run nebula
   # NOTE: The nebula certificate of the "via" node *MUST* have the "route" defined as a subnet in its certificate
+  # `mtu`: will default to tun mtu if this option is not specified
+  # `metric`: will default to 0 if this option is not specified
+  # `install`: will default to true, controls whether this route is installed in the systems routing table.
+  # This setting is reloadable.
   unsafe_routes:
     #- route: 172.16.1.0/24
     #  via: 192.168.100.99
-    #  mtu: 1300 #mtu will default to tun mtu if this option is not sepcified
+    #  mtu: 1300
+    #  metric: 100
+    #  install: true
 
+  # On linux only, set to true to manage unsafe routes directly on the system route table with gateway routes instead of
+  # in nebula configuration files. Default false, not reloadable.
+  #use_system_route_table: false
 
 # TODO
 # Configure logging level
 logging:
-  # panic, fatal, error, warning, info, or debug. Default is info
+  # panic, fatal, error, warning, info, or debug. Default is info and is reloadable.
+  #NOTE: Debug mode can log remotely controlled/untrusted data which can quickly fill a disk in some
+  # scenarios. Debug logging is also CPU intensive and will decrease performance overall.
+  # Only enable debug logging while actively investigating an issue.
   level: info
   # json or text formats currently available. Default is text
   format: text
+  # Disable timestamp logging. useful when output is redirected to logging system that already adds timestamps. Default is false
+  #disable_timestamp: true
   # timestamp format is specified in Go time format, see:
   #     https://golang.org/pkg/time/#pkg-constants
   # default when `format: json`: "2006-01-02T15:04:05Z07:00" (RFC3339)
@@ -177,33 +279,79 @@ logging:
   #subsystem: nebula
   #interval: 10s
 
-# Handshake Manger Settings
+  # enables counter metrics for meta packets
+  #   e.g.: `messages.tx.handshake`
+  # NOTE: `message.{tx,rx}.recv_error` is always emitted
+  #message_metrics: false
+
+  # enables detailed counter metrics for lighthouse packets
+  #   e.g.: `lighthouse.rx.HostQuery`
+  #lighthouse_metrics: false
+
+# Handshake Manager Settings
 #handshakes:
-  # Total time to try a handshake = sequence of `try_interval * retries`
-  # With 100ms interval and 20 retries it is 23.5 seconds
+  # Handshakes are sent to all known addresses at each interval with a linear backoff,
+  # Wait try_interval after the 1st attempt, 2 * try_interval after the 2nd, etc, until the handshake is older than timeout
+  # A 100ms interval with the default 10 retries will give a handshake 5.5 seconds to resolve before timing out
   #try_interval: 100ms
   #retries: 20
-  # wait_rotation is the number of handshake attempts to do before starting to try non-local IP addresses
-  #wait_rotation: 5
+
+  # query_buffer is the size of the buffer channel for querying lighthouses
+  #query_buffer: 64
+
+  # trigger_buffer is the size of the buffer channel for quickly sending handshakes
+  # after receiving the response for lighthouse queries
+  #trigger_buffer: 64
+
+# Tunnel manager settings
+#tunnels:
+  # drop_inactive controls whether inactive tunnels are maintained or dropped after the inactive_timeout period has
+  # elapsed.
+  # In general, it is a good idea to enable this setting. It will be enabled by default in a future release.
+  # This setting is reloadable
+  #drop_inactive: false
+
+  # inactivity_timeout controls how long a tunnel MUST NOT see any inbound or outbound traffic before being considered
+  # inactive and eligible to be dropped.
+  # This setting is reloadable
+  #inactivity_timeout: 10m
 
 # Nebula security group configuration
 firewall:
+  # Action to take when a packet is not allowed by the firewall rules.
+  # Can be one of:
+  #   `drop` (default): silently drop the packet.
+  #   `reject`: send a reject reply.
+  #     - For TCP, this will be a RST "Connection Reset" packet.
+  #     - For other protocols, this will be an ICMP port unreachable packet.
+  outbound_action: drop
+  inbound_action: drop
+
+  # Controls the default value for local_cidr. Default is true, will be deprecated after v1.9 and defaulted to false.
+  # This setting only affects nebula hosts with subnets encoded in their certificate. A nebula host acting as an
+  # unsafe router with `default_local_cidr_any: true` will expose their unsafe routes to every inbound rule regardless
+  # of the actual destination for the packet. Setting this to false requires each inbound rule to contain a `local_cidr`
+  # if the intention is to allow traffic to flow to an unsafe route.
+  #default_local_cidr_any: false
+
   conntrack:
-    tcp_timeout: 120h
+    tcp_timeout: 12m
     udp_timeout: 3m
     default_timeout: 10m
-    max_connections: 100000
 
   # The firewall is default deny. There is no way to write a deny rule.
   # Rules are comprised of a protocol, port, and one or more of host, group, or CIDR
-  # Logical evaluation is roughly: port AND proto AND (ca_sha OR ca_name) AND (host OR group OR groups OR cidr)
+  # Logical evaluation is roughly: port AND proto AND (ca_sha OR ca_name) AND (host OR group OR groups OR cidr) AND (local cidr)
   # - port: Takes `0` or `any` as any, a single number `80`, a range `200-901`, or `fragment` to match second and further fragments of fragmented packets (since there is no port available).
   #   code: same as port but makes more sense when talking about ICMP, TODO: this is not currently implemented in a way that works, use `any`
   #   proto: `any`, `tcp`, `udp`, or `icmp`
   #   host: `any` or a literal hostname, ie `test-host`
   #   group: `any` or a literal group name, ie `default-group`
   #   groups: Same as group but accepts a list of values. Multiple values are AND'd together and a certificate would have to contain all groups to pass
-  #   cidr: a CIDR, `0.0.0.0/0` is any.
+  #   cidr: a remote CIDR, `0.0.0.0/0` is any.
+  #   local_cidr: a local CIDR, `0.0.0.0/0` is any. This could be used to filter destinations when using unsafe_routes.
+  #      Default is `any` unless the certificate contains subnets and then the default is the ip issued in the certificate
+  #      if `default_local_cidr_any` is false, otherwise its `any`.
   #   ca_name: An issuing CA name
   #   ca_sha: An issuing CA shasum
 
@@ -225,3 +373,10 @@ firewall:
       groups:
         - laptop
         - home
+
+    # Expose a subnet (unsafe route) to hosts with the group remote_client
+    # This example assume you have a subnet of 192.168.100.1/24 or larger encoded in the certificate
+    - port: 8080
+      proto: tcp
+      group: remote_client
+      local_cidr: 192.168.100.1/24

examples/go_service/main.go

@@ -0,0 +1,109 @@
+package main
+
+import (
+	"bufio"
+	"fmt"
+	"log"
+	"net"
+
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/service"
+)
+
+func main() {
+	if err := run(); err != nil {
+		log.Fatalf("%+v", err)
+	}
+}
+
+func run() error {
+	configStr := `
+tun:
+  user: true
+
+static_host_map:
+  '192.168.100.1': ['localhost:4242']
+
+listen:
+  host: 0.0.0.0
+  port: 4241
+
+lighthouse:
+  am_lighthouse: false
+  interval: 60
+  hosts:
+    - '192.168.100.1'
+
+firewall:
+  outbound:
+    # Allow all outbound traffic from this node
+    - port: any
+      proto: any
+      host: any
+
+  inbound:
+    # Allow icmp between any nebula hosts
+    - port: any
+      proto: icmp
+      host: any
+    - port: any
+      proto: any
+      host: any
+
+pki:
+  ca: /home/rice/Developer/nebula-config/ca.crt
+  cert: /home/rice/Developer/nebula-config/app.crt
+  key: /home/rice/Developer/nebula-config/app.key
+`
+	var cfg config.C
+	if err := cfg.LoadString(configStr); err != nil {
+		return err
+	}
+	svc, err := service.New(&cfg)
+	if err != nil {
+		return err
+	}
+
+	ln, err := svc.Listen("tcp", ":1234")
+	if err != nil {
+		return err
+	}
+	for {
+		conn, err := ln.Accept()
+		if err != nil {
+			log.Printf("accept error: %s", err)
+			break
+		}
+		defer func(conn net.Conn) {
+			_ = conn.Close()
+		}(conn)
+
+		log.Printf("got connection")
+
+		_, err = conn.Write([]byte("hello world\n"))
+		if err != nil {
+			log.Printf("write error: %s", err)
+		}
+
+		scanner := bufio.NewScanner(conn)
+		for scanner.Scan() {
+			message := scanner.Text()
+			_, err = fmt.Fprintf(conn, "echo: %q\n", message)
+			if err != nil {
+				log.Printf("write error: %s", err)
+			}
+			log.Printf("got message %q", message)
+		}
+
+		if err := scanner.Err(); err != nil {
+			log.Printf("scanner error: %s", err)
+			break
+		}
+	}
+
+	_ = svc.Close()
+	if err := svc.Wait(); err != nil {
+		return err
+	}
+	return nil
+}

examples/quickstart-vagrant/README.md

@@ -1,139 +0,0 @@
-# Quickstart Guide
-
-This guide is intended to bring up a vagrant environment with 1 lighthouse and 2 generic hosts running nebula.
-
-## Creating the virtualenv for ansible
-
-Within the `quickstart/` directory, do the following
-
-```
-# make a virtual environment
-virtualenv venv
-
-# get into the virtualenv
-source venv/bin/activate
-
-# install ansible
-pip install -r requirements.yml
-```
-
-## Bringing up the vagrant environment
-
-A plugin that is used for the Vagrant environment is `vagrant-hostmanager`
-
-To install, run
-
-```
-vagrant plugin install vagrant-hostmanager
-```
-
-All hosts within the Vagrantfile are brought up with
-
-`vagrant up` 
-
-Once the boxes are up, go into the `ansible/` directory and deploy the playbook by running
-
-`ansible-playbook playbook.yml -i inventory -u vagrant`
-
-## Testing within the vagrant env
-
-Once the ansible run is done, hop onto a vagrant box 
-
-`vagrant ssh generic1.vagrant`
-
-or specifically
-
-`ssh vagrant@<ip-address-in-vagrant-file` (password for the vagrant user on the boxes is `vagrant`)
-
-Some quick tests once the vagrant boxes are up are to ping from `generic1.vagrant` to `generic2.vagrant` using 
-their respective nebula ip address. 
-
-```
-vagrant@generic1:~$ ping 10.168.91.220
-PING 10.168.91.220 (10.168.91.220) 56(84) bytes of data.
-64 bytes from 10.168.91.220: icmp_seq=1 ttl=64 time=241 ms
-64 bytes from 10.168.91.220: icmp_seq=2 ttl=64 time=0.704 ms
-```
-
-You can further verify that the allowed nebula firewall rules work by ssh'ing from 1 generic box to the other.
-
-`ssh vagrant@<nebula-ip-address>`  (password for the vagrant user on the boxes is `vagrant`)
-
-See `/etc/nebula/config.yml` on a box for firewall rules.
-
-To see full handshakes and hostmaps, change the logging config of `/etc/nebula/config.yml` on the vagrant boxes from 
-info to debug.
-
-You can watch nebula logs by running
-
-```
-sudo journalctl -fu nebula
-```
-
-Refer to the nebula src code directory's README for further instructions on configuring nebula.
-
-## Troubleshooting
-
-### Is nebula up and running?
-
-Run and verify that 
-
-```
-ifconfig
-``` 
-
-shows you an interface with the name `nebula1` being up.
-
-```
-vagrant@generic1:~$ ifconfig nebula1
-nebula1: flags=4305<UP,POINTOPOINT,RUNNING,NOARP,MULTICAST>  mtu 1300
-        inet 10.168.91.210  netmask 255.128.0.0  destination 10.168.91.210
-        inet6 fe80::aeaf:b105:e6dc:936c  prefixlen 64  scopeid 0x20<link>
-        unspec 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00  txqueuelen 500  (UNSPEC)
-        RX packets 2  bytes 168 (168.0 B)
-        RX errors 0  dropped 0  overruns 0  frame 0
-        TX packets 11  bytes 600 (600.0 B)
-        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
-```
-
-### Connectivity
-
-Are you able to ping other boxes on the private nebula network?
-
-The following are the private nebula ip addresses of the vagrant env 
-
-```
-generic1.vagrant [nebula_ip] 10.168.91.210
-generic2.vagrant [nebula_ip] 10.168.91.220 
-lighthouse1.vagrant [nebula_ip] 10.168.91.230
-```
-
-Try pinging generic1.vagrant to and from any other box using its nebula ip above.
-
-Double check the nebula firewall rules under /etc/nebula/config.yml to make sure that connectivity is allowed for your use-case if on a specific port.
-
-```
-vagrant@lighthouse1:~$ grep -A21 firewall /etc/nebula/config.yml 
-firewall:
-  conntrack:
-    tcp_timeout: 12m
-    udp_timeout: 3m
-    default_timeout: 10m
-    max_connections: 100,000
-
-  inbound:
-    - proto: icmp
-      port: any
-      host: any
-    - proto: any
-      port: 22
-      host: any
-    - proto: any
-      port: 53
-      host: any
-
-  outbound:
-    - proto: any
-      port: any
-      host: any
-```

examples/quickstart-vagrant/Vagrantfile

@@ -1,40 +0,0 @@
-Vagrant.require_version ">= 2.2.6"
-
-nodes = [
-  { :hostname => 'generic1.vagrant', :ip => '172.11.91.210', :box => 'bento/ubuntu-18.04', :ram => '512', :cpus => 1},
-  { :hostname => 'generic2.vagrant', :ip => '172.11.91.220', :box => 'bento/ubuntu-18.04', :ram => '512', :cpus => 1},
-  { :hostname => 'lighthouse1.vagrant', :ip => '172.11.91.230', :box => 'bento/ubuntu-18.04', :ram => '512', :cpus => 1},
-]
-
-Vagrant.configure("2") do |config|
-
-  config.ssh.insert_key = false
-
-  if Vagrant.has_plugin?('vagrant-cachier')
-    config.cache.enable :apt
-  else
-    printf("** Install vagrant-cachier plugin to speedup deploy: `vagrant plugin install vagrant-cachier`.**\n")
-  end
-
-  if Vagrant.has_plugin?('vagrant-hostmanager')
-    config.hostmanager.enabled = true
-    config.hostmanager.manage_host = true
-    config.hostmanager.include_offline = true
-  else
-    config.vagrant.plugins = "vagrant-hostmanager"
-  end
-
- nodes.each do |node|
-    config.vm.define node[:hostname] do |node_config|
-      node_config.vm.box = node[:box]
-      node_config.vm.hostname = node[:hostname]
-      node_config.vm.network :private_network, ip: node[:ip]
-      node_config.vm.provider :virtualbox do |vb|
-        vb.memory = node[:ram]
-        vb.cpus = node[:cpus]
-        vb.customize ["modifyvm", :id, "--natdnshostresolver1", "on"]
-        vb.customize ['guestproperty', 'set', :id, '/VirtualBox/GuestAdd/VBoxService/--timesync-set-threshold', 10000]
-      end
-    end
-  end
-end

examples/quickstart-vagrant/ansible/ansible.cfg

@@ -1,4 +0,0 @@
-[defaults]
-host_key_checking = False
-private_key_file = ~/.vagrant.d/insecure_private_key
-become = yes

examples/quickstart-vagrant/ansible/filter_plugins/to_nebula_ip.py

@@ -1,21 +0,0 @@
-#!/usr/bin/python
-
-
-class FilterModule(object):
-    def filters(self):
-        return {
-            'to_nebula_ip': self.to_nebula_ip,
-            'map_to_nebula_ips': self.map_to_nebula_ips,
-        }
-
-    def to_nebula_ip(self, ip_str):
-        ip_list = list(map(int, ip_str.split(".")))
-        ip_list[0] = 10
-        ip_list[1] = 168
-        ip = '.'.join(map(str, ip_list))
-        return ip
-
-    def map_to_nebula_ips(self, ip_strs):
-        ip_list = [ self.to_nebula_ip(ip_str) for ip_str in ip_strs ]
-        ips = ', '.join(ip_list)
-        return ips