v1.5.0 (#574)

Update CHANGELOG for Nebula v1.5.0

.github/workflows/gofmt.yml

@@ -14,21 +14,21 @@ jobs:
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.16
-      uses: actions/setup-go@v1
+    - name: Set up Go 1.17
+      uses: actions/setup-go@v2
       with:
-        go-version: 1.16
+        go-version: 1.17
       id: go
 
     - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
+      uses: actions/checkout@v2
 
-    - uses: actions/cache@v1
+    - uses: actions/cache@v2
       with:
         path: ~/go/pkg/mod
-        key: ${{ runner.os }}-gofmt1.16-${{ hashFiles('**/go.sum') }}
+        key: ${{ runner.os }}-gofmt1.17-${{ hashFiles('**/go.sum') }}
         restore-keys: |
-          ${{ runner.os }}-gofmt1.16-
+          ${{ runner.os }}-gofmt1.17-
 
     - name: Install goimports
       run: |

.github/workflows/release.yml

@@ -10,10 +10,10 @@ jobs:
     name: Build Linux All
     runs-on: ubuntu-latest
     steps:
-      - name: Set up Go 1.16
-        uses: actions/setup-go@v1
+      - name: Set up Go 1.17
+        uses: actions/setup-go@v2
         with:
-          go-version: 1.16
+          go-version: 1.17
 
       - name: Checkout code
         uses: actions/checkout@v2
@@ -25,7 +25,7 @@ jobs:
           mv build/*.tar.gz release
 
       - name: Upload artifacts
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v2
         with:
           name: linux-latest
           path: release
@@ -34,10 +34,10 @@ jobs:
     name: Build Windows amd64
     runs-on: windows-latest
     steps:
-      - name: Set up Go 1.16
-        uses: actions/setup-go@v1
+      - name: Set up Go 1.17
+        uses: actions/setup-go@v2
         with:
-          go-version: 1.16
+          go-version: 1.17
 
       - name: Checkout code
         uses: actions/checkout@v2
@@ -47,62 +47,77 @@ jobs:
           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\dist\windows
+          mv dist\windows\wintun build\dist\windows\
 
       - name: Upload artifacts
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v2
         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-11
     steps:
-      - name: Set up Go 1.16
-        uses: actions/setup-go@v1
+      - name: Set up Go 1.17
+        uses: actions/setup-go@v2
         with:
-          go-version: 1.16
+          go-version: 1.17
 
       - name: Checkout code
         uses: actions/checkout@v2
 
-      - name: Build
+      - name: Import certificates
+        if: env.HAS_SIGNING_CREDS == 'true'
+        uses: Apple-Actions/import-codesign-certs@v1
+        with:
+          p12-file-base64: ${{ secrets.APPLE_DEVELOPER_CERTIFICATE_P12_BASE64 }}
+          p12-password: ${{ secrets.APPLE_DEVELOPER_CERTIFICATE_PASSWORD }}
+
+      - 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
-          make BUILD_NUMBER="${GITHUB_REF#refs/tags/v}" service build/nebula-darwin-arm64.tar.gz
+          rm -rf release
           mkdir release
-          mv build/*.tar.gz release
+          make BUILD_NUMBER="${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@v2
         with:
           name: darwin-latest
-          path: release
+          path: ./release/*
 
   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
-
-      - name: Download Darwin artifacts
-        uses: actions/download-artifact@v1
-        with:
-          name: darwin-latest
-
-      - name: Download Windows artifacts
-        uses: actions/download-artifact@v1
-        with:
-          name: windows-latest
+      - name: Download artifacts
+        uses: actions/download-artifact@v2
 
       - name: Zip Windows
         run: |
           cd windows-latest
-          zip nebula-windows-amd64.zip nebula.exe nebula-cert.exe
+          zip -r nebula-windows-amd64.zip nebula.exe nebula-cert.exe dist
 
       - name: Create sha256sum
         run: |
@@ -115,12 +130,17 @@ jobs:
                 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 nebula-windows-amd64.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
@@ -150,25 +170,15 @@ jobs:
           asset_name: SHASUM256.txt
           asset_content_type: text/plain
 
-      - name: Upload darwin-amd64
+      - name: Upload darwin zip
         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 darwin-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: ./darwin-latest/nebula-darwin-arm64.tar.gz
-          asset_name: nebula-darwin-arm64.tar.gz
-          asset_content_type: application/gzip
+          asset_path: ./darwin-latest/nebula-darwin.zip
+          asset_name: nebula-darwin.zip
+          asset_content_type: application/zip
 
       - name: Upload windows-amd64
         uses: actions/upload-release-asset@v1.0.1
@@ -300,6 +310,16 @@ jobs:
           asset_name: nebula-linux-mips-softfloat.tar.gz
           asset_content_type: application/gzip
 
+      - name: Upload linux-riscv64
+        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-riscv64.tar.gz
+          asset_name: nebula-linux-riscv64.tar.gz
+          asset_content_type: application/gzip
+
       - name: Upload freebsd-amd64
         uses: actions/upload-release-asset@v1.0.1
         env:

.github/workflows/smoke.yml

@@ -18,21 +18,21 @@ jobs:
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.16
-      uses: actions/setup-go@v1
+    - name: Set up Go 1.17
+      uses: actions/setup-go@v2
       with:
-        go-version: 1.16
+        go-version: 1.17
       id: go
 
     - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
+      uses: actions/checkout@v2
 
-    - uses: actions/cache@v1
+    - uses: actions/cache@v2
       with:
         path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go1.16-${{ hashFiles('**/go.sum') }}
+        key: ${{ runner.os }}-go1.17-${{ hashFiles('**/go.sum') }}
         restore-keys: |
-          ${{ runner.os }}-go1.16-
+          ${{ runner.os }}-go1.17-
 
     - name: build
       run: make bin-docker

.github/workflows/test.yml

@@ -18,21 +18,21 @@ jobs:
     runs-on: ubuntu-latest
     steps:
 
-    - name: Set up Go 1.16
-      uses: actions/setup-go@v1
+    - name: Set up Go 1.17
+      uses: actions/setup-go@v2
       with:
-        go-version: 1.16
+        go-version: 1.17
       id: go
 
     - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
+      uses: actions/checkout@v2
 
-    - uses: actions/cache@v1
+    - uses: actions/cache@v2
       with:
         path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go1.16-${{ hashFiles('**/go.sum') }}
+        key: ${{ runner.os }}-go1.17-${{ hashFiles('**/go.sum') }}
         restore-keys: |
-          ${{ runner.os }}-go1.16-
+          ${{ runner.os }}-go1.17-
 
     - name: Build
       run: make all
@@ -40,29 +40,32 @@ jobs:
     - name: Test
       run: make test
 
+    - name: End 2 end
+      run: make e2evv
+
   test:
     name: Build and test on ${{ matrix.os }}
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [windows-latest, macOS-latest]
+        os: [windows-latest, macos-11]
     steps:
 
-    - name: Set up Go 1.16
-      uses: actions/setup-go@v1
+    - name: Set up Go 1.17
+      uses: actions/setup-go@v2
       with:
-        go-version: 1.16
+        go-version: 1.17
       id: go
 
     - name: Check out code into the Go module directory
-      uses: actions/checkout@v1
+      uses: actions/checkout@v2
 
-    - uses: actions/cache@v1
+    - uses: actions/cache@v2
       with:
         path: ~/go/pkg/mod
-        key: ${{ runner.os }}-go1.16-${{ hashFiles('**/go.sum') }}
+        key: ${{ runner.os }}-go1.17-${{ hashFiles('**/go.sum') }}
         restore-keys: |
-          ${{ runner.os }}-go1.16-
+          ${{ runner.os }}-go1.17-
 
     - name: Build nebula
       run: go build ./cmd/nebula
@@ -72,3 +75,6 @@ jobs:
 
     - name: Test
       run: go test -v ./...
+
+    - name: End 2 end
+      run: make e2evv

CHANGELOG.md

@@ -7,10 +7,131 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [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
 
-- Updated the kardianos/service go library from 1.0.0 to 1.1.0, which
-  now creates launchd plist to write stdout/stderr to files by default.
+- 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
 
@@ -185,7 +306,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Initial public release.
 
-[Unreleased]: https://github.com/slackhq/nebula/compare/v1.3.0...HEAD
+[Unreleased]: https://github.com/slackhq/nebula/compare/v1.5.0...HEAD
+[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

Makefile

@@ -1,14 +1,31 @@
-GOMINVERSION = 1.16
+GOMINVERSION = 1.17
 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
 
-# Ensure the version of go we are using is at least what is defined in GOMINVERSION at the top of this file
-GOVERSION := $(shell go version | awk '{print substr($$3, 3)}')
-GOISMIN := $(shell expr "$(GOVERSION)" ">=" "$(GOMINVERSION)")
-ifneq "$(GOISMIN)" "1"
-$(error "go version $(GOVERSION) is not supported, upgrade to $(GOMINVERSION) or above")
+# Set up OS specific bits
+ifeq ($(OS),Windows_NT)
+	#TODO: we should be able to ditch awk as well
+	GOVERSION := $(shell go version | awk "{print substr($$3, 3)}")
+	GOISMIN := $(shell IF "$(GOVERSION)" GEQ "$(GOMINVERSION)" ECHO 1)
+	NEBULA_CMD_SUFFIX = .exe
+	NULL_FILE = nul
+else
+	GOVERSION := $(shell go version | awk '{print substr($$3, 3)}')
+	GOISMIN := $(shell expr "$(GOVERSION)" ">=" "$(GOMINVERSION)")
+	NEBULA_CMD_SUFFIX =
+	NULL_FILE = /dev/null
+endif
+
+# Only defined the build number if we haven't already
+ifndef BUILD_NUMBER
+	ifeq ($(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
 
 LDFLAGS = -X main.Build=$(BUILD_NUMBER)
@@ -24,7 +41,8 @@ ALL_LINUX = linux-amd64 \
 	linux-mipsle \
 	linux-mips64 \
 	linux-mips64le \
-	linux-mips-softfloat
+	linux-mips-softfloat \
+	linux-riscv64
 
 ALL = $(ALL_LINUX) \
 	darwin-amd64 \
@@ -32,7 +50,20 @@ ALL = $(ALL_LINUX) \
 	freebsd-amd64 \
 	windows-amd64
 
+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
 
 all: $(ALL:%=build/%/nebula) $(ALL:%=build/%/nebula-cert)
 
@@ -54,8 +85,8 @@ bin-freebsd: build/freebsd-amd64/nebula build/freebsd-amd64/nebula-cert
 	mv $? .
 
 bin:
-	go build $(BUILD_ARGS) -ldflags "$(LDFLAGS)" -o ./nebula ${NEBULA_CMD_PATH}
-	go build $(BUILD_ARGS) -ldflags "$(LDFLAGS)" -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 $(BUILD_ARGS) -ldflags "$(LDFLAGS)" ${NEBULA_CMD_PATH}
@@ -113,18 +144,18 @@ 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
@@ -137,5 +168,5 @@ smoke-docker-race: BUILD_ARGS = -race
 smoke-docker-race: smoke-docker
 
 .FORCE:
-.PHONY: test test-cov-html bench bench-cpu bench-cpu-long bin proto release service smoke-docker smoke-docker-race
+.PHONY: e2e e2ev e2evv e2evvv e2evvvv test test-cov-html bench bench-cpu bench-cpu-long bin proto release service smoke-docker smoke-docker-race
 .DEFAULT_GOAL := bin

README.md

@@ -1,4 +1,4 @@
-## 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, 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.
@@ -8,21 +8,35 @@ 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://www.defined.net/nebula/introduction/).
+
 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
+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 
+- Linux - 64 and 32 bit, arm, and others
 - Windows
 - MacOS
 - Freebsd
 
+#### Distribution Packages
+
+- [Arch Linux](https://archlinux.org/packages/community/x86_64/nebula/)
+    ```
+    $ sudo pacman -S nebula
+    ```
+- [Fedora Linux](https://copr.fedorainfracloud.org/coprs/jdoss/nebula/)
+    ```
+    $ sudo dnf copr enable jdoss/nebula
+    $ sudo dnf install nebula
+    ```
+
 #### Mobile
 
 - [iOS](https://apps.apple.com/us/app/mobile-nebula/id1509587936?itsct=apps_box&itscg=30200)
@@ -36,15 +50,15 @@ 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.
 

allow_list.go

@@ -2,12 +2,29 @@ package nebula
 
 import (
 	"fmt"
+	"net"
 	"regexp"
+
+	"github.com/slackhq/nebula/cidr"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/iputil"
 )
 
 type AllowList struct {
 	// The values of this cidrTree are `bool`, signifying allow/deny
-	cidrTree *CIDRTree
+	cidrTree *cidr.Tree6
+}
+
+type RemoteAllowList struct {
+	AllowList *AllowList
+
+	// Inside Range Specific, keys of this tree are inside CIDRs and values
+	// are *AllowList
+	insideAllowLists *cidr.Tree6
+}
+
+type LocalAllowList struct {
+	AllowList *AllowList
 
 	// To avoid ambiguity, all rules must be true, or all rules must be false.
 	nameRules []AllowListNameRule
@@ -18,7 +35,224 @@ 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 := cidr.NewTree6()
+
+	// 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 := 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(ipNet, value)
+
+		maskBits, maskSize := ipNet.Mask.Size()
+
+		var rules *allowListRules
+		if maskSize == 32 {
+			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 {
+			_, zeroCIDR, _ := net.ParseCIDR("0.0.0.0/0")
+			tree.AddCIDR(zeroCIDR, !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 {
+			_, zeroCIDR, _ := net.ParseCIDR("::/0")
+			tree.AddCIDR(zeroCIDR, !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) (*cidr.Tree6, error) {
+	value := c.Get(k)
+	if value == nil {
+		return nil, nil
+	}
+
+	remoteAllowRanges := cidr.NewTree6()
+
+	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 := net.ParseCIDR(rawCIDR)
+		if err != nil {
+			return nil, fmt.Errorf("config `%s` has invalid CIDR: %s", k, rawCIDR)
+		}
+
+		remoteAllowRanges.AddCIDR(ipNet, allowList)
+	}
+
+	return remoteAllowRanges, nil
+}
+
+func (al *AllowList) Allow(ip net.IP) bool {
 	if al == nil {
 		return true
 	}
@@ -32,7 +266,42 @@ func (al *AllowList) Allow(ip uint32) bool {
 	}
 }
 
-func (al *AllowList) AllowName(name string) bool {
+func (al *AllowList) AllowIpV4(ip iputil.VpnIp) bool {
+	if al == nil {
+		return true
+	}
+
+	result := al.cidrTree.MostSpecificContainsIpV4(ip)
+	switch v := result.(type) {
+	case bool:
+		return v
+	default:
+		panic(fmt.Errorf("invalid state, allowlist returned: %T %v", result, result))
+	}
+}
+
+func (al *AllowList) AllowIpV6(hi, lo uint64) bool {
+	if al == nil {
+		return true
+	}
+
+	result := al.cidrTree.MostSpecificContainsIpV6(hi, lo)
+	switch v := result.(type) {
+	case bool:
+		return v
+	default:
+		panic(fmt.Errorf("invalid state, allowlist returned: %T %v", result, result))
+	}
+}
+
+func (al *LocalAllowList) Allow(ip net.IP) 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 +315,47 @@ 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 net.IP) bool {
+	if al == nil {
+		return true
+	}
+	return al.AllowList.Allow(ip)
+}
+
+func (al *RemoteAllowList) Allow(vpnIp iputil.VpnIp, ip net.IP) bool {
+	if !al.getInsideAllowList(vpnIp).Allow(ip) {
+		return false
+	}
+	return al.AllowList.Allow(ip)
+}
+
+func (al *RemoteAllowList) AllowIpV4(vpnIp iputil.VpnIp, ip iputil.VpnIp) bool {
+	if al == nil {
+		return true
+	}
+	if !al.getInsideAllowList(vpnIp).AllowIpV4(ip) {
+		return false
+	}
+	return al.AllowList.AllowIpV4(ip)
+}
+
+func (al *RemoteAllowList) AllowIpV6(vpnIp iputil.VpnIp, hi, lo uint64) bool {
+	if al == nil {
+		return true
+	}
+	if !al.getInsideAllowList(vpnIp).AllowIpV6(hi, lo) {
+		return false
+	}
+	return al.AllowList.AllowIpV6(hi, lo)
+}
+
+func (al *RemoteAllowList) getInsideAllowList(vpnIp iputil.VpnIp) *AllowList {
+	if al.insideAllowLists != nil {
+		inside := al.insideAllowLists.MostSpecificContainsIpV4(vpnIp)
+		if inside != nil {
+			return inside.(*AllowList)
+		}
+	}
+	return nil
+}

allow_list_test.go

@@ -5,31 +5,129 @@ import (
 	"regexp"
 	"testing"
 
+	"github.com/slackhq/nebula/cidr"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/util"
 	"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 := util.NewTestLogger()
+	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")
+	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(net.ParseIP("1.1.1.1")))
+
+	tree := cidr.NewTree6()
+	tree.AddCIDR(cidr.Parse("0.0.0.0/0"), true)
+	tree.AddCIDR(cidr.Parse("10.0.0.0/8"), false)
+	tree.AddCIDR(cidr.Parse("10.42.42.42/32"), true)
+	tree.AddCIDR(cidr.Parse("10.42.0.0/16"), true)
+	tree.AddCIDR(cidr.Parse("10.42.42.0/24"), true)
+	tree.AddCIDR(cidr.Parse("10.42.42.0/24"), false)
+	tree.AddCIDR(cidr.Parse("::1/128"), true)
+	tree.AddCIDR(cidr.Parse("::2/128"), false)
+	al := &AllowList{cidrTree: tree}
+
+	assert.Equal(t, true, al.Allow(net.ParseIP("1.1.1.1")))
+	assert.Equal(t, false, al.Allow(net.ParseIP("10.0.0.4")))
+	assert.Equal(t, true, al.Allow(net.ParseIP("10.42.42.42")))
+	assert.Equal(t, false, al.Allow(net.ParseIP("10.42.42.41")))
+	assert.Equal(t, true, al.Allow(net.ParseIP("10.42.0.1")))
+	assert.Equal(t, true, al.Allow(net.ParseIP("::1")))
+	assert.Equal(t, false, al.Allow(net.ParseIP("::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/util"
 	"github.com/stretchr/testify/assert"
 )
 
 func TestBits(t *testing.T) {
+	l := util.NewTestLogger()
 	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 := util.NewTestLogger()
 	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 := util.NewTestLogger()
 	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 := util.NewTestLogger()
 	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())

cert.go

@@ -7,11 +7,11 @@ import (
 	"strings"
 	"time"
 
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/config"
 )
 
-var trustedCAs *cert.NebulaCAPool
-
 type CertState struct {
 	certificate         *cert.NebulaCertificate
 	rawCertificate      []byte
@@ -46,7 +46,7 @@ func NewCertState(certificate *cert.NebulaCertificate, privateKey []byte) (*Cert
 	return cs, nil
 }
 
-func NewCertStateFromConfig(c *Config) (*CertState, error) {
+func NewCertStateFromConfig(c *config.C) (*CertState, error) {
 	var pemPrivateKey []byte
 	var err error
 
@@ -119,7 +119,7 @@ func NewCertStateFromConfig(c *Config) (*CertState, error) {
 	return NewCertState(nebulaCert, rawKey)
 }
 
-func loadCAFromConfig(c *Config) (*cert.NebulaCAPool, error) {
+func loadCAFromConfig(l *logrus.Logger, c *config.C) (*cert.NebulaCAPool, error) {
 	var rawCA []byte
 	var err error
 

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/cert.go

@@ -325,12 +325,26 @@ 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) {
+	if nc.Details.IsCA {
+		// 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")
+		}
+		return nil
+	}
+
+	pub, err := curve25519.X25519(key, curve25519.Basepoint)
+	if err != nil {
+		return err
+	}
+	if !bytes.Equal(pub, nc.Details.PublicKey) {
 		return fmt.Errorf("public key in cert and private key supplied don't match")
 	}
+
 	return nil
 }
 

cert/cert.pb.go

@@ -1,202 +1,298 @@
 // Code generated by protoc-gen-go. DO NOT EDIT.
+// versions:
+// 	protoc-gen-go v1.26.0
+// 	protoc        v3.14.0
 // 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
-
-// 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
+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)
+)
 
 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"`
 }
 
-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")
+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, 0xf9, 0x01, 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, 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,
 }
 
-func init() { proto.RegisterFile("cert.proto", fileDescriptor_a142e29cbef9b1cf) }
+var (
+	file_cert_proto_rawDescOnce sync.Once
+	file_cert_proto_rawDescData = file_cert_proto_rawDesc
+)
 
-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,
+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_msgTypes = make([]protoimpl.MessageInfo, 2)
+var file_cert_proto_goTypes = []interface{}{
+	(*RawNebulaCertificate)(nil),        // 0: cert.RawNebulaCertificate
+	(*RawNebulaCertificateDetails)(nil), // 1: cert.RawNebulaCertificateDetails
+}
+var file_cert_proto_depIdxs = []int32{
+	1, // 0: cert.RawNebulaCertificate.Details:type_name -> cert.RawNebulaCertificateDetails
+	1, // [1:1] is the sub-list for method output_type
+	1, // [1:1] is the sub-list for method input_type
+	1, // [1:1] is the sub-list for extension type_name
+	1, // [1:1] is the sub-list for extension extendee
+	0, // [0:1] 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
+			}
+		}
+	}
+	type x struct{}
+	out := protoimpl.TypeBuilder{
+		File: protoimpl.DescBuilder{
+			GoPackagePath: reflect.TypeOf(x{}).PkgPath(),
+			RawDescriptor: file_cert_proto_rawDesc,
+			NumEnums:      0,
+			NumMessages:   2,
+			NumExtensions: 0,
+			NumServices:   0,
+		},
+		GoTypes:           file_cert_proto_goTypes,
+		DependencyIndexes: file_cert_proto_depIdxs,
+		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

@@ -1,6 +1,8 @@
 syntax = "proto3";
 package cert;
 
+option go_package = "github.com/slackhq/nebula/cert";
+
 //import "google/protobuf/timestamp.proto";
 
 message RawNebulaCertificate {

cert/cert_test.go

@@ -375,9 +375,16 @@ 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(caKey)
+	assert.Nil(t, err)
+
+	_, _, caKey2, err := newTestCaCert(time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
+	assert.Nil(t, err)
+	err = ca.VerifyPrivateKey(caKey2)
+	assert.NotNil(t, err)
 
 	c, _, priv, err := newTestCert(ca, caKey, time.Time{}, time.Time{}, []*net.IPNet{}, []*net.IPNet{}, []string{})
 	err = c.VerifyPrivateKey(priv)
@@ -853,10 +860,15 @@ 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
 }

cidr/parse.go

@@ -0,0 +1,10 @@
+package cidr
+
+import "net"
+
+// Parse is a convenience function that returns only the IPNet
+// This function ignores errors since it is primarily a test helper, the result could be nil
+func Parse(s string) *net.IPNet {
+	_, c, _ := net.ParseCIDR(s)
+	return c
+}

cidr/tree4.go

@@ -1,39 +1,39 @@
-package nebula
+package cidr
 
 import (
-	"encoding/binary"
-	"fmt"
 	"net"
+
+	"github.com/slackhq/nebula/iputil"
 )
 
-type CIDRNode struct {
-	left   *CIDRNode
-	right  *CIDRNode
-	parent *CIDRNode
+type Node struct {
+	left   *Node
+	right  *Node
+	parent *Node
 	value  interface{}
 }
 
-type CIDRTree struct {
-	root *CIDRNode
+type Tree4 struct {
+	root *Node
 }
 
 const (
-	startbit = uint32(0x80000000)
+	startbit = iputil.VpnIp(0x80000000)
 )
 
-func NewCIDRTree() *CIDRTree {
-	tree := new(CIDRTree)
-	tree.root = &CIDRNode{}
+func NewTree4() *Tree4 {
+	tree := new(Tree4)
+	tree.root = &Node{}
 	return tree
 }
 
-func (tree *CIDRTree) AddCIDR(cidr *net.IPNet, val interface{}) {
+func (tree *Tree4) AddCIDR(cidr *net.IPNet, val interface{}) {
 	bit := startbit
 	node := tree.root
 	next := tree.root
 
-	ip := ip2int(cidr.IP)
-	mask := ip2int(cidr.Mask)
+	ip := iputil.Ip2VpnIp(cidr.IP)
+	mask := iputil.Ip2VpnIp(cidr.Mask)
 
 	// Find our last ancestor in the tree
 	for bit&mask != 0 {
@@ -59,7 +59,7 @@ func (tree *CIDRTree) AddCIDR(cidr *net.IPNet, val interface{}) {
 
 	// Build up the rest of the tree we don't already have
 	for bit&mask != 0 {
-		next = &CIDRNode{}
+		next = &Node{}
 		next.parent = node
 
 		if ip&bit != 0 {
@@ -76,8 +76,8 @@ func (tree *CIDRTree) AddCIDR(cidr *net.IPNet, val interface{}) {
 	node.value = val
 }
 
-// Finds the first match, which way be the least specific
-func (tree *CIDRTree) Contains(ip uint32) (value interface{}) {
+// Finds the first match, which may be the least specific
+func (tree *Tree4) Contains(ip iputil.VpnIp) (value interface{}) {
 	bit := startbit
 	node := tree.root
 
@@ -100,7 +100,7 @@ func (tree *CIDRTree) Contains(ip uint32) (value interface{}) {
 }
 
 // Finds the most specific match
-func (tree *CIDRTree) MostSpecificContains(ip uint32) (value interface{}) {
+func (tree *Tree4) MostSpecificContains(ip iputil.VpnIp) (value interface{}) {
 	bit := startbit
 	node := tree.root
 
@@ -116,14 +116,13 @@ func (tree *CIDRTree) MostSpecificContains(ip uint32) (value interface{}) {
 		}
 
 		bit >>= 1
-
 	}
 
 	return value
 }
 
 // Finds the most specific match
-func (tree *CIDRTree) Match(ip uint32) (value interface{}) {
+func (tree *Tree4) Match(ip iputil.VpnIp) (value interface{}) {
 	bit := startbit
 	node := tree.root
 	lastNode := node
@@ -144,27 +143,3 @@ func (tree *CIDRTree) Match(ip uint32) (value interface{}) {
 	}
 	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/tree4_test.go

@@ -0,0 +1,153 @@
+package cidr
+
+import (
+	"net"
+	"testing"
+
+	"github.com/slackhq/nebula/iputil"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestCIDRTree_Contains(t *testing.T) {
+	tree := NewTree4()
+	tree.AddCIDR(Parse("1.0.0.0/8"), "1")
+	tree.AddCIDR(Parse("2.1.0.0/16"), "2")
+	tree.AddCIDR(Parse("3.1.1.0/24"), "3")
+	tree.AddCIDR(Parse("4.1.1.0/24"), "4a")
+	tree.AddCIDR(Parse("4.1.1.1/32"), "4b")
+	tree.AddCIDR(Parse("4.1.2.1/32"), "4c")
+	tree.AddCIDR(Parse("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(iputil.Ip2VpnIp(net.ParseIP(tt.IP))))
+	}
+
+	tree = NewTree4()
+	tree.AddCIDR(Parse("1.1.1.1/0"), "cool")
+	assert.Equal(t, "cool", tree.Contains(iputil.Ip2VpnIp(net.ParseIP("0.0.0.0"))))
+	assert.Equal(t, "cool", tree.Contains(iputil.Ip2VpnIp(net.ParseIP("255.255.255.255"))))
+}
+
+func TestCIDRTree_MostSpecificContains(t *testing.T) {
+	tree := NewTree4()
+	tree.AddCIDR(Parse("1.0.0.0/8"), "1")
+	tree.AddCIDR(Parse("2.1.0.0/16"), "2")
+	tree.AddCIDR(Parse("3.1.1.0/24"), "3")
+	tree.AddCIDR(Parse("4.1.1.0/24"), "4a")
+	tree.AddCIDR(Parse("4.1.1.0/30"), "4b")
+	tree.AddCIDR(Parse("4.1.1.1/32"), "4c")
+	tree.AddCIDR(Parse("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(iputil.Ip2VpnIp(net.ParseIP(tt.IP))))
+	}
+
+	tree = NewTree4()
+	tree.AddCIDR(Parse("1.1.1.1/0"), "cool")
+	assert.Equal(t, "cool", tree.MostSpecificContains(iputil.Ip2VpnIp(net.ParseIP("0.0.0.0"))))
+	assert.Equal(t, "cool", tree.MostSpecificContains(iputil.Ip2VpnIp(net.ParseIP("255.255.255.255"))))
+}
+
+func TestCIDRTree_Match(t *testing.T) {
+	tree := NewTree4()
+	tree.AddCIDR(Parse("4.1.1.0/32"), "1a")
+	tree.AddCIDR(Parse("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(iputil.Ip2VpnIp(net.ParseIP(tt.IP))))
+	}
+
+	tree = NewTree4()
+	tree.AddCIDR(Parse("1.1.1.1/0"), "cool")
+	assert.Equal(t, "cool", tree.Contains(iputil.Ip2VpnIp(net.ParseIP("0.0.0.0"))))
+	assert.Equal(t, "cool", tree.Contains(iputil.Ip2VpnIp(net.ParseIP("255.255.255.255"))))
+}
+
+func BenchmarkCIDRTree_Contains(b *testing.B) {
+	tree := NewTree4()
+	tree.AddCIDR(Parse("1.1.0.0/16"), "1")
+	tree.AddCIDR(Parse("1.2.1.1/32"), "1")
+	tree.AddCIDR(Parse("192.2.1.1/32"), "1")
+	tree.AddCIDR(Parse("172.2.1.1/32"), "1")
+
+	ip := iputil.Ip2VpnIp(net.ParseIP("1.2.1.1"))
+	b.Run("found", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			tree.Contains(ip)
+		}
+	})
+
+	ip = iputil.Ip2VpnIp(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 := NewTree4()
+	tree.AddCIDR(Parse("1.1.0.0/16"), "1")
+	tree.AddCIDR(Parse("1.2.1.1/32"), "1")
+	tree.AddCIDR(Parse("192.2.1.1/32"), "1")
+	tree.AddCIDR(Parse("172.2.1.1/32"), "1")
+
+	ip := iputil.Ip2VpnIp(net.ParseIP("1.2.1.1"))
+	b.Run("found", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			tree.Match(ip)
+		}
+	})
+
+	ip = iputil.Ip2VpnIp(net.ParseIP("1.2.1.255"))
+	b.Run("not found", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			tree.Match(ip)
+		}
+	})
+}

cidr/tree6.go

@@ -0,0 +1,185 @@
+package cidr
+
+import (
+	"net"
+
+	"github.com/slackhq/nebula/iputil"
+)
+
+const startbit6 = uint64(1 << 63)
+
+type Tree6 struct {
+	root4 *Node
+	root6 *Node
+}
+
+func NewTree6() *Tree6 {
+	tree := new(Tree6)
+	tree.root4 = &Node{}
+	tree.root6 = &Node{}
+	return tree
+}
+
+func (tree *Tree6) AddCIDR(cidr *net.IPNet, val interface{}) {
+	var node, next *Node
+
+	cidrIP, ipv4 := isIPV4(cidr.IP)
+	if ipv4 {
+		node = tree.root4
+		next = tree.root4
+
+	} else {
+		node = tree.root6
+		next = tree.root6
+	}
+
+	for i := 0; i < len(cidrIP); i += 4 {
+		ip := iputil.Ip2VpnIp(cidrIP[i : i+4])
+		mask := iputil.Ip2VpnIp(cidr.Mask[i : i+4])
+		bit := startbit
+
+		// 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
+		}
+
+		// Build up the rest of the tree we don't already have
+		for bit&mask != 0 {
+			next = &Node{}
+			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 most specific match
+func (tree *Tree6) MostSpecificContains(ip net.IP) (value interface{}) {
+	var node *Node
+
+	wholeIP, ipv4 := isIPV4(ip)
+	if ipv4 {
+		node = tree.root4
+	} else {
+		node = tree.root6
+	}
+
+	for i := 0; i < len(wholeIP); i += 4 {
+		ip := iputil.Ip2VpnIp(wholeIP[i : i+4])
+		bit := startbit
+
+		for node != nil {
+			if node.value != nil {
+				value = node.value
+			}
+
+			if bit == 0 {
+				break
+			}
+
+			if ip&bit != 0 {
+				node = node.right
+			} else {
+				node = node.left
+			}
+
+			bit >>= 1
+		}
+	}
+
+	return value
+}
+
+func (tree *Tree6) MostSpecificContainsIpV4(ip iputil.VpnIp) (value interface{}) {
+	bit := startbit
+	node := tree.root4
+
+	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
+}
+
+func (tree *Tree6) MostSpecificContainsIpV6(hi, lo uint64) (value interface{}) {
+	ip := hi
+	node := tree.root6
+
+	for i := 0; i < 2; i++ {
+		bit := startbit6
+
+		for node != nil {
+			if node.value != nil {
+				value = node.value
+			}
+
+			if bit == 0 {
+				break
+			}
+
+			if ip&bit != 0 {
+				node = node.right
+			} else {
+				node = node.left
+			}
+
+			bit >>= 1
+		}
+
+		ip = lo
+	}
+
+	return value
+}
+
+func isIPV4(ip net.IP) (net.IP, bool) {
+	if len(ip) == net.IPv4len {
+		return ip, true
+	}
+
+	if len(ip) == net.IPv6len && isZeros(ip[0:10]) && ip[10] == 0xff && ip[11] == 0xff {
+		return ip[12:16], true
+	}
+
+	return ip, false
+}
+
+func isZeros(p net.IP) bool {
+	for i := 0; i < len(p); i++ {
+		if p[i] != 0 {
+			return false
+		}
+	}
+	return true
+}

cidr/tree6_test.go

@@ -0,0 +1,81 @@
+package cidr
+
+import (
+	"encoding/binary"
+	"net"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestCIDR6Tree_MostSpecificContains(t *testing.T) {
+	tree := NewTree6()
+	tree.AddCIDR(Parse("1.0.0.0/8"), "1")
+	tree.AddCIDR(Parse("2.1.0.0/16"), "2")
+	tree.AddCIDR(Parse("3.1.1.0/24"), "3")
+	tree.AddCIDR(Parse("4.1.1.1/24"), "4a")
+	tree.AddCIDR(Parse("4.1.1.1/30"), "4b")
+	tree.AddCIDR(Parse("4.1.1.1/32"), "4c")
+	tree.AddCIDR(Parse("254.0.0.0/4"), "5")
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/64"), "6a")
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/80"), "6b")
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/96"), "6c")
+
+	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"},
+		{"6a", "1:2:0:4:1:1:1:1"},
+		{"6b", "1:2:0:4:5:1:1:1"},
+		{"6c", "1:2:0:4:5:0:0:0"},
+		{nil, "239.0.0.0"},
+		{nil, "4.1.2.2"},
+	}
+
+	for _, tt := range tests {
+		assert.Equal(t, tt.Result, tree.MostSpecificContains(net.ParseIP(tt.IP)))
+	}
+
+	tree = NewTree6()
+	tree.AddCIDR(Parse("1.1.1.1/0"), "cool")
+	tree.AddCIDR(Parse("::/0"), "cool6")
+	assert.Equal(t, "cool", tree.MostSpecificContains(net.ParseIP("0.0.0.0")))
+	assert.Equal(t, "cool", tree.MostSpecificContains(net.ParseIP("255.255.255.255")))
+	assert.Equal(t, "cool6", tree.MostSpecificContains(net.ParseIP("::")))
+	assert.Equal(t, "cool6", tree.MostSpecificContains(net.ParseIP("1:2:3:4:5:6:7:8")))
+}
+
+func TestCIDR6Tree_MostSpecificContainsIpV6(t *testing.T) {
+	tree := NewTree6()
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/64"), "6a")
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/80"), "6b")
+	tree.AddCIDR(Parse("1:2:0:4:5:0:0:0/96"), "6c")
+
+	tests := []struct {
+		Result interface{}
+		IP     string
+	}{
+		{"6a", "1:2:0:4:1:1:1:1"},
+		{"6b", "1:2:0:4:5:1:1:1"},
+		{"6c", "1:2:0:4:5:0:0:0"},
+	}
+
+	for _, tt := range tests {
+		ip := net.ParseIP(tt.IP)
+		hi := binary.BigEndian.Uint64(ip[:8])
+		lo := binary.BigEndian.Uint64(ip[8:])
+
+		assert.Equal(t, tt.Result, tree.MostSpecificContainsIpV6(hi, lo))
+	}
+}

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_test.go

@@ -1,3 +1,4 @@
+//go:build !windows
 // +build !windows
 
 package main

cmd/nebula-cert/sign.go

@@ -92,6 +92,10 @@ 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(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)
@@ -222,12 +226,17 @@ func signCert(args []string, out io.Writer, errOut io.Writer) error {
 }
 
 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 signSummary() string {

cmd/nebula-cert/sign_test.go

@@ -1,3 +1,4 @@
+//go:build !windows
 // +build !windows
 
 package main
@@ -167,6 +168,20 @@ func Test_signCert(t *testing.T) {
 	assert.Empty(t, ob.String())
 	assert.Empty(t, eb.String())
 
+	// mismatched ca key
+	_, caPriv2, _ := ed25519.GenerateKey(rand.Reader)
+	caKeyF2, err := ioutil.TempFile("", "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), "refusing to sign, root certificate does not match private key")
+	assert.Empty(t, ob.String())
+	assert.Empty(t, eb.String())
+
 	// failed key write
 	ob.Reset()
 	eb.Reset()

cmd/nebula-cert/verify_test.go

@@ -72,7 +72,7 @@ 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,
 		},

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

@@ -7,6 +7,7 @@ import (
 
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
 )
 
 // A version string that can be set with
@@ -46,16 +47,17 @@ func main() {
 		os.Exit(1)
 	}
 
-	config := nebula.NewConfig()
-	err := config.Load(*configPath)
+	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)
 	}
 
-	l := logrus.New()
-	l.Out = os.Stdout
-	c, err := nebula.Main(config, *configTest, Build, l, nil)
+	ctrl, err := nebula.Main(c, *configTest, Build, l, nil)
 
 	switch v := err.(type) {
 	case nebula.ContextualError:
@@ -67,8 +69,8 @@ func main() {
 	}
 
 	if !*configTest {
-		c.Start()
-		c.ShutdownBlock()
+		ctrl.Start()
+		ctrl.ShutdownBlock()
 	}
 
 	os.Exit(0)

cmd/nebula-service/service.go

@@ -9,6 +9,7 @@ import (
 	"github.com/kardianos/service"
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
 )
 
 var logger service.Logger
@@ -24,15 +25,16 @@ func (p *program) Start(s service.Service) error {
 	// Start should not block.
 	logger.Info("Nebula service starting.")
 
-	config := nebula.NewConfig()
-	err := config.Load(*p.configPath)
+	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)
 	}
 
-	l := logrus.New()
-	l.Out = os.Stdout
-	p.control, err = nebula.Main(config, *p.configTest, Build, l, nil)
+	p.control, err = nebula.Main(c, *p.configTest, Build, l, nil)
 	if err != nil {
 		return err
 	}
@@ -69,6 +71,10 @@ 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)
@@ -84,6 +90,7 @@ 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)
 			}
 		}
@@ -93,6 +100,7 @@ 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

@@ -7,6 +7,7 @@ import (
 
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/config"
 )
 
 // A version string that can be set with
@@ -40,16 +41,17 @@ func main() {
 		os.Exit(1)
 	}
 
-	config := nebula.NewConfig()
-	err := config.Load(*configPath)
+	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)
 	}
 
-	l := logrus.New()
-	l.Out = os.Stdout
-	c, err := nebula.Main(config, *configTest, Build, l, nil)
+	ctrl, err := nebula.Main(c, *configTest, Build, l, nil)
 
 	switch v := err.(type) {
 	case nebula.ContextualError:
@@ -61,8 +63,8 @@ func main() {
 	}
 
 	if !*configTest {
-		c.Start()
-		c.ShutdownBlock()
+		ctrl.Start()
+		ctrl.ShutdownBlock()
 	}
 
 	os.Exit(0)

config/config.go

@@ -1,14 +1,13 @@
-package nebula
+package config
 
 import (
+	"context"
 	"errors"
 	"fmt"
 	"io/ioutil"
-	"net"
 	"os"
 	"os/signal"
 	"path/filepath"
-	"regexp"
 	"sort"
 	"strconv"
 	"strings"
@@ -20,22 +19,24 @@ import (
 	"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
 }
 
-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)
 
@@ -58,7 +59,7 @@ func (c *Config) Load(path string) error {
 	return nil
 }
 
-func (c *Config) LoadString(raw string) error {
+func (c *C) LoadString(raw string) error {
 	if raw == "" {
 		return errors.New("Empty configuration")
 	}
@@ -69,7 +70,7 @@ func (c *Config) LoadString(raw string) error {
 // 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)
 }
 
@@ -78,7 +79,7 @@ func (c *Config) RegisterReloadCallback(f func(*Config)) {
 // 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
 	}
@@ -99,12 +100,12 @@ 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)
@@ -112,19 +113,26 @@ 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) {
 	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.oldSettings = make(map[interface{}]interface{})
 	for k, v := range c.Settings {
 		c.oldSettings[k] = v
@@ -132,7 +140,7 @@ 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
 	}
 
@@ -142,7 +150,7 @@ func (c *Config) ReloadConfig() {
 }
 
 // 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
@@ -152,7 +160,7 @@ 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
@@ -172,7 +180,7 @@ 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
@@ -187,7 +195,7 @@ 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 {
@@ -198,7 +206,7 @@ func (c *Config) GetInt(k string, d int) int {
 }
 
 // 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 {
@@ -215,7 +223,7 @@ 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 {
@@ -224,138 +232,15 @@ 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{})
@@ -374,7 +259,7 @@ 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
@@ -400,7 +285,7 @@ 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" {
@@ -416,7 +301,7 @@ func (c *Config) addFile(path string, direct bool) error {
 	return nil
 }
 
-func (c *Config) parseRaw(b []byte) error {
+func (c *C) parseRaw(b []byte) error {
 	var m map[interface{}]interface{}
 
 	err := yaml.Unmarshal(b, &m)
@@ -428,7 +313,7 @@ func (c *Config) parseRaw(b []byte) error {
 	return nil
 }
 
-func (c *Config) parse() error {
+func (c *C) parse() error {
 	var m map[interface{}]interface{}
 
 	for _, path := range c.files {
@@ -471,38 +356,3 @@ 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)
-
-	disableTimestamp := c.GetBool("logging.disable_timestamp", false)
-	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,
-			DisableTimestamp: disableTimestamp,
-		}
-	case "json":
-		l.Formatter = &logrus.JSONFormatter{
-			TimestampFormat:  timestampFormat,
-			DisableTimestamp: disableTimestamp,
-		}
-	default:
-		return fmt.Errorf("unknown log format `%s`. possible formats: %s", logFormat, []string{"text", "json"})
-	}
-
-	return nil
-}

config/config_test.go

@@ -1,4 +1,4 @@
-package nebula
+package config
 
 import (
 	"io/ioutil"
@@ -7,18 +7,20 @@ import (
 	"testing"
 	"time"
 
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
 func TestConfig_Load(t *testing.T) {
+	l := util.NewTestLogger()
 	dir, err := ioutil.TempDir("", "config-test")
 	// invalid yaml
-	c := NewConfig()
+	c := NewC(l)
 	ioutil.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)
 
@@ -40,8 +42,9 @@ func TestConfig_Load(t *testing.T) {
 }
 
 func TestConfig_Get(t *testing.T) {
+	l := util.NewTestLogger()
 	// test simple type
-	c := NewConfig()
+	c := NewC(l)
 	c.Settings["firewall"] = map[interface{}]interface{}{"outbound": "hi"}
 	assert.Equal(t, "hi", c.Get("firewall.outbound"))
 
@@ -55,13 +58,15 @@ func TestConfig_Get(t *testing.T) {
 }
 
 func TestConfig_GetStringSlice(t *testing.T) {
-	c := NewConfig()
+	l := util.NewTestLogger()
+	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 := util.NewTestLogger()
+	c := NewC(l)
 	c.Settings["bool"] = true
 	assert.Equal(t, true, c.GetBool("bool", false))
 
@@ -87,91 +92,22 @@ 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 := util.NewTestLogger()
 	// 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"))
@@ -179,12 +115,13 @@ func TestConfig_HasChanged(t *testing.T) {
 }
 
 func TestConfig_ReloadConfig(t *testing.T) {
+	l := util.NewTestLogger()
 	done := make(chan bool, 1)
 	dir, err := ioutil.TempDir("", "config-test")
 	assert.Nil(t, err)
 	ioutil.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"))
@@ -193,7 +130,7 @@ func TestConfig_ReloadConfig(t *testing.T) {
 
 	ioutil.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n  inner: ho"), 0644)
 
-	c.RegisterReloadCallback(func(c *Config) {
+	c.RegisterReloadCallback(func(c *C) {
 		done <- true
 	})
 

connection_manager.go

@@ -1,10 +1,13 @@
 package nebula
 
 import (
+	"context"
 	"sync"
 	"time"
 
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
 )
 
 // TODO: incount and outcount are intended as a shortcut to locking the mutexes for every single packet
@@ -12,47 +15,49 @@ import (
 
 type connectionManager struct {
 	hostMap      *HostMap
-	in           map[uint32]struct{}
+	in           map[iputil.VpnIp]struct{}
 	inLock       *sync.RWMutex
 	inCount      int
-	out          map[uint32]struct{}
+	out          map[iputil.VpnIp]struct{}
 	outLock      *sync.RWMutex
 	outCount     int
 	TrafficTimer *SystemTimerWheel
 	intf         *Interface
 
-	pendingDeletion      map[uint32]int
+	pendingDeletion      map[iputil.VpnIp]int
 	pendingDeletionLock  *sync.RWMutex
 	pendingDeletionTimer *SystemTimerWheel
 
 	checkInterval           int
 	pendingDeletionInterval int
 
+	l *logrus.Logger
 	// I wanted to call one matLock
 }
 
-func newConnectionManager(intf *Interface, checkInterval, pendingDeletionInterval int) *connectionManager {
+func newConnectionManager(ctx context.Context, l *logrus.Logger, intf *Interface, checkInterval, pendingDeletionInterval int) *connectionManager {
 	nc := &connectionManager{
 		hostMap:                 intf.hostMap,
-		in:                      make(map[uint32]struct{}),
+		in:                      make(map[iputil.VpnIp]struct{}),
 		inLock:                  &sync.RWMutex{},
 		inCount:                 0,
-		out:                     make(map[uint32]struct{}),
+		out:                     make(map[iputil.VpnIp]struct{}),
 		outLock:                 &sync.RWMutex{},
 		outCount:                0,
 		TrafficTimer:            NewSystemTimerWheel(time.Millisecond*500, time.Second*60),
 		intf:                    intf,
-		pendingDeletion:         make(map[uint32]int),
+		pendingDeletion:         make(map[iputil.VpnIp]int),
 		pendingDeletionLock:     &sync.RWMutex{},
 		pendingDeletionTimer:    NewSystemTimerWheel(time.Millisecond*500, time.Second*60),
 		checkInterval:           checkInterval,
 		pendingDeletionInterval: pendingDeletionInterval,
+		l:                       l,
 	}
-	nc.Start()
+	nc.Start(ctx)
 	return nc
 }
 
-func (n *connectionManager) In(ip uint32) {
+func (n *connectionManager) In(ip iputil.VpnIp) {
 	n.inLock.RLock()
 	// If this already exists, return
 	if _, ok := n.in[ip]; ok {
@@ -65,7 +70,7 @@ func (n *connectionManager) In(ip uint32) {
 	n.inLock.Unlock()
 }
 
-func (n *connectionManager) Out(ip uint32) {
+func (n *connectionManager) Out(ip iputil.VpnIp) {
 	n.outLock.RLock()
 	// If this already exists, return
 	if _, ok := n.out[ip]; ok {
@@ -84,9 +89,9 @@ func (n *connectionManager) Out(ip uint32) {
 	n.outLock.Unlock()
 }
 
-func (n *connectionManager) CheckIn(vpnIP uint32) bool {
+func (n *connectionManager) CheckIn(vpnIp iputil.VpnIp) bool {
 	n.inLock.RLock()
-	if _, ok := n.in[vpnIP]; ok {
+	if _, ok := n.in[vpnIp]; ok {
 		n.inLock.RUnlock()
 		return true
 	}
@@ -94,7 +99,7 @@ func (n *connectionManager) CheckIn(vpnIP uint32) bool {
 	return false
 }
 
-func (n *connectionManager) ClearIP(ip uint32) {
+func (n *connectionManager) ClearIP(ip iputil.VpnIp) {
 	n.inLock.Lock()
 	n.outLock.Lock()
 	delete(n.in, ip)
@@ -103,13 +108,13 @@ func (n *connectionManager) ClearIP(ip uint32) {
 	n.outLock.Unlock()
 }
 
-func (n *connectionManager) ClearPendingDeletion(ip uint32) {
+func (n *connectionManager) ClearPendingDeletion(ip iputil.VpnIp) {
 	n.pendingDeletionLock.Lock()
 	delete(n.pendingDeletion, ip)
 	n.pendingDeletionLock.Unlock()
 }
 
-func (n *connectionManager) AddPendingDeletion(ip uint32) {
+func (n *connectionManager) AddPendingDeletion(ip iputil.VpnIp) {
 	n.pendingDeletionLock.Lock()
 	if _, ok := n.pendingDeletion[ip]; ok {
 		n.pendingDeletion[ip] += 1
@@ -120,7 +125,7 @@ func (n *connectionManager) AddPendingDeletion(ip uint32) {
 	n.pendingDeletionLock.Unlock()
 }
 
-func (n *connectionManager) checkPendingDeletion(ip uint32) bool {
+func (n *connectionManager) checkPendingDeletion(ip iputil.VpnIp) bool {
 	n.pendingDeletionLock.RLock()
 	if _, ok := n.pendingDeletion[ip]; ok {
 
@@ -131,23 +136,30 @@ func (n *connectionManager) checkPendingDeletion(ip uint32) bool {
 	return false
 }
 
-func (n *connectionManager) AddTrafficWatch(vpnIP uint32, seconds int) {
-	n.TrafficTimer.Add(vpnIP, time.Second*time.Duration(seconds))
+func (n *connectionManager) AddTrafficWatch(vpnIp iputil.VpnIp, seconds int) {
+	n.TrafficTimer.Add(vpnIp, time.Second*time.Duration(seconds))
 }
 
-func (n *connectionManager) Start() {
-	go n.Run()
+func (n *connectionManager) Start(ctx context.Context) {
+	go n.Run(ctx)
 }
 
-func (n *connectionManager) Run() {
-	clockSource := time.Tick(500 * time.Millisecond)
+func (n *connectionManager) Run(ctx context.Context) {
+	clockSource := time.NewTicker(500 * time.Millisecond)
+	defer clockSource.Stop()
+
 	p := []byte("")
 	nb := make([]byte, 12, 12)
 	out := make([]byte, mtu)
 
-	for now := range clockSource {
-		n.HandleMonitorTick(now, p, nb, out)
-		n.HandleDeletionTick(now)
+	for {
+		select {
+		case <-ctx.Done():
+			return
+		case now := <-clockSource.C:
+			n.HandleMonitorTick(now, p, nb, out)
+			n.HandleDeletionTick(now)
+		}
 	}
 }
 
@@ -159,44 +171,51 @@ func (n *connectionManager) HandleMonitorTick(now time.Time, p, nb, out []byte)
 			break
 		}
 
-		vpnIP := ep.(uint32)
+		vpnIp := ep.(iputil.VpnIp)
 
 		// Check for traffic coming back in from this host.
-		traf := n.CheckIn(vpnIP)
+		traf := n.CheckIn(vpnIp)
 
-		// If we saw incoming packets from this ip, just return
+		hostinfo, err := n.hostMap.QueryVpnIp(vpnIp)
+		if err != nil {
+			n.l.Debugf("Not found in hostmap: %s", vpnIp)
+
+			if !n.intf.disconnectInvalid {
+				n.ClearIP(vpnIp)
+				n.ClearPendingDeletion(vpnIp)
+				continue
+			}
+		}
+
+		if n.handleInvalidCertificate(now, vpnIp, hostinfo) {
+			continue
+		}
+
+		// If we saw an incoming packets from this ip and peer's certificate is not
+		// expired, just ignore.
 		if traf {
-			if l.Level >= logrus.DebugLevel {
-				l.WithField("vpnIp", IntIp(vpnIP)).
+			if n.l.Level >= logrus.DebugLevel {
+				n.l.WithField("vpnIp", vpnIp).
 					WithField("tunnelCheck", m{"state": "alive", "method": "passive"}).
 					Debug("Tunnel status")
 			}
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
+			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().
+		hostinfo.logger(n.l).
 			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, p, nb, out)
+			n.intf.SendMessageToVpnIp(header.Test, header.TestRequest, vpnIp, p, nb, out)
 
 		} else {
-			hostinfo.logger().Debugf("Hostinfo sadness: %s", IntIp(vpnIP))
+			hostinfo.logger(n.l).Debugf("Hostinfo sadness: %s", vpnIp)
 		}
-		n.AddPendingDeletion(vpnIP)
+		n.AddPendingDeletion(vpnIp)
 	}
 
 }
@@ -209,48 +228,88 @@ func (n *connectionManager) HandleDeletionTick(now time.Time) {
 			break
 		}
 
-		vpnIP := ep.(uint32)
+		vpnIp := ep.(iputil.VpnIp)
 
-		// 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"}).
-				Debug("Tunnel status")
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
+		hostinfo, err := n.hostMap.QueryVpnIp(vpnIp)
+		if err != nil {
+			n.l.Debugf("Not found in hostmap: %s", vpnIp)
+
+			if !n.intf.disconnectInvalid {
+				n.ClearIP(vpnIp)
+				n.ClearPendingDeletion(vpnIp)
+				continue
+			}
+		}
+
+		if n.handleInvalidCertificate(now, vpnIp, hostinfo) {
 			continue
 		}
 
-		hostinfo, err := n.hostMap.QueryVpnIP(vpnIP)
-		if err != nil {
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
-			l.Debugf("Not found in hostmap: %s", IntIp(vpnIP))
+		// If we saw an incoming packets from this ip and peer's certificate is not
+		// expired, just ignore.
+		traf := n.CheckIn(vpnIp)
+		if traf {
+			n.l.WithField("vpnIp", vpnIp).
+				WithField("tunnelCheck", m{"state": "alive", "method": "active"}).
+				Debug("Tunnel status")
+
+			n.ClearIP(vpnIp)
+			n.ClearPendingDeletion(vpnIp)
 			continue
 		}
 
 		// If it comes around on deletion wheel and hasn't resolved itself, delete
-		if n.checkPendingDeletion(vpnIP) {
+		if n.checkPendingDeletion(vpnIp) {
 			cn := ""
 			if hostinfo.ConnectionState != nil && hostinfo.ConnectionState.peerCert != nil {
 				cn = hostinfo.ConnectionState.peerCert.Details.Name
 			}
-			hostinfo.logger().
+			hostinfo.logger(n.l).
 				WithField("tunnelCheck", m{"state": "dead", "method": "active"}).
 				WithField("certName", cn).
 				Info("Tunnel status")
 
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
+			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.intf.lightHouse.DeleteVpnIp(vpnIp)
 			}
 			n.hostMap.DeleteHostInfo(hostinfo)
 		} else {
-			n.ClearIP(vpnIP)
-			n.ClearPendingDeletion(vpnIP)
+			n.ClearIP(vpnIp)
+			n.ClearPendingDeletion(vpnIp)
 		}
 	}
 }
+
+// handleInvalidCertificates will destroy a tunnel if pki.disconnect_invalid is true and the certificate is no longer valid
+func (n *connectionManager) handleInvalidCertificate(now time.Time, vpnIp iputil.VpnIp, hostinfo *HostInfo) bool {
+	if !n.intf.disconnectInvalid {
+		return false
+	}
+
+	remoteCert := hostinfo.GetCert()
+	if remoteCert == nil {
+		return false
+	}
+
+	valid, err := remoteCert.Verify(now, n.intf.caPool)
+	if valid {
+		return false
+	}
+
+	fingerprint, _ := remoteCert.Sha256Sum()
+	n.l.WithField("vpnIp", vpnIp).WithError(err).
+		WithField("certName", remoteCert.Details.Name).
+		WithField("fingerprint", fingerprint).
+		Info("Remote certificate is no longer valid, tearing down the tunnel")
+
+	// Inform the remote and close the tunnel locally
+	n.intf.sendCloseTunnel(hostinfo)
+	n.intf.closeTunnel(hostinfo, false)
+
+	n.ClearIP(vpnIp)
+	n.ClearPendingDeletion(vpnIp)
+	return true
+}

connection_manager_test.go

@@ -1,26 +1,33 @@
 package nebula
 
 import (
+	"context"
+	"crypto/ed25519"
+	"crypto/rand"
 	"net"
 	"testing"
 	"time"
 
 	"github.com/flynn/noise"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
-var vpnIP uint32
+var vpnIp iputil.VpnIp
 
 func Test_NewConnectionManagerTest(t *testing.T) {
+	l := util.NewTestLogger()
 	//_, 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"))
+	vpnIp = iputil.Ip2VpnIp(net.ParseIP("172.1.1.2"))
 	preferredRanges := []*net.IPNet{localrange}
 
 	// Very incomplete mock objects
-	hostMap := NewHostMap("test", vpncidr, preferredRanges)
+	hostMap := NewHostMap(l, "test", vpncidr, preferredRanges)
 	cs := &CertState{
 		rawCertificate:      []byte{},
 		privateKey:          []byte{},
@@ -28,35 +35,38 @@ func Test_NewConnectionManagerTest(t *testing.T) {
 		rawCertificateNoKey: []byte{},
 	}
 
-	lh := NewLightHouse(false, 0, []uint32{}, 1000, 0, &udpConn{}, false, 1, false)
+	lh := NewLightHouse(l, false, &net.IPNet{IP: net.IP{0, 0, 0, 0}, Mask: net.IPMask{0, 0, 0, 0}}, []iputil.VpnIp{}, 1000, 0, &udp.Conn{}, false, 1, false)
 	ifce := &Interface{
 		hostMap:          hostMap,
 		inside:           &Tun{},
-		outside:          &udpConn{},
+		outside:          &udp.Conn{},
 		certState:        cs,
 		firewall:         &Firewall{},
 		lightHouse:       lh,
-		handshakeManager: NewHandshakeManager(vpncidr, preferredRanges, hostMap, lh, &udpConn{}, defaultHandshakeConfig),
+		handshakeManager: NewHandshakeManager(l, vpncidr, preferredRanges, hostMap, lh, &udp.Conn{}, defaultHandshakeConfig),
+		l:                l,
 	}
 	now := time.Now()
 
 	// Create manager
-	nc := newConnectionManager(ifce, 5, 10)
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	nc := newConnectionManager(ctx, l, ifce, 5, 10)
 	p := []byte("")
 	nb := make([]byte, 12, 12)
 	out := make([]byte, mtu)
 	nc.HandleMonitorTick(now, p, nb, out)
 	// Add an ip we have established a connection w/ to hostmap
-	hostinfo := nc.hostMap.AddVpnIP(vpnIP)
+	hostinfo, _ := nc.hostMap.AddVpnIp(vpnIp, nil)
 	hostinfo.ConnectionState = &ConnectionState{
 		certState: cs,
 		H:         &noise.HandshakeState{},
 	}
 
-	// We saw traffic out to vpnIP
-	nc.Out(vpnIP)
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
+	// 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, p, nb, out)
@@ -66,26 +76,27 @@ func Test_NewConnectionManagerTest(t *testing.T) {
 	nc.HandleMonitorTick(next_tick, p, nb, out)
 	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)
+	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, p, nb, out)
 	nc.HandleDeletionTick(next_tick)
 	// The host should be evicted
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.NotContains(t, nc.hostMap.Hosts, vpnIP)
+	assert.NotContains(t, nc.pendingDeletion, vpnIp)
+	assert.NotContains(t, nc.hostMap.Hosts, vpnIp)
 
 }
 
 func Test_NewConnectionManagerTest2(t *testing.T) {
+	l := util.NewTestLogger()
 	//_, 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}
 
 	// Very incomplete mock objects
-	hostMap := NewHostMap("test", vpncidr, preferredRanges)
+	hostMap := NewHostMap(l, "test", vpncidr, preferredRanges)
 	cs := &CertState{
 		rawCertificate:      []byte{},
 		privateKey:          []byte{},
@@ -93,35 +104,38 @@ func Test_NewConnectionManagerTest2(t *testing.T) {
 		rawCertificateNoKey: []byte{},
 	}
 
-	lh := NewLightHouse(false, 0, []uint32{}, 1000, 0, &udpConn{}, false, 1, false)
+	lh := NewLightHouse(l, false, &net.IPNet{IP: net.IP{0, 0, 0, 0}, Mask: net.IPMask{0, 0, 0, 0}}, []iputil.VpnIp{}, 1000, 0, &udp.Conn{}, false, 1, false)
 	ifce := &Interface{
 		hostMap:          hostMap,
 		inside:           &Tun{},
-		outside:          &udpConn{},
+		outside:          &udp.Conn{},
 		certState:        cs,
 		firewall:         &Firewall{},
 		lightHouse:       lh,
-		handshakeManager: NewHandshakeManager(vpncidr, preferredRanges, hostMap, lh, &udpConn{}, defaultHandshakeConfig),
+		handshakeManager: NewHandshakeManager(l, vpncidr, preferredRanges, hostMap, lh, &udp.Conn{}, defaultHandshakeConfig),
+		l:                l,
 	}
 	now := time.Now()
 
 	// Create manager
-	nc := newConnectionManager(ifce, 5, 10)
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	nc := newConnectionManager(ctx, l, ifce, 5, 10)
 	p := []byte("")
 	nb := make([]byte, 12, 12)
 	out := make([]byte, mtu)
 	nc.HandleMonitorTick(now, p, nb, out)
 	// Add an ip we have established a connection w/ to hostmap
-	hostinfo := nc.hostMap.AddVpnIP(vpnIP)
+	hostinfo, _ := nc.hostMap.AddVpnIp(vpnIp, nil)
 	hostinfo.ConnectionState = &ConnectionState{
 		certState: cs,
 		H:         &noise.HandshakeState{},
 	}
 
-	// We saw traffic out to vpnIP
-	nc.Out(vpnIP)
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
+	// 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, p, nb, out)
@@ -131,16 +145,111 @@ func Test_NewConnectionManagerTest2(t *testing.T) {
 	nc.HandleMonitorTick(next_tick, p, nb, out)
 	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)
+	assert.Contains(t, nc.pendingDeletion, vpnIp)
+	assert.Contains(t, nc.hostMap.Hosts, vpnIp)
 	// We heard back this time
-	nc.In(vpnIP)
+	nc.In(vpnIp)
 	// Move ahead some more
 	next_tick = now.Add(45 * time.Second)
 	nc.HandleMonitorTick(next_tick, p, nb, out)
 	nc.HandleDeletionTick(next_tick)
 	// The host should be evicted
-	assert.NotContains(t, nc.pendingDeletion, vpnIP)
-	assert.Contains(t, nc.hostMap.Hosts, vpnIP)
+	assert.NotContains(t, nc.pendingDeletion, vpnIp)
+	assert.Contains(t, nc.hostMap.Hosts, 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 := util.NewTestLogger()
+	ipNet := net.IPNet{
+		IP:   net.IPv4(172, 1, 1, 2),
+		Mask: net.IPMask{255, 255, 255, 0},
+	}
+	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
+	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
+	preferredRanges := []*net.IPNet{localrange}
+	hostMap := NewHostMap(l, "test", vpncidr, 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,
+		},
+	}
+	caCert.Sign(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",
+		},
+	}
+	peerCert.Sign(privCA)
+
+	cs := &CertState{
+		rawCertificate:      []byte{},
+		privateKey:          []byte{},
+		certificate:         &cert.NebulaCertificate{},
+		rawCertificateNoKey: []byte{},
+	}
+
+	lh := NewLightHouse(l, false, &net.IPNet{IP: net.IP{0, 0, 0, 0}, Mask: net.IPMask{0, 0, 0, 0}}, []iputil.VpnIp{}, 1000, 0, &udp.Conn{}, false, 1, false)
+	ifce := &Interface{
+		hostMap:           hostMap,
+		inside:            &Tun{},
+		outside:           &udp.Conn{},
+		certState:         cs,
+		firewall:          &Firewall{},
+		lightHouse:        lh,
+		handshakeManager:  NewHandshakeManager(l, vpncidr, preferredRanges, hostMap, lh, &udp.Conn{}, defaultHandshakeConfig),
+		l:                 l,
+		disconnectInvalid: true,
+		caPool:            ncp,
+	}
+
+	// Create manager
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	nc := newConnectionManager(ctx, l, ifce, 5, 10)
+	ifce.connectionManager = nc
+	hostinfo, _ := nc.hostMap.AddVpnIp(vpnIp, nil)
+	hostinfo.ConnectionState = &ConnectionState{
+		certState: cs,
+		peerCert:  &peerCert,
+		H:         &noise.HandshakeState{},
+	}
+
+	// Move ahead 45s.
+	// Check if to disconnect with invalid certificate.
+	// Should be alive.
+	nextTick := now.Add(45 * time.Second)
+	destroyed := nc.handleInvalidCertificate(nextTick, vpnIp, hostinfo)
+	assert.False(t, destroyed)
+
+	// Move ahead 61s.
+	// Check if to disconnect with invalid certificate.
+	// Should be disconnected.
+	nextTick = now.Add(61 * time.Second)
+	destroyed = nc.handleInvalidCertificate(nextTick, vpnIp, hostinfo)
+	assert.True(t, destroyed)
+}

connection_state.go

@@ -7,6 +7,7 @@ import (
 	"sync/atomic"
 
 	"github.com/flynn/noise"
+	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
 )
 
@@ -26,7 +27,7 @@ type ConnectionState struct {
 	ready                bool
 }
 
-func (f *Interface) newConnectionState(initiator bool, pattern noise.HandshakePattern, psk []byte, pskStage int) *ConnectionState {
+func (f *Interface) newConnectionState(l *logrus.Logger, 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)
@@ -37,7 +38,7 @@ func (f *Interface) newConnectionState(initiator bool, pattern noise.HandshakePa
 
 	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,

control.go

@@ -1,6 +1,7 @@
 package nebula
 
 import (
+	"context"
 	"net"
 	"os"
 	"os/signal"
@@ -9,45 +10,62 @@ import (
 
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
 // 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 Control struct {
-	f *Interface
-	l *logrus.Logger
+	f          *Interface
+	l          *logrus.Logger
+	cancel     context.CancelFunc
+	sshStart   func()
+	statsStart func()
+	dnsStart   func()
 }
 
 type ControlHostInfo struct {
-	VpnIP          net.IP                  `json:"vpnIp"`
+	VpnIp          net.IP                  `json:"vpnIp"`
 	LocalIndex     uint32                  `json:"localIndex"`
 	RemoteIndex    uint32                  `json:"remoteIndex"`
-	RemoteAddrs    []udpAddr               `json:"remoteAddrs"`
+	RemoteAddrs    []*udp.Addr             `json:"remoteAddrs"`
 	CachedPackets  int                     `json:"cachedPackets"`
 	Cert           *cert.NebulaCertificate `json:"cert"`
 	MessageCounter uint64                  `json:"messageCounter"`
-	CurrentRemote  udpAddr                 `json:"currentRemote"`
+	CurrentRemote  *udp.Addr               `json:"currentRemote"`
 }
 
 // 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()
+	}
+
+	// Start reading packets.
 	c.f.run()
 }
 
 // Stop signals nebula to shutdown, returns after the shutdown is complete
 func (c *Control) Stop() {
 	//TODO: stop tun and udp routines, the lock on hostMap effectively does that though
-	//TODO: this is probably better as a function in ConnectionManager or HostMap directly
-	c.f.hostMap.Lock()
-	for _, h := range c.f.hostMap.Hosts {
-		if h.ConnectionState.ready {
-			c.f.send(closeTunnel, 0, h.ConnectionState, h, h.remote, []byte{}, make([]byte, 12, 12), make([]byte, mtu))
-			c.l.WithField("vpnIp", IntIp(h.hostId)).WithField("udpAddr", h.remote).
-				Debug("Sending close tunnel message")
-		}
+	c.CloseAllTunnels(false)
+	if err := c.f.Close(); err != nil {
+		c.l.WithError(err).Error("Close interface failed")
 	}
-	c.f.hostMap.Unlock()
+	c.cancel()
 	c.l.Info("Goodbye")
 }
 
@@ -76,27 +94,15 @@ func (c *Control) RebindUDPServer() {
 
 // ListHostmap returns details about the actual or pending (handshaking) hostmap
 func (c *Control) ListHostmap(pendingMap bool) []ControlHostInfo {
-	var hm *HostMap
 	if pendingMap {
-		hm = c.f.handshakeManager.pendingHostMap
+		return listHostMap(c.f.handshakeManager.pendingHostMap)
 	} else {
-		hm = c.f.hostMap
+		return listHostMap(c.f.hostMap)
 	}
-
-	hm.RLock()
-	hosts := make([]ControlHostInfo, len(hm.Hosts))
-	i := 0
-	for _, v := range hm.Hosts {
-		hosts[i] = copyHostInfo(v)
-		i++
-	}
-	hm.RUnlock()
-
-	return hosts
 }
 
-// GetHostInfoByVpnIP returns a single tunnels hostInfo, or nil if not found
-func (c *Control) GetHostInfoByVpnIP(vpnIP uint32, pending bool) *ControlHostInfo {
+// GetHostInfoByVpnIp returns a single tunnels hostInfo, or nil if not found
+func (c *Control) GetHostInfoByVpnIp(vpnIp iputil.VpnIp, pending bool) *ControlHostInfo {
 	var hm *HostMap
 	if pending {
 		hm = c.f.handshakeManager.pendingHostMap
@@ -104,37 +110,37 @@ func (c *Control) GetHostInfoByVpnIP(vpnIP uint32, pending bool) *ControlHostInf
 		hm = c.f.hostMap
 	}
 
-	h, err := hm.QueryVpnIP(vpnIP)
+	h, err := hm.QueryVpnIp(vpnIp)
 	if err != nil {
 		return nil
 	}
 
-	ch := copyHostInfo(h)
+	ch := copyHostInfo(h, c.f.hostMap.preferredRanges)
 	return &ch
 }
 
 // SetRemoteForTunnel forces a tunnel to use a specific remote
-func (c *Control) SetRemoteForTunnel(vpnIP uint32, addr udpAddr) *ControlHostInfo {
-	hostInfo, err := c.f.hostMap.QueryVpnIP(vpnIP)
+func (c *Control) SetRemoteForTunnel(vpnIp iputil.VpnIp, addr udp.Addr) *ControlHostInfo {
+	hostInfo, err := c.f.hostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return nil
 	}
 
 	hostInfo.SetRemote(addr.Copy())
-	ch := copyHostInfo(hostInfo)
+	ch := copyHostInfo(hostInfo, c.f.hostMap.preferredRanges)
 	return &ch
 }
 
 // CloseTunnel closes a fully established tunnel. If localOnly is false it will notify the remote end as well.
-func (c *Control) CloseTunnel(vpnIP uint32, localOnly bool) bool {
-	hostInfo, err := c.f.hostMap.QueryVpnIP(vpnIP)
+func (c *Control) CloseTunnel(vpnIp iputil.VpnIp, localOnly bool) bool {
+	hostInfo, err := c.f.hostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return false
 	}
 
 	if !localOnly {
 		c.f.send(
-			closeTunnel,
+			header.CloseTunnel,
 			0,
 			hostInfo.ConnectionState,
 			hostInfo,
@@ -145,19 +151,46 @@ func (c *Control) CloseTunnel(vpnIP uint32, localOnly bool) bool {
 		)
 	}
 
-	c.f.closeTunnel(hostInfo)
+	c.f.closeTunnel(hostInfo, false)
 	return true
 }
 
-func copyHostInfo(h *HostInfo) ControlHostInfo {
-	addrs := h.RemoteUDPAddrs()
+// 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
+	c.f.hostMap.Lock()
+	for _, h := range c.f.hostMap.Hosts {
+		if excludeLighthouses {
+			if _, ok := c.f.lightHouse.lighthouses[h.vpnIp]; ok {
+				continue
+			}
+		}
+
+		if h.ConnectionState.ready {
+			c.f.send(header.CloseTunnel, 0, h.ConnectionState, h, h.remote, []byte{}, make([]byte, 12, 12), make([]byte, mtu))
+			c.f.closeTunnel(h, true)
+
+			c.l.WithField("vpnIp", h.vpnIp).WithField("udpAddr", h.remote).
+				Debug("Sending close tunnel message")
+			closed++
+		}
+	}
+	c.f.hostMap.Unlock()
+	return
+}
+
+func copyHostInfo(h *HostInfo, preferredRanges []*net.IPNet) ControlHostInfo {
 	chi := ControlHostInfo{
-		VpnIP:          int2ip(h.hostId),
-		LocalIndex:     h.localIndexId,
-		RemoteIndex:    h.remoteIndexId,
-		RemoteAddrs:    make([]udpAddr, len(addrs), len(addrs)),
-		CachedPackets:  len(h.packetStore),
-		MessageCounter: atomic.LoadUint64(&h.ConnectionState.atomicMessageCounter),
+		VpnIp:         h.vpnIp.ToIP(),
+		LocalIndex:    h.localIndexId,
+		RemoteIndex:   h.remoteIndexId,
+		RemoteAddrs:   h.remotes.CopyAddrs(preferredRanges),
+		CachedPackets: len(h.packetStore),
+	}
+
+	if h.ConnectionState != nil {
+		chi.MessageCounter = atomic.LoadUint64(&h.ConnectionState.atomicMessageCounter)
 	}
 
 	if c := h.GetCert(); c != nil {
@@ -165,12 +198,21 @@ func copyHostInfo(h *HostInfo) ControlHostInfo {
 	}
 
 	if h.remote != nil {
-		chi.CurrentRemote = *h.remote
-	}
-
-	for i, addr := range addrs {
-		chi.RemoteAddrs[i] = addr.Copy()
+		chi.CurrentRemote = h.remote.Copy()
 	}
 
 	return chi
 }
+
+func listHostMap(hm *HostMap) []ControlHostInfo {
+	hm.RLock()
+	hosts := make([]ControlHostInfo, len(hm.Hosts))
+	i := 0
+	for _, v := range hm.Hosts {
+		hosts[i] = copyHostInfo(v, hm.preferredRanges)
+		i++
+	}
+	hm.RUnlock()
+
+	return hosts
+}

control_test.go

@@ -8,23 +8,26 @@ import (
 
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
-func TestControl_GetHostInfoByVpnIP(t *testing.T) {
+func TestControl_GetHostInfoByVpnIp(t *testing.T) {
+	l := util.NewTestLogger()
 	// 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("test", &net.IPNet{}, make([]*net.IPNet, 0))
-	remote1 := NewUDPAddr(100, 4444)
-	remote2 := NewUDPAddr(101, 4444)
+	hm := NewHostMap(l, "test", &net.IPNet{}, make([]*net.IPNet, 0))
+	remote1 := udp.NewAddr(net.ParseIP("0.0.0.100"), 4444)
+	remote2 := udp.NewAddr(net.ParseIP("1:2:3:4:5:6:7:8"), 4444)
 	ipNet := net.IPNet{
 		IP:   net.IPv4(1, 2, 3, 4),
 		Mask: net.IPMask{255, 255, 255, 0},
 	}
 
 	ipNet2 := net.IPNet{
-		IP:   net.IPv4(1, 2, 3, 5),
+		IP:   net.ParseIP("1:2:3:4:5:6:7:8"),
 		Mask: net.IPMask{255, 255, 255, 0},
 	}
 
@@ -44,27 +47,29 @@ func TestControl_GetHostInfoByVpnIP(t *testing.T) {
 		Signature: []byte{1, 2, 1, 2, 1, 3},
 	}
 
-	remotes := []*HostInfoDest{NewHostInfoDest(remote1), NewHostInfoDest(remote2)}
-	hm.Add(ip2int(ipNet.IP), &HostInfo{
+	remotes := NewRemoteList()
+	remotes.unlockedPrependV4(0, NewIp4AndPort(remote1.IP, uint32(remote1.Port)))
+	remotes.unlockedPrependV6(0, NewIp6AndPort(remote2.IP, uint32(remote2.Port)))
+	hm.Add(iputil.Ip2VpnIp(ipNet.IP), &HostInfo{
 		remote:  remote1,
-		Remotes: remotes,
+		remotes: remotes,
 		ConnectionState: &ConnectionState{
 			peerCert: crt,
 		},
 		remoteIndexId: 200,
 		localIndexId:  201,
-		hostId:        ip2int(ipNet.IP),
+		vpnIp:         iputil.Ip2VpnIp(ipNet.IP),
 	})
 
-	hm.Add(ip2int(ipNet2.IP), &HostInfo{
+	hm.Add(iputil.Ip2VpnIp(ipNet2.IP), &HostInfo{
 		remote:  remote1,
-		Remotes: remotes,
+		remotes: remotes,
 		ConnectionState: &ConnectionState{
 			peerCert: nil,
 		},
 		remoteIndexId: 200,
 		localIndexId:  201,
-		hostId:        ip2int(ipNet2.IP),
+		vpnIp:         iputil.Ip2VpnIp(ipNet2.IP),
 	})
 
 	c := Control{
@@ -74,26 +79,26 @@ func TestControl_GetHostInfoByVpnIP(t *testing.T) {
 		l: logrus.New(),
 	}
 
-	thi := c.GetHostInfoByVpnIP(ip2int(ipNet.IP), false)
+	thi := c.GetHostInfoByVpnIp(iputil.Ip2VpnIp(ipNet.IP), false)
 
 	expectedInfo := ControlHostInfo{
-		VpnIP:          net.IPv4(1, 2, 3, 4).To4(),
+		VpnIp:          net.IPv4(1, 2, 3, 4).To4(),
 		LocalIndex:     201,
 		RemoteIndex:    200,
-		RemoteAddrs:    []udpAddr{*remote1, *remote2},
+		RemoteAddrs:    []*udp.Addr{remote2, remote1},
 		CachedPackets:  0,
 		Cert:           crt.Copy(),
 		MessageCounter: 0,
-		CurrentRemote:  *NewUDPAddr(100, 4444),
+		CurrentRemote:  udp.NewAddr(net.ParseIP("0.0.0.100"), 4444),
 	}
 
 	// Make sure we don't have any unexpected fields
-	assertFields(t, []string{"VpnIP", "LocalIndex", "RemoteIndex", "RemoteAddrs", "CachedPackets", "Cert", "MessageCounter", "CurrentRemote"}, thi)
+	assertFields(t, []string{"VpnIp", "LocalIndex", "RemoteIndex", "RemoteAddrs", "CachedPackets", "Cert", "MessageCounter", "CurrentRemote"}, thi)
 	util.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(ip2int(ipNet2.IP), false)
+		thi = c.GetHostInfoByVpnIp(iputil.Ip2VpnIp(ipNet2.IP), false)
 	})
 }
 

control_tester.go

@@ -0,0 +1,132 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package nebula
+
+import (
+	"net"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+)
+
+// WaitForTypeByIndex 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.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.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 net.IP, toAddr *net.UDPAddr) {
+	c.f.lightHouse.Lock()
+	remoteList := c.f.lightHouse.unlockedGetRemoteList(iputil.Ip2VpnIp(vpnIp))
+	remoteList.Lock()
+	defer remoteList.Unlock()
+	c.f.lightHouse.Unlock()
+
+	iVpnIp := iputil.Ip2VpnIp(vpnIp)
+	if v4 := toAddr.IP.To4(); v4 != nil {
+		remoteList.unlockedPrependV4(iVpnIp, NewIp4AndPort(v4, uint32(toAddr.Port)))
+	} else {
+		remoteList.unlockedPrependV6(iVpnIp, NewIp6AndPort(toAddr.IP, uint32(toAddr.Port)))
+	}
+}
+
+// GetFromTun will pull a packet off the tun side of nebula
+func (c *Control) GetFromTun(block bool) []byte {
+	return c.f.inside.(*Tun).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.Get(block)
+}
+
+func (c *Control) GetUDPTxChan() <-chan *udp.Packet {
+	return c.f.outside.TxPackets
+}
+
+func (c *Control) GetTunTxChan() <-chan []byte {
+	return c.f.inside.(*Tun).txPackets
+}
+
+// InjectUDPPacket will inject a packet into the udp side of nebula
+func (c *Control) InjectUDPPacket(p *udp.Packet) {
+	c.f.outside.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 net.IP, toPort uint16, fromPort uint16, data []byte) {
+	ip := layers.IPv4{
+		Version:  4,
+		TTL:      64,
+		Protocol: layers.IPProtocolUDP,
+		SrcIP:    c.f.inside.CidrNet().IP,
+		DstIP:    toIp,
+	}
+
+	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.(*Tun).Send(buffer.Bytes())
+}
+
+func (c *Control) GetUDPAddr() string {
+	return c.f.outside.Addr.String()
+}
+
+func (c *Control) KillPendingTunnel(vpnIp net.IP) bool {
+	hostinfo, ok := c.f.handshakeManager.pendingHostMap.Hosts[iputil.Ip2VpnIp(vpnIp)]
+	if !ok {
+		return false
+	}
+
+	c.f.handshakeManager.pendingHostMap.DeleteHostInfo(hostinfo)
+	return true
+}

dist/arch/nebula.service

@@ -5,8 +5,6 @@ After=basic.target network.target network-online.target
 
 [Service]
 SyslogIdentifier=nebula
-StandardOutput=syslog
-StandardError=syslog
 ExecReload=/bin/kill -HUP $MAINPID
 ExecStart=/usr/bin/nebula -config /etc/nebula/config.yml
 Restart=always

dist/fedora/nebula.service

@@ -0,0 +1,15 @@
+[Unit]
+Description=Nebula overlay networking tool
+
+After=basic.target network.target network-online.target
+Before=sshd.service
+Wants=basic.target network-online.target
+
+[Service]
+ExecReload=/bin/kill -HUP $MAINPID
+ExecStart=/usr/bin/nebula -config /etc/nebula/config.yml
+Restart=always
+SyslogIdentifier=nebula
+
+[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

@@ -7,6 +7,9 @@ import (
 	"sync"
 
 	"github.com/miekg/dns"
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/iputil"
 )
 
 // This whole thing should be rewritten to use context
@@ -43,8 +46,8 @@ func (d *dnsRecords) QueryCert(data string) string {
 	if ip == nil {
 		return ""
 	}
-	iip := ip2int(ip)
-	hostinfo, err := d.hostMap.QueryVpnIP(iip)
+	iip := iputil.Ip2VpnIp(ip)
+	hostinfo, err := d.hostMap.QueryVpnIp(iip)
 	if err != nil {
 		return ""
 	}
@@ -63,7 +66,7 @@ func (d *dnsRecords) Add(host, data string) {
 	d.Unlock()
 }
 
-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:
@@ -95,37 +98,44 @@ func parseQuery(m *dns.Msg, w dns.ResponseWriter) {
 	}
 }
 
-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 {
+func getDnsServerAddr(c *config.C) string {
 	return c.GetString("lighthouse.dns.host", "") + ":" + 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).Infof("Starting DNS responder")
 	err := dnsServer.ListenAndServe()
 	defer dnsServer.Shutdown()
 	if err != nil {
@@ -133,7 +143,7 @@ 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 +151,5 @@ func reloadDns(c *Config) {
 
 	l.Debug("Restarting DNS server")
 	dnsServer.Shutdown()
-	go startDns(c)
+	go startDns(l, c)
 }

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/handshakes_test.go

@@ -0,0 +1,185 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package e2e
+
+import (
+	"net"
+	"testing"
+	"time"
+
+	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/e2e/router"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestGoodHandshake(t *testing.T) {
+	ca, _, caKey, _ := newTestCaCert(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	myControl, myVpnIp, myUdpAddr := newSimpleServer(ca, caKey, "me", net.IP{10, 0, 0, 1})
+	theirControl, theirVpnIp, theirUdpAddr := newSimpleServer(ca, caKey, "them", net.IP{10, 0, 0, 2})
+
+	// Put their info in our lighthouse
+	myControl.InjectLightHouseAddr(theirVpnIp, theirUdpAddr)
+
+	// Start the servers
+	myControl.Start()
+	theirControl.Start()
+
+	t.Log("Send a udp packet through to begin standing up the tunnel, this should come out the other side")
+	myControl.InjectTunUDPPacket(theirVpnIp, 80, 80, []byte("Hi from me"))
+
+	t.Log("Have them consume my stage 0 packet. They have a tunnel now")
+	theirControl.InjectUDPPacket(myControl.GetFromUDP(true))
+
+	t.Log("Get their stage 1 packet so that we can play with it")
+	stage1Packet := theirControl.GetFromUDP(true)
+
+	t.Log("I consume a garbage packet with a proper nebula header for our tunnel")
+	// this should log a statement and get ignored, allowing the real handshake packet to complete the tunnel
+	badPacket := stage1Packet.Copy()
+	badPacket.Data = badPacket.Data[:len(badPacket.Data)-header.Len]
+	myControl.InjectUDPPacket(badPacket)
+
+	t.Log("Have me consume their real stage 1 packet. I have a tunnel now")
+	myControl.InjectUDPPacket(stage1Packet)
+
+	t.Log("Wait until we see my cached packet come through")
+	myControl.WaitForType(1, 0, theirControl)
+
+	t.Log("Make sure our host infos are correct")
+	assertHostInfoPair(t, myUdpAddr, theirUdpAddr, myVpnIp, theirVpnIp, myControl, theirControl)
+
+	t.Log("Get that cached packet and make sure it looks right")
+	myCachedPacket := theirControl.GetFromTun(true)
+	assertUdpPacket(t, []byte("Hi from me"), myCachedPacket, myVpnIp, theirVpnIp, 80, 80)
+
+	t.Log("Do a bidirectional tunnel test")
+	assertTunnel(t, myVpnIp, theirVpnIp, myControl, theirControl, router.NewR(myControl, theirControl))
+
+	myControl.Stop()
+	theirControl.Stop()
+	//TODO: assert hostmaps
+}
+
+func TestWrongResponderHandshake(t *testing.T) {
+	ca, _, caKey, _ := newTestCaCert(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+
+	// The IPs here are chosen on purpose:
+	// The current remote handling will sort by preference, public, and then lexically.
+	// So we need them to have a higher address than evil (we could apply a preference though)
+	myControl, myVpnIp, myUdpAddr := newSimpleServer(ca, caKey, "me", net.IP{10, 0, 0, 100})
+	theirControl, theirVpnIp, theirUdpAddr := newSimpleServer(ca, caKey, "them", net.IP{10, 0, 0, 99})
+	evilControl, evilVpnIp, evilUdpAddr := newSimpleServer(ca, caKey, "evil", net.IP{10, 0, 0, 2})
+
+	// Add their real udp addr, which should be tried after evil.
+	myControl.InjectLightHouseAddr(theirVpnIp, theirUdpAddr)
+
+	// Put the evil udp addr in for their vpn Ip, this is a case of being lied to by the lighthouse.
+	myControl.InjectLightHouseAddr(theirVpnIp, evilUdpAddr)
+
+	// Build a router so we don't have to reason who gets which packet
+	r := router.NewR(myControl, theirControl, evilControl)
+
+	// Start the servers
+	myControl.Start()
+	theirControl.Start()
+	evilControl.Start()
+
+	t.Log("Start the handshake process, we will route until we see our cached packet get sent to them")
+	myControl.InjectTunUDPPacket(theirVpnIp, 80, 80, []byte("Hi from me"))
+	r.RouteForAllExitFunc(func(p *udp.Packet, c *nebula.Control) router.ExitType {
+		h := &header.H{}
+		err := h.Parse(p.Data)
+		if err != nil {
+			panic(err)
+		}
+
+		if p.ToIp.Equal(theirUdpAddr.IP) && p.ToPort == uint16(theirUdpAddr.Port) && h.Type == 1 {
+			return router.RouteAndExit
+		}
+
+		return router.KeepRouting
+	})
+
+	//TODO: Assert pending hostmap - I should have a correct hostinfo for them now
+
+	t.Log("My cached packet should be received by them")
+	myCachedPacket := theirControl.GetFromTun(true)
+	assertUdpPacket(t, []byte("Hi from me"), myCachedPacket, myVpnIp, theirVpnIp, 80, 80)
+
+	t.Log("Test the tunnel with them")
+	assertHostInfoPair(t, myUdpAddr, theirUdpAddr, myVpnIp, theirVpnIp, myControl, theirControl)
+	assertTunnel(t, myVpnIp, theirVpnIp, myControl, theirControl, r)
+
+	t.Log("Flush all packets from all controllers")
+	r.FlushAll()
+
+	t.Log("Ensure ensure I don't have any hostinfo artifacts from evil")
+	assert.Nil(t, myControl.GetHostInfoByVpnIp(iputil.Ip2VpnIp(evilVpnIp), true), "My pending hostmap should not contain evil")
+	assert.Nil(t, myControl.GetHostInfoByVpnIp(iputil.Ip2VpnIp(evilVpnIp), false), "My main hostmap should not contain evil")
+	//NOTE: if evil lost the handshake race it may still have a tunnel since me would reject the handshake since the tunnel is complete
+
+	//TODO: assert hostmaps for everyone
+	t.Log("Success!")
+	myControl.Stop()
+	theirControl.Stop()
+}
+
+func Test_Case1_Stage1Race(t *testing.T) {
+	ca, _, caKey, _ := newTestCaCert(time.Now(), time.Now().Add(10*time.Minute), []*net.IPNet{}, []*net.IPNet{}, []string{})
+	myControl, myVpnIp, myUdpAddr := newSimpleServer(ca, caKey, "me  ", net.IP{10, 0, 0, 1})
+	theirControl, theirVpnIp, theirUdpAddr := newSimpleServer(ca, caKey, "them", net.IP{10, 0, 0, 2})
+
+	// Put their info in our lighthouse and vice versa
+	myControl.InjectLightHouseAddr(theirVpnIp, theirUdpAddr)
+	theirControl.InjectLightHouseAddr(myVpnIp, myUdpAddr)
+
+	// Build a router so we don't have to reason who gets which packet
+	r := router.NewR(myControl, theirControl)
+
+	// Start the servers
+	myControl.Start()
+	theirControl.Start()
+
+	t.Log("Trigger a handshake to start on both me and them")
+	myControl.InjectTunUDPPacket(theirVpnIp, 80, 80, []byte("Hi from me"))
+	theirControl.InjectTunUDPPacket(myVpnIp, 80, 80, []byte("Hi from them"))
+
+	t.Log("Get both stage 1 handshake packets")
+	myHsForThem := myControl.GetFromUDP(true)
+	theirHsForMe := theirControl.GetFromUDP(true)
+
+	t.Log("Now inject both stage 1 handshake packets")
+	myControl.InjectUDPPacket(theirHsForMe)
+	theirControl.InjectUDPPacket(myHsForThem)
+	//TODO: they should win, grab their index for me and make sure I use it in the end.
+
+	t.Log("They should not have a stage 2 (won the race) but I should send one")
+	theirControl.InjectUDPPacket(myControl.GetFromUDP(true))
+
+	t.Log("Route for me until I send a message packet to them")
+	myControl.WaitForType(1, 0, theirControl)
+
+	t.Log("My cached packet should be received by them")
+	myCachedPacket := theirControl.GetFromTun(true)
+	assertUdpPacket(t, []byte("Hi from me"), myCachedPacket, myVpnIp, theirVpnIp, 80, 80)
+
+	t.Log("Route for them until I send a message packet to me")
+	theirControl.WaitForType(1, 0, myControl)
+
+	t.Log("Their cached packet should be received by me")
+	theirCachedPacket := myControl.GetFromTun(true)
+	assertUdpPacket(t, []byte("Hi from them"), theirCachedPacket, theirVpnIp, myVpnIp, 80, 80)
+
+	t.Log("Do a bidirectional tunnel test")
+	assertTunnel(t, myVpnIp, theirVpnIp, myControl, theirControl, r)
+
+	myControl.Stop()
+	theirControl.Stop()
+	//TODO: assert hostmaps
+}
+
+//TODO: add a test with many lies

e2e/helpers_test.go

@@ -0,0 +1,311 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package e2e
+
+import (
+	"crypto/rand"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"net"
+	"os"
+	"testing"
+	"time"
+
+	"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/slackhq/nebula/iputil"
+	"github.com/stretchr/testify/assert"
+	"golang.org/x/crypto/curve25519"
+	"golang.org/x/crypto/ed25519"
+	"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, udpIp net.IP) (*nebula.Control, net.IP, *net.UDPAddr) {
+	l := NewTestLogger()
+
+	vpnIpNet := &net.IPNet{IP: make([]byte, len(udpIp)), Mask: net.IPMask{255, 255, 255, 0}}
+	copy(vpnIpNet.IP, udpIp)
+	vpnIpNet.IP[1] += 128
+	udpAddr := net.UDPAddr{
+		IP:   udpIp,
+		Port: 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.IP.String(),
+			"port": udpAddr.Port,
+		},
+		"logging": m{
+			"timestamp_format": fmt.Sprintf("%v 15:04:05.000000", name),
+			"level":            l.Level.String(),
+		},
+	}
+	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.IP, &udpAddr
+}
+
+// newTestCaCert will generate a CA cert
+func newTestCaCert(before, after time.Time, ips, subnets []*net.IPNet, 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 = ips
+	}
+
+	if len(subnets) > 0 {
+		nc.Details.Subnets = subnets
+	}
+
+	if len(groups) > 0 {
+		nc.Details.Groups = groups
+	}
+
+	err = nc.Sign(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 *net.IPNet, subnets []*net.IPNet, 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()
+
+	nc := &cert.NebulaCertificate{
+		Details: cert.NebulaCertificateDetails{
+			Name:           name,
+			Ips:            []*net.IPNet{ip},
+			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(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
+}
+
+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 net.IP, 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.RouteUntilTxTun(controlB, 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.RouteUntilTxTun(controlA, controlB)
+	assertUdpPacket(t, []byte("Hello from A"), aPacket, vpnIpA, vpnIpB, 90, 80)
+}
+
+func assertHostInfoPair(t *testing.T, addrA, addrB *net.UDPAddr, vpnIpA, vpnIpB net.IP, controlA, controlB *nebula.Control) {
+	// Get both host infos
+	hBinA := controlA.GetHostInfoByVpnIp(iputil.Ip2VpnIp(vpnIpB), false)
+	assert.NotNil(t, hBinA, "Host B was not found by vpnIp in controlA")
+
+	hAinB := controlB.GetHostInfoByVpnIp(iputil.Ip2VpnIp(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.IP.To16(), hBinA.CurrentRemote.IP.To16(), "Host B remote ip is wrong in control A")
+	assert.Equal(t, addrA.IP.To16(), hAinB.CurrentRemote.IP.To16(), "Host A remote ip is wrong in control B")
+
+	assert.Equal(t, addrB.Port, int(hBinA.CurrentRemote.Port), "Host B remote port is wrong in control A")
+	assert.Equal(t, addrA.Port, int(hAinB.CurrentRemote.Port), "Host A remote port 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 net.IP, 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, v4.SrcIP, "Source ip was incorrect")
+	assert.Equal(t, toIp, 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(ioutil.Discard)
+		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/router.go

@@ -0,0 +1,323 @@
+//go:build e2e_testing
+// +build e2e_testing
+
+package router
+
+import (
+	"fmt"
+	"net"
+	"reflect"
+	"strconv"
+	"sync"
+
+	"github.com/slackhq/nebula"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/udp"
+)
+
+type R struct {
+	// Simple map of the ip:port registered on a control to the control
+	// Basically a router, right?
+	controls map[string]*nebula.Control
+
+	// A map for inbound packets for a control that doesn't know about this address
+	inNat map[string]*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]net.UDPAddr
+
+	// All interactions are locked to help serialize behavior
+	sync.Mutex
+}
+
+type ExitType int
+
+const (
+	// Keeps routing, the function will get called again on the next packet
+	KeepRouting ExitType = 0
+	// Does not route this packet and exits immediately
+	ExitNow ExitType = 1
+	// Routes this packet and exits immediately afterwards
+	RouteAndExit ExitType = 2
+)
+
+type ExitFunc func(packet *udp.Packet, receiver *nebula.Control) ExitType
+
+func NewR(controls ...*nebula.Control) *R {
+	r := &R{
+		controls: make(map[string]*nebula.Control),
+		inNat:    make(map[string]*nebula.Control),
+		outNat:   make(map[string]net.UDPAddr),
+	}
+
+	for _, c := range controls {
+		addr := c.GetUDPAddr()
+		if _, ok := r.controls[addr]; ok {
+			panic("Duplicate listen address: " + addr)
+		}
+		r.controls[addr] = c
+	}
+
+	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 net.IP, port uint16, c *nebula.Control) {
+	r.Lock()
+	defer r.Unlock()
+
+	inAddr := net.JoinHostPort(ip.String(), fmt.Sprintf("%v", port))
+	if _, ok := r.inNat[inAddr]; ok {
+		panic("Duplicate listen address inNat: " + inAddr)
+	}
+	r.inNat[inAddr] = c
+}
+
+// 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:
+			return b
+
+		// Nope, lets push the sender along
+		case p := <-udpTx:
+			outAddr := sender.GetUDPAddr()
+			r.Lock()
+			inAddr := net.JoinHostPort(p.ToIp.String(), fmt.Sprintf("%v", p.ToPort))
+			c := r.getControl(outAddr, inAddr, p)
+			if c == nil {
+				r.Unlock()
+				panic("No control for udp tx")
+			}
+
+			c.InjectUDPPacket(p)
+			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)
+		}
+
+		outAddr := sender.GetUDPAddr()
+		inAddr := net.JoinHostPort(p.ToIp.String(), fmt.Sprintf("%v", p.ToPort))
+		receiver := r.getControl(outAddr, inAddr, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't route for host: " + inAddr)
+		}
+
+		e := whatDo(p, receiver)
+		switch e {
+		case ExitNow:
+			r.Unlock()
+			return
+
+		case RouteAndExit:
+			receiver.InjectUDPPacket(p)
+			r.Unlock()
+			return
+
+		case KeepRouting:
+			receiver.InjectUDPPacket(p)
+
+		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
+	})
+}
+
+// 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 *net.UDPAddr, finish ExitType) {
+	if finish == KeepRouting {
+		finish = RouteAndExit
+	}
+
+	r.RouteExitFunc(sender, func(p *udp.Packet, r *nebula.Control) ExitType {
+		if p.ToIp.Equal(toAddr.IP) && p.ToPort == uint16(toAddr.Port) {
+			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)
+
+		outAddr := cm[x].GetUDPAddr()
+		inAddr := net.JoinHostPort(p.ToIp.String(), fmt.Sprintf("%v", p.ToPort))
+		receiver := r.getControl(outAddr, inAddr, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't route for host: " + inAddr)
+		}
+
+		e := whatDo(p, receiver)
+		switch e {
+		case ExitNow:
+			r.Unlock()
+			return
+
+		case RouteAndExit:
+			receiver.InjectUDPPacket(p)
+			r.Unlock()
+			return
+
+		case KeepRouting:
+			receiver.InjectUDPPacket(p)
+
+		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)
+
+		outAddr := cm[x].GetUDPAddr()
+		inAddr := net.JoinHostPort(p.ToIp.String(), fmt.Sprintf("%v", p.ToPort))
+		receiver := r.getControl(outAddr, inAddr, p)
+		if receiver == nil {
+			r.Unlock()
+			panic("Can't route for host: " + inAddr)
+		}
+		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 string, p *udp.Packet) *nebula.Control {
+	if newAddr, ok := r.outNat[fromAddr+":"+toAddr]; ok {
+		p.FromIp = newAddr.IP
+		p.FromPort = uint16(newAddr.Port)
+	}
+
+	c, ok := r.inNat[toAddr]
+	if ok {
+		sHost, sPort, err := net.SplitHostPort(toAddr)
+		if err != nil {
+			panic(err)
+		}
+
+		port, err := strconv.Atoi(sPort)
+		if err != nil {
+			panic(err)
+		}
+
+		r.outNat[c.GetUDPAddr()+":"+fromAddr] = net.UDPAddr{
+			IP:   net.ParseIP(sHost),
+			Port: port,
+		}
+		return c
+	}
+
+	return r.controls[toAddr]
+}

examples/config.yml

@@ -7,9 +7,11 @@ pki:
   ca: /etc/nebula/ca.crt
   cert: /etc/nebula/host.crt
   key: /etc/nebula/host.key
-  #blocklist is a list of certificate fingerprints that we will refuse to talk to
+  # 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: false
 
 # 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.
@@ -56,6 +58,14 @@ lighthouse:
     #"10.0.0.0/8": false
     #"10.42.42.0/24": true
 
+  # EXPERIMENTAL: This option my 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
@@ -74,6 +84,7 @@ lighthouse:
 # 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)
@@ -111,9 +122,11 @@ punchy:
 # IMPORTANT: this value must be identical on ALL NODES/LIGHTHOUSES. We do not/will not support use of different ciphers simultaneously!
 #cipher: chachapoly
 
-# 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"
+# NOTE: the previous option "local_range" only allowed definition of a single range
+# and has been deprecated for "preferred_ranges"
+#preferred_ranges: ["172.16.0.0/24"]
 
 # sshd can expose informational and administrative functions via ssh this is a
 #sshd:
@@ -135,7 +148,9 @@ punchy:
 tun:
   # When tun is disabled, a lighthouse can be started without a local tun interface (and therefore without root)
   disabled: false
-  # Name of the device
+  # 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 FreeBSD: 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
@@ -152,10 +167,13 @@ tun:
   # 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
   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
 
 
 # TODO
@@ -199,18 +217,18 @@ logging:
   #   e.g.: `lighthouse.rx.HostQuery`
   #lighthouse_metrics: false
 
-# Handshake Manger Settings
+# 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
   # trigger_buffer is the size of the buffer channel for quickly sending handshakes
   # after receiving the response for lighthouse queries
   #trigger_buffer: 64
 
+
 # Nebula security group configuration
 firewall:
   conntrack:

examples/quickstart-vagrant/ansible/roles/nebula/files/systemd.nebula.service

@@ -5,8 +5,6 @@ After=basic.target network.target
 
 [Service]
 SyslogIdentifier=nebula
-StandardOutput=syslog
-StandardError=syslog
 ExecReload=/bin/kill -HUP $MAINPID
 ExecStart=/usr/local/bin/nebula -config /etc/nebula/config.yml
 Restart=always

examples/service_scripts/nebula.service

@@ -6,8 +6,6 @@ Before=sshd.service
 
 [Service]
 SyslogIdentifier=nebula
-StandardOutput=syslog
-StandardError=syslog
 ExecReload=/bin/kill -HUP $MAINPID
 ExecStart=/usr/local/bin/nebula -config /etc/nebula/config.yml
 Restart=always

firewall.go

@@ -4,7 +4,6 @@ import (
 	"crypto/sha256"
 	"encoding/binary"
 	"encoding/hex"
-	"encoding/json"
 	"errors"
 	"fmt"
 	"net"
@@ -12,22 +11,14 @@ import (
 	"strconv"
 	"strings"
 	"sync"
-	"sync/atomic"
 	"time"
 
 	"github.com/rcrowley/go-metrics"
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
-)
-
-const (
-	fwProtoAny  = 0 // When we want to handle HOPOPT (0) we can change this, if ever
-	fwProtoTCP  = 6
-	fwProtoUDP  = 17
-	fwProtoICMP = 1
-
-	fwPortAny      = 0  // Special value for matching `port: any`
-	fwPortFragment = -1 // Special value for matching `port: fragment`
+	"github.com/slackhq/nebula/cidr"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/firewall"
 )
 
 const tcpACK = 0x10
@@ -63,19 +54,29 @@ type Firewall struct {
 	DefaultTimeout time.Duration //linux: 600s
 
 	// Used to ensure we don't emit local packets for ips we don't own
-	localIps *CIDRTree
+	localIps *cidr.Tree4
 
 	rules        string
 	rulesVersion uint16
 
-	trackTCPRTT  bool
-	metricTCPRTT metrics.Histogram
+	trackTCPRTT     bool
+	metricTCPRTT    metrics.Histogram
+	incomingMetrics firewallMetrics
+	outgoingMetrics firewallMetrics
+
+	l *logrus.Logger
+}
+
+type firewallMetrics struct {
+	droppedLocalIP  metrics.Counter
+	droppedRemoteIP metrics.Counter
+	droppedNoRule   metrics.Counter
 }
 
 type FirewallConntrack struct {
 	sync.Mutex
 
-	Conns      map[FirewallPacket]*conn
+	Conns      map[firewall.Packet]*conn
 	TimerWheel *TimerWheel
 }
 
@@ -106,57 +107,15 @@ type FirewallRule struct {
 	Any    bool
 	Hosts  map[string]struct{}
 	Groups [][]string
-	CIDR   *CIDRTree
+	CIDR   *cidr.Tree4
 }
 
 // Even though ports are uint16, int32 maps are faster for lookup
 // Plus we can use `-1` for fragment rules
 type firewallPort map[int32]*FirewallCA
 
-type FirewallPacket struct {
-	LocalIP    uint32
-	RemoteIP   uint32
-	LocalPort  uint16
-	RemotePort uint16
-	Protocol   uint8
-	Fragment   bool
-}
-
-func (fp *FirewallPacket) Copy() *FirewallPacket {
-	return &FirewallPacket{
-		LocalIP:    fp.LocalIP,
-		RemoteIP:   fp.RemoteIP,
-		LocalPort:  fp.LocalPort,
-		RemotePort: fp.RemotePort,
-		Protocol:   fp.Protocol,
-		Fragment:   fp.Fragment,
-	}
-}
-
-func (fp FirewallPacket) MarshalJSON() ([]byte, error) {
-	var proto string
-	switch fp.Protocol {
-	case fwProtoTCP:
-		proto = "tcp"
-	case fwProtoICMP:
-		proto = "icmp"
-	case fwProtoUDP:
-		proto = "udp"
-	default:
-		proto = fmt.Sprintf("unknown %v", fp.Protocol)
-	}
-	return json.Marshal(m{
-		"LocalIP":    int2ip(fp.LocalIP).String(),
-		"RemoteIP":   int2ip(fp.RemoteIP).String(),
-		"LocalPort":  fp.LocalPort,
-		"RemotePort": fp.RemotePort,
-		"Protocol":   proto,
-		"Fragment":   fp.Fragment,
-	})
-}
-
 // NewFirewall creates a new Firewall object. A TimerWheel is created for you from the provided timeouts.
-func NewFirewall(tcpTimeout, UDPTimeout, defaultTimeout time.Duration, c *cert.NebulaCertificate) *Firewall {
+func NewFirewall(l *logrus.Logger, tcpTimeout, UDPTimeout, defaultTimeout time.Duration, c *cert.NebulaCertificate) *Firewall {
 	//TODO: error on 0 duration
 	var min, max time.Duration
 
@@ -174,7 +133,7 @@ func NewFirewall(tcpTimeout, UDPTimeout, defaultTimeout time.Duration, c *cert.N
 		max = defaultTimeout
 	}
 
-	localIps := NewCIDRTree()
+	localIps := cidr.NewTree4()
 	for _, ip := range c.Details.Ips {
 		localIps.AddCIDR(&net.IPNet{IP: ip.IP, Mask: net.IPMask{255, 255, 255, 255}}, struct{}{})
 	}
@@ -185,7 +144,7 @@ func NewFirewall(tcpTimeout, UDPTimeout, defaultTimeout time.Duration, c *cert.N
 
 	return &Firewall{
 		Conntrack: &FirewallConntrack{
-			Conns:      make(map[FirewallPacket]*conn),
+			Conns:      make(map[firewall.Packet]*conn),
 			TimerWheel: NewTimerWheel(min, max),
 		},
 		InRules:        newFirewallTable(),
@@ -194,12 +153,25 @@ func NewFirewall(tcpTimeout, UDPTimeout, defaultTimeout time.Duration, c *cert.N
 		UDPTimeout:     UDPTimeout,
 		DefaultTimeout: defaultTimeout,
 		localIps:       localIps,
-		metricTCPRTT:   metrics.GetOrRegisterHistogram("network.tcp.rtt", nil, metrics.NewExpDecaySample(1028, 0.015)),
+		l:              l,
+
+		metricTCPRTT: metrics.GetOrRegisterHistogram("network.tcp.rtt", nil, metrics.NewExpDecaySample(1028, 0.015)),
+		incomingMetrics: firewallMetrics{
+			droppedLocalIP:  metrics.GetOrRegisterCounter("firewall.incoming.dropped.local_ip", nil),
+			droppedRemoteIP: metrics.GetOrRegisterCounter("firewall.incoming.dropped.remote_ip", nil),
+			droppedNoRule:   metrics.GetOrRegisterCounter("firewall.incoming.dropped.no_rule", nil),
+		},
+		outgoingMetrics: firewallMetrics{
+			droppedLocalIP:  metrics.GetOrRegisterCounter("firewall.outgoing.dropped.local_ip", nil),
+			droppedRemoteIP: metrics.GetOrRegisterCounter("firewall.outgoing.dropped.remote_ip", nil),
+			droppedNoRule:   metrics.GetOrRegisterCounter("firewall.outgoing.dropped.no_rule", nil),
+		},
 	}
 }
 
-func NewFirewallFromConfig(nc *cert.NebulaCertificate, c *Config) (*Firewall, error) {
+func NewFirewallFromConfig(l *logrus.Logger, nc *cert.NebulaCertificate, c *config.C) (*Firewall, error) {
 	fw := NewFirewall(
+		l,
 		c.GetDuration("firewall.conntrack.tcp_timeout", time.Minute*12),
 		c.GetDuration("firewall.conntrack.udp_timeout", time.Minute*3),
 		c.GetDuration("firewall.conntrack.default_timeout", time.Minute*10),
@@ -207,12 +179,12 @@ func NewFirewallFromConfig(nc *cert.NebulaCertificate, c *Config) (*Firewall, er
 		//TODO: max_connections
 	)
 
-	err := AddFirewallRulesFromConfig(false, c, fw)
+	err := AddFirewallRulesFromConfig(l, false, c, fw)
 	if err != nil {
 		return nil, err
 	}
 
-	err = AddFirewallRulesFromConfig(true, c, fw)
+	err = AddFirewallRulesFromConfig(l, true, c, fw)
 	if err != nil {
 		return nil, err
 	}
@@ -240,7 +212,7 @@ func (f *Firewall) AddRule(incoming bool, proto uint8, startPort int32, endPort
 	if !incoming {
 		direction = "outgoing"
 	}
-	l.WithField("firewallRule", m{"direction": direction, "proto": proto, "startPort": startPort, "endPort": endPort, "groups": groups, "host": host, "ip": sIp, "caName": caName, "caSha": caSha}).
+	f.l.WithField("firewallRule", m{"direction": direction, "proto": proto, "startPort": startPort, "endPort": endPort, "groups": groups, "host": host, "ip": sIp, "caName": caName, "caSha": caSha}).
 		Info("Firewall rule added")
 
 	var (
@@ -255,13 +227,13 @@ func (f *Firewall) AddRule(incoming bool, proto uint8, startPort int32, endPort
 	}
 
 	switch proto {
-	case fwProtoTCP:
+	case firewall.ProtoTCP:
 		fp = ft.TCP
-	case fwProtoUDP:
+	case firewall.ProtoUDP:
 		fp = ft.UDP
-	case fwProtoICMP:
+	case firewall.ProtoICMP:
 		fp = ft.ICMP
-	case fwProtoAny:
+	case firewall.ProtoAny:
 		fp = ft.AnyProto
 	default:
 		return fmt.Errorf("unknown protocol %v", proto)
@@ -276,7 +248,7 @@ func (f *Firewall) GetRuleHash() string {
 	return hex.EncodeToString(sum[:])
 }
 
-func AddFirewallRulesFromConfig(inbound bool, config *Config, fw FirewallInterface) error {
+func AddFirewallRulesFromConfig(l *logrus.Logger, inbound bool, c *config.C, fw FirewallInterface) error {
 	var table string
 	if inbound {
 		table = "firewall.inbound"
@@ -284,7 +256,7 @@ func AddFirewallRulesFromConfig(inbound bool, config *Config, fw FirewallInterfa
 		table = "firewall.outbound"
 	}
 
-	r := config.Get(table)
+	r := c.Get(table)
 	if r == nil {
 		return nil
 	}
@@ -296,7 +268,7 @@ func AddFirewallRulesFromConfig(inbound bool, config *Config, fw FirewallInterfa
 
 	for i, t := range rs {
 		var groups []string
-		r, err := convertRule(t, table, i)
+		r, err := convertRule(l, t, table, i)
 		if err != nil {
 			return fmt.Errorf("%s rule #%v; %s", table, i, err)
 		}
@@ -339,13 +311,13 @@ func AddFirewallRulesFromConfig(inbound bool, config *Config, fw FirewallInterfa
 		var proto uint8
 		switch r.Proto {
 		case "any":
-			proto = fwProtoAny
+			proto = firewall.ProtoAny
 		case "tcp":
-			proto = fwProtoTCP
+			proto = firewall.ProtoTCP
 		case "udp":
-			proto = fwProtoUDP
+			proto = firewall.ProtoUDP
 		case "icmp":
-			proto = fwProtoICMP
+			proto = firewall.ProtoICMP
 		default:
 			return fmt.Errorf("%s rule #%v; proto was not understood; `%s`", table, i, r.Proto)
 		}
@@ -373,7 +345,7 @@ var ErrNoMatchingRule = errors.New("no matching rule in firewall table")
 
 // Drop returns an error if the packet should be dropped, explaining why. It
 // returns nil if the packet should not be dropped.
-func (f *Firewall) Drop(packet []byte, fp FirewallPacket, incoming bool, h *HostInfo, caPool *cert.NebulaCAPool, localCache ConntrackCache) error {
+func (f *Firewall) Drop(packet []byte, fp firewall.Packet, incoming bool, h *HostInfo, caPool *cert.NebulaCAPool, localCache firewall.ConntrackCache) error {
 	// Check if we spoke to this tuple, if we did then allow this packet
 	if f.inConns(packet, fp, incoming, h, caPool, localCache) {
 		return nil
@@ -382,17 +354,20 @@ func (f *Firewall) Drop(packet []byte, fp FirewallPacket, incoming bool, h *Host
 	// Make sure remote address matches nebula certificate
 	if remoteCidr := h.remoteCidr; remoteCidr != nil {
 		if remoteCidr.Contains(fp.RemoteIP) == nil {
+			f.metrics(incoming).droppedRemoteIP.Inc(1)
 			return ErrInvalidRemoteIP
 		}
 	} else {
 		// Simple case: Certificate has one IP and no subnets
-		if fp.RemoteIP != h.hostId {
+		if fp.RemoteIP != h.vpnIp {
+			f.metrics(incoming).droppedRemoteIP.Inc(1)
 			return ErrInvalidRemoteIP
 		}
 	}
 
 	// Make sure we are supposed to be handling this local ip address
 	if f.localIps.Contains(fp.LocalIP) == nil {
+		f.metrics(incoming).droppedLocalIP.Inc(1)
 		return ErrInvalidLocalIP
 	}
 
@@ -403,6 +378,7 @@ func (f *Firewall) Drop(packet []byte, fp FirewallPacket, incoming bool, h *Host
 
 	// We now know which firewall table to check against
 	if !table.match(fp, incoming, h.ConnectionState.peerCert, caPool) {
+		f.metrics(incoming).droppedNoRule.Inc(1)
 		return ErrNoMatchingRule
 	}
 
@@ -412,6 +388,14 @@ func (f *Firewall) Drop(packet []byte, fp FirewallPacket, incoming bool, h *Host
 	return nil
 }
 
+func (f *Firewall) metrics(incoming bool) firewallMetrics {
+	if incoming {
+		return f.incomingMetrics
+	} else {
+		return f.outgoingMetrics
+	}
+}
+
 // Destroy cleans up any known cyclical references so the object can be free'd my GC. This should be called if a new
 // firewall object is created
 func (f *Firewall) Destroy() {
@@ -427,7 +411,7 @@ func (f *Firewall) EmitStats() {
 	metrics.GetOrRegisterGauge("firewall.rules.version", nil).Update(int64(f.rulesVersion))
 }
 
-func (f *Firewall) inConns(packet []byte, fp FirewallPacket, incoming bool, h *HostInfo, caPool *cert.NebulaCAPool, localCache ConntrackCache) bool {
+func (f *Firewall) inConns(packet []byte, fp firewall.Packet, incoming bool, h *HostInfo, caPool *cert.NebulaCAPool, localCache firewall.ConntrackCache) bool {
 	if localCache != nil {
 		if _, ok := localCache[fp]; ok {
 			return true
@@ -459,8 +443,8 @@ func (f *Firewall) inConns(packet []byte, fp FirewallPacket, incoming bool, h *H
 
 		// We now know which firewall table to check against
 		if !table.match(fp, c.incoming, h.ConnectionState.peerCert, caPool) {
-			if l.Level >= logrus.DebugLevel {
-				h.logger().
+			if f.l.Level >= logrus.DebugLevel {
+				h.logger(f.l).
 					WithField("fwPacket", fp).
 					WithField("incoming", c.incoming).
 					WithField("rulesVersion", f.rulesVersion).
@@ -472,8 +456,8 @@ func (f *Firewall) inConns(packet []byte, fp FirewallPacket, incoming bool, h *H
 			return false
 		}
 
-		if l.Level >= logrus.DebugLevel {
-			h.logger().
+		if f.l.Level >= logrus.DebugLevel {
+			h.logger(f.l).
 				WithField("fwPacket", fp).
 				WithField("incoming", c.incoming).
 				WithField("rulesVersion", f.rulesVersion).
@@ -485,14 +469,14 @@ func (f *Firewall) inConns(packet []byte, fp FirewallPacket, incoming bool, h *H
 	}
 
 	switch fp.Protocol {
-	case fwProtoTCP:
+	case firewall.ProtoTCP:
 		c.Expires = time.Now().Add(f.TCPTimeout)
 		if incoming {
 			f.checkTCPRTT(c, packet)
 		} else {
 			setTCPRTTTracking(c, packet)
 		}
-	case fwProtoUDP:
+	case firewall.ProtoUDP:
 		c.Expires = time.Now().Add(f.UDPTimeout)
 	default:
 		c.Expires = time.Now().Add(f.DefaultTimeout)
@@ -507,17 +491,17 @@ func (f *Firewall) inConns(packet []byte, fp FirewallPacket, incoming bool, h *H
 	return true
 }
 
-func (f *Firewall) addConn(packet []byte, fp FirewallPacket, incoming bool) {
+func (f *Firewall) addConn(packet []byte, fp firewall.Packet, incoming bool) {
 	var timeout time.Duration
 	c := &conn{}
 
 	switch fp.Protocol {
-	case fwProtoTCP:
+	case firewall.ProtoTCP:
 		timeout = f.TCPTimeout
 		if !incoming {
 			setTCPRTTTracking(c, packet)
 		}
-	case fwProtoUDP:
+	case firewall.ProtoUDP:
 		timeout = f.UDPTimeout
 	default:
 		timeout = f.DefaultTimeout
@@ -540,7 +524,7 @@ func (f *Firewall) addConn(packet []byte, fp FirewallPacket, incoming bool) {
 
 // Evict checks if a conntrack entry has expired, if so it is removed, if not it is re-added to the wheel
 // Caller must own the connMutex lock!
-func (f *Firewall) evict(p FirewallPacket) {
+func (f *Firewall) evict(p firewall.Packet) {
 	//TODO: report a stat if the tcp rtt tracking was never resolved?
 	// Are we still tracking this conn?
 	conntrack := f.Conntrack
@@ -561,21 +545,21 @@ func (f *Firewall) evict(p FirewallPacket) {
 	delete(conntrack.Conns, p)
 }
 
-func (ft *FirewallTable) match(p FirewallPacket, incoming bool, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
+func (ft *FirewallTable) match(p firewall.Packet, incoming bool, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
 	if ft.AnyProto.match(p, incoming, c, caPool) {
 		return true
 	}
 
 	switch p.Protocol {
-	case fwProtoTCP:
+	case firewall.ProtoTCP:
 		if ft.TCP.match(p, incoming, c, caPool) {
 			return true
 		}
-	case fwProtoUDP:
+	case firewall.ProtoUDP:
 		if ft.UDP.match(p, incoming, c, caPool) {
 			return true
 		}
-	case fwProtoICMP:
+	case firewall.ProtoICMP:
 		if ft.ICMP.match(p, incoming, c, caPool) {
 			return true
 		}
@@ -605,7 +589,7 @@ func (fp firewallPort) addRule(startPort int32, endPort int32, groups []string,
 	return nil
 }
 
-func (fp firewallPort) match(p FirewallPacket, incoming bool, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
+func (fp firewallPort) match(p firewall.Packet, incoming bool, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
 	// We don't have any allowed ports, bail
 	if fp == nil {
 		return false
@@ -614,7 +598,7 @@ func (fp firewallPort) match(p FirewallPacket, incoming bool, c *cert.NebulaCert
 	var port int32
 
 	if p.Fragment {
-		port = fwPortFragment
+		port = firewall.PortFragment
 	} else if incoming {
 		port = int32(p.LocalPort)
 	} else {
@@ -625,7 +609,7 @@ func (fp firewallPort) match(p FirewallPacket, incoming bool, c *cert.NebulaCert
 		return true
 	}
 
-	return fp[fwPortAny].match(p, c, caPool)
+	return fp[firewall.PortAny].match(p, c, caPool)
 }
 
 func (fc *FirewallCA) addRule(groups []string, host string, ip *net.IPNet, caName, caSha string) error {
@@ -633,7 +617,7 @@ func (fc *FirewallCA) addRule(groups []string, host string, ip *net.IPNet, caNam
 		return &FirewallRule{
 			Hosts:  make(map[string]struct{}),
 			Groups: make([][]string, 0),
-			CIDR:   NewCIDRTree(),
+			CIDR:   cidr.NewTree4(),
 		}
 	}
 
@@ -668,7 +652,7 @@ func (fc *FirewallCA) addRule(groups []string, host string, ip *net.IPNet, caNam
 	return nil
 }
 
-func (fc *FirewallCA) match(p FirewallPacket, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
+func (fc *FirewallCA) match(p firewall.Packet, c *cert.NebulaCertificate, caPool *cert.NebulaCAPool) bool {
 	if fc == nil {
 		return false
 	}
@@ -701,7 +685,7 @@ func (fr *FirewallRule) addRule(groups []string, host string, ip *net.IPNet) err
 		// If it's any we need to wipe out any pre-existing rules to save on memory
 		fr.Groups = make([][]string, 0)
 		fr.Hosts = make(map[string]struct{})
-		fr.CIDR = NewCIDRTree()
+		fr.CIDR = cidr.NewTree4()
 	} else {
 		if len(groups) > 0 {
 			fr.Groups = append(fr.Groups, groups)
@@ -741,7 +725,7 @@ func (fr *FirewallRule) isAny(groups []string, host string, ip *net.IPNet) bool
 	return false
 }
 
-func (fr *FirewallRule) match(p FirewallPacket, c *cert.NebulaCertificate) bool {
+func (fr *FirewallRule) match(p firewall.Packet, c *cert.NebulaCertificate) bool {
 	if fr == nil {
 		return false
 	}
@@ -795,7 +779,7 @@ type rule struct {
 	CASha  string
 }
 
-func convertRule(p interface{}, table string, i int) (rule, error) {
+func convertRule(l *logrus.Logger, p interface{}, table string, i int) (rule, error) {
 	r := rule{}
 
 	m, ok := p.(map[interface{}]interface{})
@@ -850,12 +834,12 @@ func convertRule(p interface{}, table string, i int) (rule, error) {
 
 func parsePort(s string) (startPort, endPort int32, err error) {
 	if s == "any" {
-		startPort = fwPortAny
-		endPort = fwPortAny
+		startPort = firewall.PortAny
+		endPort = firewall.PortAny
 
 	} else if s == "fragment" {
-		startPort = fwPortFragment
-		endPort = fwPortFragment
+		startPort = firewall.PortFragment
+		endPort = firewall.PortFragment
 
 	} else if strings.Contains(s, `-`) {
 		sPorts := strings.SplitN(s, `-`, 2)
@@ -879,8 +863,8 @@ func parsePort(s string) (startPort, endPort int32, err error) {
 		startPort = int32(rStartPort)
 		endPort = int32(rEndPort)
 
-		if startPort == fwPortAny {
-			endPort = fwPortAny
+		if startPort == firewall.PortAny {
+			endPort = firewall.PortAny
 		}
 
 	} else {
@@ -933,54 +917,3 @@ func (f *Firewall) checkTCPRTT(c *conn, p []byte) bool {
 	c.Seq = 0
 	return true
 }
-
-// ConntrackCache is used as a local routine cache to know if a given flow
-// has been seen in the conntrack table.
-type ConntrackCache map[FirewallPacket]struct{}
-
-type ConntrackCacheTicker struct {
-	cacheV    uint64
-	cacheTick uint64
-
-	cache ConntrackCache
-}
-
-func NewConntrackCacheTicker(d time.Duration) *ConntrackCacheTicker {
-	if d == 0 {
-		return nil
-	}
-
-	c := &ConntrackCacheTicker{
-		cache: ConntrackCache{},
-	}
-
-	go c.tick(d)
-
-	return c
-}
-
-func (c *ConntrackCacheTicker) tick(d time.Duration) {
-	for {
-		time.Sleep(d)
-		atomic.AddUint64(&c.cacheTick, 1)
-	}
-}
-
-// Get checks if the cache ticker has moved to the next version before returning
-// the map. If it has moved, we reset the map.
-func (c *ConntrackCacheTicker) Get() ConntrackCache {
-	if c == nil {
-		return nil
-	}
-	if tick := atomic.LoadUint64(&c.cacheTick); tick != c.cacheV {
-		c.cacheV = tick
-		if ll := len(c.cache); ll > 0 {
-			if l.GetLevel() == logrus.DebugLevel {
-				l.WithField("len", ll).Debug("resetting conntrack cache")
-			}
-			c.cache = make(ConntrackCache, ll)
-		}
-	}
-
-	return c.cache
-}

firewall/cache.go

@@ -0,0 +1,59 @@
+package firewall
+
+import (
+	"sync/atomic"
+	"time"
+
+	"github.com/sirupsen/logrus"
+)
+
+// ConntrackCache is used as a local routine cache to know if a given flow
+// has been seen in the conntrack table.
+type ConntrackCache map[Packet]struct{}
+
+type ConntrackCacheTicker struct {
+	cacheV    uint64
+	cacheTick uint64
+
+	cache ConntrackCache
+}
+
+func NewConntrackCacheTicker(d time.Duration) *ConntrackCacheTicker {
+	if d == 0 {
+		return nil
+	}
+
+	c := &ConntrackCacheTicker{
+		cache: ConntrackCache{},
+	}
+
+	go c.tick(d)
+
+	return c
+}
+
+func (c *ConntrackCacheTicker) tick(d time.Duration) {
+	for {
+		time.Sleep(d)
+		atomic.AddUint64(&c.cacheTick, 1)
+	}
+}
+
+// Get checks if the cache ticker has moved to the next version before returning
+// the map. If it has moved, we reset the map.
+func (c *ConntrackCacheTicker) Get(l *logrus.Logger) ConntrackCache {
+	if c == nil {
+		return nil
+	}
+	if tick := atomic.LoadUint64(&c.cacheTick); tick != c.cacheV {
+		c.cacheV = tick
+		if ll := len(c.cache); ll > 0 {
+			if l.Level == logrus.DebugLevel {
+				l.WithField("len", ll).Debug("resetting conntrack cache")
+			}
+			c.cache = make(ConntrackCache, ll)
+		}
+	}
+
+	return c.cache
+}

firewall/packet.go

@@ -0,0 +1,62 @@
+package firewall
+
+import (
+	"encoding/json"
+	"fmt"
+
+	"github.com/slackhq/nebula/iputil"
+)
+
+type m map[string]interface{}
+
+const (
+	ProtoAny  = 0 // When we want to handle HOPOPT (0) we can change this, if ever
+	ProtoTCP  = 6
+	ProtoUDP  = 17
+	ProtoICMP = 1
+
+	PortAny      = 0  // Special value for matching `port: any`
+	PortFragment = -1 // Special value for matching `port: fragment`
+)
+
+type Packet struct {
+	LocalIP    iputil.VpnIp
+	RemoteIP   iputil.VpnIp
+	LocalPort  uint16
+	RemotePort uint16
+	Protocol   uint8
+	Fragment   bool
+}
+
+func (fp *Packet) Copy() *Packet {
+	return &Packet{
+		LocalIP:    fp.LocalIP,
+		RemoteIP:   fp.RemoteIP,
+		LocalPort:  fp.LocalPort,
+		RemotePort: fp.RemotePort,
+		Protocol:   fp.Protocol,
+		Fragment:   fp.Fragment,
+	}
+}
+
+func (fp Packet) MarshalJSON() ([]byte, error) {
+	var proto string
+	switch fp.Protocol {
+	case ProtoTCP:
+		proto = "tcp"
+	case ProtoICMP:
+		proto = "icmp"
+	case ProtoUDP:
+		proto = "udp"
+	default:
+		proto = fmt.Sprintf("unknown %v", fp.Protocol)
+	}
+	return json.Marshal(m{
+		"LocalIP":    fp.LocalIP.String(),
+		"RemoteIP":   fp.RemoteIP.String(),
+		"LocalPort":  fp.LocalPort,
+		"RemotePort": fp.RemotePort,
+		"Protocol":   proto,
+		"Fragment":   fp.Fragment,
+	})
+}

firewall_test.go

@@ -11,12 +11,17 @@ import (
 
 	"github.com/rcrowley/go-metrics"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/firewall"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
 func TestNewFirewall(t *testing.T) {
+	l := util.NewTestLogger()
 	c := &cert.NebulaCertificate{}
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, c)
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, c)
 	conntrack := fw.Conntrack
 	assert.NotNil(t, conntrack)
 	assert.NotNil(t, conntrack.Conns)
@@ -31,127 +36,113 @@ func TestNewFirewall(t *testing.T) {
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 
-	fw = NewFirewall(time.Second, time.Hour, time.Minute, c)
+	fw = NewFirewall(l, time.Second, time.Hour, time.Minute, c)
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 
-	fw = NewFirewall(time.Hour, time.Second, time.Minute, c)
+	fw = NewFirewall(l, time.Hour, time.Second, time.Minute, c)
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 
-	fw = NewFirewall(time.Hour, time.Minute, time.Second, c)
+	fw = NewFirewall(l, time.Hour, time.Minute, time.Second, c)
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 
-	fw = NewFirewall(time.Minute, time.Hour, time.Second, c)
+	fw = NewFirewall(l, time.Minute, time.Hour, time.Second, c)
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 
-	fw = NewFirewall(time.Minute, time.Second, time.Hour, c)
+	fw = NewFirewall(l, time.Minute, time.Second, time.Hour, c)
 	assert.Equal(t, time.Hour, conntrack.TimerWheel.wheelDuration)
 	assert.Equal(t, 3601, conntrack.TimerWheel.wheelLen)
 }
 
 func TestFirewall_AddRule(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
 	c := &cert.NebulaCertificate{}
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, c)
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, c)
 	assert.NotNil(t, fw.InRules)
 	assert.NotNil(t, fw.OutRules)
 
 	_, ti, _ := net.ParseCIDR("1.2.3.4/32")
 
-	assert.Nil(t, fw.AddRule(true, fwProtoTCP, 1, 1, []string{}, "", nil, "", ""))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoTCP, 1, 1, []string{}, "", nil, "", ""))
 	// An empty rule is any
 	assert.True(t, fw.InRules.TCP[1].Any.Any)
 	assert.Empty(t, fw.InRules.TCP[1].Any.Groups)
 	assert.Empty(t, fw.InRules.TCP[1].Any.Hosts)
-	assert.Nil(t, fw.InRules.TCP[1].Any.CIDR.root.left)
-	assert.Nil(t, fw.InRules.TCP[1].Any.CIDR.root.right)
-	assert.Nil(t, fw.InRules.TCP[1].Any.CIDR.root.value)
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(true, fwProtoUDP, 1, 1, []string{"g1"}, "", nil, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoUDP, 1, 1, []string{"g1"}, "", nil, "", ""))
 	assert.False(t, fw.InRules.UDP[1].Any.Any)
 	assert.Contains(t, fw.InRules.UDP[1].Any.Groups[0], "g1")
 	assert.Empty(t, fw.InRules.UDP[1].Any.Hosts)
-	assert.Nil(t, fw.InRules.UDP[1].Any.CIDR.root.left)
-	assert.Nil(t, fw.InRules.UDP[1].Any.CIDR.root.right)
-	assert.Nil(t, fw.InRules.UDP[1].Any.CIDR.root.value)
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(true, fwProtoICMP, 1, 1, []string{}, "h1", nil, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoICMP, 1, 1, []string{}, "h1", nil, "", ""))
 	assert.False(t, fw.InRules.ICMP[1].Any.Any)
 	assert.Empty(t, fw.InRules.ICMP[1].Any.Groups)
 	assert.Contains(t, fw.InRules.ICMP[1].Any.Hosts, "h1")
-	assert.Nil(t, fw.InRules.ICMP[1].Any.CIDR.root.left)
-	assert.Nil(t, fw.InRules.ICMP[1].Any.CIDR.root.right)
-	assert.Nil(t, fw.InRules.ICMP[1].Any.CIDR.root.value)
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 1, 1, []string{}, "", ti, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 1, 1, []string{}, "", ti, "", ""))
 	assert.False(t, fw.OutRules.AnyProto[1].Any.Any)
 	assert.Empty(t, fw.OutRules.AnyProto[1].Any.Groups)
 	assert.Empty(t, fw.OutRules.AnyProto[1].Any.Hosts)
-	assert.NotNil(t, fw.OutRules.AnyProto[1].Any.CIDR.Match(ip2int(ti.IP)))
+	assert.NotNil(t, fw.OutRules.AnyProto[1].Any.CIDR.Match(iputil.Ip2VpnIp(ti.IP)))
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(true, fwProtoUDP, 1, 1, []string{"g1"}, "", nil, "ca-name", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoUDP, 1, 1, []string{"g1"}, "", nil, "ca-name", ""))
 	assert.Contains(t, fw.InRules.UDP[1].CANames, "ca-name")
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(true, fwProtoUDP, 1, 1, []string{"g1"}, "", nil, "", "ca-sha"))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoUDP, 1, 1, []string{"g1"}, "", nil, "", "ca-sha"))
 	assert.Contains(t, fw.InRules.UDP[1].CAShas, "ca-sha")
 
 	// Set any and clear fields
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 0, 0, []string{"g1", "g2"}, "h1", ti, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{"g1", "g2"}, "h1", ti, "", ""))
 	assert.Equal(t, []string{"g1", "g2"}, fw.OutRules.AnyProto[0].Any.Groups[0])
 	assert.Contains(t, fw.OutRules.AnyProto[0].Any.Hosts, "h1")
-	assert.NotNil(t, fw.OutRules.AnyProto[0].Any.CIDR.Match(ip2int(ti.IP)))
+	assert.NotNil(t, fw.OutRules.AnyProto[0].Any.CIDR.Match(iputil.Ip2VpnIp(ti.IP)))
 
 	// run twice just to make sure
 	//TODO: these ANY rules should clear the CA firewall portion
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 0, 0, []string{}, "any", nil, "", ""))
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{}, "any", nil, "", ""))
 	assert.True(t, fw.OutRules.AnyProto[0].Any.Any)
 	assert.Empty(t, fw.OutRules.AnyProto[0].Any.Groups)
 	assert.Empty(t, fw.OutRules.AnyProto[0].Any.Hosts)
-	assert.Nil(t, fw.OutRules.AnyProto[0].Any.CIDR.root.left)
-	assert.Nil(t, fw.OutRules.AnyProto[0].Any.CIDR.root.right)
-	assert.Nil(t, fw.OutRules.AnyProto[0].Any.CIDR.root.value)
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 0, 0, []string{}, "any", nil, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{}, "any", nil, "", ""))
 	assert.True(t, fw.OutRules.AnyProto[0].Any.Any)
 
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
 	_, anyIp, _ := net.ParseCIDR("0.0.0.0/0")
-	assert.Nil(t, fw.AddRule(false, fwProtoAny, 0, 0, []string{}, "", anyIp, "", ""))
+	assert.Nil(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{}, "", anyIp, "", ""))
 	assert.True(t, fw.OutRules.AnyProto[0].Any.Any)
 
 	// Test error conditions
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, c)
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, c)
 	assert.Error(t, fw.AddRule(true, math.MaxUint8, 0, 0, []string{}, "", nil, "", ""))
-	assert.Error(t, fw.AddRule(true, fwProtoAny, 10, 0, []string{}, "", nil, "", ""))
+	assert.Error(t, fw.AddRule(true, firewall.ProtoAny, 10, 0, []string{}, "", nil, "", ""))
 }
 
 func TestFirewall_Drop(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
-	p := FirewallPacket{
-		ip2int(net.IPv4(1, 2, 3, 4)),
-		ip2int(net.IPv4(1, 2, 3, 4)),
+	p := firewall.Packet{
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
 		10,
 		90,
-		fwProtoUDP,
+		firewall.ProtoUDP,
 		false,
 	}
 
@@ -173,12 +164,12 @@ func TestFirewall_Drop(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h.CreateRemoteCIDR(&c)
 
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
 	cp := cert.NewCAPool()
 
 	// Drop outbound
@@ -191,34 +182,34 @@ func TestFirewall_Drop(t *testing.T) {
 
 	// test remote mismatch
 	oldRemote := p.RemoteIP
-	p.RemoteIP = ip2int(net.IPv4(1, 2, 3, 10))
+	p.RemoteIP = iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 10))
 	assert.Equal(t, fw.Drop([]byte{}, p, false, &h, cp, nil), ErrInvalidRemoteIP)
 	p.RemoteIP = oldRemote
 
 	// ensure signer doesn't get in the way of group checks
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"nope"}, "", nil, "", "signer-shasum"))
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"default-group"}, "", nil, "", "signer-shasum-bad"))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"nope"}, "", nil, "", "signer-shasum"))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"default-group"}, "", nil, "", "signer-shasum-bad"))
 	assert.Equal(t, fw.Drop([]byte{}, p, true, &h, cp, nil), ErrNoMatchingRule)
 
 	// test caSha doesn't drop on match
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"nope"}, "", nil, "", "signer-shasum-bad"))
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"default-group"}, "", nil, "", "signer-shasum"))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"nope"}, "", nil, "", "signer-shasum-bad"))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"default-group"}, "", nil, "", "signer-shasum"))
 	assert.NoError(t, fw.Drop([]byte{}, p, true, &h, cp, nil))
 
 	// ensure ca name doesn't get in the way of group checks
 	cp.CAs["signer-shasum"] = &cert.NebulaCertificate{Details: cert.NebulaCertificateDetails{Name: "ca-good"}}
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"nope"}, "", nil, "ca-good", ""))
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"default-group"}, "", nil, "ca-good-bad", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"nope"}, "", nil, "ca-good", ""))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"default-group"}, "", nil, "ca-good-bad", ""))
 	assert.Equal(t, fw.Drop([]byte{}, p, true, &h, cp, nil), ErrNoMatchingRule)
 
 	// test caName doesn't drop on match
 	cp.CAs["signer-shasum"] = &cert.NebulaCertificate{Details: cert.NebulaCertificateDetails{Name: "ca-good"}}
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"nope"}, "", nil, "ca-good-bad", ""))
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"default-group"}, "", nil, "ca-good", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"nope"}, "", nil, "ca-good-bad", ""))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"default-group"}, "", nil, "ca-good", ""))
 	assert.NoError(t, fw.Drop([]byte{}, p, true, &h, cp, nil))
 }
 
@@ -238,14 +229,14 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 	b.Run("fail on proto", func(b *testing.B) {
 		c := &cert.NebulaCertificate{}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoUDP}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoUDP}, true, c, cp)
 		}
 	})
 
 	b.Run("fail on port", func(b *testing.B) {
 		c := &cert.NebulaCertificate{}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 1}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 1}, true, c, cp)
 		}
 	})
 
@@ -259,7 +250,7 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 			},
 		}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 10}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 10}, true, c, cp)
 		}
 	})
 
@@ -271,7 +262,7 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 			},
 		}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 10}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 10}, true, c, cp)
 		}
 	})
 
@@ -283,12 +274,12 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 			},
 		}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 10}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 10}, true, c, cp)
 		}
 	})
 
 	b.Run("pass on ip", func(b *testing.B) {
-		ip := ip2int(net.IPv4(172, 1, 1, 1))
+		ip := iputil.Ip2VpnIp(net.IPv4(172, 1, 1, 1))
 		c := &cert.NebulaCertificate{
 			Details: cert.NebulaCertificateDetails{
 				InvertedGroups: map[string]struct{}{"nope": {}},
@@ -296,14 +287,14 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 			},
 		}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 10, RemoteIP: ip}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 10, RemoteIP: ip}, true, c, cp)
 		}
 	})
 
 	_ = ft.TCP.addRule(0, 0, []string{"good-group"}, "good-host", n, "", "")
 
 	b.Run("pass on ip with any port", func(b *testing.B) {
-		ip := ip2int(net.IPv4(172, 1, 1, 1))
+		ip := iputil.Ip2VpnIp(net.IPv4(172, 1, 1, 1))
 		c := &cert.NebulaCertificate{
 			Details: cert.NebulaCertificateDetails{
 				InvertedGroups: map[string]struct{}{"nope": {}},
@@ -311,23 +302,22 @@ func BenchmarkFirewallTable_match(b *testing.B) {
 			},
 		}
 		for n := 0; n < b.N; n++ {
-			ft.match(FirewallPacket{Protocol: fwProtoTCP, LocalPort: 100, RemoteIP: ip}, true, c, cp)
+			ft.match(firewall.Packet{Protocol: firewall.ProtoTCP, LocalPort: 100, RemoteIP: ip}, true, c, cp)
 		}
 	})
 }
 
 func TestFirewall_Drop2(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
-	p := FirewallPacket{
-		ip2int(net.IPv4(1, 2, 3, 4)),
-		ip2int(net.IPv4(1, 2, 3, 4)),
+	p := firewall.Packet{
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
 		10,
 		90,
-		fwProtoUDP,
+		firewall.ProtoUDP,
 		false,
 	}
 
@@ -347,7 +337,7 @@ func TestFirewall_Drop2(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h.CreateRemoteCIDR(&c)
 
@@ -365,8 +355,8 @@ func TestFirewall_Drop2(t *testing.T) {
 	}
 	h1.CreateRemoteCIDR(&c1)
 
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"default-group", "test-group"}, "", nil, "", ""))
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"default-group", "test-group"}, "", nil, "", ""))
 	cp := cert.NewCAPool()
 
 	// h1/c1 lacks the proper groups
@@ -377,17 +367,16 @@ func TestFirewall_Drop2(t *testing.T) {
 }
 
 func TestFirewall_Drop3(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
-	p := FirewallPacket{
-		ip2int(net.IPv4(1, 2, 3, 4)),
-		ip2int(net.IPv4(1, 2, 3, 4)),
+	p := firewall.Packet{
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
 		1,
 		1,
-		fwProtoUDP,
+		firewall.ProtoUDP,
 		false,
 	}
 
@@ -414,7 +403,7 @@ func TestFirewall_Drop3(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c1,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h1.CreateRemoteCIDR(&c1)
 
@@ -429,7 +418,7 @@ func TestFirewall_Drop3(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c2,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h2.CreateRemoteCIDR(&c2)
 
@@ -444,13 +433,13 @@ func TestFirewall_Drop3(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c3,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h3.CreateRemoteCIDR(&c3)
 
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 1, 1, []string{}, "host1", nil, "", ""))
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 1, 1, []string{}, "", nil, "", "signer-sha"))
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 1, 1, []string{}, "host1", nil, "", ""))
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 1, 1, []string{}, "", nil, "", "signer-sha"))
 	cp := cert.NewCAPool()
 
 	// c1 should pass because host match
@@ -464,17 +453,16 @@ func TestFirewall_Drop3(t *testing.T) {
 }
 
 func TestFirewall_DropConntrackReload(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
-	p := FirewallPacket{
-		ip2int(net.IPv4(1, 2, 3, 4)),
-		ip2int(net.IPv4(1, 2, 3, 4)),
+	p := firewall.Packet{
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
+		iputil.Ip2VpnIp(net.IPv4(1, 2, 3, 4)),
 		10,
 		90,
-		fwProtoUDP,
+		firewall.ProtoUDP,
 		false,
 	}
 
@@ -496,12 +484,12 @@ func TestFirewall_DropConntrackReload(t *testing.T) {
 		ConnectionState: &ConnectionState{
 			peerCert: &c,
 		},
-		hostId: ip2int(ipNet.IP),
+		vpnIp: iputil.Ip2VpnIp(ipNet.IP),
 	}
 	h.CreateRemoteCIDR(&c)
 
-	fw := NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
+	fw := NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"any"}, "", nil, "", ""))
 	cp := cert.NewCAPool()
 
 	// Drop outbound
@@ -513,8 +501,8 @@ func TestFirewall_DropConntrackReload(t *testing.T) {
 	assert.NoError(t, fw.Drop([]byte{}, p, false, &h, cp, nil))
 
 	oldFw := fw
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 10, 10, []string{"any"}, "", nil, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 10, 10, []string{"any"}, "", nil, "", ""))
 	fw.Conntrack = oldFw.Conntrack
 	fw.rulesVersion = oldFw.rulesVersion + 1
 
@@ -522,8 +510,8 @@ func TestFirewall_DropConntrackReload(t *testing.T) {
 	assert.NoError(t, fw.Drop([]byte{}, p, false, &h, cp, nil))
 
 	oldFw = fw
-	fw = NewFirewall(time.Second, time.Minute, time.Hour, &c)
-	assert.Nil(t, fw.AddRule(true, fwProtoAny, 11, 11, []string{"any"}, "", nil, "", ""))
+	fw = NewFirewall(l, time.Second, time.Minute, time.Hour, &c)
+	assert.Nil(t, fw.AddRule(true, firewall.ProtoAny, 11, 11, []string{"any"}, "", nil, "", ""))
 	fw.Conntrack = oldFw.Conntrack
 	fw.rulesVersion = oldFw.rulesVersion + 1
 
@@ -647,124 +635,126 @@ func Test_parsePort(t *testing.T) {
 }
 
 func TestNewFirewallFromConfig(t *testing.T) {
+	l := util.NewTestLogger()
 	// Test a bad rule definition
 	c := &cert.NebulaCertificate{}
-	conf := NewConfig()
+	conf := config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": "asdf"}
-	_, err := NewFirewallFromConfig(c, conf)
+	_, err := NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound failed to parse, should be an array of rules")
 
 	// Test both port and code
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"port": "1", "code": "2"}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; only one of port or code should be provided")
 
 	// Test missing host, group, cidr, ca_name and ca_sha
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; at least one of host, group, cidr, ca_name, or ca_sha must be provided")
 
 	// Test code/port error
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"code": "a", "host": "testh"}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; code was not a number; `a`")
 
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"port": "a", "host": "testh"}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; port was not a number; `a`")
 
 	// Test proto error
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"code": "1", "host": "testh"}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; proto was not understood; ``")
 
 	// Test cidr parse error
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"code": "1", "cidr": "testh", "proto": "any"}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.outbound rule #0; cidr did not parse; invalid CIDR address: testh")
 
 	// Test both group and groups
-	conf = NewConfig()
+	conf = config.NewC(l)
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "group": "a", "groups": []string{"b", "c"}}}}
-	_, err = NewFirewallFromConfig(c, conf)
+	_, err = NewFirewallFromConfig(l, c, conf)
 	assert.EqualError(t, err, "firewall.inbound rule #0; only one of group or groups should be defined, both provided")
 }
 
 func TestAddFirewallRulesFromConfig(t *testing.T) {
+	l := util.NewTestLogger()
 	// Test adding tcp rule
-	conf := NewConfig()
+	conf := config.NewC(l)
 	mf := &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "tcp", "host": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(false, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: false, proto: fwProtoTCP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, false, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: false, proto: firewall.ProtoTCP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
 
 	// Test adding udp rule
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "udp", "host": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(false, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: false, proto: fwProtoUDP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, false, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: false, proto: firewall.ProtoUDP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
 
 	// Test adding icmp rule
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"outbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "icmp", "host": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(false, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: false, proto: fwProtoICMP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, false, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: false, proto: firewall.ProtoICMP, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
 
 	// Test adding any rule
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "host": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: nil, host: "a", ip: nil}, mf.lastCall)
 
 	// Test adding rule with ca_sha
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "ca_sha": "12312313123"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: nil, ip: nil, caSha: "12312313123"}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: nil, ip: nil, caSha: "12312313123"}, mf.lastCall)
 
 	// Test adding rule with ca_name
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "ca_name": "root01"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: nil, ip: nil, caName: "root01"}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: nil, ip: nil, caName: "root01"}, mf.lastCall)
 
 	// Test single group
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "group": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: []string{"a"}, ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: []string{"a"}, ip: nil}, mf.lastCall)
 
 	// Test single groups
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "groups": "a"}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: []string{"a"}, ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: []string{"a"}, ip: nil}, mf.lastCall)
 
 	// Test multiple AND groups
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "groups": []string{"a", "b"}}}}
-	assert.Nil(t, AddFirewallRulesFromConfig(true, conf, mf))
-	assert.Equal(t, addRuleCall{incoming: true, proto: fwProtoAny, startPort: 1, endPort: 1, groups: []string{"a", "b"}, ip: nil}, mf.lastCall)
+	assert.Nil(t, AddFirewallRulesFromConfig(l, true, conf, mf))
+	assert.Equal(t, addRuleCall{incoming: true, proto: firewall.ProtoAny, startPort: 1, endPort: 1, groups: []string{"a", "b"}, ip: nil}, mf.lastCall)
 
 	// Test Add error
-	conf = NewConfig()
+	conf = config.NewC(l)
 	mf = &mockFirewall{}
 	mf.nextCallReturn = errors.New("test error")
 	conf.Settings["firewall"] = map[interface{}]interface{}{"inbound": []interface{}{map[interface{}]interface{}{"port": "1", "proto": "any", "host": "a"}}}
-	assert.EqualError(t, AddFirewallRulesFromConfig(true, conf, mf), "firewall.inbound rule #0; `test error`")
+	assert.EqualError(t, AddFirewallRulesFromConfig(l, true, conf, mf), "firewall.inbound rule #0; `test error`")
 }
 
 func TestTCPRTTTracking(t *testing.T) {
@@ -859,17 +849,16 @@ func TestTCPRTTTracking(t *testing.T) {
 }
 
 func TestFirewall_convertRule(t *testing.T) {
+	l := util.NewTestLogger()
 	ob := &bytes.Buffer{}
-	out := l.Out
 	l.SetOutput(ob)
-	defer l.SetOutput(out)
 
 	// Ensure group array of 1 is converted and a warning is printed
 	c := map[interface{}]interface{}{
 		"group": []interface{}{"group1"},
 	}
 
-	r, err := convertRule(c, "test", 1)
+	r, err := convertRule(l, c, "test", 1)
 	assert.Contains(t, ob.String(), "test rule #1; group was an array with a single value, converting to simple value")
 	assert.Nil(t, err)
 	assert.Equal(t, "group1", r.Group)
@@ -880,7 +869,7 @@ func TestFirewall_convertRule(t *testing.T) {
 		"group": []interface{}{"group1", "group2"},
 	}
 
-	r, err = convertRule(c, "test", 1)
+	r, err = convertRule(l, c, "test", 1)
 	assert.Equal(t, "", ob.String())
 	assert.Error(t, err, "group should contain a single value, an array with more than one entry was provided")
 
@@ -890,7 +879,7 @@ func TestFirewall_convertRule(t *testing.T) {
 		"group": "group1",
 	}
 
-	r, err = convertRule(c, "test", 1)
+	r, err = convertRule(l, c, "test", 1)
 	assert.Nil(t, err)
 	assert.Equal(t, "group1", r.Group)
 }
@@ -932,6 +921,6 @@ func (mf *mockFirewall) AddRule(incoming bool, proto uint8, startPort int32, end
 
 func resetConntrack(fw *Firewall) {
 	fw.Conntrack.Lock()
-	fw.Conntrack.Conns = map[FirewallPacket]*conn{}
+	fw.Conntrack.Conns = map[firewall.Packet]*conn{}
 	fw.Conntrack.Unlock()
 }

go.mod

@@ -1,34 +1,45 @@
 module github.com/slackhq/nebula
 
-go 1.16
+go 1.17
 
 require (
-	github.com/anmitsu/go-shlex v0.0.0-20161002113705-648efa622239
+	github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be
 	github.com/armon/go-radix v1.0.0
-	github.com/cespare/xxhash/v2 v2.1.1 // indirect
+	github.com/cespare/xxhash/v2 v2.1.2 // indirect
 	github.com/cyberdelia/go-metrics-graphite v0.0.0-20161219230853-39f87cc3b432
-	github.com/flynn/go-shlex v0.0.0-20150515145356-3f9db97f8568 // indirect
-	github.com/flynn/noise v0.0.0-20180327030543-2492fe189ae6
-	github.com/golang/protobuf v1.3.2
+	github.com/flynn/noise v1.0.0
+	github.com/gogo/protobuf v1.3.2
+	github.com/golang/protobuf v1.5.2
+	github.com/google/gopacket v1.1.19
 	github.com/imdario/mergo v0.3.8
-	github.com/kardianos/service v1.1.0
-	github.com/konsorten/go-windows-terminal-sequences v1.0.2 // indirect
-	github.com/kr/pretty v0.1.0 // indirect
-	github.com/miekg/dns v1.1.25
-	github.com/nbrownus/go-metrics-prometheus v0.0.0-20180622211546-6e6d5173d99c
-	github.com/prometheus/client_golang v1.2.1
-	github.com/prometheus/client_model v0.0.0-20191202183732-d1d2010b5bee // indirect
-	github.com/prometheus/procfs v0.0.8 // indirect
-	github.com/rcrowley/go-metrics v0.0.0-20190826022208-cac0b30c2563
-	github.com/sirupsen/logrus v1.4.2
+	github.com/kardianos/service v1.2.0
+	github.com/miekg/dns v1.1.43
+	github.com/nbrownus/go-metrics-prometheus v0.0.0-20210712211119-974a6260965f
+	github.com/prometheus/client_golang v1.11.0
+	github.com/prometheus/client_model v0.2.0 // indirect
+	github.com/prometheus/procfs v0.7.3 // indirect
+	github.com/rcrowley/go-metrics v0.0.0-20201227073835-cf1acfcdf475
+	github.com/sirupsen/logrus v1.8.1
 	github.com/skip2/go-qrcode v0.0.0-20200617195104-da1b6568686e
-	github.com/songgao/water v0.0.0-20190725173103-fd331bda3f4b
-	github.com/stretchr/testify v1.6.1
-	github.com/vishvananda/netlink v1.0.1-0.20190522153524-00009fb8606a
-	github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df // indirect
-	golang.org/x/crypto v0.0.0-20200220183623-bac4c82f6975
-	golang.org/x/net v0.0.0-20191209160850-c0dbc17a3553
-	golang.org/x/sys v0.0.0-20191210023423-ac6580df4449
-	gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 // indirect
-	gopkg.in/yaml.v2 v2.2.7
+	github.com/songgao/water v0.0.0-20200317203138-2b4b6d7c09d8
+	github.com/stretchr/testify v1.7.0
+	github.com/vishvananda/netlink v1.1.0
+	github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74 // indirect
+	golang.org/x/crypto v0.0.0-20210921155107-089bfa567519
+	golang.org/x/net v0.0.0-20211101193420-4a448f8816b3
+	golang.org/x/sys v0.0.0-20211103235746-7861aae1554b
+	golang.zx2c4.com/wintun v0.0.0-20211104114900-415007cec224
+	golang.zx2c4.com/wireguard/windows v0.5.1
+	google.golang.org/protobuf v1.27.1
+	gopkg.in/yaml.v2 v2.4.0
+)
+
+require (
+	github.com/beorn7/perks v1.0.1 // indirect
+	github.com/davecgh/go-spew v1.1.1 // indirect
+	github.com/matttproud/golang_protobuf_extensions v1.0.1 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
+	github.com/prometheus/common v0.32.1 // indirect
+	golang.org/x/term v0.0.0-20210927222741-03fcf44c2211 // indirect
+	gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c // indirect
 )

go.sum

@@ -1,160 +1,544 @@
+cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go v0.34.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go v0.38.0/go.mod h1:990N+gfupTy94rShfmMCWGDn0LpTmnzTp2qbd1dvSRU=
+cloud.google.com/go v0.44.1/go.mod h1:iSa0KzasP4Uvy3f1mN/7PiObzGgflwredwwASm/v6AU=
+cloud.google.com/go v0.44.2/go.mod h1:60680Gw3Yr4ikxnPRS/oxxkBccT6SA1yMk63TGekxKY=
+cloud.google.com/go v0.45.1/go.mod h1:RpBamKRgapWJb87xiFSdk4g1CME7QZg3uwTez+TSTjc=
+cloud.google.com/go v0.46.3/go.mod h1:a6bKKbmY7er1mI7TEI4lsAkts/mkhTSZK8w33B4RAg0=
+cloud.google.com/go v0.50.0/go.mod h1:r9sluTvynVuxRIOHXQEHMFffphuXHOMZMycpNR5e6To=
+cloud.google.com/go v0.52.0/go.mod h1:pXajvRH/6o3+F9jDHZWQ5PbGhn+o8w9qiu/CffaVdO4=
+cloud.google.com/go v0.53.0/go.mod h1:fp/UouUEsRkN6ryDKNW/Upv/JBKnv6WDthjR6+vze6M=
+cloud.google.com/go v0.54.0/go.mod h1:1rq2OEkV3YMf6n/9ZvGWI3GWw0VoqH/1x2nd8Is/bPc=
+cloud.google.com/go v0.56.0/go.mod h1:jr7tqZxxKOVYizybht9+26Z/gUq7tiRzu+ACVAMbKVk=
+cloud.google.com/go v0.57.0/go.mod h1:oXiQ6Rzq3RAkkY7N6t3TcE6jE+CIBBbA36lwQ1JyzZs=
+cloud.google.com/go v0.62.0/go.mod h1:jmCYTdRCQuc1PHIIJ/maLInMho30T/Y0M4hTdTShOYc=
+cloud.google.com/go v0.65.0/go.mod h1:O5N8zS7uWy9vkA9vayVHs65eM1ubvY4h553ofrNHObY=
+cloud.google.com/go/bigquery v1.0.1/go.mod h1:i/xbL2UlR5RvWAURpBYZTtm/cXjCha9lbfbpx4poX+o=
+cloud.google.com/go/bigquery v1.3.0/go.mod h1:PjpwJnslEMmckchkHFfq+HTD2DmtT67aNFKH1/VBDHE=
+cloud.google.com/go/bigquery v1.4.0/go.mod h1:S8dzgnTigyfTmLBfrtrhyYhwRxG72rYxvftPBK2Dvzc=
+cloud.google.com/go/bigquery v1.5.0/go.mod h1:snEHRnqQbz117VIFhE8bmtwIDY80NLUZUMb4Nv6dBIg=
+cloud.google.com/go/bigquery v1.7.0/go.mod h1://okPTzCYNXSlb24MZs83e2Do+h+VXtc4gLoIoXIAPc=
+cloud.google.com/go/bigquery v1.8.0/go.mod h1:J5hqkt3O0uAFnINi6JXValWIb1v0goeZM77hZzJN/fQ=
+cloud.google.com/go/datastore v1.0.0/go.mod h1:LXYbyblFSglQ5pkeyhO+Qmw7ukd3C+pD7TKLgZqpHYE=
+cloud.google.com/go/datastore v1.1.0/go.mod h1:umbIZjpQpHh4hmRpGhH4tLFup+FVzqBi1b3c64qFpCk=
+cloud.google.com/go/pubsub v1.0.1/go.mod h1:R0Gpsv3s54REJCy4fxDixWD93lHJMoZTyQ2kNxGRt3I=
+cloud.google.com/go/pubsub v1.1.0/go.mod h1:EwwdRX2sKPjnvnqCa270oGRyludottCI76h+R3AArQw=
+cloud.google.com/go/pubsub v1.2.0/go.mod h1:jhfEVHT8odbXTkndysNHCcx0awwzvfOlguIAii9o8iA=
+cloud.google.com/go/pubsub v1.3.1/go.mod h1:i+ucay31+CNRpDW4Lu78I4xXG+O1r/MAHgjpRVR+TSU=
+cloud.google.com/go/storage v1.0.0/go.mod h1:IhtSnM/ZTZV8YYJWCY8RULGVqBDmpoyjwiyrjsg+URw=
+cloud.google.com/go/storage v1.5.0/go.mod h1:tpKbwo567HUNpVclU5sGELwQWBDZ8gh0ZeosJ0Rtdos=
+cloud.google.com/go/storage v1.6.0/go.mod h1:N7U0C8pVQ/+NIKOBQyamJIeKQKkZ+mxpohlUTyfDhBk=
+cloud.google.com/go/storage v1.8.0/go.mod h1:Wv1Oy7z6Yz3DshWRJFhqM/UCfaWIRTdp0RXyy7KQOVs=
+cloud.google.com/go/storage v1.10.0/go.mod h1:FLPqc6j+Ki4BU591ie1oL6qBQGu2Bl/tZ9ullr3+Kg0=
+dmitri.shuralyov.com/gpu/mtl v0.0.0-20190408044501-666a987793e9/go.mod h1:H6x//7gZCb22OMCxBHrMx7a5I7Hp++hsVxbQ4BYO7hU=
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
+github.com/BurntSushi/xgb v0.0.0-20160522181843-27f122750802/go.mod h1:IVnqGOEym/WlBOVXweHU+Q+/VP0lqqI8lqeDx9IjBqo=
 github.com/alecthomas/template v0.0.0-20160405071501-a0175ee3bccc/go.mod h1:LOuyumcjzFXgccqObfd/Ljyb9UuFJ6TxHnclSeseNhc=
 github.com/alecthomas/template v0.0.0-20190718012654-fb15b899a751/go.mod h1:LOuyumcjzFXgccqObfd/Ljyb9UuFJ6TxHnclSeseNhc=
 github.com/alecthomas/units v0.0.0-20151022065526-2efee857e7cf/go.mod h1:ybxpYRFXyAe+OPACYpWeL0wqObRcbAqCMya13uyzqw0=
 github.com/alecthomas/units v0.0.0-20190717042225-c3de453c63f4/go.mod h1:ybxpYRFXyAe+OPACYpWeL0wqObRcbAqCMya13uyzqw0=
-github.com/anmitsu/go-shlex v0.0.0-20161002113705-648efa622239 h1:kFOfPq6dUM1hTo4JG6LR5AXSUEsOjtdm0kw0FtQtMJA=
-github.com/anmitsu/go-shlex v0.0.0-20161002113705-648efa622239/go.mod h1:2FmKhYUyUczH0OGQWaF5ceTx0UBShxjsH6f8oGKYe2c=
+github.com/alecthomas/units v0.0.0-20190924025748-f65c72e2690d/go.mod h1:rBZYJk541a8SKzHPHnH3zbiI+7dagKZ0cgpgrD7Fyho=
+github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be h1:9AeTilPcZAjCFIImctFaOjnTIavg87rW78vTPkQqLI8=
+github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be/go.mod h1:ySMOLuWl6zY27l47sB3qLNK6tF2fkHG55UZxx8oIVo4=
 github.com/armon/go-radix v1.0.0 h1:F4z6KzEeeQIMeLFa97iZU6vupzoecKdU5TX24SNppXI=
 github.com/armon/go-radix v1.0.0/go.mod h1:ufUuZ+zHj4x4TnLV4JWEpy2hxWSpsRywHrMgIH9cCH8=
 github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973/go.mod h1:Dwedo/Wpr24TaqPxmxbtue+5NUziq4I4S80YR8gNf3Q=
-github.com/beorn7/perks v1.0.0 h1:HWo1m869IqiPhD389kmkxeTalrjNbbJTC8LXupb+sl0=
 github.com/beorn7/perks v1.0.0/go.mod h1:KWe93zE9D1o94FZ5RNwFwVgaQK1VOXiVxmqh+CedLV8=
 github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
 github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
-github.com/cespare/xxhash/v2 v2.1.0 h1:yTUvW7Vhb89inJ+8irsUqiWjh8iT6sQPZiQzI6ReGkA=
-github.com/cespare/xxhash/v2 v2.1.0/go.mod h1:dgIUBU3pDso/gPgZ1osOZ0iQf77oPR28Tjxl5dIMyVM=
-github.com/cespare/xxhash/v2 v2.1.1 h1:6MnRN8NT7+YBpUIWxHtefFZOKTAPgGjpQSxqLNn0+qY=
+github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
 github.com/cespare/xxhash/v2 v2.1.1/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/cespare/xxhash/v2 v2.1.2 h1:YRXhKfTDauu4ajMg1TPgFO5jnlC2HCbmLXMcTG5cbYE=
+github.com/cespare/xxhash/v2 v2.1.2/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/chzyer/logex v1.1.10/go.mod h1:+Ywpsq7O8HXn0nuIou7OrIPyXbp3wmkHB+jjWRnGsAI=
+github.com/chzyer/readline v0.0.0-20180603132655-2972be24d48e/go.mod h1:nSuG5e5PlCu98SY8svDHJxuZscDgtXS6KTTbou5AhLI=
+github.com/chzyer/test v0.0.0-20180213035817-a1ea475d72b1/go.mod h1:Q3SI9o4m/ZMnBNeIyt5eFwwo7qiLfzFZmjNmxjkiQlU=
+github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
+github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
 github.com/cyberdelia/go-metrics-graphite v0.0.0-20161219230853-39f87cc3b432 h1:M5QgkYacWj0Xs8MhpIK/5uwU02icXpEoSo9sM2aRCps=
 github.com/cyberdelia/go-metrics-graphite v0.0.0-20161219230853-39f87cc3b432/go.mod h1:xwIwAxMvYnVrGJPe2FKx5prTrnAjGOD8zvDOnxnrrkM=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/flynn/go-shlex v0.0.0-20150515145356-3f9db97f8568 h1:BHsljHzVlRcyQhjrss6TZTdY2VfCqZPbv5k3iBFa2ZQ=
-github.com/flynn/go-shlex v0.0.0-20150515145356-3f9db97f8568/go.mod h1:xEzjJPgXI435gkrCt3MPfRiAkVrwSbHsst4LCFVfpJc=
-github.com/flynn/noise v0.0.0-20180327030543-2492fe189ae6 h1:u/UEqS66A5ckRmS4yNpjmVH56sVtS/RfclBAYocb4as=
-github.com/flynn/noise v0.0.0-20180327030543-2492fe189ae6/go.mod h1:1i71OnUq3iUe1ma7Lr6yG6/rjvM3emb6yoL7xLFzcVQ=
+github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
+github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
+github.com/flynn/noise v1.0.0 h1:DlTHqmzmvcEiKj+4RYo/imoswx/4r6iBlCMfVtrMXpQ=
+github.com/flynn/noise v1.0.0/go.mod h1:xbMo+0i6+IGbYdJhF31t2eR1BIU0CYc12+BNAKwUTag=
+github.com/go-gl/glfw v0.0.0-20190409004039-e6da0acd62b1/go.mod h1:vR7hzQXu2zJy9AVAgeJqvqgH9Q5CA+iKCZ2gyEVpxRU=
+github.com/go-gl/glfw/v3.3/glfw v0.0.0-20191125211704-12ad95a8df72/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
+github.com/go-gl/glfw/v3.3/glfw v0.0.0-20200222043503-6f7a984d4dc4/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
 github.com/go-kit/kit v0.8.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
 github.com/go-kit/kit v0.9.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
+github.com/go-kit/log v0.1.0/go.mod h1:zbhenjAZHb184qTLMA9ZjW7ThYL0H2mk7Q6pNt4vbaY=
 github.com/go-logfmt/logfmt v0.3.0/go.mod h1:Qt1PoO58o5twSAckw1HlFXLmHsOX5/0LbT9GBnD5lWE=
 github.com/go-logfmt/logfmt v0.4.0/go.mod h1:3RMwSq7FuexP4Kalkev3ejPJsZTpXXBr9+V4qmtdjCk=
+github.com/go-logfmt/logfmt v0.5.0/go.mod h1:wCYkCAKZfumFQihp8CzCvQ3paCTfi41vtzG1KdI/P7A=
 github.com/go-stack/stack v1.8.0/go.mod h1:v0f6uXyyMGvRgIKkXu+yp6POWl0qKG85gN/melR3HDY=
 github.com/gogo/protobuf v1.1.1/go.mod h1:r8qH/GZQm5c6nD/R0oafs1akxWv10x8SbQlK7atdtwQ=
+github.com/gogo/protobuf v1.3.2 h1:Ov1cvc58UF3b5XjBnZv7+opcTcQFZebYjWzi34vdm4Q=
+github.com/gogo/protobuf v1.3.2/go.mod h1:P1XiOD3dCwIKUDQYPy72D8LYyHL2YPYrpS2s69NZV8Q=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
+github.com/golang/groupcache v0.0.0-20190702054246-869f871628b6/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20191227052852-215e87163ea7/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.2.0/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.3.1/go.mod h1:sBzyDLLjw3U8JLTeZvSv8jJB+tU5PVekmnlKIyFUx0Y=
+github.com/golang/mock v1.4.0/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.1/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.3/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.4/go.mod h1:l3mdAwkq5BuhzHwde/uurv3sEJeZMXNpwsxVWU71h+4=
 github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
-github.com/golang/protobuf v1.3.1 h1:YF8+flBXS5eO826T4nzqPrxfhQThhXl0YzfuUPu4SBg=
 github.com/golang/protobuf v1.3.1/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
-github.com/golang/protobuf v1.3.2 h1:6nsPYzhq5kReh6QImI3k5qWzO4PEbvbIW2cwSfR/6xs=
 github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.3/go.mod h1:vzj43D7+SQXF/4pzW/hwtAqwc6iTitCiVSaWz5lYuqw=
+github.com/golang/protobuf v1.3.4/go.mod h1:vzj43D7+SQXF/4pzW/hwtAqwc6iTitCiVSaWz5lYuqw=
+github.com/golang/protobuf v1.3.5/go.mod h1:6O5/vntMXwX2lRkT1hjjk0nAC1IDOTvTlVgjlRvqsdk=
+github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
+github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
+github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
+github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
+github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
+github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
+github.com/golang/protobuf v1.4.2/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/golang/protobuf v1.5.0/go.mod h1:FsONVRAS9T7sI+LIUmWTfcYkHO4aIWwzhcaSAoJOfIk=
+github.com/golang/protobuf v1.5.2 h1:ROPKBNFfQgOUMifHyP+KYbvpjbdoFNs+aK7DXlji0Tw=
+github.com/golang/protobuf v1.5.2/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiuN0vRsmY=
+github.com/google/btree v0.0.0-20180813153112-4030bb1f1f0c/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ=
+github.com/google/btree v1.0.0/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ=
+github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
 github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
-github.com/google/go-cmp v0.3.1 h1:Xye71clBPdm5HgqGwUkwhbynsUJZhDbS20FvLhQ2izg=
 github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.4.1/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.1/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.4/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.5 h1:Khx7svrCpmxxtHBq5j2mp/xVjsi8hQMfNLvJFAlrGgU=
+github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
+github.com/google/gopacket v1.1.19 h1:ves8RnFZPGiFnTS0uPQStjwru6uO6h+nlr9j6fL7kF8=
+github.com/google/gopacket v1.1.19/go.mod h1:iJ8V8n6KS+z2U1A8pUwu8bW5SyEMkXJB8Yo/Vo+TKTo=
+github.com/google/martian v2.1.0+incompatible/go.mod h1:9I4somxYTbIHy5NJKHRl3wXiIaQGbYVAs8BPL6v8lEs=
+github.com/google/martian/v3 v3.0.0/go.mod h1:y5Zk1BBys9G+gd6Jrk0W3cC1+ELVxBWuIGO+w/tUAp0=
+github.com/google/pprof v0.0.0-20181206194817-3ea8567a2e57/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
+github.com/google/pprof v0.0.0-20190515194954-54271f7e092f/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
+github.com/google/pprof v0.0.0-20191218002539-d4f498aebedc/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200212024743-f11f1df84d12/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200229191704-1ebb73c60ed3/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200430221834-fc25d7d30c6d/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200708004538-1a94d8640e99/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/renameio v0.1.0/go.mod h1:KWCgfxg9yswjAJkECMjeO8J8rahYeXnNhOm40UhjYkI=
+github.com/googleapis/gax-go/v2 v2.0.4/go.mod h1:0Wqv26UfaUD9n4G6kQubkQ+KchISgw+vpHVxEJEs9eg=
+github.com/googleapis/gax-go/v2 v2.0.5/go.mod h1:DWXyrwAJ9X0FpwwEdw+IPEYBICEFu5mhpdKc/us6bOk=
+github.com/hashicorp/golang-lru v0.5.0/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
+github.com/hashicorp/golang-lru v0.5.1/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
+github.com/ianlancetaylor/demangle v0.0.0-20181102032728-5e5cf60278f6/go.mod h1:aSSvb/t6k1mPoxDqO4vJh6VOCGPwU4O0C2/Eqndh1Sc=
 github.com/imdario/mergo v0.3.8 h1:CGgOkSJeqMRmt0D9XLWExdT4m4F1vd3FV3VPt+0VxkQ=
 github.com/imdario/mergo v0.3.8/go.mod h1:2EnlNZ0deacrJVfApfmtdGgDfMuh/nq6Ok1EcJh5FfA=
+github.com/jpillora/backoff v1.0.0/go.mod h1:J/6gKK9jxlEcS3zixgDgUAsiuZ7yrSoa/FX5e0EB2j4=
 github.com/json-iterator/go v1.1.6/go.mod h1:+SdeFBvtyEkXs7REEP0seUULqWtbJapLOCVDaaPEHmU=
-github.com/json-iterator/go v1.1.7/go.mod h1:KdQUCv79m/52Kvf8AW2vK1V8akMuk1QjK/uOdHXbAo4=
+github.com/json-iterator/go v1.1.10/go.mod h1:KdQUCv79m/52Kvf8AW2vK1V8akMuk1QjK/uOdHXbAo4=
+github.com/json-iterator/go v1.1.11/go.mod h1:KdQUCv79m/52Kvf8AW2vK1V8akMuk1QjK/uOdHXbAo4=
+github.com/jstemmer/go-junit-report v0.0.0-20190106144839-af01ea7f8024/go.mod h1:6v2b51hI/fHJwM22ozAgKL4VKDeJcHhJFhtBdhmNjmU=
+github.com/jstemmer/go-junit-report v0.9.1/go.mod h1:Brl9GWCQeLvo8nXZwPNNblvFj/XSXhF0NWZEnDohbsk=
 github.com/julienschmidt/httprouter v1.2.0/go.mod h1:SYymIcj16QtmaHHD7aYtjjsJG7VTCxuUUipMqKk8s4w=
-github.com/kardianos/service v1.0.0 h1:HgQS3mFfOlyntWX8Oke98JcJLqt1DBcHR4kxShpYef0=
-github.com/kardianos/service v1.0.0/go.mod h1:8CzDhVuCuugtsHyZoTvsOBuvonN/UDBvl0kH+BUxvbo=
-github.com/kardianos/service v1.1.0 h1:QV2SiEeWK42P0aEmGcsAgjApw/lRxkwopvT+Gu6t1/0=
-github.com/kardianos/service v1.1.0/go.mod h1:RrJI2xn5vve/r32U5suTbeaSGoMU6GbNPoj36CVYcHc=
+github.com/julienschmidt/httprouter v1.3.0/go.mod h1:JR6WtHb+2LUe8TCKY3cZOxFyyO8IZAc4RVcycCCAKdM=
+github.com/kardianos/service v1.2.0 h1:bGuZ/epo3vrt8IPC7mnKQolqFeYJb7Cs8Rk4PSOBB/g=
+github.com/kardianos/service v1.2.0/go.mod h1:CIMRFEJVL+0DS1a3Nx06NaMn4Dz63Ng6O7dl0qH0zVM=
+github.com/kisielk/errcheck v1.5.0/go.mod h1:pFxgyoBC7bSaBwPgfKdkLd5X25qrDl4LWUI2bnpBCr8=
+github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
 github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
-github.com/konsorten/go-windows-terminal-sequences v1.0.2 h1:DB17ag19krx9CFsz4o3enTrPXyIXCl+2iCXH/aMAp9s=
-github.com/konsorten/go-windows-terminal-sequences v1.0.2/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
+github.com/konsorten/go-windows-terminal-sequences v1.0.3/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
 github.com/kr/logfmt v0.0.0-20140226030751-b84e30acd515/go.mod h1:+0opPa2QZZtGFBFZlji/RkVcI2GknAs/DXo4wKdlNEc=
-github.com/kr/pretty v0.1.0 h1:L/CwN0zerZDmRFUapSPitk6f+Q3+0za1rQkzVuMiMFI=
 github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
+github.com/kr/pretty v0.2.1 h1:Fmg33tUaq4/8ym9TJN1x7sLJnHVwhP33CNkpYV/7rwI=
+github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
 github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
 github.com/kr/text v0.1.0 h1:45sCR5RtlFHMR4UwH9sdQ5TC8v0qDQCHnXt+kaKSTVE=
 github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
+github.com/lxn/walk v0.0.0-20210112085537-c389da54e794/go.mod h1:E23UucZGqpuUANJooIbHWCufXvOcT6E7Stq81gU+CSQ=
+github.com/lxn/win v0.0.0-20210218163916-a377121e959e/go.mod h1:KxxjdtRkfNoYDCUP5ryK7XJJNTnpC8atvtmTheChOtk=
 github.com/matttproud/golang_protobuf_extensions v1.0.1 h1:4hp9jkHxhMHkqkrB3Ix0jegS5sx/RkqARlsWZ6pIwiU=
 github.com/matttproud/golang_protobuf_extensions v1.0.1/go.mod h1:D8He9yQNgCq6Z5Ld7szi9bcBfOoFv/3dc6xSMkL2PC0=
-github.com/miekg/dns v1.1.25 h1:dFwPR6SfLtrSwgDcIq2bcU/gVutB4sNApq2HBdqcakg=
-github.com/miekg/dns v1.1.25/go.mod h1:bPDLeHnStXmXAq1m/Ch/hvfNHr14JKNPMBo3VZKjuso=
+github.com/miekg/dns v1.1.43 h1:JKfpVSCB84vrAmHzyrsxB5NAr5kLoMXZArPSw7Qlgyg=
+github.com/miekg/dns v1.1.43/go.mod h1:+evo5L0630/F6ca/Z9+GAqzhjGyn8/c+TBaOyfEl0V4=
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
-github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/reflect2 v0.0.0-20180701023420-4b7aa43c6742/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
-github.com/modern-go/reflect2 v1.0.1 h1:9f412s+6RmYXLWZSEzVVgPGK7C2PphHj5RJrvfx9AWI=
 github.com/modern-go/reflect2 v1.0.1/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
 github.com/mwitkow/go-conntrack v0.0.0-20161129095857-cc309e4a2223/go.mod h1:qRWi+5nqEBWmkhHvq77mSJWrCKwh8bxhgT7d/eI7P4U=
-github.com/nbrownus/go-metrics-prometheus v0.0.0-20180622211546-6e6d5173d99c h1:G/mfx/MWYuaaGlHkZQBBXFAJiYnRt/GaOVxnRHjlxg4=
-github.com/nbrownus/go-metrics-prometheus v0.0.0-20180622211546-6e6d5173d99c/go.mod h1:1yMri853KAI2pPAUnESjaqZj9JeImOUM+6A4GuuPmTs=
+github.com/mwitkow/go-conntrack v0.0.0-20190716064945-2f068394615f/go.mod h1:qRWi+5nqEBWmkhHvq77mSJWrCKwh8bxhgT7d/eI7P4U=
+github.com/nbrownus/go-metrics-prometheus v0.0.0-20210712211119-974a6260965f h1:8dM0ilqKL0Uzl42GABzzC4Oqlc3kGRILz0vgoff7nwg=
+github.com/nbrownus/go-metrics-prometheus v0.0.0-20210712211119-974a6260965f/go.mod h1:nwPd6pDNId/Xi16qtKrFHrauSwMNuvk+zcjk89wrnlA=
 github.com/pkg/errors v0.8.0/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pkg/errors v0.8.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
+github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/prometheus/client_golang v0.9.1/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXPKyh/dDVn+NZz0KFw=
 github.com/prometheus/client_golang v1.0.0/go.mod h1:db9x61etRT2tGnBNRi70OPL5FsnadC4Ky3P0J6CfImo=
-github.com/prometheus/client_golang v1.2.1 h1:JnMpQc6ppsNgw9QPAGF6Dod479itz7lvlsMzzNayLOI=
-github.com/prometheus/client_golang v1.2.1/go.mod h1:XMU6Z2MjaRKVu/dC1qupJI9SiNkDYzz3xecMgSW/F+U=
+github.com/prometheus/client_golang v1.7.1/go.mod h1:PY5Wy2awLA44sXw4AOSfFBetzPP4j5+D6mVACh+pe2M=
+github.com/prometheus/client_golang v1.11.0 h1:HNkLOAEQMIDv/K+04rukrLx6ch7msSRwf3/SASFAGtQ=
+github.com/prometheus/client_golang v1.11.0/go.mod h1:Z6t4BnS23TR94PD6BsDNk8yVqroYurpAkEiz0P2BEV0=
 github.com/prometheus/client_model v0.0.0-20180712105110-5c3871d89910/go.mod h1:MbSGuTsp3dbXC40dX6PRTWyKYBIrTGTE9sqQNg2J8bo=
-github.com/prometheus/client_model v0.0.0-20190129233127-fd36f4220a90 h1:S/YWwWx/RA8rT8tKFRuGUZhuA90OyIBpPCXkcbwU8DE=
 github.com/prometheus/client_model v0.0.0-20190129233127-fd36f4220a90/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
-github.com/prometheus/client_model v0.0.0-20191202183732-d1d2010b5bee h1:iBZPTYkGLvdu6+A5TsMUJQkQX9Ad4aCEnSQtdxPuTCQ=
-github.com/prometheus/client_model v0.0.0-20191202183732-d1d2010b5bee/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
-github.com/prometheus/common v0.4.1 h1:K0MGApIoQvMw27RTdJkPbr3JZ7DNbtxQNyi5STVM6Kw=
+github.com/prometheus/client_model v0.2.0 h1:uq5h0d+GuxiXLJLNABMgp2qUWDPiLvgCzz2dUR+/W/M=
+github.com/prometheus/client_model v0.2.0/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/prometheus/common v0.4.1/go.mod h1:TNfzLD0ON7rHzMJeJkieUDPYmFC7Snx/y86RQel1bk4=
-github.com/prometheus/common v0.7.0 h1:L+1lyG48J1zAQXA3RBX/nG/B3gjlHq0zTt2tlbJLyCY=
-github.com/prometheus/common v0.7.0/go.mod h1:DjGbpBbp5NYNiECxcL/VnbXCCaQpKd3tt26CguLLsqA=
+github.com/prometheus/common v0.10.0/go.mod h1:Tlit/dnDKsSWFlCLTWaA1cyBgKHSMdTB80sz/V91rCo=
+github.com/prometheus/common v0.26.0/go.mod h1:M7rCNAaPfAosfx8veZJCuw84e35h3Cfd9VFqTh1DIvc=
+github.com/prometheus/common v0.32.1 h1:hWIdL3N2HoUx3B8j3YN9mWor0qhY/NlEKZEaXxuIRh4=
+github.com/prometheus/common v0.32.1/go.mod h1:vu+V0TpY+O6vW9J44gczi3Ap/oXXR10b+M/gUGO4Hls=
 github.com/prometheus/procfs v0.0.0-20181005140218-185b4288413d/go.mod h1:c3At6R/oaqEKCNdg8wHV1ftS6bRYblBhIjjI8uT2IGk=
 github.com/prometheus/procfs v0.0.2/go.mod h1:TjEm7ze935MbeOT/UhFTIMYKhuLP4wbCsTZCD3I8kEA=
-github.com/prometheus/procfs v0.0.5/go.mod h1:4A/X28fw3Fc593LaREMrKMqOKvUAntwMDaekg4FpcdQ=
-github.com/prometheus/procfs v0.0.8 h1:+fpWZdT24pJBiqJdAwYBjPSk+5YmQzYNPYzQsdzLkt8=
-github.com/prometheus/procfs v0.0.8/go.mod h1:7Qr8sr6344vo1JqZ6HhLceV9o3AJ1Ff+GxbHq6oeK9A=
-github.com/rcrowley/go-metrics v0.0.0-20190826022208-cac0b30c2563 h1:dY6ETXrvDG7Sa4vE8ZQG4yqWg6UnOcbqTAahkV813vQ=
-github.com/rcrowley/go-metrics v0.0.0-20190826022208-cac0b30c2563/go.mod h1:bCqnVzQkZxMG4s8nGwiZ5l3QUCyqpo9Y+/ZMZ9VjZe4=
+github.com/prometheus/procfs v0.1.3/go.mod h1:lV6e/gmhEcM9IjHGsFOCxxuZ+z1YqCvr4OA4YeYWdaU=
+github.com/prometheus/procfs v0.6.0/go.mod h1:cz+aTbrPOrUb4q7XlbU9ygM+/jj0fzG6c1xBZuNvfVA=
+github.com/prometheus/procfs v0.7.3 h1:4jVXhlkAyzOScmCkXBTOLRLTz8EeU+eyjrwB/EPq0VU=
+github.com/prometheus/procfs v0.7.3/go.mod h1:cz+aTbrPOrUb4q7XlbU9ygM+/jj0fzG6c1xBZuNvfVA=
+github.com/rcrowley/go-metrics v0.0.0-20201227073835-cf1acfcdf475 h1:N/ElC8H3+5XpJzTSTfLsJV/mx9Q9g7kxmchpfZyxgzM=
+github.com/rcrowley/go-metrics v0.0.0-20201227073835-cf1acfcdf475/go.mod h1:bCqnVzQkZxMG4s8nGwiZ5l3QUCyqpo9Y+/ZMZ9VjZe4=
+github.com/rogpeppe/go-internal v1.3.0/go.mod h1:M8bDsm7K2OlrFYOpmOWEs/qY81heoFRclV5y23lUDJ4=
 github.com/sirupsen/logrus v1.2.0/go.mod h1:LxeOpSwHxABJmUn/MG1IvRgCAasNZTLOkJPxbbu5VWo=
-github.com/sirupsen/logrus v1.4.2 h1:SPIRibHv4MatM3XXNO2BJeFLZwZ2LvZgfQ5+UNI2im4=
 github.com/sirupsen/logrus v1.4.2/go.mod h1:tLMulIdttU9McNUspp0xgXVQah82FyeX6MwdIuYE2rE=
+github.com/sirupsen/logrus v1.6.0/go.mod h1:7uNnSEd1DgxDLC74fIahvMZmmYsHGZGEOFrfsX/uA88=
+github.com/sirupsen/logrus v1.8.1 h1:dJKuHgqk1NNQlqoA6BTlM1Wf9DOH3NBjQyu0h9+AZZE=
+github.com/sirupsen/logrus v1.8.1/go.mod h1:yWOB1SBYBC5VeMP7gHvWumXLIWorT60ONWic61uBYv0=
 github.com/skip2/go-qrcode v0.0.0-20200617195104-da1b6568686e h1:MRM5ITcdelLK2j1vwZ3Je0FKVCfqOLp5zO6trqMLYs0=
 github.com/skip2/go-qrcode v0.0.0-20200617195104-da1b6568686e/go.mod h1:XV66xRDqSt+GTGFMVlhk3ULuV0y9ZmzeVGR4mloJI3M=
-github.com/songgao/water v0.0.0-20190725173103-fd331bda3f4b h1:+y4hCMc/WKsDbAPsOQZgBSaSZ26uh2afyaWeVg/3s/c=
-github.com/songgao/water v0.0.0-20190725173103-fd331bda3f4b/go.mod h1:P5HUIBuIWKbyjl083/loAegFkfbFNx5i2qEP4CNbm7E=
+github.com/songgao/water v0.0.0-20200317203138-2b4b6d7c09d8 h1:TG/diQgUe0pntT/2D9tmUCz4VNwm9MfrtPr0SU2qSX8=
+github.com/songgao/water v0.0.0-20200317203138-2b4b6d7c09d8/go.mod h1:P5HUIBuIWKbyjl083/loAegFkfbFNx5i2qEP4CNbm7E=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/objx v0.1.1/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
-github.com/stretchr/testify v1.3.0 h1:TivCn/peBQ7UY8ooIcPgZFpTNSz0Q2U6UrFlUfqbe0Q=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
-github.com/stretchr/testify v1.6.1 h1:hDPOHmpOpP40lSULcqw7IrRb/u7w6RpDC9399XyoNd0=
-github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/vishvananda/netlink v1.0.1-0.20190522153524-00009fb8606a h1:Bt1IVPhiCDMqwGrc2nnbIN4QKvJGx6SK2NzWBmW00ao=
-github.com/vishvananda/netlink v1.0.1-0.20190522153524-00009fb8606a/go.mod h1:+SR5DhBJrl6ZM7CoCKvpw5BKroDKQ+PJqOg65H/2ktk=
-github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df h1:OviZH7qLw/7ZovXvuNyL3XQl8UFofeikI1NW1Gypu7k=
+github.com/stretchr/testify v1.4.0/go.mod h1:j7eGeouHqKxXV5pUuKE4zz7dFj8WfuZ+81PSLYec5m4=
+github.com/stretchr/testify v1.7.0 h1:nwc3DEeHmmLAfoZucVR881uASk0Mfjw8xYJ99tb5CcY=
+github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
+github.com/vishvananda/netlink v1.1.0 h1:1iyaYNBLmP6L0220aDnYQpo1QEV4t4hJ+xEEhhJH8j0=
+github.com/vishvananda/netlink v1.1.0/go.mod h1:cTgwzPIzzgDAYoQrMm0EdrjRUBkTqKYppBueQtXaqoE=
 github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df/go.mod h1:JP3t17pCcGlemwknint6hfoeCVQrEMVwxRLRjXpq+BU=
+github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74 h1:gga7acRE695APm9hlsSMoOoE65U4/TcqNj90mc69Rlg=
+github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74/go.mod h1:DD4vA1DwXk04H54A1oHXtwZmA0grkVMdPxx/VGLCah0=
+github.com/yuin/goldmark v1.1.25/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.1.27/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.1.32/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.4.0/go.mod h1:mwnBkeHKe2W/ZEtQ+71ViKU8L12m81fl3OWwC1Zlc8k=
+go.opencensus.io v0.21.0/go.mod h1:mSImk1erAIZhrmZN+AvHh14ztQfjbGwt4TtuofqLduU=
+go.opencensus.io v0.22.0/go.mod h1:+kGneAE2xo2IficOXnaByMWTGM9T73dGwxeWcUqIpI8=
+go.opencensus.io v0.22.2/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
+go.opencensus.io v0.22.3/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
+go.opencensus.io v0.22.4/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
 golang.org/x/crypto v0.0.0-20180904163835-0709b304e793/go.mod h1:6SG95UA2DQfeDnfUPMdvaQW0Q7yPrPDi9nlGo2tz2b4=
 golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
-golang.org/x/crypto v0.0.0-20190923035154-9ee001bba392/go.mod h1:/lpIB1dKB+9EgE3H3cr1v9wB50oz8l4C4h62xy7jSTY=
-golang.org/x/crypto v0.0.0-20200220183623-bac4c82f6975 h1:/Tl7pH94bvbAAHBdZJT947M/+gp0+CqQXDtMRC0fseo=
-golang.org/x/crypto v0.0.0-20200220183623-bac4c82f6975/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/crypto v0.0.0-20190510104115-cbcb75029529/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20190605123033-f99c8df09eb5/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/crypto v0.0.0-20210322153248-0c34fe9e7dc2/go.mod h1:T9bdIzuCu7OtxOm1hfPfRQxPLYneinmdGuTeoZ9dtd4=
+golang.org/x/crypto v0.0.0-20210921155107-089bfa567519 h1:7I4JAnoQBe7ZtJcBaYHi5UtiO8tQHbUSXxL+pnGRANg=
+golang.org/x/crypto v0.0.0-20210921155107-089bfa567519/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
+golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
+golang.org/x/exp v0.0.0-20190306152737-a1d7652674e8/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
+golang.org/x/exp v0.0.0-20190510132918-efd6b22b2522/go.mod h1:ZjyILWgesfNpC6sMxTJOJm9Kp84zZh5NQWvqDGG3Qr8=
+golang.org/x/exp v0.0.0-20190829153037-c13cbed26979/go.mod h1:86+5VVa7VpoJ4kLfm080zCjGlMRFzhUhsZKEZO7MGek=
+golang.org/x/exp v0.0.0-20191030013958-a1ab85dbe136/go.mod h1:JXzH8nQsPlswgeRAPE3MuO9GYsAcnJvJ4vnMwN/5qkY=
+golang.org/x/exp v0.0.0-20191129062945-2f5052295587/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20191227195350-da58074b4299/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20200119233911-0405dc783f0a/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20200207192155-f17229e696bd/go.mod h1:J/WKrq2StrnmMY6+EHIKF9dgMWnmCNThgcyBT1FY9mM=
+golang.org/x/exp v0.0.0-20200224162631-6cc2880d07d6/go.mod h1:3jZMyOhIsHpP37uCMkUooju7aAi5cS1Q23tOzKc+0MU=
+golang.org/x/image v0.0.0-20190227222117-0694c2d4d067/go.mod h1:kZ7UVZpmo3dzQBMxlp+ypCbDeSB+sBbTgSJuh5dn5js=
+golang.org/x/image v0.0.0-20190802002840-cff245a6509b/go.mod h1:FeLwcggjj3mMvU+oOTbSwawSJRM1uh48EjtB4UJZlP0=
+golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
+golang.org/x/lint v0.0.0-20190301231843-5614ed5bae6f/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190409202823-959b441ac422/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190909230951-414d861bb4ac/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190930215403-16217165b5de/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20191125180803-fdd1cda4f05f/go.mod h1:5qLYkcX4OjUUV8bRuDixDT3tpyyb+LUpUlRWLxfhWrs=
+golang.org/x/lint v0.0.0-20200130185559-910be7a94367/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
+golang.org/x/lint v0.0.0-20200302205851-738671d3881b/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
+golang.org/x/mobile v0.0.0-20190312151609-d3739f865fa6/go.mod h1:z+o9i4GpDbdi3rU15maQ/Ox0txvL9dWGYEHz965HBQE=
+golang.org/x/mobile v0.0.0-20190719004257-d2bd2a29d028/go.mod h1:E/iHnbuqvinMTCcRqshq8CkpyQDoeVncDDYHnLhea+o=
+golang.org/x/mod v0.0.0-20190513183733-4bf6d317e70e/go.mod h1:mXi4GBBbnImb6dmsKGUJ2LatrhH/nqhxcFungHvyanc=
+golang.org/x/mod v0.1.0/go.mod h1:0QHyrYULN0/3qlju5TqG8bIK38QM8yzMo5ekMj3DlcY=
+golang.org/x/mod v0.1.1-0.20191105210325-c90efee705ee/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
+golang.org/x/mod v0.1.1-0.20191107180719-034126e5016b/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
+golang.org/x/mod v0.2.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/mod v0.3.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/mod v0.4.2/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
 golang.org/x/net v0.0.0-20181114220301-adae6a3d119a/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190108225652-1e06a53dbb7e/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
 golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190501004415-9ce7a6920f09/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190503192946-f4e77d36d62c/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190603091049-60506f45cf65/go.mod h1:HSz+uSET+XFnRR8LxR5pz3Of3rY3CfYBVs4xY44aLks=
 golang.org/x/net v0.0.0-20190613194153-d28f0bde5980/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
 golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20190923162816-aa69164e4478/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20191209160850-c0dbc17a3553 h1:efeOvDhwQ29Dj3SdAV/MJf8oukgn+8D8WgaCaRMchF8=
+golang.org/x/net v0.0.0-20190628185345-da137c7871d7/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20190724013045-ca1201d0de80/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
 golang.org/x/net v0.0.0-20191209160850-c0dbc17a3553/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200114155413-6afb5195e5aa/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200202094626-16171245cfb2/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200222125558-5a598a2470a0/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200226121028-0de0cce0169b/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200301022130-244492dfa37a/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200324143707-d3edc9973b7e/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200501053045-e0ff5e5a1de5/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200506145744-7e3656a0809f/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200513185701-a91f0712d120/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200520182314-0ba52f642ac2/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200625001655-4c5254603344/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20200707034311-ab3426394381/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20200822124328-c89045814202/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
+golang.org/x/net v0.0.0-20210525063256-abc453219eb5/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/net v0.0.0-20210805182204-aaa1db679c0d/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/net v0.0.0-20211020060615-d418f374d309/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/net v0.0.0-20211101193420-4a448f8816b3 h1:VrJZAjbekhoRn7n5FBujY31gboH+iB3pdLxn3gE9FjU=
+golang.org/x/net v0.0.0-20211101193420-4a448f8816b3/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
+golang.org/x/oauth2 v0.0.0-20190226205417-e64efc72b421/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20190604053449-0f29369cfe45/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20191202225959-858c2ad4c8b6/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20200107190931-bf48bf16ab8d/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20210514164344-f6687ab2804c/go.mod h1:KelEdhl1UZF7XfJ4dDtk6s++YSgaE7mD/BuKKDLBl4A=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20181221193216-37e7f081c4d4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.0.0-20190423024810-112230192c58 h1:8gQV6CLnAEikrhgkHFbMAEhagSSnXWGV915qUMm9mrU=
+golang.org/x/sync v0.0.0-20190227155943-e225da77a7e6/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e h1:vcxGaoTs7kV8m5Np9uUNQin4BrLOthgV7252N8V+FwY=
 golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20200317015054-43a5402ce75a/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20200625203802-6e8e738ad208/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20201020160332-67f06af15bc9/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20201207232520-09787c993a3a/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20210220032951-036812b2e83c h1:5KslGYwFpkhGh+Q16bwMP3cOontH8FOep7tGV86Y7SQ=
+golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20181116152217-5ac8a444bdc5/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
-golang.org/x/sys v0.0.0-20190204203706-41f3e6584952/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190312061237-fead79001313/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20190422165155-953cdadca894/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190922100055-0a153f010e69/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190924154521-2837fb4f24fe/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20191010194322-b09406accb47/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20191210023423-ac6580df4449 h1:gSbV7h1NRL2G1xTg/owz62CST1oJBmxy4QpMMregXVQ=
-golang.org/x/sys v0.0.0-20191210023423-ac6580df4449/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190502145724-3ef323f4f1fd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190507160741-ecd444e8653b/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190606165138-5da285871e9c/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190606203320-7fc4e5ec1444/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190624142023-c5567b49c5d0/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190726091711-fc99dfbffb4e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191001151750-bb3f8db39f24/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191026070338-33540a1f6037/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191204072324-ce4227a45e2e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191228213918-04cbcbbfeed8/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200106162015-b016eb3dc98e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200113162924-86b910548bc1/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200122134326-e047566fdf82/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200202164722-d101bd2416d5/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200212091648-12a6c2dcc1e4/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200217220822-9197077df867/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200223170610-d5e6a3e2c0ae/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200302150141-5c8b2ff67527/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200323222414-85ca7c5b95cd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200331124033-c3d80250170d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200501052902-10377860bb8e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200511232937-7e40ca221e25/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200515095857-1151b9dac4a9/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200523222454-059865788121/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200615200032-f1bc736245b1/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200625212154-ddb9806d33ae/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200803210538-64077c9b5642/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20201015000850-e3ed0017c211/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20201018230417-eeed37f84f13/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210124154548-22da62e12c0c/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210303074136-134d130e1a04/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210603081109-ebe580a85c40/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20210809222454-d867a43fc93e/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20211025201205-69cdffdb9359/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20211103235746-7861aae1554b h1:1VkfZQv42XQlA/jchYumAnv1UPo6RgF9rJFkTgZIxO4=
+golang.org/x/sys v0.0.0-20211103235746-7861aae1554b/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
+golang.org/x/term v0.0.0-20210927222741-03fcf44c2211 h1:JGgROgKl9N8DuW20oFS5gxc+lE67/N3FcwmBPMe7ArY=
+golang.org/x/term v0.0.0-20210927222741-03fcf44c2211/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
+golang.org/x/text v0.0.0-20170915032832-14c0d48ead0c/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.1-0.20180807135948-17ff2d5776d2/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
+golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/text v0.3.8-0.20211004125949-5bd84dd9b33b/go.mod h1:EFNZuWvGYxIRUEX+K8UmCFwYmZjqcrnq15ZuVldZkZ0=
+golang.org/x/time v0.0.0-20181108054448-85acf8d2951c/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
+golang.org/x/time v0.0.0-20190308202827-9d24e82272b4/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
+golang.org/x/time v0.0.0-20191024005414-555d28b269f0/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.0.0-20190907020128-2ca718005c18/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
+golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190312151545-0bb0c0a6e846/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190312170243-e65039ee4138/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190425150028-36563e24a262/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190506145303-2d16b83fe98c/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190606124116-d0a3d012864b/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190621195816-6e04913cbbac/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190628153133-6cdbf07be9d0/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190816200558-6889da9d5479/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20190911174233-4f2ddba30aff/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191012152004-8de300cfc20a/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191113191852-77e3bb0ad9e7/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191115202509-3a792d9c32b2/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191125144606-a911d9008d1f/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191130070609-6e064ea0cf2d/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191216173652-a0e659d51361/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20191227053925-7b8e75db28f4/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200117161641-43d50277825c/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200122220014-bf1340f18c4a/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200130002326-2f3ba24bd6e7/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200204074204-1cc6d1ef6c74/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200207183749-b753a1ba74fa/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200212150539-ea181f53ac56/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200224181240-023911ca70b2/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200227222343-706bc42d1f0d/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200304193943-95d2e580d8eb/go.mod h1:o4KQGtdN14AW+yjsvvwRTJJuXz8XRtIHtEnmAXLyFUw=
+golang.org/x/tools v0.0.0-20200312045724-11d5b4c81c7d/go.mod h1:o4KQGtdN14AW+yjsvvwRTJJuXz8XRtIHtEnmAXLyFUw=
+golang.org/x/tools v0.0.0-20200331025713-a30bf2db82d4/go.mod h1:Sl4aGygMT6LrqrWclx+PTx3U+LnKx/seiNR+3G19Ar8=
+golang.org/x/tools v0.0.0-20200501065659-ab2804fb9c9d/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200512131952-2bc93b1c0c88/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200515010526-7d3b6ebf133d/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200618134242-20370b0cb4b2/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200619180055-7c47624df98f/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200729194436-6467de6f59a7/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20200804011535-6c149bb5ef0d/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20200825202427-b303f430e36d/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20210106214847-113979e3529a/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.1.7/go.mod h1:LGqMHiF4EqQNHR1JncWGqT5BVaXmza+X+BDGol+dOxo=
 golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20191011141410-1b5146add898/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 h1:go1bK/D/BFZV2I8cIQd1NKEZ+0owSTG1fDTci4IqFcE=
+golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.zx2c4.com/wintun v0.0.0-20211104114900-415007cec224 h1:Ug9qvr1myri/zFN6xL17LSCBGFDnphBBhzmILHsM5TY=
+golang.zx2c4.com/wintun v0.0.0-20211104114900-415007cec224/go.mod h1:deeaetjYA+DHMHg+sMSMI58GrEteJUUzzw7en6TJQcI=
+golang.zx2c4.com/wireguard/windows v0.5.1 h1:OnYw96PF+CsIMrqWo5QP3Q59q5hY1rFErk/yN3cS+JQ=
+golang.zx2c4.com/wireguard/windows v0.5.1/go.mod h1:EApyTk/ZNrkbZjurHL1nleDYnsPpJYBO7LZEBCyDAHk=
+google.golang.org/api v0.4.0/go.mod h1:8k5glujaEP+g9n7WNsDg8QP6cUVNI86fCNMcbazEtwE=
+google.golang.org/api v0.7.0/go.mod h1:WtwebWUNSVBH/HAw79HIFXZNqEvBhG+Ra+ax0hx3E3M=
+google.golang.org/api v0.8.0/go.mod h1:o4eAsZoiT+ibD93RtjEohWalFOjRDx6CVaqeizhEnKg=
+google.golang.org/api v0.9.0/go.mod h1:o4eAsZoiT+ibD93RtjEohWalFOjRDx6CVaqeizhEnKg=
+google.golang.org/api v0.13.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.14.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.15.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.17.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.18.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.19.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.20.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.22.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.24.0/go.mod h1:lIXQywCXRcnZPGlsd8NbLnOjtAoL6em04bJ9+z0MncE=
+google.golang.org/api v0.28.0/go.mod h1:lIXQywCXRcnZPGlsd8NbLnOjtAoL6em04bJ9+z0MncE=
+google.golang.org/api v0.29.0/go.mod h1:Lcubydp8VUV7KeIHD9z2Bys/sm/vGKnG1UHuDBSrHWM=
+google.golang.org/api v0.30.0/go.mod h1:QGmEvQ87FHZNiUVJkT14jQNYJ4ZJjdRF23ZXz5138Fc=
+google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
+google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/appengine v1.5.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/appengine v1.6.1/go.mod h1:i06prIuMbXzDqacNJfV5OdTW448YApPu5ww/cMBSeb0=
+google.golang.org/appengine v1.6.5/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
+google.golang.org/appengine v1.6.6/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
+google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
+google.golang.org/genproto v0.0.0-20190307195333-5fe7a883aa19/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190418145605-e7d98fc518a7/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190425155659-357c62f0e4bb/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190502173448-54afdca5d873/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190801165951-fa694d86fc64/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20190911173649-1774047e7e51/go.mod h1:IbNlFCBrqXvoKpeg0TB2l7cyZUmoaFKYIwrEpbDKLA8=
+google.golang.org/genproto v0.0.0-20191108220845-16a3f7862a1a/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191115194625-c23dd37a84c9/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191216164720-4f79533eabd1/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191230161307-f3c370f40bfb/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200115191322-ca5a22157cba/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200122232147-0452cf42e150/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200204135345-fa8e72b47b90/go.mod h1:GmwEX6Z4W5gMy59cAlVYjN9JhxgbQH6Gn+gFDQe2lzA=
+google.golang.org/genproto v0.0.0-20200212174721-66ed5ce911ce/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200224152610-e50cd9704f63/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200228133532-8c2c7df3a383/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200305110556-506484158171/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200312145019-da6875a35672/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200331122359-1ee6d9798940/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200430143042-b979b6f78d84/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200511104702-f5ebc3bea380/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200515170657-fc4c6c6a6587/go.mod h1:YsZOwe1myG/8QRHRsmBRE1LrgQY60beZKjly0O1fX9U=
+google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
+google.golang.org/genproto v0.0.0-20200618031413-b414f8b61790/go.mod h1:jDfRM7FcilCzHH/e9qn6dsT145K34l5v+OpcnNgKAAA=
+google.golang.org/genproto v0.0.0-20200729003335-053ba62fc06f/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20200804131852-c06518451d9c/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20200825200019-8632dd797987/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
+google.golang.org/grpc v1.20.1/go.mod h1:10oTOabMzJvdu6/UiuZezV6QK5dSlG84ov/aaiqXj38=
+google.golang.org/grpc v1.21.1/go.mod h1:oYelfM1adQP15Ek0mdvEgi9Df8B9CZIaU1084ijfRaM=
+google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
+google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=
+google.golang.org/grpc v1.26.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.27.1/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.28.0/go.mod h1:rpkK4SK4GF4Ach/+MFLZUBavHOvF2JJB5uozKKal+60=
+google.golang.org/grpc v1.29.1/go.mod h1:itym6AZVZYACWQqET3MqgPpjcuV5QH3BxFS3IjizoKk=
+google.golang.org/grpc v1.30.0/go.mod h1:N36X2cJ7JwdamYAgDz+s+rVMFjt3numwzf/HckM8pak=
+google.golang.org/grpc v1.31.0/go.mod h1:N36X2cJ7JwdamYAgDz+s+rVMFjt3numwzf/HckM8pak=
+google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
+google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
+google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
+google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
+google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
+google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.24.0/go.mod h1:r/3tXBNzIEhYS9I1OUVjXDlt8tc493IdKGjtUeSXeh4=
+google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
+google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=
+google.golang.org/protobuf v1.26.0/go.mod h1:9q0QmTI4eRPtz6boOQmLYwt+qCgq0jsYwAQnmE0givc=
+google.golang.org/protobuf v1.27.1 h1:SnqbnDw1V7RiZcXPx5MEeqPv2s79L9i7BJUlG/+RurQ=
+google.golang.org/protobuf v1.27.1/go.mod h1:9q0QmTI4eRPtz6boOQmLYwt+qCgq0jsYwAQnmE0givc=
 gopkg.in/alecthomas/kingpin.v2 v2.2.6/go.mod h1:FMv+mEhP44yOT+4EoQTLFTRgOQ1FBLkstjWtayDeSgw=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
-gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 h1:qIbj1fsPNlZgppZ+VLlY7N33q108Sa+fhmuc+sWQYwY=
 gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20190902080502-41f04d3bba15/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
+gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
 gopkg.in/yaml.v2 v2.2.1/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v2 v2.2.2 h1:ZCJp+EgiOT7lHqUV2J862kp8Qj64Jo6az82+3Td9dZw=
 gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v2 v2.2.7 h1:VUgggvou5XRW9mHwD/yXxIYSMtY0zoKQf/v226p2nyo=
-gopkg.in/yaml.v2 v2.2.7/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.2.4/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.2.5/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.3.0/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
+gopkg.in/yaml.v2 v2.4.0/go.mod h1:RDklbk79AGWmwhnvt/jBztapEOGDOx6ZbXqjP6csGnQ=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c h1:dUUwHk2QECo/6vqA44rthZ8ie2QXMNeKRTHCNY2nXvo=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
+honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190106161140-3f1c8253044a/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190418001031-e561f6794a2a/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.1-2019.2.3/go.mod h1:a3bituU0lyd329TUQxRnasdCoJDkEUEAqEt0JzvZhAg=
+honnef.co/go/tools v0.0.1-2020.1.3/go.mod h1:X/FiERA/W4tHapMX5mGpAtMSVEeEUOyHaw9vFzvIQ3k=
+honnef.co/go/tools v0.0.1-2020.1.4/go.mod h1:X/FiERA/W4tHapMX5mGpAtMSVEeEUOyHaw9vFzvIQ3k=
+rsc.io/binaryregexp v0.2.0/go.mod h1:qTv7/COck+e2FymRvadv62gMdZztPaShugOCi3I+8D8=
+rsc.io/quote/v3 v3.1.0/go.mod h1:yEA65RcK8LyAZtP9Kv3t0HmxON59tX3rD+tICJqUlj0=
+rsc.io/sampler v1.3.0/go.mod h1:T1hPZKmBbMNahiBKFy5HrXp6adAjACjK9JXDnKaTXpA=

handshake.go

@@ -1,18 +1,19 @@
 package nebula
 
-const (
-	handshakeIXPSK0 = 0
-	handshakeXXPSK0 = 1
+import (
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/udp"
 )
 
-func HandleIncomingHandshake(f *Interface, addr *udpAddr, packet []byte, h *Header, hostinfo *HostInfo) {
-	if !f.lightHouse.remoteAllowList.Allow(udp2ipInt(addr)) {
-		l.WithField("udpAddr", addr).Debug("lighthouse.remote_allow_list denied incoming handshake")
+func HandleIncomingHandshake(f *Interface, addr *udp.Addr, packet []byte, h *header.H, hostinfo *HostInfo) {
+	// First remote allow list check before we know the vpnIp
+	if !f.lightHouse.remoteAllowList.AllowUnknownVpnIp(addr.IP) {
+		f.l.WithField("udpAddr", addr).Debug("lighthouse.remote_allow_list denied incoming handshake")
 		return
 	}
 
 	switch h.Subtype {
-	case handshakeIXPSK0:
+	case header.HandshakeIXPSK0:
 		switch h.MessageCounter {
 		case 1:
 			ixHandshakeStage1(f, addr, packet, h)

handshake_ix.go

@@ -1,33 +1,31 @@
 package nebula
 
 import (
-	"bytes"
 	"sync/atomic"
 	"time"
 
 	"github.com/flynn/noise"
 	"github.com/golang/protobuf/proto"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
 // NOISE IX Handshakes
 
 // This function constructs a handshake packet, but does not actually send it
 // Sending is done by the handshake manager
-func ixHandshakeStage0(f *Interface, vpnIp uint32, hostinfo *HostInfo) {
+func ixHandshakeStage0(f *Interface, vpnIp iputil.VpnIp, hostinfo *HostInfo) {
 	// This queries the lighthouse if we don't know a remote for the host
+	// We do it here to provoke the lighthouse to preempt our timer wheel and trigger the stage 1 packet to send
+	// more quickly, effect is a quicker handshake.
 	if hostinfo.remote == nil {
-		ips, err := f.lightHouse.Query(vpnIp, f)
-		if err != nil {
-			//l.Debugln(err)
-		}
-		for _, ip := range ips {
-			hostinfo.AddRemote(ip)
-		}
+		f.lightHouse.QueryServer(vpnIp, f)
 	}
 
 	err := f.handshakeManager.AddIndexHostInfo(hostinfo)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(vpnIp)).
+		f.l.WithError(err).WithField("vpnIp", vpnIp).
 			WithField("handshake", m{"stage": 0, "style": "ix_psk0"}).Error("Failed to generate index")
 		return
 	}
@@ -36,7 +34,7 @@ func ixHandshakeStage0(f *Interface, vpnIp uint32, hostinfo *HostInfo) {
 
 	hsProto := &NebulaHandshakeDetails{
 		InitiatorIndex: hostinfo.localIndexId,
-		Time:           uint64(time.Now().Unix()),
+		Time:           uint64(time.Now().UnixNano()),
 		Cert:           ci.certState.rawCertificateNoKey,
 	}
 
@@ -48,39 +46,38 @@ func ixHandshakeStage0(f *Interface, vpnIp uint32, hostinfo *HostInfo) {
 	hsBytes, err = proto.Marshal(hs)
 
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(vpnIp)).
+		f.l.WithError(err).WithField("vpnIp", vpnIp).
 			WithField("handshake", m{"stage": 0, "style": "ix_psk0"}).Error("Failed to marshal handshake message")
 		return
 	}
 
-	header := HeaderEncode(make([]byte, HeaderLen), Version, uint8(handshake), handshakeIXPSK0, 0, 1)
+	h := header.Encode(make([]byte, header.Len), header.Version, header.Handshake, header.HandshakeIXPSK0, 0, 1)
 	atomic.AddUint64(&ci.atomicMessageCounter, 1)
 
-	msg, _, _, err := ci.H.WriteMessage(header, hsBytes)
+	msg, _, _, err := ci.H.WriteMessage(h, hsBytes)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(vpnIp)).
+		f.l.WithError(err).WithField("vpnIp", vpnIp).
 			WithField("handshake", m{"stage": 0, "style": "ix_psk0"}).Error("Failed to call noise.WriteMessage")
 		return
 	}
 
 	// We are sending handshake packet 1, so we don't expect to receive
 	// handshake packet 1 from the responder
-	ci.window.Update(1)
+	ci.window.Update(f.l, 1)
 
 	hostinfo.HandshakePacket[0] = msg
 	hostinfo.HandshakeReady = true
 	hostinfo.handshakeStart = time.Now()
-
 }
 
-func ixHandshakeStage1(f *Interface, addr *udpAddr, packet []byte, h *Header) {
-	ci := f.newConnectionState(false, noise.HandshakeIX, []byte{}, 0)
+func ixHandshakeStage1(f *Interface, addr *udp.Addr, packet []byte, h *header.H) {
+	ci := f.newConnectionState(f.l, false, noise.HandshakeIX, []byte{}, 0)
 	// Mark packet 1 as seen so it doesn't show up as missed
-	ci.window.Update(1)
+	ci.window.Update(f.l, 1)
 
-	msg, _, _, err := ci.H.ReadMessage(nil, packet[HeaderLen:])
+	msg, _, _, err := ci.H.ReadMessage(nil, packet[header.Len:])
 	if err != nil {
-		l.WithError(err).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Failed to call noise.ReadMessage")
 		return
 	}
@@ -91,85 +88,102 @@ func ixHandshakeStage1(f *Interface, addr *udpAddr, packet []byte, h *Header) {
 		l.Debugln("GOT INDEX: ", hs.Details.InitiatorIndex)
 	*/
 	if err != nil || hs.Details == nil {
-		l.WithError(err).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Failed unmarshal handshake message")
 		return
 	}
 
-	remoteCert, err := RecombineCertAndValidate(ci.H, hs.Details.Cert)
+	remoteCert, err := RecombineCertAndValidate(ci.H, hs.Details.Cert, f.caPool)
 	if err != nil {
-		l.WithError(err).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).WithField("cert", remoteCert).
 			Info("Invalid certificate from host")
 		return
 	}
-	vpnIP := ip2int(remoteCert.Details.Ips[0].IP)
+	vpnIp := iputil.Ip2VpnIp(remoteCert.Details.Ips[0].IP)
 	certName := remoteCert.Details.Name
 	fingerprint, _ := remoteCert.Sha256Sum()
+	issuer := remoteCert.Details.Issuer
 
-	if vpnIP == ip2int(f.certState.certificate.Details.Ips[0].IP) {
-		l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+	if vpnIp == f.myVpnIp {
+		f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Refusing to handshake with myself")
 		return
 	}
 
-	myIndex, err := generateIndex()
+	if !f.lightHouse.remoteAllowList.Allow(vpnIp, addr.IP) {
+		f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).Debug("lighthouse.remote_allow_list denied incoming handshake")
+		return
+	}
+
+	myIndex, err := generateIndex(f.l)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Failed to generate index")
 		return
 	}
 
 	hostinfo := &HostInfo{
-		ConnectionState: ci,
-		Remotes:         []*HostInfoDest{},
-		localIndexId:    myIndex,
-		remoteIndexId:   hs.Details.InitiatorIndex,
-		hostId:          vpnIP,
-		HandshakePacket: make(map[uint8][]byte, 0),
+		ConnectionState:   ci,
+		localIndexId:      myIndex,
+		remoteIndexId:     hs.Details.InitiatorIndex,
+		vpnIp:             vpnIp,
+		HandshakePacket:   make(map[uint8][]byte, 0),
+		lastHandshakeTime: hs.Details.Time,
 	}
 
-	l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+	hostinfo.Lock()
+	defer hostinfo.Unlock()
+
+	f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 		WithField("certName", certName).
 		WithField("fingerprint", fingerprint).
+		WithField("issuer", issuer).
 		WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 		WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
 		Info("Handshake message received")
 
 	hs.Details.ResponderIndex = myIndex
 	hs.Details.Cert = ci.certState.rawCertificateNoKey
+	// Update the time in case their clock is way off from ours
+	hs.Details.Time = uint64(time.Now().UnixNano())
 
 	hsBytes, err := proto.Marshal(hs)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Failed to marshal handshake message")
 		return
 	}
 
-	header := HeaderEncode(make([]byte, HeaderLen), Version, uint8(handshake), handshakeIXPSK0, hs.Details.InitiatorIndex, 2)
-	msg, dKey, eKey, err := ci.H.WriteMessage(header, hsBytes)
+	nh := header.Encode(make([]byte, header.Len), header.Version, header.Handshake, header.HandshakeIXPSK0, hs.Details.InitiatorIndex, 2)
+	msg, dKey, eKey, err := ci.H.WriteMessage(nh, hsBytes)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Failed to call noise.WriteMessage")
 		return
 	} else if dKey == nil || eKey == nil {
-		l.WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).Error("Noise did not arrive at a key")
 		return
 	}
 
-	hostinfo.HandshakePacket[0] = make([]byte, len(packet[HeaderLen:]))
-	copy(hostinfo.HandshakePacket[0], packet[HeaderLen:])
+	hostinfo.HandshakePacket[0] = make([]byte, len(packet[header.Len:]))
+	copy(hostinfo.HandshakePacket[0], packet[header.Len:])
 
 	// Regardless of whether you are the sender or receiver, you should arrive here
 	// and complete standing up the connection.
@@ -178,69 +192,92 @@ func ixHandshakeStage1(f *Interface, addr *udpAddr, packet []byte, h *Header) {
 
 	// We are sending handshake packet 2, so we don't expect to receive
 	// handshake packet 2 from the initiator.
-	ci.window.Update(2)
+	ci.window.Update(f.l, 2)
 
 	ci.peerCert = remoteCert
 	ci.dKey = NewNebulaCipherState(dKey)
 	ci.eKey = NewNebulaCipherState(eKey)
-	//l.Debugln("got symmetric pairs")
 
-	//hostinfo.ClearRemotes()
-	hostinfo.AddRemote(*addr)
-	hostinfo.ForcePromoteBest(f.hostMap.preferredRanges)
+	hostinfo.remotes = f.lightHouse.QueryCache(vpnIp)
+	hostinfo.SetRemote(addr)
 	hostinfo.CreateRemoteCIDR(remoteCert)
 
-	hostinfo.Lock()
-	defer hostinfo.Unlock()
-
 	// Only overwrite existing record if we should win the handshake race
-	overwrite := vpnIP > ip2int(f.certState.certificate.Details.Ips[0].IP)
+	overwrite := vpnIp > f.myVpnIp
 	existing, err := f.handshakeManager.CheckAndComplete(hostinfo, 0, overwrite, f)
 	if err != nil {
 		switch err {
 		case ErrAlreadySeen:
+			// Update remote if preferred (Note we have to switch to locking
+			// the existing hostinfo, and then switch back so the defer Unlock
+			// higher in this function still works)
+			hostinfo.Unlock()
+			existing.Lock()
+			// Update remote if preferred
+			if existing.SetRemoteIfPreferred(f.hostMap, addr) {
+				// Send a test packet to ensure the other side has also switched to
+				// the preferred remote
+				f.SendMessageToVpnIp(header.Test, header.TestRequest, vpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
+			}
+			existing.Unlock()
+			hostinfo.Lock()
+
 			msg = existing.HandshakePacket[2]
-			f.messageMetrics.Tx(handshake, NebulaMessageSubType(msg[1]), 1)
+			f.messageMetrics.Tx(header.Handshake, header.MessageSubType(msg[1]), 1)
 			err := f.outside.WriteTo(msg, addr)
 			if err != nil {
-				l.WithField("vpnIp", IntIp(existing.hostId)).WithField("udpAddr", addr).
+				f.l.WithField("vpnIp", existing.vpnIp).WithField("udpAddr", addr).
 					WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).WithField("cached", true).
 					WithError(err).Error("Failed to send handshake message")
 			} else {
-				l.WithField("vpnIp", IntIp(existing.hostId)).WithField("udpAddr", addr).
+				f.l.WithField("vpnIp", existing.vpnIp).WithField("udpAddr", addr).
 					WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).WithField("cached", true).
 					Info("Handshake message sent")
 			}
 			return
 		case ErrExistingHostInfo:
-			// This means there was an existing tunnel and we didn't win
-			// handshake avoidance
-			l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+			// This means there was an existing tunnel and this handshake was older than the one we are currently based on
+			f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 				WithField("certName", certName).
+				WithField("oldHandshakeTime", existing.lastHandshakeTime).
+				WithField("newHandshakeTime", hostinfo.lastHandshakeTime).
 				WithField("fingerprint", fingerprint).
+				WithField("issuer", issuer).
 				WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 				WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
-				Info("Prevented a handshake race")
+				Info("Handshake too old")
 
 			// Send a test packet to trigger an authenticated tunnel test, this should suss out any lingering tunnel issues
-			f.SendMessageToVpnIp(test, testRequest, vpnIP, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
+			f.SendMessageToVpnIp(header.Test, header.TestRequest, vpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
 			return
 		case ErrLocalIndexCollision:
 			// This means we failed to insert because of collision on localIndexId. Just let the next handshake packet retry
-			l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+			f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 				WithField("certName", certName).
 				WithField("fingerprint", fingerprint).
+				WithField("issuer", issuer).
 				WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 				WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
-				WithField("localIndex", hostinfo.localIndexId).WithField("collision", IntIp(existing.hostId)).
+				WithField("localIndex", hostinfo.localIndexId).WithField("collision", existing.vpnIp).
 				Error("Failed to add HostInfo due to localIndex collision")
 			return
+		case ErrExistingHandshake:
+			// We have a race where both parties think they are an initiator and this tunnel lost, let the other one finish
+			f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
+				WithField("certName", certName).
+				WithField("fingerprint", fingerprint).
+				WithField("issuer", issuer).
+				WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
+				WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
+				Error("Prevented a pending handshake race")
+			return
 		default:
 			// Shouldn't happen, but just in case someone adds a new error type to CheckAndComplete
 			// And we forget to update it here
-			l.WithError(err).WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+			f.l.WithError(err).WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 				WithField("certName", certName).
 				WithField("fingerprint", fingerprint).
+				WithField("issuer", issuer).
 				WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 				WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
 				Error("Failed to add HostInfo to HostMap")
@@ -249,53 +286,66 @@ func ixHandshakeStage1(f *Interface, addr *udpAddr, packet []byte, h *Header) {
 	}
 
 	// Do the send
-	f.messageMetrics.Tx(handshake, NebulaMessageSubType(msg[1]), 1)
+	f.messageMetrics.Tx(header.Handshake, header.MessageSubType(msg[1]), 1)
 	err = f.outside.WriteTo(msg, addr)
 	if err != nil {
-		l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+		f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 			WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
 			WithError(err).Error("Failed to send handshake")
 	} else {
-		l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+		f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 			WithField("certName", certName).
 			WithField("fingerprint", fingerprint).
+			WithField("issuer", issuer).
 			WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 			WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
+			WithField("sentCachedPackets", len(hostinfo.packetStore)).
 			Info("Handshake message sent")
 	}
 
-	hostinfo.handshakeComplete()
+	hostinfo.handshakeComplete(f.l, f.cachedPacketMetrics)
 
 	return
 }
 
-func ixHandshakeStage2(f *Interface, addr *udpAddr, hostinfo *HostInfo, packet []byte, h *Header) bool {
+func ixHandshakeStage2(f *Interface, addr *udp.Addr, hostinfo *HostInfo, packet []byte, h *header.H) bool {
 	if hostinfo == nil {
+		// Nothing here to tear down, got a bogus stage 2 packet
 		return true
 	}
+
 	hostinfo.Lock()
 	defer hostinfo.Unlock()
 
-	if bytes.Equal(hostinfo.HandshakePacket[2], packet[HeaderLen:]) {
-		l.WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
-			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).WithField("header", h).
-			Info("Already seen this handshake packet")
+	if !f.lightHouse.remoteAllowList.Allow(hostinfo.vpnIp, addr.IP) {
+		f.l.WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).Debug("lighthouse.remote_allow_list denied incoming handshake")
 		return false
 	}
 
 	ci := hostinfo.ConnectionState
-	// Mark packet 2 as seen so it doesn't show up as missed
-	ci.window.Update(2)
+	if ci.ready {
+		f.l.WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
+			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).WithField("header", h).
+			Info("Handshake is already complete")
 
-	hostinfo.HandshakePacket[2] = make([]byte, len(packet[HeaderLen:]))
-	copy(hostinfo.HandshakePacket[2], packet[HeaderLen:])
+		// Update remote if preferred
+		if hostinfo.SetRemoteIfPreferred(f.hostMap, addr) {
+			// Send a test packet to ensure the other side has also switched to
+			// the preferred remote
+			f.SendMessageToVpnIp(header.Test, header.TestRequest, hostinfo.vpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
+		}
 
-	msg, eKey, dKey, err := ci.H.ReadMessage(nil, packet[HeaderLen:])
+		// We already have a complete tunnel, there is nothing that can be done by processing further stage 1 packets
+		return false
+	}
+
+	msg, eKey, dKey, err := ci.H.ReadMessage(nil, packet[header.Len:])
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).WithField("header", h).
 			Error("Failed to call noise.ReadMessage")
 
@@ -304,65 +354,112 @@ func ixHandshakeStage2(f *Interface, addr *udpAddr, hostinfo *HostInfo, packet [
 		// near future
 		return false
 	} else if dKey == nil || eKey == nil {
-		l.WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
 			Error("Noise did not arrive at a key")
+
+		// This should be impossible in IX but just in case, if we get here then there is no chance to recover
+		// the handshake state machine. Tear it down
 		return true
 	}
 
 	hs := &NebulaHandshake{}
 	err = proto.Unmarshal(msg, hs)
 	if err != nil || hs.Details == nil {
-		l.WithError(err).WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).Error("Failed unmarshal handshake message")
+
+		// The handshake state machine is complete, if things break now there is no chance to recover. Tear down and start again
 		return true
 	}
 
-	remoteCert, err := RecombineCertAndValidate(ci.H, hs.Details.Cert)
+	remoteCert, err := RecombineCertAndValidate(ci.H, hs.Details.Cert, f.caPool)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(hostinfo.hostId)).WithField("udpAddr", addr).
+		f.l.WithError(err).WithField("vpnIp", hostinfo.vpnIp).WithField("udpAddr", addr).
 			WithField("cert", remoteCert).WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
 			Error("Invalid certificate from host")
+
+		// The handshake state machine is complete, if things break now there is no chance to recover. Tear down and start again
 		return true
 	}
-	vpnIP := ip2int(remoteCert.Details.Ips[0].IP)
+
+	vpnIp := iputil.Ip2VpnIp(remoteCert.Details.Ips[0].IP)
 	certName := remoteCert.Details.Name
 	fingerprint, _ := remoteCert.Sha256Sum()
+	issuer := remoteCert.Details.Issuer
+
+	// Ensure the right host responded
+	if vpnIp != hostinfo.vpnIp {
+		f.l.WithField("intendedVpnIp", hostinfo.vpnIp).WithField("haveVpnIp", vpnIp).
+			WithField("udpAddr", addr).WithField("certName", certName).
+			WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
+			Info("Incorrect host responded to handshake")
+
+		// Release our old handshake from pending, it should not continue
+		f.handshakeManager.pendingHostMap.DeleteHostInfo(hostinfo)
+
+		// Create a new hostinfo/handshake for the intended vpn ip
+		//TODO: this adds it to the timer wheel in a way that aggressively retries
+		newHostInfo := f.getOrHandshake(hostinfo.vpnIp)
+		newHostInfo.Lock()
+
+		// Block the current used address
+		newHostInfo.remotes = hostinfo.remotes
+		newHostInfo.remotes.BlockRemote(addr)
+
+		// Get the correct remote list for the host we did handshake with
+		hostinfo.remotes = f.lightHouse.QueryCache(vpnIp)
+
+		f.l.WithField("blockedUdpAddrs", newHostInfo.remotes.CopyBlockedRemotes()).WithField("vpnIp", vpnIp).
+			WithField("remotes", newHostInfo.remotes.CopyAddrs(f.hostMap.preferredRanges)).
+			Info("Blocked addresses for handshakes")
+
+		// Swap the packet store to benefit the original intended recipient
+		hostinfo.ConnectionState.queueLock.Lock()
+		newHostInfo.packetStore = hostinfo.packetStore
+		hostinfo.packetStore = []*cachedPacket{}
+		hostinfo.ConnectionState.queueLock.Unlock()
+
+		// Finally, put the correct vpn ip in the host info, tell them to close the tunnel, and return true to tear down
+		hostinfo.vpnIp = vpnIp
+		f.sendCloseTunnel(hostinfo)
+		newHostInfo.Unlock()
+
+		return true
+	}
+
+	// Mark packet 2 as seen so it doesn't show up as missed
+	ci.window.Update(f.l, 2)
 
 	duration := time.Since(hostinfo.handshakeStart).Nanoseconds()
-	l.WithField("vpnIp", IntIp(vpnIP)).WithField("udpAddr", addr).
+	f.l.WithField("vpnIp", vpnIp).WithField("udpAddr", addr).
 		WithField("certName", certName).
 		WithField("fingerprint", fingerprint).
+		WithField("issuer", issuer).
 		WithField("initiatorIndex", hs.Details.InitiatorIndex).WithField("responderIndex", hs.Details.ResponderIndex).
 		WithField("remoteIndex", h.RemoteIndex).WithField("handshake", m{"stage": 2, "style": "ix_psk0"}).
 		WithField("durationNs", duration).
+		WithField("sentCachedPackets", len(hostinfo.packetStore)).
 		Info("Handshake message received")
 
-	//ci.remoteIndex = hs.ResponderIndex
 	hostinfo.remoteIndexId = hs.Details.ResponderIndex
-	hs.Details.Cert = ci.certState.rawCertificateNoKey
-
-	/*
-		hsBytes, err := proto.Marshal(hs)
-		if err != nil {
-			l.Debugln("Failed to marshal handshake: ", err)
-			return
-		}
-	*/
-
-	// Regardless of whether you are the sender or receiver, you should arrive here
-	// and complete standing up the connection.
+	hostinfo.lastHandshakeTime = hs.Details.Time
 
+	// Store their cert and our symmetric keys
 	ci.peerCert = remoteCert
 	ci.dKey = NewNebulaCipherState(dKey)
 	ci.eKey = NewNebulaCipherState(eKey)
-	//l.Debugln("got symmetric pairs")
 
-	hostinfo.SetRemote(*addr)
+	// Make sure the current udpAddr being used is set for responding
+	hostinfo.SetRemote(addr)
+
+	// Build up the radix for the firewall if we have subnets in the cert
 	hostinfo.CreateRemoteCIDR(remoteCert)
 
+	// Complete our handshake and update metrics, this will replace any existing tunnels for this vpnIp
+	//TODO: Complete here does not do a race avoidance, it will just take the new tunnel. Is this ok?
 	f.handshakeManager.Complete(hostinfo, f)
-	hostinfo.handshakeComplete()
+	hostinfo.handshakeComplete(f.l, f.cachedPacketMetrics)
 	f.metricHandshakes.Update(duration)
 
 	return false

handshake_manager.go

@@ -2,22 +2,23 @@ package nebula
 
 import (
 	"bytes"
+	"context"
 	"crypto/rand"
 	"encoding/binary"
 	"errors"
 	"net"
 	"time"
 
+	"github.com/rcrowley/go-metrics"
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
 const (
-	// Total time to try a handshake = sequence of HandshakeTryInterval * HandshakeRetries
-	// With 100ms interval and 20 retries is 23.5 seconds
-	DefaultHandshakeTryInterval = time.Millisecond * 100
-	DefaultHandshakeRetries     = 20
-	// DefaultHandshakeWaitRotation is the number of handshake attempts to do before starting to use other ips addresses
-	DefaultHandshakeWaitRotation  = 5
+	DefaultHandshakeTryInterval   = time.Millisecond * 100
+	DefaultHandshakeRetries       = 10
 	DefaultHandshakeTriggerBuffer = 64
 )
 
@@ -25,7 +26,6 @@ var (
 	defaultHandshakeConfig = HandshakeConfig{
 		tryInterval:   DefaultHandshakeTryInterval,
 		retries:       DefaultHandshakeRetries,
-		waitRotation:  DefaultHandshakeWaitRotation,
 		triggerBuffer: DefaultHandshakeTriggerBuffer,
 	}
 )
@@ -33,166 +33,171 @@ var (
 type HandshakeConfig struct {
 	tryInterval   time.Duration
 	retries       int
-	waitRotation  int
 	triggerBuffer int
 
 	messageMetrics *MessageMetrics
 }
 
 type HandshakeManager struct {
-	pendingHostMap *HostMap
-	mainHostMap    *HostMap
-	lightHouse     *LightHouse
-	outside        *udpConn
-	config         HandshakeConfig
-
-	// can be used to trigger outbound handshake for the given vpnIP
-	trigger chan uint32
-
+	pendingHostMap         *HostMap
+	mainHostMap            *HostMap
+	lightHouse             *LightHouse
+	outside                *udp.Conn
+	config                 HandshakeConfig
 	OutboundHandshakeTimer *SystemTimerWheel
-	InboundHandshakeTimer  *SystemTimerWheel
+	messageMetrics         *MessageMetrics
+	metricInitiated        metrics.Counter
+	metricTimedOut         metrics.Counter
+	l                      *logrus.Logger
 
-	messageMetrics *MessageMetrics
+	// can be used to trigger outbound handshake for the given vpnIp
+	trigger chan iputil.VpnIp
 }
 
-func NewHandshakeManager(tunCidr *net.IPNet, preferredRanges []*net.IPNet, mainHostMap *HostMap, lightHouse *LightHouse, outside *udpConn, config HandshakeConfig) *HandshakeManager {
+func NewHandshakeManager(l *logrus.Logger, tunCidr *net.IPNet, preferredRanges []*net.IPNet, mainHostMap *HostMap, lightHouse *LightHouse, outside *udp.Conn, config HandshakeConfig) *HandshakeManager {
 	return &HandshakeManager{
-		pendingHostMap: NewHostMap("pending", tunCidr, preferredRanges),
-		mainHostMap:    mainHostMap,
-		lightHouse:     lightHouse,
-		outside:        outside,
-
-		config: config,
-
-		trigger: make(chan uint32, config.triggerBuffer),
-
-		OutboundHandshakeTimer: NewSystemTimerWheel(config.tryInterval, config.tryInterval*time.Duration(config.retries)),
-		InboundHandshakeTimer:  NewSystemTimerWheel(config.tryInterval, config.tryInterval*time.Duration(config.retries)),
-
-		messageMetrics: config.messageMetrics,
+		pendingHostMap:         NewHostMap(l, "pending", tunCidr, preferredRanges),
+		mainHostMap:            mainHostMap,
+		lightHouse:             lightHouse,
+		outside:                outside,
+		config:                 config,
+		trigger:                make(chan iputil.VpnIp, config.triggerBuffer),
+		OutboundHandshakeTimer: NewSystemTimerWheel(config.tryInterval, hsTimeout(config.retries, config.tryInterval)),
+		messageMetrics:         config.messageMetrics,
+		metricInitiated:        metrics.GetOrRegisterCounter("handshake_manager.initiated", nil),
+		metricTimedOut:         metrics.GetOrRegisterCounter("handshake_manager.timed_out", nil),
+		l:                      l,
 	}
 }
 
-func (c *HandshakeManager) Run(f EncWriter) {
-	clockSource := time.Tick(c.config.tryInterval)
+func (c *HandshakeManager) Run(ctx context.Context, f udp.EncWriter) {
+	clockSource := time.NewTicker(c.config.tryInterval)
+	defer clockSource.Stop()
+
 	for {
 		select {
+		case <-ctx.Done():
+			return
 		case vpnIP := <-c.trigger:
-			l.WithField("vpnIp", IntIp(vpnIP)).Debug("HandshakeManager: triggered")
+			c.l.WithField("vpnIp", vpnIP).Debug("HandshakeManager: triggered")
 			c.handleOutbound(vpnIP, f, true)
-		case now := <-clockSource:
+		case now := <-clockSource.C:
 			c.NextOutboundHandshakeTimerTick(now, f)
-			c.NextInboundHandshakeTimerTick(now)
 		}
 	}
 }
 
-func (c *HandshakeManager) NextOutboundHandshakeTimerTick(now time.Time, f EncWriter) {
+func (c *HandshakeManager) NextOutboundHandshakeTimerTick(now time.Time, f udp.EncWriter) {
 	c.OutboundHandshakeTimer.advance(now)
 	for {
 		ep := c.OutboundHandshakeTimer.Purge()
 		if ep == nil {
 			break
 		}
-		vpnIP := ep.(uint32)
-		c.handleOutbound(vpnIP, f, false)
+		vpnIp := ep.(iputil.VpnIp)
+		c.handleOutbound(vpnIp, f, false)
 	}
 }
 
-func (c *HandshakeManager) handleOutbound(vpnIP uint32, f EncWriter, lighthouseTriggered bool) {
-	hostinfo, err := c.pendingHostMap.QueryVpnIP(vpnIP)
+func (c *HandshakeManager) handleOutbound(vpnIp iputil.VpnIp, f udp.EncWriter, lighthouseTriggered bool) {
+	hostinfo, err := c.pendingHostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return
 	}
 	hostinfo.Lock()
 	defer hostinfo.Unlock()
 
-	// If we haven't finished the handshake and we haven't hit max retries, query
-	// lighthouse and then send the handshake packet again.
-	if hostinfo.HandshakeCounter < c.config.retries && !hostinfo.HandshakeComplete {
-		if hostinfo.remote == nil {
-			// We continue to query the lighthouse because hosts may
-			// come online during handshake retries. If the query
-			// succeeds (no error), add the lighthouse info to hostinfo
-			ips := c.lightHouse.QueryCache(vpnIP)
-			// If we have no responses yet, or only one IP (the host hadn't
-			// finished reporting its own IPs yet), then send another query to
-			// the LH.
-			if len(ips) <= 1 {
-				ips, err = c.lightHouse.Query(vpnIP, f)
-			}
-			if err == nil {
-				for _, ip := range ips {
-					hostinfo.AddRemote(ip)
-				}
-				hostinfo.ForcePromoteBest(c.mainHostMap.preferredRanges)
-			}
-		} else if lighthouseTriggered {
-			// We were triggered by a lighthouse HostQueryReply packet, but
-			// we have already picked a remote for this host (this can happen
-			// if we are configured with multiple lighthouses). So we can skip
-			// this trigger and let the timerwheel handle the rest of the
-			// process
-			return
-		}
-
-		hostinfo.HandshakeCounter++
-
-		// We want to use the "best" calculated ip for the first 5 attempts, after that we just blindly rotate through
-		// all the others until we can stand up a connection.
-		if hostinfo.HandshakeCounter > c.config.waitRotation {
-			hostinfo.rotateRemote()
-		}
-
-		// Ensure the handshake is ready to avoid a race in timer tick and stage 0 handshake generation
-		if hostinfo.HandshakeReady && hostinfo.remote != nil {
-			c.messageMetrics.Tx(handshake, NebulaMessageSubType(hostinfo.HandshakePacket[0][1]), 1)
-			err := c.outside.WriteTo(hostinfo.HandshakePacket[0], hostinfo.remote)
-			if err != nil {
-				hostinfo.logger().WithField("udpAddr", hostinfo.remote).
-					WithField("initiatorIndex", hostinfo.localIndexId).
-					WithField("remoteIndex", hostinfo.remoteIndexId).
-					WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
-					WithError(err).Error("Failed to send handshake message")
-			} else {
-				//TODO: this log line is assuming a lot of stuff around the cached stage 0 handshake packet, we should
-				// keep the real packet struct around for logging purposes
-				hostinfo.logger().WithField("udpAddr", hostinfo.remote).
-					WithField("initiatorIndex", hostinfo.localIndexId).
-					WithField("remoteIndex", hostinfo.remoteIndexId).
-					WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
-					Info("Handshake message sent")
-			}
-		}
-
-		// Readd to the timer wheel so we continue trying wait HandshakeTryInterval * counter longer for next try
-		if !lighthouseTriggered {
-			//l.Infoln("Interval: ", HandshakeTryInterval*time.Duration(hostinfo.HandshakeCounter))
-			c.OutboundHandshakeTimer.Add(vpnIP, c.config.tryInterval*time.Duration(hostinfo.HandshakeCounter))
-		}
-	} else {
+	// We may have raced to completion but now that we have a lock we should ensure we have not yet completed.
+	if hostinfo.HandshakeComplete {
+		// Ensure we don't exist in the pending hostmap anymore since we have completed
 		c.pendingHostMap.DeleteHostInfo(hostinfo)
+		return
 	}
-}
 
-func (c *HandshakeManager) NextInboundHandshakeTimerTick(now time.Time) {
-	c.InboundHandshakeTimer.advance(now)
-	for {
-		ep := c.InboundHandshakeTimer.Purge()
-		if ep == nil {
-			break
+	// Check if we have a handshake packet to transmit yet
+	if !hostinfo.HandshakeReady {
+		// There is currently a slight race in getOrHandshake due to ConnectionState not being part of the HostInfo directly
+		// Our hostinfo here was added to the pending map and the wheel may have ticked to us before we created ConnectionState
+		c.OutboundHandshakeTimer.Add(vpnIp, c.config.tryInterval*time.Duration(hostinfo.HandshakeCounter))
+		return
+	}
+
+	// If we are out of time, clean up
+	if hostinfo.HandshakeCounter >= c.config.retries {
+		hostinfo.logger(c.l).WithField("udpAddrs", hostinfo.remotes.CopyAddrs(c.pendingHostMap.preferredRanges)).
+			WithField("initiatorIndex", hostinfo.localIndexId).
+			WithField("remoteIndex", hostinfo.remoteIndexId).
+			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
+			WithField("durationNs", time.Since(hostinfo.handshakeStart).Nanoseconds()).
+			Info("Handshake timed out")
+		c.metricTimedOut.Inc(1)
+		c.pendingHostMap.DeleteHostInfo(hostinfo)
+		return
+	}
+
+	// We only care about a lighthouse trigger before the first handshake transmit attempt. This is a very specific
+	// optimization for a fast lighthouse reply
+	//TODO: it would feel better to do this once, anytime, as our delay increases over time
+	if lighthouseTriggered && hostinfo.HandshakeCounter > 0 {
+		// If we didn't return here a lighthouse could cause us to aggressively send handshakes
+		return
+	}
+
+	// Get a remotes object if we don't already have one.
+	// This is mainly to protect us as this should never be the case
+	if hostinfo.remotes == nil {
+		hostinfo.remotes = c.lightHouse.QueryCache(vpnIp)
+	}
+
+	//TODO: this will generate a load of queries for hosts with only 1 ip (i'm not using a lighthouse, static mapped)
+	if hostinfo.remotes.Len(c.pendingHostMap.preferredRanges) <= 1 {
+		// If we only have 1 remote it is highly likely our query raced with the other host registered within the lighthouse
+		// Our vpnIp here has a tunnel with a lighthouse but has yet to send a host update packet there so we only know about
+		// the learned public ip for them. Query again to short circuit the promotion counter
+		c.lightHouse.QueryServer(vpnIp, f)
+	}
+
+	// Send a the handshake to all known ips, stage 2 takes care of assigning the hostinfo.remote based on the first to reply
+	var sentTo []*udp.Addr
+	hostinfo.remotes.ForEach(c.pendingHostMap.preferredRanges, func(addr *udp.Addr, _ bool) {
+		c.messageMetrics.Tx(header.Handshake, header.MessageSubType(hostinfo.HandshakePacket[0][1]), 1)
+		err = c.outside.WriteTo(hostinfo.HandshakePacket[0], addr)
+		if err != nil {
+			hostinfo.logger(c.l).WithField("udpAddr", addr).
+				WithField("initiatorIndex", hostinfo.localIndexId).
+				WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
+				WithError(err).Error("Failed to send handshake message")
+
+		} else {
+			sentTo = append(sentTo, addr)
 		}
-		index := ep.(uint32)
+	})
 
-		c.pendingHostMap.DeleteIndex(index)
+	// Don't be too noisy or confusing if we fail to send a handshake - if we don't get through we'll eventually log a timeout
+	if len(sentTo) > 0 {
+		hostinfo.logger(c.l).WithField("udpAddrs", sentTo).
+			WithField("initiatorIndex", hostinfo.localIndexId).
+			WithField("handshake", m{"stage": 1, "style": "ix_psk0"}).
+			Info("Handshake message sent")
+	}
+
+	// Increment the counter to increase our delay, linear backoff
+	hostinfo.HandshakeCounter++
+
+	// If a lighthouse triggered this attempt then we are still in the timer wheel and do not need to re-add
+	if !lighthouseTriggered {
+		//TODO: feel like we dupe handshake real fast in a tight loop, why?
+		c.OutboundHandshakeTimer.Add(vpnIp, c.config.tryInterval*time.Duration(hostinfo.HandshakeCounter))
 	}
 }
 
-func (c *HandshakeManager) AddVpnIP(vpnIP uint32) *HostInfo {
-	hostinfo := c.pendingHostMap.AddVpnIP(vpnIP)
-	// We lock here and use an array to insert items to prevent locking the
-	// main receive thread for very long by waiting to add items to the pending map
-	c.OutboundHandshakeTimer.Add(vpnIP, c.config.tryInterval)
+func (c *HandshakeManager) AddVpnIp(vpnIp iputil.VpnIp, init func(*HostInfo)) *HostInfo {
+	hostinfo, created := c.pendingHostMap.AddVpnIp(vpnIp, init)
+
+	if created {
+		c.OutboundHandshakeTimer.Add(vpnIp, c.config.tryInterval)
+		c.metricInitiated.Inc(1)
+	}
 
 	return hostinfo
 }
@@ -201,33 +206,40 @@ var (
 	ErrExistingHostInfo    = errors.New("existing hostinfo")
 	ErrAlreadySeen         = errors.New("already seen")
 	ErrLocalIndexCollision = errors.New("local index collision")
+	ErrExistingHandshake   = errors.New("existing handshake")
 )
 
 // CheckAndComplete checks for any conflicts in the main and pending hostmap
 // before adding hostinfo to main. If err is nil, it was added. Otherwise err will be:
-
+//
 // ErrAlreadySeen if we already have an entry in the hostmap that has seen the
 // exact same handshake packet
 //
 // ErrExistingHostInfo if we already have an entry in the hostmap for this
-// VpnIP and overwrite was false.
+// VpnIp and the new handshake was older than the one we currently have
 //
 // ErrLocalIndexCollision if we already have an entry in the main or pending
 // hostmap for the hostinfo.localIndexId.
 func (c *HandshakeManager) CheckAndComplete(hostinfo *HostInfo, handshakePacket uint8, overwrite bool, f *Interface) (*HostInfo, error) {
-	c.pendingHostMap.RLock()
-	defer c.pendingHostMap.RUnlock()
+	c.pendingHostMap.Lock()
+	defer c.pendingHostMap.Unlock()
 	c.mainHostMap.Lock()
 	defer c.mainHostMap.Unlock()
 
-	existingHostInfo, found := c.mainHostMap.Hosts[hostinfo.hostId]
+	// Check if we already have a tunnel with this vpn ip
+	existingHostInfo, found := c.mainHostMap.Hosts[hostinfo.vpnIp]
 	if found && existingHostInfo != nil {
+		// Is it just a delayed handshake packet?
 		if bytes.Equal(hostinfo.HandshakePacket[handshakePacket], existingHostInfo.HandshakePacket[handshakePacket]) {
 			return existingHostInfo, ErrAlreadySeen
 		}
-		if !overwrite {
+
+		// Is this a newer handshake?
+		if existingHostInfo.lastHandshakeTime >= hostinfo.lastHandshakeTime {
 			return existingHostInfo, ErrExistingHostInfo
 		}
+
+		existingHostInfo.logger(c.l).Info("Taking new handshake")
 	}
 
 	existingIndex, found := c.mainHostMap.Indexes[hostinfo.localIndexId]
@@ -235,6 +247,7 @@ func (c *HandshakeManager) CheckAndComplete(hostinfo *HostInfo, handshakePacket
 		// We have a collision, but for a different hostinfo
 		return existingIndex, ErrLocalIndexCollision
 	}
+
 	existingIndex, found = c.pendingHostMap.Indexes[hostinfo.localIndexId]
 	if found && existingIndex != hostinfo {
 		// We have a collision, but for a different hostinfo
@@ -242,17 +255,33 @@ func (c *HandshakeManager) CheckAndComplete(hostinfo *HostInfo, handshakePacket
 	}
 
 	existingRemoteIndex, found := c.mainHostMap.RemoteIndexes[hostinfo.remoteIndexId]
-	if found && existingRemoteIndex != nil && existingRemoteIndex.hostId != hostinfo.hostId {
+	if found && existingRemoteIndex != nil && existingRemoteIndex.vpnIp != hostinfo.vpnIp {
 		// We have a collision, but this can happen since we can't control
 		// the remote ID. Just log about the situation as a note.
-		hostinfo.logger().
-			WithField("remoteIndex", hostinfo.remoteIndexId).WithField("collision", IntIp(existingRemoteIndex.hostId)).
+		hostinfo.logger(c.l).
+			WithField("remoteIndex", hostinfo.remoteIndexId).WithField("collision", existingRemoteIndex.vpnIp).
 			Info("New host shadows existing host remoteIndex")
 	}
 
+	// Check if we are also handshaking with this vpn ip
+	pendingHostInfo, found := c.pendingHostMap.Hosts[hostinfo.vpnIp]
+	if found && pendingHostInfo != nil {
+		if !overwrite {
+			// We won, let our pending handshake win
+			return pendingHostInfo, ErrExistingHandshake
+		}
+
+		// We lost, take this handshake and move any cached packets over so they get sent
+		pendingHostInfo.ConnectionState.queueLock.Lock()
+		hostinfo.packetStore = append(hostinfo.packetStore, pendingHostInfo.packetStore...)
+		c.pendingHostMap.unlockedDeleteHostInfo(pendingHostInfo)
+		pendingHostInfo.ConnectionState.queueLock.Unlock()
+		pendingHostInfo.logger(c.l).Info("Handshake race lost, replacing pending handshake with completed tunnel")
+	}
+
 	if existingHostInfo != nil {
 		// We are going to overwrite this entry, so remove the old references
-		delete(c.mainHostMap.Hosts, existingHostInfo.hostId)
+		delete(c.mainHostMap.Hosts, existingHostInfo.vpnIp)
 		delete(c.mainHostMap.Indexes, existingHostInfo.localIndexId)
 		delete(c.mainHostMap.RemoteIndexes, existingHostInfo.remoteIndexId)
 	}
@@ -265,13 +294,15 @@ func (c *HandshakeManager) CheckAndComplete(hostinfo *HostInfo, handshakePacket
 // won't have a localIndexId collision because we already have an entry in the
 // pendingHostMap
 func (c *HandshakeManager) Complete(hostinfo *HostInfo, f *Interface) {
+	c.pendingHostMap.Lock()
+	defer c.pendingHostMap.Unlock()
 	c.mainHostMap.Lock()
 	defer c.mainHostMap.Unlock()
 
-	existingHostInfo, found := c.mainHostMap.Hosts[hostinfo.hostId]
+	existingHostInfo, found := c.mainHostMap.Hosts[hostinfo.vpnIp]
 	if found && existingHostInfo != nil {
 		// We are going to overwrite this entry, so remove the old references
-		delete(c.mainHostMap.Hosts, existingHostInfo.hostId)
+		delete(c.mainHostMap.Hosts, existingHostInfo.vpnIp)
 		delete(c.mainHostMap.Indexes, existingHostInfo.localIndexId)
 		delete(c.mainHostMap.RemoteIndexes, existingHostInfo.remoteIndexId)
 	}
@@ -280,12 +311,13 @@ func (c *HandshakeManager) Complete(hostinfo *HostInfo, f *Interface) {
 	if found && existingRemoteIndex != nil {
 		// We have a collision, but this can happen since we can't control
 		// the remote ID. Just log about the situation as a note.
-		hostinfo.logger().
-			WithField("remoteIndex", hostinfo.remoteIndexId).WithField("collision", IntIp(existingRemoteIndex.hostId)).
+		hostinfo.logger(c.l).
+			WithField("remoteIndex", hostinfo.remoteIndexId).WithField("collision", existingRemoteIndex.vpnIp).
 			Info("New host shadows existing host remoteIndex")
 	}
 
 	c.mainHostMap.addHostInfo(hostinfo, f)
+	c.pendingHostMap.unlockedDeleteHostInfo(hostinfo)
 }
 
 // AddIndexHostInfo generates a unique localIndexId for this HostInfo
@@ -298,7 +330,7 @@ func (c *HandshakeManager) AddIndexHostInfo(h *HostInfo) error {
 	defer c.mainHostMap.RUnlock()
 
 	for i := 0; i < 32; i++ {
-		index, err := generateIndex()
+		index, err := generateIndex(c.l)
 		if err != nil {
 			return err
 		}
@@ -336,7 +368,7 @@ func (c *HandshakeManager) EmitStats() {
 
 // Utility functions below
 
-func generateIndex() (uint32, error) {
+func generateIndex(l *logrus.Logger) (uint32, error) {
 	b := make([]byte, 4)
 
 	// Let zero mean we don't know the ID, so don't generate zero
@@ -357,3 +389,7 @@ func generateIndex() (uint32, error) {
 	}
 	return index, nil
 }
+
+func hsTimeout(tries int, interval time.Duration) time.Duration {
+	return time.Duration(tries / 2 * ((2 * int(interval)) + (tries-1)*int(interval)))
+}

handshake_manager_test.go

@@ -5,148 +5,103 @@ import (
 	"testing"
 	"time"
 
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
-//var ips []uint32 = []uint32{9000, 9999999, 3, 292394923}
-var ips []uint32
-
-func Test_NewHandshakeManagerIndex(t *testing.T) {
-
+func Test_NewHandshakeManagerVpnIp(t *testing.T) {
+	l := util.NewTestLogger()
 	_, tuncidr, _ := net.ParseCIDR("172.1.1.1/24")
 	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
 	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	ips = []uint32{ip2int(net.ParseIP("172.1.1.2"))}
-	preferredRanges := []*net.IPNet{localrange}
-	mainHM := NewHostMap("test", vpncidr, preferredRanges)
-
-	blah := NewHandshakeManager(tuncidr, preferredRanges, mainHM, &LightHouse{}, &udpConn{}, defaultHandshakeConfig)
-
-	now := time.Now()
-	blah.NextInboundHandshakeTimerTick(now)
-
-	var indexes = make([]uint32, 4)
-	var hostinfo = make([]*HostInfo, len(indexes))
-	for i := range indexes {
-		hostinfo[i] = &HostInfo{ConnectionState: &ConnectionState{}}
-	}
-
-	// Add four indexes
-	for i := range indexes {
-		err := blah.AddIndexHostInfo(hostinfo[i])
-		assert.NoError(t, err)
-		indexes[i] = hostinfo[i].localIndexId
-		blah.InboundHandshakeTimer.Add(indexes[i], time.Second*10)
-	}
-	// Confirm they are in the pending index list
-	for _, v := range indexes {
-		assert.Contains(t, blah.pendingHostMap.Indexes, uint32(v))
-	}
-	// Adding something to pending should not affect the main hostmap
-	assert.Len(t, mainHM.Indexes, 0)
-	// Jump ahead 8 seconds
-	for i := 1; i <= DefaultHandshakeRetries; i++ {
-		next_tick := now.Add(DefaultHandshakeTryInterval * time.Duration(i))
-		blah.NextInboundHandshakeTimerTick(next_tick)
-	}
-	// Confirm they are still in the pending index list
-	for _, v := range indexes {
-		assert.Contains(t, blah.pendingHostMap.Indexes, uint32(v))
-	}
-	// Jump ahead 4 more seconds
-	next_tick := now.Add(12 * time.Second)
-	blah.NextInboundHandshakeTimerTick(next_tick)
-	// Confirm they have been removed
-	for _, v := range indexes {
-		assert.NotContains(t, blah.pendingHostMap.Indexes, uint32(v))
-	}
-}
-
-func Test_NewHandshakeManagerVpnIP(t *testing.T) {
-	_, tuncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	ips = []uint32{ip2int(net.ParseIP("172.1.1.2"))}
+	ip := iputil.Ip2VpnIp(net.ParseIP("172.1.1.2"))
 	preferredRanges := []*net.IPNet{localrange}
 	mw := &mockEncWriter{}
-	mainHM := NewHostMap("test", vpncidr, preferredRanges)
+	mainHM := NewHostMap(l, "test", vpncidr, preferredRanges)
 
-	blah := NewHandshakeManager(tuncidr, preferredRanges, mainHM, &LightHouse{}, &udpConn{}, defaultHandshakeConfig)
+	blah := NewHandshakeManager(l, tuncidr, preferredRanges, mainHM, &LightHouse{}, &udp.Conn{}, defaultHandshakeConfig)
 
 	now := time.Now()
 	blah.NextOutboundHandshakeTimerTick(now, mw)
 
-	// Add four "IPs" - which are just uint32s
-	for _, v := range ips {
-		blah.AddVpnIP(v)
-	}
-	// Adding something to pending should not affect the main hostmap
-	assert.Len(t, mainHM.Hosts, 0)
-	// Confirm they are in the pending index list
-	for _, v := range ips {
-		assert.Contains(t, blah.pendingHostMap.Hosts, uint32(v))
+	var initCalled bool
+	initFunc := func(*HostInfo) {
+		initCalled = true
 	}
 
-	// Jump ahead `HandshakeRetries` ticks
-	cumulative := time.Duration(0)
-	for i := 0; i <= DefaultHandshakeRetries+1; i++ {
-		cumulative += time.Duration(i)*DefaultHandshakeTryInterval + 1
-		next_tick := now.Add(cumulative)
-		//l.Infoln(next_tick)
-		blah.NextOutboundHandshakeTimerTick(next_tick, mw)
+	i := blah.AddVpnIp(ip, initFunc)
+	assert.True(t, initCalled)
+
+	initCalled = false
+	i2 := blah.AddVpnIp(ip, initFunc)
+	assert.False(t, initCalled)
+	assert.Same(t, i, i2)
+
+	i.remotes = NewRemoteList()
+	i.HandshakeReady = true
+
+	// Adding something to pending should not affect the main hostmap
+	assert.Len(t, mainHM.Hosts, 0)
+
+	// Confirm they are in the pending index list
+	assert.Contains(t, blah.pendingHostMap.Hosts, ip)
+
+	// Jump ahead `HandshakeRetries` ticks, offset by one to get the sleep logic right
+	for i := 1; i <= DefaultHandshakeRetries+1; i++ {
+		now = now.Add(time.Duration(i) * DefaultHandshakeTryInterval)
+		blah.NextOutboundHandshakeTimerTick(now, mw)
 	}
 
 	// Confirm they are still in the pending index list
-	for _, v := range ips {
-		assert.Contains(t, blah.pendingHostMap.Hosts, uint32(v))
-	}
-	// Jump ahead 1 more second
-	cumulative += time.Duration(DefaultHandshakeRetries+1) * DefaultHandshakeTryInterval
-	next_tick := now.Add(cumulative)
-	//l.Infoln(next_tick)
-	blah.NextOutboundHandshakeTimerTick(next_tick, mw)
+	assert.Contains(t, blah.pendingHostMap.Hosts, ip)
+
+	// Tick 1 more time, a minute will certainly flush it out
+	blah.NextOutboundHandshakeTimerTick(now.Add(time.Minute), mw)
+
 	// Confirm they have been removed
-	for _, v := range ips {
-		assert.NotContains(t, blah.pendingHostMap.Hosts, uint32(v))
-	}
+	assert.NotContains(t, blah.pendingHostMap.Hosts, ip)
 }
 
 func Test_NewHandshakeManagerTrigger(t *testing.T) {
+	l := util.NewTestLogger()
 	_, tuncidr, _ := net.ParseCIDR("172.1.1.1/24")
 	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
 	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	ip := ip2int(net.ParseIP("172.1.1.2"))
+	ip := iputil.Ip2VpnIp(net.ParseIP("172.1.1.2"))
 	preferredRanges := []*net.IPNet{localrange}
 	mw := &mockEncWriter{}
-	mainHM := NewHostMap("test", vpncidr, preferredRanges)
-	lh := &LightHouse{}
+	mainHM := NewHostMap(l, "test", vpncidr, preferredRanges)
+	lh := &LightHouse{addrMap: make(map[iputil.VpnIp]*RemoteList), l: l}
 
-	blah := NewHandshakeManager(tuncidr, preferredRanges, mainHM, lh, &udpConn{}, defaultHandshakeConfig)
+	blah := NewHandshakeManager(l, tuncidr, preferredRanges, mainHM, lh, &udp.Conn{}, defaultHandshakeConfig)
 
 	now := time.Now()
 	blah.NextOutboundHandshakeTimerTick(now, mw)
 
 	assert.Equal(t, 0, testCountTimerWheelEntries(blah.OutboundHandshakeTimer))
 
-	blah.AddVpnIP(ip)
+	hi := blah.AddVpnIp(ip, nil)
+	hi.HandshakeReady = true
+	assert.Equal(t, 1, testCountTimerWheelEntries(blah.OutboundHandshakeTimer))
+	assert.Equal(t, 0, hi.HandshakeCounter, "Should not have attempted a handshake yet")
 
+	// Trigger the same method the channel will but, this should set our remotes pointer
+	blah.handleOutbound(ip, mw, true)
+	assert.Equal(t, 1, hi.HandshakeCounter, "Trigger should have done a handshake attempt")
+	assert.NotNil(t, hi.remotes, "Manager should have set my remotes pointer")
+
+	// Make sure the trigger doesn't double schedule the timer entry
 	assert.Equal(t, 1, testCountTimerWheelEntries(blah.OutboundHandshakeTimer))
 
-	// Trigger the same method the channel will
+	uaddr := udp.NewAddrFromString("10.1.1.1:4242")
+	hi.remotes.unlockedPrependV4(ip, NewIp4AndPort(uaddr.IP, uint32(uaddr.Port)))
+
+	// We now have remotes but only the first trigger should have pushed things forward
 	blah.handleOutbound(ip, mw, true)
-
-	// Make sure the trigger doesn't schedule another timer entry
-	assert.Equal(t, 1, testCountTimerWheelEntries(blah.OutboundHandshakeTimer))
-	hi := blah.pendingHostMap.Hosts[ip]
-	assert.Nil(t, hi.remote)
-
-	lh.addrMap = map[uint32][]udpAddr{
-		ip: {*NewUDPAddrFromString("10.1.1.1:4242")},
-	}
-
-	// This should trigger the hostmap to populate the hostinfo
-	blah.handleOutbound(ip, mw, true)
-	assert.NotNil(t, hi.remote)
+	assert.Equal(t, 1, hi.HandshakeCounter, "Trigger should have not done a handshake attempt")
 	assert.Equal(t, 1, testCountTimerWheelEntries(blah.OutboundHandshakeTimer))
 }
 
@@ -161,98 +116,9 @@ func testCountTimerWheelEntries(tw *SystemTimerWheel) (c int) {
 	return c
 }
 
-func Test_NewHandshakeManagerVpnIPcleanup(t *testing.T) {
-	_, tuncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, 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}
-	mw := &mockEncWriter{}
-	mainHM := NewHostMap("test", vpncidr, preferredRanges)
-
-	blah := NewHandshakeManager(tuncidr, preferredRanges, mainHM, &LightHouse{}, &udpConn{}, defaultHandshakeConfig)
-
-	now := time.Now()
-	blah.NextOutboundHandshakeTimerTick(now, mw)
-
-	hostinfo := blah.AddVpnIP(vpnIP)
-	// Pretned we have an index too
-	err := blah.AddIndexHostInfo(hostinfo)
-	assert.NoError(t, err)
-	blah.InboundHandshakeTimer.Add(hostinfo.localIndexId, time.Second*10)
-	assert.NotZero(t, hostinfo.localIndexId)
-	assert.Contains(t, blah.pendingHostMap.Indexes, hostinfo.localIndexId)
-
-	// Jump ahead `HandshakeRetries` ticks. Eviction should happen in pending
-	// but not main hostmap
-	cumulative := time.Duration(0)
-	for i := 1; i <= DefaultHandshakeRetries+2; i++ {
-		cumulative += DefaultHandshakeTryInterval * time.Duration(i)
-		next_tick := now.Add(cumulative)
-		blah.NextOutboundHandshakeTimerTick(next_tick, mw)
-	}
-	/*
-		for i := 0; i <= HandshakeRetries+1; i++ {
-			next_tick := now.Add(cumulative)
-			//l.Infoln(next_tick)
-			blah.NextOutboundHandshakeTimerTick(next_tick)
-		}
-	*/
-	/*
-		for i := 0; i <= HandshakeRetries+1; i++ {
-			next_tick := now.Add(time.Duration(i) * time.Second)
-			blah.NextOutboundHandshakeTimerTick(next_tick)
-		}
-	*/
-
-	/*
-		cumulative += HandshakeTryInterval*time.Duration(HandshakeRetries) + 3
-		next_tick := now.Add(cumulative)
-		l.Infoln(cumulative, next_tick)
-		blah.NextOutboundHandshakeTimerTick(next_tick)
-	*/
-	assert.NotContains(t, blah.pendingHostMap.Hosts, uint32(vpnIP))
-	assert.NotContains(t, blah.pendingHostMap.Indexes, uint32(12341234))
-}
-
-func Test_NewHandshakeManagerIndexcleanup(t *testing.T) {
-	_, tuncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, vpncidr, _ := net.ParseCIDR("172.1.1.1/24")
-	_, localrange, _ := net.ParseCIDR("10.1.1.1/24")
-	preferredRanges := []*net.IPNet{localrange}
-	mainHM := NewHostMap("test", vpncidr, preferredRanges)
-
-	blah := NewHandshakeManager(tuncidr, preferredRanges, mainHM, &LightHouse{}, &udpConn{}, defaultHandshakeConfig)
-
-	now := time.Now()
-	blah.NextInboundHandshakeTimerTick(now)
-
-	hostinfo := &HostInfo{ConnectionState: &ConnectionState{}}
-	err := blah.AddIndexHostInfo(hostinfo)
-	assert.NoError(t, err)
-	blah.InboundHandshakeTimer.Add(hostinfo.localIndexId, time.Second*10)
-	// Pretned we have an index too
-	blah.pendingHostMap.AddVpnIPHostInfo(101010, hostinfo)
-	assert.Contains(t, blah.pendingHostMap.Hosts, uint32(101010))
-
-	for i := 1; i <= DefaultHandshakeRetries+2; i++ {
-		next_tick := now.Add(DefaultHandshakeTryInterval * time.Duration(i))
-		blah.NextInboundHandshakeTimerTick(next_tick)
-	}
-
-	next_tick := now.Add(DefaultHandshakeTryInterval*DefaultHandshakeRetries + 3)
-	blah.NextInboundHandshakeTimerTick(next_tick)
-	assert.NotContains(t, blah.pendingHostMap.Hosts, uint32(101010))
-	assert.NotContains(t, blah.pendingHostMap.Indexes, uint32(hostinfo.localIndexId))
-}
-
 type mockEncWriter struct {
 }
 
-func (mw *mockEncWriter) SendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte) {
-	return
-}
-
-func (mw *mockEncWriter) SendMessageToAll(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte) {
+func (mw *mockEncWriter) SendMessageToVpnIp(t header.MessageType, st header.MessageSubType, vpnIp iputil.VpnIp, p, nb, out []byte) {
 	return
 }

header/header.go

@@ -1,4 +1,4 @@
-package nebula
+package header
 
 import (
 	"encoding/binary"
@@ -19,82 +19,78 @@ import (
 // |-----------------------------------------------------------------------|
 // |                               payload...                              |
 
-const (
-	Version   uint8 = 1
-	HeaderLen       = 16
-)
-
-type NebulaMessageType uint8
-type NebulaMessageSubType uint8
+type m map[string]interface{}
 
 const (
-	handshake   NebulaMessageType = 0
-	message     NebulaMessageType = 1
-	recvError   NebulaMessageType = 2
-	lightHouse  NebulaMessageType = 3
-	test        NebulaMessageType = 4
-	closeTunnel NebulaMessageType = 5
-
-	//TODO These are deprecated as of 06/12/2018 - NB
-	testRemote      NebulaMessageType = 6
-	testRemoteReply NebulaMessageType = 7
+	Version uint8 = 1
+	Len           = 16
 )
 
-var typeMap = map[NebulaMessageType]string{
-	handshake:   "handshake",
-	message:     "message",
-	recvError:   "recvError",
-	lightHouse:  "lightHouse",
-	test:        "test",
-	closeTunnel: "closeTunnel",
+type MessageType uint8
+type MessageSubType uint8
 
-	//TODO These are deprecated as of 06/12/2018 - NB
-	testRemote:      "testRemote",
-	testRemoteReply: "testRemoteReply",
+const (
+	Handshake   MessageType = 0
+	Message     MessageType = 1
+	RecvError   MessageType = 2
+	LightHouse  MessageType = 3
+	Test        MessageType = 4
+	CloseTunnel MessageType = 5
+)
+
+var typeMap = map[MessageType]string{
+	Handshake:   "handshake",
+	Message:     "message",
+	RecvError:   "recvError",
+	LightHouse:  "lightHouse",
+	Test:        "test",
+	CloseTunnel: "closeTunnel",
 }
 
 const (
-	testRequest NebulaMessageSubType = 0
-	testReply   NebulaMessageSubType = 1
+	TestRequest MessageSubType = 0
+	TestReply   MessageSubType = 1
 )
 
-var eHeaderTooShort = errors.New("header is too short")
+const (
+	HandshakeIXPSK0 MessageSubType = 0
+	HandshakeXXPSK0 MessageSubType = 1
+)
 
-var subTypeTestMap = map[NebulaMessageSubType]string{
-	testRequest: "testRequest",
-	testReply:   "testReply",
+var ErrHeaderTooShort = errors.New("header is too short")
+
+var subTypeTestMap = map[MessageSubType]string{
+	TestRequest: "testRequest",
+	TestReply:   "testReply",
 }
 
-var subTypeNoneMap = map[NebulaMessageSubType]string{0: "none"}
+var subTypeNoneMap = map[MessageSubType]string{0: "none"}
 
-var subTypeMap = map[NebulaMessageType]*map[NebulaMessageSubType]string{
-	message:     &subTypeNoneMap,
-	recvError:   &subTypeNoneMap,
-	lightHouse:  &subTypeNoneMap,
-	test:        &subTypeTestMap,
-	closeTunnel: &subTypeNoneMap,
-	handshake: {
-		handshakeIXPSK0: "ix_psk0",
+var subTypeMap = map[MessageType]*map[MessageSubType]string{
+	Message:     &subTypeNoneMap,
+	RecvError:   &subTypeNoneMap,
+	LightHouse:  &subTypeNoneMap,
+	Test:        &subTypeTestMap,
+	CloseTunnel: &subTypeNoneMap,
+	Handshake: {
+		HandshakeIXPSK0: "ix_psk0",
 	},
-	//TODO: these are deprecated
-	testRemote:      &subTypeNoneMap,
-	testRemoteReply: &subTypeNoneMap,
 }
 
-type Header struct {
+type H struct {
 	Version        uint8
-	Type           NebulaMessageType
-	Subtype        NebulaMessageSubType
+	Type           MessageType
+	Subtype        MessageSubType
 	Reserved       uint16
 	RemoteIndex    uint32
 	MessageCounter uint64
 }
 
-// HeaderEncode uses the provided byte array to encode the provided header values into.
+// Encode uses the provided byte array to encode the provided header values into.
 // Byte array must be capped higher than HeaderLen or this will panic
-func HeaderEncode(b []byte, v uint8, t uint8, st uint8, ri uint32, c uint64) []byte {
-	b = b[:HeaderLen]
-	b[0] = byte(v<<4 | (t & 0x0f))
+func Encode(b []byte, v uint8, t MessageType, st MessageSubType, ri uint32, c uint64) []byte {
+	b = b[:Len]
+	b[0] = v<<4 | byte(t&0x0f)
 	b[1] = byte(st)
 	binary.BigEndian.PutUint16(b[2:4], 0)
 	binary.BigEndian.PutUint32(b[4:8], ri)
@@ -103,7 +99,7 @@ func HeaderEncode(b []byte, v uint8, t uint8, st uint8, ri uint32, c uint64) []b
 }
 
 // String creates a readable string representation of a header
-func (h *Header) String() string {
+func (h *H) String() string {
 	if h == nil {
 		return "<nil>"
 	}
@@ -112,7 +108,7 @@ func (h *Header) String() string {
 }
 
 // MarshalJSON creates a json string representation of a header
-func (h *Header) MarshalJSON() ([]byte, error) {
+func (h *H) MarshalJSON() ([]byte, error) {
 	return json.Marshal(m{
 		"version":        h.Version,
 		"type":           h.TypeName(),
@@ -124,24 +120,24 @@ func (h *Header) MarshalJSON() ([]byte, error) {
 }
 
 // Encode turns header into bytes
-func (h *Header) Encode(b []byte) ([]byte, error) {
+func (h *H) Encode(b []byte) ([]byte, error) {
 	if h == nil {
 		return nil, errors.New("nil header")
 	}
 
-	return HeaderEncode(b, h.Version, uint8(h.Type), uint8(h.Subtype), h.RemoteIndex, h.MessageCounter), nil
+	return Encode(b, h.Version, h.Type, h.Subtype, h.RemoteIndex, h.MessageCounter), nil
 }
 
 // Parse is a helper function to parses given bytes into new Header struct
-func (h *Header) Parse(b []byte) error {
-	if len(b) < HeaderLen {
-		return eHeaderTooShort
+func (h *H) Parse(b []byte) error {
+	if len(b) < Len {
+		return ErrHeaderTooShort
 	}
 	// get upper 4 bytes
 	h.Version = uint8((b[0] >> 4) & 0x0f)
 	// get lower 4 bytes
-	h.Type = NebulaMessageType(b[0] & 0x0f)
-	h.Subtype = NebulaMessageSubType(b[1])
+	h.Type = MessageType(b[0] & 0x0f)
+	h.Subtype = MessageSubType(b[1])
 	h.Reserved = binary.BigEndian.Uint16(b[2:4])
 	h.RemoteIndex = binary.BigEndian.Uint32(b[4:8])
 	h.MessageCounter = binary.BigEndian.Uint64(b[8:16])
@@ -149,12 +145,12 @@ func (h *Header) Parse(b []byte) error {
 }
 
 // TypeName will transform the headers message type into a human string
-func (h *Header) TypeName() string {
+func (h *H) TypeName() string {
 	return TypeName(h.Type)
 }
 
 // TypeName will transform a nebula message type into a human string
-func TypeName(t NebulaMessageType) string {
+func TypeName(t MessageType) string {
 	if n, ok := typeMap[t]; ok {
 		return n
 	}
@@ -163,12 +159,12 @@ func TypeName(t NebulaMessageType) string {
 }
 
 // SubTypeName will transform the headers message sub type into a human string
-func (h *Header) SubTypeName() string {
+func (h *H) SubTypeName() string {
 	return SubTypeName(h.Type, h.Subtype)
 }
 
 // SubTypeName will transform a nebula message sub type into a human string
-func SubTypeName(t NebulaMessageType, s NebulaMessageSubType) string {
+func SubTypeName(t MessageType, s MessageSubType) string {
 	if n, ok := subTypeMap[t]; ok {
 		if x, ok := (*n)[s]; ok {
 			return x
@@ -179,8 +175,8 @@ func SubTypeName(t NebulaMessageType, s NebulaMessageSubType) string {
 }
 
 // NewHeader turns bytes into a header
-func NewHeader(b []byte) (*Header, error) {
-	h := new(Header)
+func NewHeader(b []byte) (*H, error) {
+	h := new(H)
 	if err := h.Parse(b); err != nil {
 		return nil, err
 	}

header/header_test.go

@@ -0,0 +1,115 @@
+package header
+
+import (
+	"reflect"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+type headerTest struct {
+	expectedBytes []byte
+	*H
+}
+
+// 0001 0010 00010010
+var headerBigEndianTests = []headerTest{{
+	expectedBytes: []byte{0x54, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0xa, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x9},
+	// 1010 0000
+	H: &H{
+		// 1111 1+2+4+8 = 15
+		Version:        5,
+		Type:           4,
+		Subtype:        0,
+		Reserved:       0,
+		RemoteIndex:    10,
+		MessageCounter: 9,
+	},
+},
+}
+
+func TestEncode(t *testing.T) {
+	for _, tt := range headerBigEndianTests {
+		b, err := tt.Encode(make([]byte, Len))
+		if err != nil {
+			t.Fatal(err)
+		}
+
+		assert.Equal(t, tt.expectedBytes, b)
+	}
+}
+
+func TestParse(t *testing.T) {
+	for _, tt := range headerBigEndianTests {
+		b := tt.expectedBytes
+		parsedHeader := &H{}
+		parsedHeader.Parse(b)
+
+		if !reflect.DeepEqual(tt.H, parsedHeader) {
+			t.Fatalf("got %#v; want %#v", parsedHeader, tt.H)
+		}
+	}
+}
+
+func TestTypeName(t *testing.T) {
+	assert.Equal(t, "test", TypeName(Test))
+	assert.Equal(t, "test", (&H{Type: Test}).TypeName())
+
+	assert.Equal(t, "unknown", TypeName(99))
+	assert.Equal(t, "unknown", (&H{Type: 99}).TypeName())
+}
+
+func TestSubTypeName(t *testing.T) {
+	assert.Equal(t, "testRequest", SubTypeName(Test, TestRequest))
+	assert.Equal(t, "testRequest", (&H{Type: Test, Subtype: TestRequest}).SubTypeName())
+
+	assert.Equal(t, "unknown", SubTypeName(99, TestRequest))
+	assert.Equal(t, "unknown", (&H{Type: 99, Subtype: TestRequest}).SubTypeName())
+
+	assert.Equal(t, "unknown", SubTypeName(Test, 99))
+	assert.Equal(t, "unknown", (&H{Type: Test, Subtype: 99}).SubTypeName())
+
+	assert.Equal(t, "none", SubTypeName(Message, 0))
+	assert.Equal(t, "none", (&H{Type: Message, Subtype: 0}).SubTypeName())
+}
+
+func TestTypeMap(t *testing.T) {
+	// Force people to document this stuff
+	assert.Equal(t, map[MessageType]string{
+		Handshake:   "handshake",
+		Message:     "message",
+		RecvError:   "recvError",
+		LightHouse:  "lightHouse",
+		Test:        "test",
+		CloseTunnel: "closeTunnel",
+	}, typeMap)
+
+	assert.Equal(t, map[MessageType]*map[MessageSubType]string{
+		Message:     &subTypeNoneMap,
+		RecvError:   &subTypeNoneMap,
+		LightHouse:  &subTypeNoneMap,
+		Test:        &subTypeTestMap,
+		CloseTunnel: &subTypeNoneMap,
+		Handshake: {
+			HandshakeIXPSK0: "ix_psk0",
+		},
+	}, subTypeMap)
+}
+
+func TestHeader_String(t *testing.T) {
+	assert.Equal(
+		t,
+		"ver=100 type=test subtype=testRequest reserved=0x63 remoteindex=98 messagecounter=97",
+		(&H{100, Test, TestRequest, 99, 98, 97}).String(),
+	)
+}
+
+func TestHeader_MarshalJSON(t *testing.T) {
+	b, err := (&H{100, Test, TestRequest, 99, 98, 97}).MarshalJSON()
+	assert.Nil(t, err)
+	assert.Equal(
+		t,
+		"{\"messageCounter\":97,\"remoteIndex\":98,\"reserved\":99,\"subType\":\"testRequest\",\"type\":\"test\",\"version\":100}",
+		string(b),
+	)
+}

header_test.go

@@ -1,119 +0,0 @@
-package nebula
-
-import (
-	"reflect"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-)
-
-type headerTest struct {
-	expectedBytes []byte
-	*Header
-}
-
-// 0001 0010 00010010
-var headerBigEndianTests = []headerTest{{
-	expectedBytes: []byte{0x54, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0xa, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x9},
-	// 1010 0000
-	Header: &Header{
-		// 1111 1+2+4+8 = 15
-		Version:        5,
-		Type:           4,
-		Subtype:        0,
-		Reserved:       0,
-		RemoteIndex:    10,
-		MessageCounter: 9,
-	},
-},
-}
-
-func TestEncode(t *testing.T) {
-	for _, tt := range headerBigEndianTests {
-		b, err := tt.Encode(make([]byte, HeaderLen))
-		if err != nil {
-			t.Fatal(err)
-		}
-
-		assert.Equal(t, tt.expectedBytes, b)
-	}
-}
-
-func TestParse(t *testing.T) {
-	for _, tt := range headerBigEndianTests {
-		b := tt.expectedBytes
-		parsedHeader := &Header{}
-		parsedHeader.Parse(b)
-
-		if !reflect.DeepEqual(tt.Header, parsedHeader) {
-			t.Fatalf("got %#v; want %#v", parsedHeader, tt.Header)
-		}
-	}
-}
-
-func TestTypeName(t *testing.T) {
-	assert.Equal(t, "test", TypeName(test))
-	assert.Equal(t, "test", (&Header{Type: test}).TypeName())
-
-	assert.Equal(t, "unknown", TypeName(99))
-	assert.Equal(t, "unknown", (&Header{Type: 99}).TypeName())
-}
-
-func TestSubTypeName(t *testing.T) {
-	assert.Equal(t, "testRequest", SubTypeName(test, testRequest))
-	assert.Equal(t, "testRequest", (&Header{Type: test, Subtype: testRequest}).SubTypeName())
-
-	assert.Equal(t, "unknown", SubTypeName(99, testRequest))
-	assert.Equal(t, "unknown", (&Header{Type: 99, Subtype: testRequest}).SubTypeName())
-
-	assert.Equal(t, "unknown", SubTypeName(test, 99))
-	assert.Equal(t, "unknown", (&Header{Type: test, Subtype: 99}).SubTypeName())
-
-	assert.Equal(t, "none", SubTypeName(message, 0))
-	assert.Equal(t, "none", (&Header{Type: message, Subtype: 0}).SubTypeName())
-}
-
-func TestTypeMap(t *testing.T) {
-	// Force people to document this stuff
-	assert.Equal(t, map[NebulaMessageType]string{
-		handshake:       "handshake",
-		message:         "message",
-		recvError:       "recvError",
-		lightHouse:      "lightHouse",
-		test:            "test",
-		closeTunnel:     "closeTunnel",
-		testRemote:      "testRemote",
-		testRemoteReply: "testRemoteReply",
-	}, typeMap)
-
-	assert.Equal(t, map[NebulaMessageType]*map[NebulaMessageSubType]string{
-		message:     &subTypeNoneMap,
-		recvError:   &subTypeNoneMap,
-		lightHouse:  &subTypeNoneMap,
-		test:        &subTypeTestMap,
-		closeTunnel: &subTypeNoneMap,
-		handshake: {
-			handshakeIXPSK0: "ix_psk0",
-		},
-		testRemote:      &subTypeNoneMap,
-		testRemoteReply: &subTypeNoneMap,
-	}, subTypeMap)
-}
-
-func TestHeader_String(t *testing.T) {
-	assert.Equal(
-		t,
-		"ver=100 type=test subtype=testRequest reserved=0x63 remoteindex=98 messagecounter=97",
-		(&Header{100, test, testRequest, 99, 98, 97}).String(),
-	)
-}
-
-func TestHeader_MarshalJSON(t *testing.T) {
-	b, err := (&Header{100, test, testRequest, 99, 98, 97}).MarshalJSON()
-	assert.Nil(t, err)
-	assert.Equal(
-		t,
-		"{\"messageCounter\":97,\"remoteIndex\":98,\"reserved\":99,\"subType\":\"testRequest\",\"type\":\"test\",\"version\":100}",
-		string(b),
-	)
-}

hostmap.go

@@ -1,7 +1,7 @@
 package nebula
 
 import (
-	"encoding/json"
+	"context"
 	"errors"
 	"fmt"
 	"net"
@@ -12,10 +12,15 @@ import (
 	"github.com/rcrowley/go-metrics"
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/cidr"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
 //const ProbeLen = 100
 const PromoteEvery = 1000
+const ReQueryEvery = 5000
 const MaxRemotes = 10
 
 // How long we should prevent roaming back to the previous IP.
@@ -27,64 +32,63 @@ type HostMap struct {
 	name            string
 	Indexes         map[uint32]*HostInfo
 	RemoteIndexes   map[uint32]*HostInfo
-	Hosts           map[uint32]*HostInfo
+	Hosts           map[iputil.VpnIp]*HostInfo
 	preferredRanges []*net.IPNet
 	vpnCIDR         *net.IPNet
-	defaultRoute    uint32
-	unsafeRoutes    *CIDRTree
+	unsafeRoutes    *cidr.Tree4
 	metricsEnabled  bool
+	l               *logrus.Logger
 }
 
 type HostInfo struct {
 	sync.RWMutex
 
-	remote            *udpAddr
-	Remotes           []*HostInfoDest
+	remote            *udp.Addr
+	remotes           *RemoteList
 	promoteCounter    uint32
 	ConnectionState   *ConnectionState
-	handshakeStart    time.Time
-	HandshakeReady    bool
-	HandshakeCounter  int
-	HandshakeComplete bool
-	HandshakePacket   map[uint8][]byte
-	packetStore       []*cachedPacket
+	handshakeStart    time.Time        //todo: this an entry in the handshake manager
+	HandshakeReady    bool             //todo: being in the manager means you are ready
+	HandshakeCounter  int              //todo: another handshake manager entry
+	HandshakeComplete bool             //todo: this should go away in favor of ConnectionState.ready
+	HandshakePacket   map[uint8][]byte //todo: this is other handshake manager entry
+	packetStore       []*cachedPacket  //todo: this is other handshake manager entry
 	remoteIndexId     uint32
 	localIndexId      uint32
-	hostId            uint32
+	vpnIp             iputil.VpnIp
 	recvError         int
-	remoteCidr        *CIDRTree
+	remoteCidr        *cidr.Tree4
 
 	// lastRebindCount is the other side of Interface.rebindCount, if these values don't match then we need to ask LH
 	// for a punch from the remote end of this tunnel. The goal being to prime their conntrack for our traffic just like
 	// with a handshake
 	lastRebindCount int8
 
+	// lastHandshakeTime records the time the remote side told us about at the stage when the handshake was completed locally
+	// Stage 1 packet will contain it if I am a responder, stage 2 packet if I am an initiator
+	// This is used to avoid an attack where a handshake packet is replayed after some time
+	lastHandshakeTime uint64
+
 	lastRoam       time.Time
-	lastRoamRemote *udpAddr
+	lastRoamRemote *udp.Addr
 }
 
 type cachedPacket struct {
-	messageType    NebulaMessageType
-	messageSubType NebulaMessageSubType
+	messageType    header.MessageType
+	messageSubType header.MessageSubType
 	callback       packetCallback
 	packet         []byte
 }
 
-type packetCallback func(t NebulaMessageType, st NebulaMessageSubType, h *HostInfo, p, nb, out []byte)
+type packetCallback func(t header.MessageType, st header.MessageSubType, h *HostInfo, p, nb, out []byte)
 
-type HostInfoDest struct {
-	addr *udpAddr
-	//probes       [ProbeLen]bool
-	probeCounter int
+type cachedPacketMetrics struct {
+	sent    metrics.Counter
+	dropped metrics.Counter
 }
 
-type Probe struct {
-	Addr    *net.UDPAddr
-	Counter int
-}
-
-func NewHostMap(name string, vpnCIDR *net.IPNet, preferredRanges []*net.IPNet) *HostMap {
-	h := map[uint32]*HostInfo{}
+func NewHostMap(l *logrus.Logger, name string, vpnCIDR *net.IPNet, preferredRanges []*net.IPNet) *HostMap {
+	h := map[iputil.VpnIp]*HostInfo{}
 	i := map[uint32]*HostInfo{}
 	r := map[uint32]*HostInfo{}
 	m := HostMap{
@@ -94,8 +98,8 @@ func NewHostMap(name string, vpnCIDR *net.IPNet, preferredRanges []*net.IPNet) *
 		Hosts:           h,
 		preferredRanges: preferredRanges,
 		vpnCIDR:         vpnCIDR,
-		defaultRoute:    0,
-		unsafeRoutes:    NewCIDRTree(),
+		unsafeRoutes:    cidr.NewTree4(),
+		l:               l,
 	}
 	return &m
 }
@@ -113,9 +117,9 @@ func (hm *HostMap) EmitStats(name string) {
 	metrics.GetOrRegisterGauge("hostmap."+name+".remoteIndexes", nil).Update(int64(remoteIndexLen))
 }
 
-func (hm *HostMap) GetIndexByVpnIP(vpnIP uint32) (uint32, error) {
+func (hm *HostMap) GetIndexByVpnIp(vpnIp iputil.VpnIp) (uint32, error) {
 	hm.RLock()
-	if i, ok := hm.Hosts[vpnIP]; ok {
+	if i, ok := hm.Hosts[vpnIp]; ok {
 		index := i.localIndexId
 		hm.RUnlock()
 		return index, nil
@@ -124,44 +128,44 @@ func (hm *HostMap) GetIndexByVpnIP(vpnIP uint32) (uint32, error) {
 	return 0, errors.New("vpn IP not found")
 }
 
-func (hm *HostMap) Add(ip uint32, hostinfo *HostInfo) {
+func (hm *HostMap) Add(ip iputil.VpnIp, hostinfo *HostInfo) {
 	hm.Lock()
 	hm.Hosts[ip] = hostinfo
 	hm.Unlock()
 }
 
-func (hm *HostMap) AddVpnIP(vpnIP uint32) *HostInfo {
-	h := &HostInfo{}
+func (hm *HostMap) AddVpnIp(vpnIp iputil.VpnIp, init func(hostinfo *HostInfo)) (hostinfo *HostInfo, created bool) {
 	hm.RLock()
-	if _, ok := hm.Hosts[vpnIP]; !ok {
+	if h, ok := hm.Hosts[vpnIp]; !ok {
 		hm.RUnlock()
 		h = &HostInfo{
-			Remotes:         []*HostInfoDest{},
 			promoteCounter:  0,
-			hostId:          vpnIP,
+			vpnIp:           vpnIp,
 			HandshakePacket: make(map[uint8][]byte, 0),
 		}
+		if init != nil {
+			init(h)
+		}
 		hm.Lock()
-		hm.Hosts[vpnIP] = h
+		hm.Hosts[vpnIp] = h
 		hm.Unlock()
-		return h
+		return h, true
 	} else {
-		h = hm.Hosts[vpnIP]
 		hm.RUnlock()
-		return h
+		return h, false
 	}
 }
 
-func (hm *HostMap) DeleteVpnIP(vpnIP uint32) {
+func (hm *HostMap) DeleteVpnIp(vpnIp iputil.VpnIp) {
 	hm.Lock()
-	delete(hm.Hosts, vpnIP)
+	delete(hm.Hosts, vpnIp)
 	if len(hm.Hosts) == 0 {
-		hm.Hosts = map[uint32]*HostInfo{}
+		hm.Hosts = map[iputil.VpnIp]*HostInfo{}
 	}
 	hm.Unlock()
 
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": IntIp(vpnIP), "mapTotalSize": len(hm.Hosts)}).
+	if hm.l.Level >= logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": vpnIp, "mapTotalSize": len(hm.Hosts)}).
 			Debug("Hostmap vpnIp deleted")
 	}
 }
@@ -173,24 +177,24 @@ func (hm *HostMap) addRemoteIndexHostInfo(index uint32, h *HostInfo) {
 	hm.RemoteIndexes[index] = h
 	hm.Unlock()
 
-	if l.Level > logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes),
-			"hostinfo": m{"existing": true, "localIndexId": h.localIndexId, "hostId": IntIp(h.hostId)}}).
+	if hm.l.Level > logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes),
+			"hostinfo": m{"existing": true, "localIndexId": h.localIndexId, "hostId": h.vpnIp}}).
 			Debug("Hostmap remoteIndex added")
 	}
 }
 
-func (hm *HostMap) AddVpnIPHostInfo(vpnIP uint32, h *HostInfo) {
+func (hm *HostMap) AddVpnIpHostInfo(vpnIp iputil.VpnIp, h *HostInfo) {
 	hm.Lock()
-	h.hostId = vpnIP
-	hm.Hosts[vpnIP] = h
+	h.vpnIp = vpnIp
+	hm.Hosts[vpnIp] = h
 	hm.Indexes[h.localIndexId] = h
 	hm.RemoteIndexes[h.remoteIndexId] = h
 	hm.Unlock()
 
-	if l.Level > logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": IntIp(vpnIP), "mapTotalSize": len(hm.Hosts),
-			"hostinfo": m{"existing": true, "localIndexId": h.localIndexId, "hostId": IntIp(h.hostId)}}).
+	if hm.l.Level > logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": vpnIp, "mapTotalSize": len(hm.Hosts),
+			"hostinfo": m{"existing": true, "localIndexId": h.localIndexId, "vpnIp": h.vpnIp}}).
 			Debug("Hostmap vpnIp added")
 	}
 }
@@ -205,15 +209,15 @@ func (hm *HostMap) DeleteIndex(index uint32) {
 
 		// Check if we have an entry under hostId that matches the same hostinfo
 		// instance. Clean it up as well if we do.
-		hostinfo2, ok := hm.Hosts[hostinfo.hostId]
+		hostinfo2, ok := hm.Hosts[hostinfo.vpnIp]
 		if ok && hostinfo2 == hostinfo {
-			delete(hm.Hosts, hostinfo.hostId)
+			delete(hm.Hosts, hostinfo.vpnIp)
 		}
 	}
 	hm.Unlock()
 
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes)}).
+	if hm.l.Level >= logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes)}).
 			Debug("Hostmap index deleted")
 	}
 }
@@ -229,35 +233,39 @@ func (hm *HostMap) DeleteReverseIndex(index uint32) {
 		// Check if we have an entry under hostId that matches the same hostinfo
 		// instance. Clean it up as well if we do (they might not match in pendingHostmap)
 		var hostinfo2 *HostInfo
-		hostinfo2, ok = hm.Hosts[hostinfo.hostId]
+		hostinfo2, ok = hm.Hosts[hostinfo.vpnIp]
 		if ok && hostinfo2 == hostinfo {
-			delete(hm.Hosts, hostinfo.hostId)
+			delete(hm.Hosts, hostinfo.vpnIp)
 		}
 	}
 	hm.Unlock()
 
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes)}).
+	if hm.l.Level >= logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "indexNumber": index, "mapTotalSize": len(hm.Indexes)}).
 			Debug("Hostmap remote index deleted")
 	}
 }
 
 func (hm *HostMap) DeleteHostInfo(hostinfo *HostInfo) {
 	hm.Lock()
+	defer hm.Unlock()
+	hm.unlockedDeleteHostInfo(hostinfo)
+}
 
+func (hm *HostMap) unlockedDeleteHostInfo(hostinfo *HostInfo) {
 	// Check if this same hostId is in the hostmap with a different instance.
 	// This could happen if we have an entry in the pending hostmap with different
 	// index values than the one in the main hostmap.
-	hostinfo2, ok := hm.Hosts[hostinfo.hostId]
+	hostinfo2, ok := hm.Hosts[hostinfo.vpnIp]
 	if ok && hostinfo2 != hostinfo {
-		delete(hm.Hosts, hostinfo2.hostId)
+		delete(hm.Hosts, hostinfo2.vpnIp)
 		delete(hm.Indexes, hostinfo2.localIndexId)
 		delete(hm.RemoteIndexes, hostinfo2.remoteIndexId)
 	}
 
-	delete(hm.Hosts, hostinfo.hostId)
+	delete(hm.Hosts, hostinfo.vpnIp)
 	if len(hm.Hosts) == 0 {
-		hm.Hosts = map[uint32]*HostInfo{}
+		hm.Hosts = map[iputil.VpnIp]*HostInfo{}
 	}
 	delete(hm.Indexes, hostinfo.localIndexId)
 	if len(hm.Indexes) == 0 {
@@ -267,11 +275,10 @@ func (hm *HostMap) DeleteHostInfo(hostinfo *HostInfo) {
 	if len(hm.RemoteIndexes) == 0 {
 		hm.RemoteIndexes = map[uint32]*HostInfo{}
 	}
-	hm.Unlock()
 
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "mapTotalSize": len(hm.Hosts),
-			"vpnIp": IntIp(hostinfo.hostId), "indexNumber": hostinfo.localIndexId, "remoteIndexNumber": hostinfo.remoteIndexId}).
+	if hm.l.Level >= logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "mapTotalSize": len(hm.Hosts),
+			"vpnIp": hostinfo.vpnIp, "indexNumber": hostinfo.localIndexId, "remoteIndexNumber": hostinfo.remoteIndexId}).
 			Debug("Hostmap hostInfo deleted")
 	}
 }
@@ -299,64 +306,36 @@ func (hm *HostMap) QueryReverseIndex(index uint32) (*HostInfo, error) {
 	}
 }
 
-func (hm *HostMap) AddRemote(vpnIp uint32, remote *udpAddr) *HostInfo {
-	hm.Lock()
-	i, v := hm.Hosts[vpnIp]
-	if v {
-		i.AddRemote(*remote)
-	} else {
-		i = &HostInfo{
-			Remotes:         []*HostInfoDest{NewHostInfoDest(remote)},
-			promoteCounter:  0,
-			hostId:          vpnIp,
-			HandshakePacket: make(map[uint8][]byte, 0),
-		}
-		i.remote = i.Remotes[0].addr
-		hm.Hosts[vpnIp] = i
-		l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": IntIp(vpnIp), "udpAddr": remote, "mapTotalSize": len(hm.Hosts)}).
-			Debug("Hostmap remote ip added")
-	}
-	i.ForcePromoteBest(hm.preferredRanges)
-	hm.Unlock()
-	return i
+func (hm *HostMap) QueryVpnIp(vpnIp iputil.VpnIp) (*HostInfo, error) {
+	return hm.queryVpnIp(vpnIp, nil)
 }
 
-func (hm *HostMap) QueryVpnIP(vpnIp uint32) (*HostInfo, error) {
-	return hm.queryVpnIP(vpnIp, nil)
-}
-
-// PromoteBestQueryVpnIP will attempt to lazily switch to the best remote every
+// PromoteBestQueryVpnIp will attempt to lazily switch to the best remote every
 // `PromoteEvery` calls to this function for a given host.
-func (hm *HostMap) PromoteBestQueryVpnIP(vpnIp uint32, ifce *Interface) (*HostInfo, error) {
-	return hm.queryVpnIP(vpnIp, ifce)
+func (hm *HostMap) PromoteBestQueryVpnIp(vpnIp iputil.VpnIp, ifce *Interface) (*HostInfo, error) {
+	return hm.queryVpnIp(vpnIp, ifce)
 }
 
-func (hm *HostMap) queryVpnIP(vpnIp uint32, promoteIfce *Interface) (*HostInfo, error) {
+func (hm *HostMap) queryVpnIp(vpnIp iputil.VpnIp, promoteIfce *Interface) (*HostInfo, error) {
 	hm.RLock()
 	if h, ok := hm.Hosts[vpnIp]; ok {
-		if promoteIfce != nil {
+		hm.RUnlock()
+		// Do not attempt promotion if you are a lighthouse
+		if promoteIfce != nil && !promoteIfce.lightHouse.amLighthouse {
 			h.TryPromoteBest(hm.preferredRanges, promoteIfce)
 		}
-		//fmt.Println(h.remote)
-		hm.RUnlock()
 		return h, nil
-	} else {
-		//return &net.UDPAddr{}, nil, errors.New("Unable to find host")
-		hm.RUnlock()
-		/*
-			if lightHouse != nil {
-				lightHouse.Query(vpnIp)
-				return nil, errors.New("Unable to find host")
-			}
-		*/
-		return nil, errors.New("unable to find host")
+
 	}
+
+	hm.RUnlock()
+	return nil, errors.New("unable to find host")
 }
 
-func (hm *HostMap) queryUnsafeRoute(ip uint32) uint32 {
+func (hm *HostMap) queryUnsafeRoute(ip iputil.VpnIp) iputil.VpnIp {
 	r := hm.unsafeRoutes.MostSpecificContains(ip)
 	if r != nil {
-		return r.(uint32)
+		return r.(iputil.VpnIp)
 	} else {
 		return 0
 	}
@@ -365,58 +344,38 @@ func (hm *HostMap) queryUnsafeRoute(ip uint32) uint32 {
 // We already have the hm Lock when this is called, so make sure to not call
 // any other methods that might try to grab it again
 func (hm *HostMap) addHostInfo(hostinfo *HostInfo, f *Interface) {
-	remoteCert := hostinfo.ConnectionState.peerCert
-	ip := ip2int(remoteCert.Details.Ips[0].IP)
-
-	f.lightHouse.AddRemoteAndReset(ip, hostinfo.remote)
 	if f.serveDns {
+		remoteCert := hostinfo.ConnectionState.peerCert
 		dnsR.Add(remoteCert.Details.Name+".", remoteCert.Details.Ips[0].IP.String())
 	}
 
-	hm.Hosts[hostinfo.hostId] = hostinfo
+	hm.Hosts[hostinfo.vpnIp] = hostinfo
 	hm.Indexes[hostinfo.localIndexId] = hostinfo
 	hm.RemoteIndexes[hostinfo.remoteIndexId] = hostinfo
 
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": IntIp(hostinfo.hostId), "mapTotalSize": len(hm.Hosts),
-			"hostinfo": m{"existing": true, "localIndexId": hostinfo.localIndexId, "hostId": IntIp(hostinfo.hostId)}}).
+	if hm.l.Level >= logrus.DebugLevel {
+		hm.l.WithField("hostMap", m{"mapName": hm.name, "vpnIp": hostinfo.vpnIp, "mapTotalSize": len(hm.Hosts),
+			"hostinfo": m{"existing": true, "localIndexId": hostinfo.localIndexId, "hostId": hostinfo.vpnIp}}).
 			Debug("Hostmap vpnIp added")
 	}
 }
 
-func (hm *HostMap) ClearRemotes(vpnIP uint32) {
-	hm.Lock()
-	i := hm.Hosts[vpnIP]
-	if i == nil {
-		hm.Unlock()
-		return
-	}
-	i.remote = nil
-	i.Remotes = nil
-	hm.Unlock()
-}
-
-func (hm *HostMap) SetDefaultRoute(ip uint32) {
-	hm.defaultRoute = ip
-}
-
-func (hm *HostMap) PunchList() []*udpAddr {
-	var list []*udpAddr
+// punchList assembles a list of all non nil RemoteList pointer entries in this hostmap
+// The caller can then do the its work outside of the read lock
+func (hm *HostMap) punchList(rl []*RemoteList) []*RemoteList {
 	hm.RLock()
+	defer hm.RUnlock()
+
 	for _, v := range hm.Hosts {
-		for _, r := range v.Remotes {
-			list = append(list, r.addr)
+		if v.remotes != nil {
+			rl = append(rl, v.remotes)
 		}
-		//	if h, ok := hm.Hosts[vpnIp]; ok {
-		//		hm.Hosts[vpnIp].PromoteBest(hm.preferredRanges, false)
-		//fmt.Println(h.remote)
-		//	}
 	}
-	hm.RUnlock()
-	return list
+	return rl
 }
 
-func (hm *HostMap) Punchy(conn *udpConn) {
+// Punchy iterates through the result of punchList() to assemble all known addresses and sends a hole punch packet to them
+func (hm *HostMap) Punchy(ctx context.Context, conn *udp.Conn) {
 	var metricsTxPunchy metrics.Counter
 	if hm.metricsEnabled {
 		metricsTxPunchy = metrics.GetOrRegisterCounter("messages.tx.punchy", nil)
@@ -424,148 +383,77 @@ func (hm *HostMap) Punchy(conn *udpConn) {
 		metricsTxPunchy = metrics.NilCounter{}
 	}
 
+	var remotes []*RemoteList
+	b := []byte{1}
+
+	clockSource := time.NewTicker(time.Second * 10)
+	defer clockSource.Stop()
+
 	for {
-		for _, addr := range hm.PunchList() {
-			metricsTxPunchy.Inc(1)
-			conn.WriteTo([]byte{1}, addr)
+		remotes = hm.punchList(remotes[:0])
+		for _, rl := range remotes {
+			//TODO: CopyAddrs generates garbage but ForEach locks for the work here, figure out which way is better
+			for _, addr := range rl.CopyAddrs(hm.preferredRanges) {
+				metricsTxPunchy.Inc(1)
+				conn.WriteTo(b, addr)
+			}
+		}
+
+		select {
+		case <-ctx.Done():
+			return
+		case <-clockSource.C:
+			continue
 		}
-		time.Sleep(time.Second * 30)
 	}
 }
 
 func (hm *HostMap) addUnsafeRoutes(routes *[]route) {
 	for _, r := range *routes {
-		l.WithField("route", r.route).WithField("via", r.via).Warn("Adding UNSAFE Route")
-		hm.unsafeRoutes.AddCIDR(r.route, ip2int(*r.via))
+		hm.l.WithField("route", r.route).WithField("via", r.via).Warn("Adding UNSAFE Route")
+		hm.unsafeRoutes.AddCIDR(r.route, iputil.Ip2VpnIp(*r.via))
 	}
 }
 
-func (i *HostInfo) MarshalJSON() ([]byte, error) {
-	return json.Marshal(m{
-		"remote":             i.remote,
-		"remotes":            i.Remotes,
-		"promote_counter":    i.promoteCounter,
-		"connection_state":   i.ConnectionState,
-		"handshake_start":    i.handshakeStart,
-		"handshake_ready":    i.HandshakeReady,
-		"handshake_counter":  i.HandshakeCounter,
-		"handshake_complete": i.HandshakeComplete,
-		"handshake_packet":   i.HandshakePacket,
-		"packet_store":       i.packetStore,
-		"remote_index":       i.remoteIndexId,
-		"local_index":        i.localIndexId,
-		"host_id":            int2ip(i.hostId),
-		"receive_errors":     i.recvError,
-		"last_roam":          i.lastRoam,
-		"last_roam_remote":   i.lastRoamRemote,
-	})
-}
-
 func (i *HostInfo) BindConnectionState(cs *ConnectionState) {
 	i.ConnectionState = cs
 }
 
+// TryPromoteBest handles re-querying lighthouses and probing for better paths
+// NOTE: It is an error to call this if you are a lighthouse since they should not roam clients!
 func (i *HostInfo) TryPromoteBest(preferredRanges []*net.IPNet, ifce *Interface) {
-	if i.remote == nil {
-		i.ForcePromoteBest(preferredRanges)
-		return
-	}
+	c := atomic.AddUint32(&i.promoteCounter, 1)
+	if c%PromoteEvery == 0 {
+		// The lock here is currently protecting i.remote access
+		i.RLock()
+		defer i.RUnlock()
 
-	if atomic.AddUint32(&i.promoteCounter, 1)&PromoteEvery == 0 {
 		// return early if we are already on a preferred remote
-		rIP := udp2ip(i.remote)
+		rIP := i.remote.IP
 		for _, l := range preferredRanges {
 			if l.Contains(rIP) {
 				return
 			}
 		}
 
-		// We re-query the lighthouse periodically while sending packets, so
-		// check for new remotes in our local lighthouse cache
-		ips := ifce.lightHouse.QueryCache(i.hostId)
-		for _, ip := range ips {
-			i.AddRemote(ip)
-		}
+		i.remotes.ForEach(preferredRanges, func(addr *udp.Addr, preferred bool) {
+			if addr == nil || !preferred {
+				return
+			}
 
-		best, preferred := i.getBestRemote(preferredRanges)
-		if preferred && !best.Equals(i.remote) {
 			// Try to send a test packet to that host, this should
 			// cause it to detect a roaming event and switch remotes
-			ifce.send(test, testRequest, i.ConnectionState, i, best, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
-		}
+			ifce.send(header.Test, header.TestRequest, i.ConnectionState, i, addr, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
+		})
+	}
+
+	// Re query our lighthouses for new remotes occasionally
+	if c%ReQueryEvery == 0 && ifce.lightHouse != nil {
+		ifce.lightHouse.QueryServer(i.vpnIp, ifce)
 	}
 }
 
-func (i *HostInfo) ForcePromoteBest(preferredRanges []*net.IPNet) {
-	best, _ := i.getBestRemote(preferredRanges)
-	if best != nil {
-		i.remote = best
-	}
-}
-
-func (i *HostInfo) getBestRemote(preferredRanges []*net.IPNet) (best *udpAddr, preferred bool) {
-	if len(i.Remotes) > 0 {
-		for _, r := range i.Remotes {
-			rIP := udp2ip(r.addr)
-
-			for _, l := range preferredRanges {
-				if l.Contains(rIP) {
-					return r.addr, true
-				}
-			}
-
-			if best == nil || !PrivateIP(rIP) {
-				best = r.addr
-			}
-			/*
-				for _, r := range i.Remotes {
-					// Must have > 80% probe success to be considered.
-					//fmt.Println("GRADE:", r.addr.IP, r.Grade())
-					if r.Grade() > float64(.8) {
-						if localToMe.Contains(r.addr.IP) == true {
-							best = r.addr
-							break
-							//i.remote = i.Remotes[c].addr
-						} else {
-								//}
-					}
-			*/
-		}
-		return best, false
-	}
-
-	return nil, false
-}
-
-// rotateRemote will move remote to the next ip in the list of remote ips for this host
-// This is different than PromoteBest in that what is algorithmically best may not actually work.
-// Only known use case is when sending a stage 0 handshake.
-// It may be better to just send stage 0 handshakes to all known ips and sort it out in the receiver.
-func (i *HostInfo) rotateRemote() {
-	// We have 0, can't rotate
-	if len(i.Remotes) < 1 {
-		return
-	}
-
-	if i.remote == nil {
-		i.remote = i.Remotes[0].addr
-		return
-	}
-
-	// We want to look at all but the very last entry since that is handled at the end
-	for x := 0; x < len(i.Remotes)-1; x++ {
-		// Find our current position and move to the next one in the list
-		if i.Remotes[x].addr.Equals(i.remote) {
-			i.remote = i.Remotes[x+1].addr
-			return
-		}
-	}
-
-	// Our current position was likely the last in the list, start over at 0
-	i.remote = i.Remotes[0].addr
-}
-
-func (i *HostInfo) cachePacket(t NebulaMessageType, st NebulaMessageSubType, packet []byte, f packetCallback) {
+func (i *HostInfo) cachePacket(l *logrus.Logger, t header.MessageType, st header.MessageSubType, packet []byte, f packetCallback, m *cachedPacketMetrics) {
 	//TODO: return the error so we can log with more context
 	if len(i.packetStore) < 100 {
 		tempPacket := make([]byte, len(packet))
@@ -573,14 +461,15 @@ func (i *HostInfo) cachePacket(t NebulaMessageType, st NebulaMessageSubType, pac
 		//l.WithField("trace", string(debug.Stack())).Error("Caching packet", tempPacket)
 		i.packetStore = append(i.packetStore, &cachedPacket{t, st, f, tempPacket})
 		if l.Level >= logrus.DebugLevel {
-			i.logger().
+			i.logger(l).
 				WithField("length", len(i.packetStore)).
 				WithField("stored", true).
 				Debugf("Packet store")
 		}
 
 	} else if l.Level >= logrus.DebugLevel {
-		i.logger().
+		m.dropped.Inc(1)
+		i.logger(l).
 			WithField("length", len(i.packetStore)).
 			WithField("stored", false).
 			Debugf("Packet store")
@@ -588,7 +477,7 @@ func (i *HostInfo) cachePacket(t NebulaMessageType, st NebulaMessageSubType, pac
 }
 
 // handshakeComplete will set the connection as ready to communicate, as well as flush any stored packets
-func (i *HostInfo) handshakeComplete() {
+func (i *HostInfo) handshakeComplete(l *logrus.Logger, m *cachedPacketMetrics) {
 	//TODO: I'm not certain the distinction between handshake complete and ConnectionState being ready matters because:
 	//TODO: HandshakeComplete means send stored packets and ConnectionState.ready means we are ready to send
 	//TODO: if the transition from HandhsakeComplete to ConnectionState.ready happens all within this function they are identical
@@ -598,26 +487,27 @@ func (i *HostInfo) handshakeComplete() {
 	//TODO: this should be managed by the handshake state machine to set it based on how many handshake were seen.
 	// Clamping it to 2 gets us out of the woods for now
 	atomic.StoreUint64(&i.ConnectionState.atomicMessageCounter, 2)
-	i.logger().Debugf("Sending %d stored packets", len(i.packetStore))
-	nb := make([]byte, 12, 12)
-	out := make([]byte, mtu)
-	for _, cp := range i.packetStore {
-		cp.callback(cp.messageType, cp.messageSubType, i, cp.packet, nb, out)
+
+	if l.Level >= logrus.DebugLevel {
+		i.logger(l).Debugf("Sending %d stored packets", len(i.packetStore))
 	}
+
+	if len(i.packetStore) > 0 {
+		nb := make([]byte, 12, 12)
+		out := make([]byte, mtu)
+		for _, cp := range i.packetStore {
+			cp.callback(cp.messageType, cp.messageSubType, i, cp.packet, nb, out)
+		}
+		m.sent.Inc(int64(len(i.packetStore)))
+	}
+
+	i.remotes.ResetBlockedRemotes()
 	i.packetStore = make([]*cachedPacket, 0)
 	i.ConnectionState.ready = true
 	i.ConnectionState.queueLock.Unlock()
 	i.ConnectionState.certState = nil
 }
 
-func (i *HostInfo) RemoteUDPAddrs() []*udpAddr {
-	var addrs []*udpAddr
-	for _, r := range i.Remotes {
-		addrs = append(addrs, r.addr)
-	}
-	return addrs
-}
-
 func (i *HostInfo) GetCert() *cert.NebulaCertificate {
 	if i.ConnectionState != nil {
 		return i.ConnectionState.peerCert
@@ -625,31 +515,48 @@ func (i *HostInfo) GetCert() *cert.NebulaCertificate {
 	return nil
 }
 
-func (i *HostInfo) AddRemote(r udpAddr) *udpAddr {
-	remote := &r
-	//add := true
-	for _, r := range i.Remotes {
-		if r.addr.Equals(remote) {
-			return r.addr
-			//add = false
+func (i *HostInfo) SetRemote(remote *udp.Addr) {
+	// We copy here because we likely got this remote from a source that reuses the object
+	if !i.remote.Equals(remote) {
+		i.remote = remote.Copy()
+		i.remotes.LearnRemote(i.vpnIp, remote.Copy())
+	}
+}
+
+// SetRemoteIfPreferred returns true if the remote was changed. The lastRoam
+// time on the HostInfo will also be updated.
+func (i *HostInfo) SetRemoteIfPreferred(hm *HostMap, newRemote *udp.Addr) bool {
+	currentRemote := i.remote
+	if currentRemote == nil {
+		i.SetRemote(newRemote)
+		return true
+	}
+
+	// NOTE: We do this loop here instead of calling `isPreferred` in
+	// remote_list.go so that we only have to loop over preferredRanges once.
+	newIsPreferred := false
+	for _, l := range hm.preferredRanges {
+		// return early if we are already on a preferred remote
+		if l.Contains(currentRemote.IP) {
+			return false
+		}
+
+		if l.Contains(newRemote.IP) {
+			newIsPreferred = true
 		}
 	}
-	// Trim this down if necessary
-	if len(i.Remotes) > MaxRemotes {
-		i.Remotes = i.Remotes[len(i.Remotes)-MaxRemotes:]
+
+	if newIsPreferred {
+		// Consider this a roaming event
+		i.lastRoam = time.Now()
+		i.lastRoamRemote = currentRemote.Copy()
+
+		i.SetRemote(newRemote)
+
+		return true
 	}
-	i.Remotes = append(i.Remotes, NewHostInfoDest(remote))
-	return remote
-	//l.Debugf("Added remote %s for vpn ip", remote)
-}
 
-func (i *HostInfo) SetRemote(remote udpAddr) {
-	i.remote = i.AddRemote(remote)
-}
-
-func (i *HostInfo) ClearRemotes() {
-	i.remote = nil
-	i.Remotes = []*HostInfoDest{}
+	return false
 }
 
 func (i *HostInfo) ClearConnectionState() {
@@ -670,7 +577,7 @@ func (i *HostInfo) CreateRemoteCIDR(c *cert.NebulaCertificate) {
 		return
 	}
 
-	remoteCidr := NewCIDRTree()
+	remoteCidr := cidr.NewTree4()
 	for _, ip := range c.Details.Ips {
 		remoteCidr.AddCIDR(&net.IPNet{IP: ip.IP, Mask: net.IPMask{255, 255, 255, 255}}, struct{}{})
 	}
@@ -681,13 +588,12 @@ func (i *HostInfo) CreateRemoteCIDR(c *cert.NebulaCertificate) {
 	i.remoteCidr = remoteCidr
 }
 
-func (i *HostInfo) logger() *logrus.Entry {
+func (i *HostInfo) logger(l *logrus.Logger) *logrus.Entry {
 	if i == nil {
 		return logrus.NewEntry(l)
 	}
 
-	li := l.WithField("vpnIp", IntIp(i.hostId))
-
+	li := l.WithField("vpnIp", i.vpnIp)
 	if connState := i.ConnectionState; connState != nil {
 		if peerCert := connState.peerCert; peerCert != nil {
 			li = li.WithField("certName", peerCert.Details.Name)
@@ -697,112 +603,18 @@ func (i *HostInfo) logger() *logrus.Entry {
 	return li
 }
 
-//########################
-
-func NewHostInfoDest(addr *udpAddr) *HostInfoDest {
-	i := &HostInfoDest{
-		addr: addr,
-	}
-	return i
-}
-
-func (hid *HostInfoDest) MarshalJSON() ([]byte, error) {
-	return json.Marshal(m{
-		"address":     hid.addr,
-		"probe_count": hid.probeCounter,
-	})
-}
-
-/*
-
-func (hm *HostMap) DebugRemotes(vpnIp uint32) string {
-	s := "\n"
-	for _, h := range hm.Hosts {
-		for _, r := range h.Remotes {
-			s += fmt.Sprintf("%s : %d ## %v\n", r.addr.IP.String(), r.addr.Port, r.probes)
-		}
-	}
-	return s
-}
-
-
-func (d *HostInfoDest) Grade() float64 {
-	c1 := ProbeLen
-	for n := len(d.probes) - 1; n >= 0; n-- {
-		if d.probes[n] == true {
-			c1 -= 1
-		}
-	}
-	return float64(c1) / float64(ProbeLen)
-}
-
-func (d *HostInfoDest) Grade() (float64, float64, float64) {
-	c1 := ProbeLen
-	c2 := ProbeLen / 2
-	c2c := ProbeLen - ProbeLen/2
-	c3 := ProbeLen / 5
-	c3c := ProbeLen - ProbeLen/5
-	for n := len(d.probes) - 1; n >= 0; n-- {
-		if d.probes[n] == true {
-			c1 -= 1
-			if n >= c2c {
-				c2 -= 1
-				if n >= c3c {
-					c3 -= 1
-				}
-			}
-		}
-		//if n >= d {
-	}
-	return float64(c3) / float64(ProbeLen/5), float64(c2) / float64(ProbeLen/2), float64(c1) / float64(ProbeLen)
-	//return float64(c1) / float64(ProbeLen), float64(c2) / float64(ProbeLen/2), float64(c3) / float64(ProbeLen/5)
-}
-
-
-func (i *HostInfo) HandleReply(addr *net.UDPAddr, counter int) {
-	for _, r := range i.Remotes {
-		if r.addr.IP.Equal(addr.IP) && r.addr.Port == addr.Port {
-			r.ProbeReceived(counter)
-		}
-	}
-}
-
-func (i *HostInfo) Probes() []*Probe {
-	p := []*Probe{}
-	for _, d := range i.Remotes {
-		p = append(p, &Probe{Addr: d.addr, Counter: d.Probe()})
-	}
-	return p
-}
-
-
-func (d *HostInfoDest) Probe() int {
-	//d.probes = append(d.probes, true)
-	d.probeCounter++
-	d.probes[d.probeCounter%ProbeLen] = true
-	return d.probeCounter
-	//return d.probeCounter
-}
-
-func (d *HostInfoDest) ProbeReceived(probeCount int) {
-	if probeCount >= (d.probeCounter - ProbeLen) {
-		//fmt.Println("PROBE WORKED", probeCount)
-		//fmt.Println(d.addr, d.Grade())
-		d.probes[probeCount%ProbeLen] = false
-	}
-}
-
-*/
-
 // Utility functions
 
-func localIps(allowList *AllowList) *[]net.IP {
+func localIps(l *logrus.Logger, allowList *LocalAllowList) *[]net.IP {
 	//FIXME: This function is pretty garbage
 	var ips []net.IP
 	ifaces, _ := net.Interfaces()
 	for _, i := range ifaces {
 		allow := allowList.AllowName(i.Name)
-		l.WithField("interfaceName", i.Name).WithField("allow", allow).Debug("localAllowList.AllowName")
+		if l.Level >= logrus.TraceLevel {
+			l.WithField("interfaceName", i.Name).WithField("allow", allow).Trace("localAllowList.AllowName")
+		}
+
 		if !allow {
 			continue
 		}
@@ -816,9 +628,14 @@ func localIps(allowList *AllowList) *[]net.IP {
 			case *net.IPAddr:
 				ip = v.IP
 			}
-			if ip.To4() != nil && ip.IsLoopback() == false {
-				allow := allowList.Allow(ip2int(ip))
-				l.WithField("localIp", ip).WithField("allow", allow).Debug("localAllowList.Allow")
+
+			//TODO: Filtering out link local for now, this is probably the most correct thing
+			//TODO: Would be nice to filter out SLAAC MAC based ips as well
+			if ip.IsLoopback() == false && !ip.IsLinkLocalUnicast() {
+				allow := allowList.Allow(ip)
+				if l.Level >= logrus.TraceLevel {
+					l.WithField("localIp", ip).WithField("allow", allow).Trace("localAllowList.Allow")
+				}
 				if !allow {
 					continue
 				}
@@ -829,12 +646,3 @@ func localIps(allowList *AllowList) *[]net.IP {
 	}
 	return &ips
 }
-
-func PrivateIP(ip net.IP) bool {
-	private := false
-	_, private24BitBlock, _ := net.ParseCIDR("10.0.0.0/8")
-	_, private20BitBlock, _ := net.ParseCIDR("172.16.0.0/12")
-	_, private16BitBlock, _ := net.ParseCIDR("192.168.0.0/16")
-	private = private24BitBlock.Contains(ip) || private20BitBlock.Contains(ip) || private16BitBlock.Contains(ip)
-	return private
-}

hostmap_test.go

@@ -1,164 +1 @@
 package nebula
-
-import (
-	"net"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-)
-
-/*
-func TestHostInfoDestProbe(t *testing.T) {
-	a, _ := net.ResolveUDPAddr("udp", "1.0.0.1:22222")
-	d := NewHostInfoDest(a)
-
-	// 999 probes that all return should give a 100% success rate
-	for i := 0; i < 999; i++ {
-		meh := d.Probe()
-		d.ProbeReceived(meh)
-	}
-	assert.Equal(t, d.Grade(), float64(1))
-
-	// 999 probes of which only half return should give a 50% success rate
-	for i := 0; i < 999; i++ {
-		meh := d.Probe()
-		if i%2 == 0 {
-			d.ProbeReceived(meh)
-		}
-	}
-	assert.Equal(t, d.Grade(), float64(.5))
-
-	// 999 probes of which none return should give a 0% success rate
-	for i := 0; i < 999; i++ {
-		d.Probe()
-	}
-	assert.Equal(t, d.Grade(), float64(0))
-
-	// 999 probes of which only 1/4 return should give a 25% success rate
-	for i := 0; i < 999; i++ {
-		meh := d.Probe()
-		if i%4 == 0 {
-			d.ProbeReceived(meh)
-		}
-	}
-	assert.Equal(t, d.Grade(), float64(.25))
-
-	// 999 probes of which only half return and are duplicates should give a 50% success rate
-	for i := 0; i < 999; i++ {
-		meh := d.Probe()
-		if i%2 == 0 {
-			d.ProbeReceived(meh)
-			d.ProbeReceived(meh)
-		}
-	}
-	assert.Equal(t, d.Grade(), float64(.5))
-
-	// 999 probes of which only way old replies return should give a 0% success rate
-	for i := 0; i < 999; i++ {
-		meh := d.Probe()
-		d.ProbeReceived(meh - 101)
-	}
-	assert.Equal(t, d.Grade(), float64(0))
-
-}
-*/
-
-func TestHostmap(t *testing.T) {
-	_, myNet, _ := net.ParseCIDR("10.128.0.0/16")
-	_, localToMe, _ := net.ParseCIDR("192.168.1.0/24")
-	myNets := []*net.IPNet{myNet}
-	preferredRanges := []*net.IPNet{localToMe}
-
-	m := NewHostMap("test", myNet, preferredRanges)
-
-	a := NewUDPAddrFromString("10.127.0.3:11111")
-	b := NewUDPAddrFromString("1.0.0.1:22222")
-	y := NewUDPAddrFromString("10.128.0.3:11111")
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), a)
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), b)
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), y)
-
-	info, _ := m.QueryVpnIP(ip2int(net.ParseIP("10.128.1.1")))
-
-	// There should be three remotes in the host map
-	assert.Equal(t, 3, len(info.Remotes))
-
-	// Adding an identical remote should not change the count
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), y)
-	assert.Equal(t, 3, len(info.Remotes))
-
-	// Adding a fresh remote should add one
-	y = NewUDPAddrFromString("10.18.0.3:11111")
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), y)
-	assert.Equal(t, 4, len(info.Remotes))
-
-	// Query and reference remote should get the first one (and not nil)
-	info, _ = m.QueryVpnIP(ip2int(net.ParseIP("10.128.1.1")))
-	assert.NotNil(t, info.remote)
-
-	// Promotion should ensure that the best remote is chosen (y)
-	info.ForcePromoteBest(myNets)
-	assert.True(t, myNet.Contains(udp2ip(info.remote)))
-
-}
-
-func TestHostmapdebug(t *testing.T) {
-	_, myNet, _ := net.ParseCIDR("10.128.0.0/16")
-	_, localToMe, _ := net.ParseCIDR("192.168.1.0/24")
-	preferredRanges := []*net.IPNet{localToMe}
-	m := NewHostMap("test", myNet, preferredRanges)
-
-	a := NewUDPAddrFromString("10.127.0.3:11111")
-	b := NewUDPAddrFromString("1.0.0.1:22222")
-	y := NewUDPAddrFromString("10.128.0.3:11111")
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), a)
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), b)
-	m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), y)
-
-	//t.Errorf("%s", m.DebugRemotes(1))
-}
-
-func TestHostMap_rotateRemote(t *testing.T) {
-	h := HostInfo{}
-	// 0 remotes, no panic
-	h.rotateRemote()
-	assert.Nil(t, h.remote)
-
-	// 1 remote, no panic
-	h.AddRemote(*NewUDPAddr(ip2int(net.IP{1, 1, 1, 1}), 0))
-	h.rotateRemote()
-	assert.Equal(t, udp2ipInt(h.remote), ip2int(net.IP{1, 1, 1, 1}))
-
-	h.AddRemote(*NewUDPAddr(ip2int(net.IP{1, 1, 1, 2}), 0))
-	h.AddRemote(*NewUDPAddr(ip2int(net.IP{1, 1, 1, 3}), 0))
-	h.AddRemote(*NewUDPAddr(ip2int(net.IP{1, 1, 1, 4}), 0))
-
-	// Rotate through those 3
-	h.rotateRemote()
-	assert.Equal(t, udp2ipInt(h.remote), ip2int(net.IP{1, 1, 1, 2}))
-
-	h.rotateRemote()
-	assert.Equal(t, udp2ipInt(h.remote), ip2int(net.IP{1, 1, 1, 3}))
-
-	h.rotateRemote()
-	assert.Equal(t, udp2ipInt(h.remote), ip2int(net.IP{1, 1, 1, 4}))
-
-	// Finally, we should start over
-	h.rotateRemote()
-	assert.Equal(t, udp2ipInt(h.remote), ip2int(net.IP{1, 1, 1, 1}))
-}
-
-func BenchmarkHostmappromote2(b *testing.B) {
-	for n := 0; n < b.N; n++ {
-		_, myNet, _ := net.ParseCIDR("10.128.0.0/16")
-		_, localToMe, _ := net.ParseCIDR("192.168.1.0/24")
-		preferredRanges := []*net.IPNet{localToMe}
-		m := NewHostMap("test", myNet, preferredRanges)
-		y := NewUDPAddrFromString("10.128.0.3:11111")
-		a := NewUDPAddrFromString("10.127.0.3:11111")
-		g := NewUDPAddrFromString("1.0.0.1:22222")
-		m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), a)
-		m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), g)
-		m.AddRemote(ip2int(net.ParseIP("10.128.1.1")), y)
-	}
-}

inside.go

@@ -5,12 +5,16 @@ import (
 
 	"github.com/flynn/noise"
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/firewall"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
-func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *FirewallPacket, nb, out []byte, q int, localCache ConntrackCache) {
+func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *firewall.Packet, nb, out []byte, q int, localCache firewall.ConntrackCache) {
 	err := newPacket(packet, false, fwPacket)
 	if err != nil {
-		l.WithField("packet", packet).Debugf("Error while validating outbound packet: %s", err)
+		f.l.WithField("packet", packet).Debugf("Error while validating outbound packet: %s", err)
 		return
 	}
 
@@ -20,7 +24,7 @@ func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *FirewallPacket,
 	}
 
 	// Ignore packets from self to self
-	if fwPacket.RemoteIP == f.lightHouse.myIp {
+	if fwPacket.RemoteIP == f.myVpnIp {
 		return
 	}
 
@@ -31,8 +35,8 @@ func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *FirewallPacket,
 
 	hostinfo := f.getOrHandshake(fwPacket.RemoteIP)
 	if hostinfo == nil {
-		if l.Level >= logrus.DebugLevel {
-			l.WithField("vpnIp", IntIp(fwPacket.RemoteIP)).
+		if f.l.Level >= logrus.DebugLevel {
+			f.l.WithField("vpnIp", fwPacket.RemoteIP).
 				WithField("fwPacket", fwPacket).
 				Debugln("dropping outbound packet, vpnIp not in our CIDR or in unsafe routes")
 		}
@@ -45,22 +49,19 @@ func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *FirewallPacket,
 		// the packet queue.
 		ci.queueLock.Lock()
 		if !ci.ready {
-			hostinfo.cachePacket(message, 0, packet, f.sendMessageNow)
+			hostinfo.cachePacket(f.l, header.Message, 0, packet, f.sendMessageNow, f.cachedPacketMetrics)
 			ci.queueLock.Unlock()
 			return
 		}
 		ci.queueLock.Unlock()
 	}
 
-	dropReason := f.firewall.Drop(packet, *fwPacket, false, hostinfo, trustedCAs, localCache)
+	dropReason := f.firewall.Drop(packet, *fwPacket, false, hostinfo, f.caPool, localCache)
 	if dropReason == nil {
-		mc := f.sendNoMetrics(message, 0, ci, hostinfo, hostinfo.remote, packet, nb, out, q)
-		if f.lightHouse != nil && mc%5000 == 0 {
-			f.lightHouse.Query(fwPacket.RemoteIP, f)
-		}
+		f.sendNoMetrics(header.Message, 0, ci, hostinfo, hostinfo.remote, packet, nb, out, q)
 
-	} else if l.Level >= logrus.DebugLevel {
-		hostinfo.logger().
+	} else if f.l.Level >= logrus.DebugLevel {
+		hostinfo.logger(f.l).
 			WithField("fwPacket", fwPacket).
 			WithField("reason", dropReason).
 			Debugln("dropping outbound packet")
@@ -68,31 +69,30 @@ func (f *Interface) consumeInsidePacket(packet []byte, fwPacket *FirewallPacket,
 }
 
 // getOrHandshake returns nil if the vpnIp is not routable
-func (f *Interface) getOrHandshake(vpnIp uint32) *HostInfo {
-	if f.hostMap.vpnCIDR.Contains(int2ip(vpnIp)) == false {
+func (f *Interface) getOrHandshake(vpnIp iputil.VpnIp) *HostInfo {
+	//TODO: we can find contains without converting back to bytes
+	if f.hostMap.vpnCIDR.Contains(vpnIp.ToIP()) == false {
 		vpnIp = f.hostMap.queryUnsafeRoute(vpnIp)
 		if vpnIp == 0 {
 			return nil
 		}
 	}
-	hostinfo, err := f.hostMap.PromoteBestQueryVpnIP(vpnIp, f)
+	hostinfo, err := f.hostMap.PromoteBestQueryVpnIp(vpnIp, f)
 
 	//if err != nil || hostinfo.ConnectionState == nil {
 	if err != nil {
-		hostinfo, err = f.handshakeManager.pendingHostMap.QueryVpnIP(vpnIp)
+		hostinfo, err = f.handshakeManager.pendingHostMap.QueryVpnIp(vpnIp)
 		if err != nil {
-			hostinfo = f.handshakeManager.AddVpnIP(vpnIp)
+			hostinfo = f.handshakeManager.AddVpnIp(vpnIp, f.initHostInfo)
 		}
 	}
-
 	ci := hostinfo.ConnectionState
 
 	if ci != nil && ci.eKey != nil && ci.ready {
 		return hostinfo
 	}
 
-	// Handshake is not ready, we need to grab the lock now before we start
-	// the handshake process
+	// Handshake is not ready, we need to grab the lock now before we start the handshake process
 	hostinfo.Lock()
 	defer hostinfo.Unlock()
 
@@ -102,16 +102,6 @@ func (f *Interface) getOrHandshake(vpnIp uint32) *HostInfo {
 		return hostinfo
 	}
 
-	if ci == nil {
-		// if we don't have a connection state, then send a handshake initiation
-		ci = f.newConnectionState(true, noise.HandshakeIX, []byte{}, 0)
-		// FIXME: Maybe make XX selectable, but probably not since psk makes it nearly pointless for us.
-		//ci = f.newConnectionState(true, noise.HandshakeXX, []byte{}, 0)
-		hostinfo.ConnectionState = ci
-	} else if ci.eKey == nil {
-		// if we don't have any state at all, create it
-	}
-
 	// If we have already created the handshake packet, we don't want to call the function at all.
 	if !hostinfo.HandshakeReady {
 		ixHandshakeStage0(f, vpnIp, hostinfo)
@@ -131,37 +121,40 @@ func (f *Interface) getOrHandshake(vpnIp uint32) *HostInfo {
 	return hostinfo
 }
 
-func (f *Interface) sendMessageNow(t NebulaMessageType, st NebulaMessageSubType, hostInfo *HostInfo, p, nb, out []byte) {
-	fp := &FirewallPacket{}
+// initHostInfo is the init function to pass to (*HandshakeManager).AddVpnIP that
+// will create the initial Noise ConnectionState
+func (f *Interface) initHostInfo(hostinfo *HostInfo) {
+	hostinfo.ConnectionState = f.newConnectionState(f.l, true, noise.HandshakeIX, []byte{}, 0)
+}
+
+func (f *Interface) sendMessageNow(t header.MessageType, st header.MessageSubType, hostInfo *HostInfo, p, nb, out []byte) {
+	fp := &firewall.Packet{}
 	err := newPacket(p, false, fp)
 	if err != nil {
-		l.Warnf("error while parsing outgoing packet for firewall check; %v", err)
+		f.l.Warnf("error while parsing outgoing packet for firewall check; %v", err)
 		return
 	}
 
 	// check if packet is in outbound fw rules
-	dropReason := f.firewall.Drop(p, *fp, false, hostInfo, trustedCAs, nil)
+	dropReason := f.firewall.Drop(p, *fp, false, hostInfo, f.caPool, nil)
 	if dropReason != nil {
-		if l.Level >= logrus.DebugLevel {
-			l.WithField("fwPacket", fp).
+		if f.l.Level >= logrus.DebugLevel {
+			f.l.WithField("fwPacket", fp).
 				WithField("reason", dropReason).
 				Debugln("dropping cached packet")
 		}
 		return
 	}
 
-	messageCounter := f.sendNoMetrics(message, st, hostInfo.ConnectionState, hostInfo, hostInfo.remote, p, nb, out, 0)
-	if f.lightHouse != nil && messageCounter%5000 == 0 {
-		f.lightHouse.Query(fp.RemoteIP, f)
-	}
+	f.sendNoMetrics(header.Message, st, hostInfo.ConnectionState, hostInfo, hostInfo.remote, p, nb, out, 0)
 }
 
 // SendMessageToVpnIp handles real ip:port lookup and sends to the current best known address for vpnIp
-func (f *Interface) SendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte) {
+func (f *Interface) SendMessageToVpnIp(t header.MessageType, st header.MessageSubType, vpnIp iputil.VpnIp, p, nb, out []byte) {
 	hostInfo := f.getOrHandshake(vpnIp)
 	if hostInfo == nil {
-		if l.Level >= logrus.DebugLevel {
-			l.WithField("vpnIp", IntIp(vpnIp)).
+		if f.l.Level >= logrus.DebugLevel {
+			f.l.WithField("vpnIp", vpnIp).
 				Debugln("dropping SendMessageToVpnIp, vpnIp not in our CIDR or in unsafe routes")
 		}
 		return
@@ -172,7 +165,7 @@ func (f *Interface) SendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubT
 		// the packet queue.
 		hostInfo.ConnectionState.queueLock.Lock()
 		if !hostInfo.ConnectionState.ready {
-			hostInfo.cachePacket(t, st, p, f.sendMessageToVpnIp)
+			hostInfo.cachePacket(f.l, t, st, p, f.sendMessageToVpnIp, f.cachedPacketMetrics)
 			hostInfo.ConnectionState.queueLock.Unlock()
 			return
 		}
@@ -183,52 +176,19 @@ func (f *Interface) SendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubT
 	return
 }
 
-func (f *Interface) sendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubType, hostInfo *HostInfo, p, nb, out []byte) {
+func (f *Interface) sendMessageToVpnIp(t header.MessageType, st header.MessageSubType, hostInfo *HostInfo, p, nb, out []byte) {
 	f.send(t, st, hostInfo.ConnectionState, hostInfo, hostInfo.remote, p, nb, out)
 }
 
-// SendMessageToAll handles real ip:port lookup and sends to all known addresses for vpnIp
-func (f *Interface) SendMessageToAll(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte) {
-	hostInfo := f.getOrHandshake(vpnIp)
-	if hostInfo == nil {
-		if l.Level >= logrus.DebugLevel {
-			l.WithField("vpnIp", IntIp(vpnIp)).
-				Debugln("dropping SendMessageToAll, vpnIp not in our CIDR or in unsafe routes")
-		}
-		return
-	}
-
-	if hostInfo.ConnectionState.ready == false {
-		// Because we might be sending stored packets, lock here to stop new things going to
-		// the packet queue.
-		hostInfo.ConnectionState.queueLock.Lock()
-		if !hostInfo.ConnectionState.ready {
-			hostInfo.cachePacket(t, st, p, f.sendMessageToAll)
-			hostInfo.ConnectionState.queueLock.Unlock()
-			return
-		}
-		hostInfo.ConnectionState.queueLock.Unlock()
-	}
-
-	f.sendMessageToAll(t, st, hostInfo, p, nb, out)
-	return
-}
-
-func (f *Interface) sendMessageToAll(t NebulaMessageType, st NebulaMessageSubType, hostInfo *HostInfo, p, nb, b []byte) {
-	for _, r := range hostInfo.RemoteUDPAddrs() {
-		f.send(t, st, hostInfo.ConnectionState, hostInfo, r, p, nb, b)
-	}
-}
-
-func (f *Interface) send(t NebulaMessageType, st NebulaMessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote *udpAddr, p, nb, out []byte) {
+func (f *Interface) send(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote *udp.Addr, p, nb, out []byte) {
 	f.messageMetrics.Tx(t, st, 1)
 	f.sendNoMetrics(t, st, ci, hostinfo, remote, p, nb, out, 0)
 }
 
-func (f *Interface) sendNoMetrics(t NebulaMessageType, st NebulaMessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote *udpAddr, p, nb, out []byte, q int) uint64 {
+func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType, ci *ConnectionState, hostinfo *HostInfo, remote *udp.Addr, p, nb, out []byte, q int) {
 	if ci.eKey == nil {
 		//TODO: log warning
-		return 0
+		return
 	}
 
 	var err error
@@ -237,18 +197,18 @@ func (f *Interface) sendNoMetrics(t NebulaMessageType, st NebulaMessageSubType,
 	c := atomic.AddUint64(&ci.atomicMessageCounter, 1)
 
 	//l.WithField("trace", string(debug.Stack())).Error("out Header ", &Header{Version, t, st, 0, hostinfo.remoteIndexId, c}, p)
-	out = HeaderEncode(out, Version, uint8(t), uint8(st), hostinfo.remoteIndexId, c)
-	f.connectionManager.Out(hostinfo.hostId)
+	out = header.Encode(out, header.Version, t, st, hostinfo.remoteIndexId, c)
+	f.connectionManager.Out(hostinfo.vpnIp)
 
 	// Query our LH if we haven't since the last time we've been rebound, this will cause the remote to punch against
 	// all our IPs and enable a faster roaming.
-	if hostinfo.lastRebindCount != f.rebindCount {
+	if t != header.CloseTunnel && hostinfo.lastRebindCount != f.rebindCount {
 		//NOTE: there is an update hole if a tunnel isn't used and exactly 256 rebinds occur before the tunnel is
 		// finally used again. This tunnel would eventually be torn down and recreated if this action didn't help.
-		f.lightHouse.Query(hostinfo.hostId, f)
+		f.lightHouse.QueryServer(hostinfo.vpnIp, f)
 		hostinfo.lastRebindCount = f.rebindCount
-		if l.Level >= logrus.DebugLevel {
-			l.WithField("vpnIp", hostinfo.hostId).Debug("Lighthouse update triggered for punch due to rebind counter")
+		if f.l.Level >= logrus.DebugLevel {
+			f.l.WithField("vpnIp", hostinfo.vpnIp).Debug("Lighthouse update triggered for punch due to rebind counter")
 		}
 	}
 
@@ -256,22 +216,22 @@ func (f *Interface) sendNoMetrics(t NebulaMessageType, st NebulaMessageSubType,
 	//TODO: see above note on lock
 	//ci.writeLock.Unlock()
 	if err != nil {
-		hostinfo.logger().WithError(err).
+		hostinfo.logger(f.l).WithError(err).
 			WithField("udpAddr", remote).WithField("counter", c).
 			WithField("attemptedCounter", c).
 			Error("Failed to encrypt outgoing packet")
-		return c
+		return
 	}
 
 	err = f.writers[q].WriteTo(out, remote)
 	if err != nil {
-		hostinfo.logger().WithError(err).
+		hostinfo.logger(f.l).WithError(err).
 			WithField("udpAddr", remote).Error("Failed to write outgoing packet")
 	}
-	return c
+	return
 }
 
-func isMulticast(ip uint32) bool {
+func isMulticast(ip iputil.VpnIp) bool {
 	// Class D multicast
 	if (((ip >> 24) & 0xff) & 0xf0) == 0xe0 {
 		return true

interface.go

@@ -1,14 +1,22 @@
 package nebula
 
 import (
+	"context"
 	"errors"
 	"io"
 	"net"
 	"os"
 	"runtime"
+	"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/firewall"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
 const mtu = 9001
@@ -24,7 +32,7 @@ type Inside interface {
 
 type InterfaceConfig struct {
 	HostMap                 *HostMap
-	Outside                 *udpConn
+	Outside                 *udp.Conn
 	Inside                  Inside
 	certState               *CertState
 	Cipher                  string
@@ -36,17 +44,19 @@ type InterfaceConfig struct {
 	pendingDeletionInterval int
 	DropLocalBroadcast      bool
 	DropMulticast           bool
-	UDPBatchSize            int
 	routines                int
 	MessageMetrics          *MessageMetrics
 	version                 string
+	caPool                  *cert.NebulaCAPool
+	disconnectInvalid       bool
 
 	ConntrackCacheTimeout time.Duration
+	l                     *logrus.Logger
 }
 
 type Interface struct {
 	hostMap            *HostMap
-	outside            *udpConn
+	outside            *udp.Conn
 	inside             Inside
 	certState          *CertState
 	cipher             string
@@ -56,11 +66,14 @@ type Interface struct {
 	serveDns           bool
 	createTime         time.Time
 	lightHouse         *LightHouse
-	localBroadcast     uint32
+	localBroadcast     iputil.VpnIp
+	myVpnIp            iputil.VpnIp
 	dropLocalBroadcast bool
 	dropMulticast      bool
-	udpBatchSize       int
 	routines           int
+	caPool             *cert.NebulaCAPool
+	disconnectInvalid  bool
+	closed             int32
 
 	// rebindCount is used to decide if an active tunnel should trigger a punch notification through a lighthouse
 	rebindCount int8
@@ -68,14 +81,17 @@ type Interface struct {
 
 	conntrackCacheTimeout time.Duration
 
-	writers []*udpConn
+	writers []*udp.Conn
 	readers []io.ReadWriteCloser
 
-	metricHandshakes metrics.Histogram
-	messageMetrics   *MessageMetrics
+	metricHandshakes    metrics.Histogram
+	messageMetrics      *MessageMetrics
+	cachedPacketMetrics *cachedPacketMetrics
+
+	l *logrus.Logger
 }
 
-func NewInterface(c *InterfaceConfig) (*Interface, error) {
+func NewInterface(ctx context.Context, c *InterfaceConfig) (*Interface, error) {
 	if c.Outside == nil {
 		return nil, errors.New("no outside connection")
 	}
@@ -89,6 +105,7 @@ func NewInterface(c *InterfaceConfig) (*Interface, error) {
 		return nil, errors.New("no firewall rules")
 	}
 
+	myVpnIp := iputil.Ip2VpnIp(c.certState.certificate.Details.Ips[0].IP)
 	ifce := &Interface{
 		hostMap:            c.HostMap,
 		outside:            c.Outside,
@@ -100,35 +117,46 @@ func NewInterface(c *InterfaceConfig) (*Interface, error) {
 		handshakeManager:   c.HandshakeManager,
 		createTime:         time.Now(),
 		lightHouse:         c.lightHouse,
-		localBroadcast:     ip2int(c.certState.certificate.Details.Ips[0].IP) | ^ip2int(c.certState.certificate.Details.Ips[0].Mask),
+		localBroadcast:     myVpnIp | ^iputil.Ip2VpnIp(c.certState.certificate.Details.Ips[0].Mask),
 		dropLocalBroadcast: c.DropLocalBroadcast,
 		dropMulticast:      c.DropMulticast,
-		udpBatchSize:       c.UDPBatchSize,
 		routines:           c.routines,
 		version:            c.version,
-		writers:            make([]*udpConn, c.routines),
+		writers:            make([]*udp.Conn, c.routines),
 		readers:            make([]io.ReadWriteCloser, c.routines),
+		caPool:             c.caPool,
+		disconnectInvalid:  c.disconnectInvalid,
+		myVpnIp:            myVpnIp,
 
 		conntrackCacheTimeout: c.ConntrackCacheTimeout,
 
 		metricHandshakes: metrics.GetOrRegisterHistogram("handshakes", nil, metrics.NewExpDecaySample(1028, 0.015)),
 		messageMetrics:   c.MessageMetrics,
+		cachedPacketMetrics: &cachedPacketMetrics{
+			sent:    metrics.GetOrRegisterCounter("hostinfo.cached_packets.sent", nil),
+			dropped: metrics.GetOrRegisterCounter("hostinfo.cached_packets.dropped", nil),
+		},
+
+		l: c.l,
 	}
 
-	ifce.connectionManager = newConnectionManager(ifce, c.checkInterval, c.pendingDeletionInterval)
+	ifce.connectionManager = newConnectionManager(ctx, c.l, ifce, c.checkInterval, c.pendingDeletionInterval)
 
 	return ifce, nil
 }
 
-func (f *Interface) run() {
+// activate creates the interface on the host. After the interface is created, any
+// other services that want to bind listeners to its IP may do so successfully. However,
+// the interface isn't going to process anything until run() is called.
+func (f *Interface) activate() {
 	// actually turn on tun dev
 
 	addr, err := f.outside.LocalAddr()
 	if err != nil {
-		l.WithError(err).Error("Failed to get udp listen address")
+		f.l.WithError(err).Error("Failed to get udp listen address")
 	}
 
-	l.WithField("interface", f.inside.DeviceName()).WithField("network", f.inside.CidrNet().String()).
+	f.l.WithField("interface", f.inside.DeviceName()).WithField("network", f.inside.CidrNet().String()).
 		WithField("build", f.version).WithField("udpAddr", addr).
 		Info("Nebula interface is active")
 
@@ -140,16 +168,19 @@ func (f *Interface) run() {
 		if i > 0 {
 			reader, err = f.inside.NewMultiQueueReader()
 			if err != nil {
-				l.Fatal(err)
+				f.l.Fatal(err)
 			}
 		}
 		f.readers[i] = reader
 	}
 
 	if err := f.inside.Activate(); err != nil {
-		l.Fatal(err)
+		f.inside.Close()
+		f.l.Fatal(err)
 	}
+}
 
+func (f *Interface) run() {
 	// Launch n queues to read packets from udp
 	for i := 0; i < f.routines; i++ {
 		go f.listenOut(i)
@@ -164,14 +195,17 @@ func (f *Interface) run() {
 func (f *Interface) listenOut(i int) {
 	runtime.LockOSThread()
 
-	var li *udpConn
+	var li *udp.Conn
 	// TODO clean this up with a coherent interface for each outside connection
 	if i > 0 {
 		li = f.writers[i]
 	} else {
 		li = f.outside
 	}
-	li.ListenOut(f, i)
+
+	lhh := f.lightHouse.NewRequestHandler()
+	conntrackCache := firewall.NewConntrackCacheTicker(f.conntrackCacheTimeout)
+	li.ListenOut(f.readOutsidePackets, lhh.HandleRequest, conntrackCache, i)
 }
 
 func (f *Interface) listenIn(reader io.ReadWriteCloser, i int) {
@@ -179,50 +213,54 @@ func (f *Interface) listenIn(reader io.ReadWriteCloser, i int) {
 
 	packet := make([]byte, mtu)
 	out := make([]byte, mtu)
-	fwPacket := &FirewallPacket{}
+	fwPacket := &firewall.Packet{}
 	nb := make([]byte, 12, 12)
 
-	conntrackCache := NewConntrackCacheTicker(f.conntrackCacheTimeout)
+	conntrackCache := firewall.NewConntrackCacheTicker(f.conntrackCacheTimeout)
 
 	for {
 		n, err := reader.Read(packet)
 		if err != nil {
-			l.WithError(err).Error("Error while reading outbound packet")
+			if errors.Is(err, os.ErrClosed) && atomic.LoadInt32(&f.closed) != 0 {
+				return
+			}
+
+			f.l.WithError(err).Error("Error while reading outbound packet")
 			// This only seems to happen when something fatal happens to the fd, so exit.
 			os.Exit(2)
 		}
 
-		f.consumeInsidePacket(packet[:n], fwPacket, nb, out, i, conntrackCache.Get())
+		f.consumeInsidePacket(packet[:n], fwPacket, nb, out, i, conntrackCache.Get(f.l))
 	}
 }
 
-func (f *Interface) RegisterConfigChangeCallbacks(c *Config) {
+func (f *Interface) RegisterConfigChangeCallbacks(c *config.C) {
 	c.RegisterReloadCallback(f.reloadCA)
 	c.RegisterReloadCallback(f.reloadCertKey)
 	c.RegisterReloadCallback(f.reloadFirewall)
 	for _, udpConn := range f.writers {
-		c.RegisterReloadCallback(udpConn.reloadConfig)
+		c.RegisterReloadCallback(udpConn.ReloadConfig)
 	}
 }
 
-func (f *Interface) reloadCA(c *Config) {
+func (f *Interface) reloadCA(c *config.C) {
 	// reload and check regardless
 	// todo: need mutex?
-	newCAs, err := loadCAFromConfig(c)
+	newCAs, err := loadCAFromConfig(f.l, c)
 	if err != nil {
-		l.WithError(err).Error("Could not refresh trusted CA certificates")
+		f.l.WithError(err).Error("Could not refresh trusted CA certificates")
 		return
 	}
 
-	trustedCAs = newCAs
-	l.WithField("fingerprints", trustedCAs.GetFingerprints()).Info("Trusted CA certificates refreshed")
+	f.caPool = newCAs
+	f.l.WithField("fingerprints", f.caPool.GetFingerprints()).Info("Trusted CA certificates refreshed")
 }
 
-func (f *Interface) reloadCertKey(c *Config) {
+func (f *Interface) reloadCertKey(c *config.C) {
 	// reload and check in all cases
 	cs, err := NewCertStateFromConfig(c)
 	if err != nil {
-		l.WithError(err).Error("Could not refresh client cert")
+		f.l.WithError(err).Error("Could not refresh client cert")
 		return
 	}
 
@@ -230,24 +268,24 @@ func (f *Interface) reloadCertKey(c *Config) {
 	oldIPs := f.certState.certificate.Details.Ips
 	newIPs := cs.certificate.Details.Ips
 	if len(oldIPs) > 0 && len(newIPs) > 0 && oldIPs[0].String() != newIPs[0].String() {
-		l.WithField("new_ip", newIPs[0]).WithField("old_ip", oldIPs[0]).Error("IP in new cert was different from old")
+		f.l.WithField("new_ip", newIPs[0]).WithField("old_ip", oldIPs[0]).Error("IP in new cert was different from old")
 		return
 	}
 
 	f.certState = cs
-	l.WithField("cert", cs.certificate).Info("Client cert refreshed from disk")
+	f.l.WithField("cert", cs.certificate).Info("Client cert refreshed from disk")
 }
 
-func (f *Interface) reloadFirewall(c *Config) {
+func (f *Interface) reloadFirewall(c *config.C) {
 	//TODO: need to trigger/detect if the certificate changed too
 	if c.HasChanged("firewall") == false {
-		l.Debug("No firewall config change detected")
+		f.l.Debug("No firewall config change detected")
 		return
 	}
 
-	fw, err := NewFirewallFromConfig(f.certState.certificate, c)
+	fw, err := NewFirewallFromConfig(f.l, f.certState.certificate, c)
 	if err != nil {
-		l.WithError(err).Error("Error while creating firewall during reload")
+		f.l.WithError(err).Error("Error while creating firewall during reload")
 		return
 	}
 
@@ -260,7 +298,7 @@ func (f *Interface) reloadFirewall(c *Config) {
 	// If rulesVersion is back to zero, we have wrapped all the way around. Be
 	// safe and just reset conntrack in this case.
 	if fw.rulesVersion == 0 {
-		l.WithField("firewallHash", fw.GetRuleHash()).
+		f.l.WithField("firewallHash", fw.GetRuleHash()).
 			WithField("oldFirewallHash", oldFw.GetRuleHash()).
 			WithField("rulesVersion", fw.rulesVersion).
 			Warn("firewall rulesVersion has overflowed, resetting conntrack")
@@ -271,21 +309,33 @@ func (f *Interface) reloadFirewall(c *Config) {
 	f.firewall = fw
 
 	oldFw.Destroy()
-	l.WithField("firewallHash", fw.GetRuleHash()).
+	f.l.WithField("firewallHash", fw.GetRuleHash()).
 		WithField("oldFirewallHash", oldFw.GetRuleHash()).
 		WithField("rulesVersion", fw.rulesVersion).
 		Info("New firewall has been installed")
 }
 
-func (f *Interface) emitStats(i time.Duration) {
+func (f *Interface) emitStats(ctx context.Context, i time.Duration) {
 	ticker := time.NewTicker(i)
+	defer ticker.Stop()
 
-	udpStats := NewUDPStatsEmitter(f.writers)
+	udpStats := udp.NewUDPStatsEmitter(f.writers)
 
-	for range ticker.C {
-		f.firewall.EmitStats()
-		f.handshakeManager.EmitStats()
-
-		udpStats()
+	for {
+		select {
+		case <-ctx.Done():
+			return
+		case <-ticker.C:
+			f.firewall.EmitStats()
+			f.handshakeManager.EmitStats()
+			udpStats()
+		}
 	}
 }
+
+func (f *Interface) Close() error {
+	atomic.StoreInt32(&f.closed, 1)
+
+	// Release the tun device
+	return f.inside.Close()
+}

iputil/util.go

@@ -0,0 +1,66 @@
+package iputil
+
+import (
+	"encoding/binary"
+	"fmt"
+	"net"
+)
+
+type VpnIp uint32
+
+const maxIPv4StringLen = len("255.255.255.255")
+
+func (ip VpnIp) String() string {
+	b := make([]byte, maxIPv4StringLen)
+
+	n := ubtoa(b, 0, byte(ip>>24))
+	b[n] = '.'
+	n++
+
+	n += ubtoa(b, n, byte(ip>>16&255))
+	b[n] = '.'
+	n++
+
+	n += ubtoa(b, n, byte(ip>>8&255))
+	b[n] = '.'
+	n++
+
+	n += ubtoa(b, n, byte(ip&255))
+	return string(b[:n])
+}
+
+func (ip VpnIp) MarshalJSON() ([]byte, error) {
+	return []byte(fmt.Sprintf("\"%s\"", ip.String())), nil
+}
+
+func (ip VpnIp) ToIP() net.IP {
+	nip := make(net.IP, 4)
+	binary.BigEndian.PutUint32(nip, uint32(ip))
+	return nip
+}
+
+func Ip2VpnIp(ip []byte) VpnIp {
+	if len(ip) == 16 {
+		return VpnIp(binary.BigEndian.Uint32(ip[12:16]))
+	}
+	return VpnIp(binary.BigEndian.Uint32(ip))
+}
+
+// ubtoa encodes the string form of the integer v to dst[start:] and
+// returns the number of bytes written to dst. The caller must ensure
+// that dst has sufficient length.
+func ubtoa(dst []byte, start int, v byte) int {
+	if v < 10 {
+		dst[start] = v + '0'
+		return 1
+	} else if v < 100 {
+		dst[start+1] = v%10 + '0'
+		dst[start] = v/10 + '0'
+		return 2
+	}
+
+	dst[start+2] = v%10 + '0'
+	dst[start+1] = (v/10)%10 + '0'
+	dst[start] = v/100 + '0'
+	return 3
+}

iputil/util_test.go

@@ -0,0 +1,17 @@
+package iputil
+
+import (
+	"net"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestVpnIp_String(t *testing.T) {
+	assert.Equal(t, "255.255.255.255", Ip2VpnIp(net.ParseIP("255.255.255.255")).String())
+	assert.Equal(t, "1.255.255.255", Ip2VpnIp(net.ParseIP("1.255.255.255")).String())
+	assert.Equal(t, "1.1.255.255", Ip2VpnIp(net.ParseIP("1.1.255.255")).String())
+	assert.Equal(t, "1.1.1.255", Ip2VpnIp(net.ParseIP("1.1.1.255")).String())
+	assert.Equal(t, "1.1.1.1", Ip2VpnIp(net.ParseIP("1.1.1.1")).String())
+	assert.Equal(t, "0.0.0.0", Ip2VpnIp(net.ParseIP("0.0.0.0")).String())
+}

lighthouse.go

@@ -1,6 +1,8 @@
 package nebula
 
 import (
+	"context"
+	"encoding/binary"
 	"errors"
 	"fmt"
 	"net"
@@ -9,63 +11,71 @@ import (
 
 	"github.com/golang/protobuf/proto"
 	"github.com/rcrowley/go-metrics"
-	"github.com/slackhq/nebula/cert"
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 )
 
+//TODO: if a lighthouse doesn't have an answer, clients AGGRESSIVELY REQUERY.. why? handshake manager and/or getOrHandshake?
+//TODO: nodes are roaming lighthouses, this is bad. How are they learning?
+
 var ErrHostNotKnown = errors.New("host not known")
 
 type LightHouse struct {
+	//TODO: We need a timer wheel to kick out vpnIps that haven't reported in a long time
 	sync.RWMutex //Because we concurrently read and write to our maps
 	amLighthouse bool
-	myIp         uint32
-	punchConn    *udpConn
+	myVpnIp      iputil.VpnIp
+	myVpnZeros   iputil.VpnIp
+	punchConn    *udp.Conn
 
 	// Local cache of answers from light houses
-	addrMap map[uint32][]udpAddr
+	// map of vpn Ip to answers
+	addrMap map[iputil.VpnIp]*RemoteList
 
 	// filters remote addresses allowed for each host
 	// - When we are a lighthouse, this filters what addresses we store and
 	// respond with.
 	// - When we are not a lighthouse, this filters which addresses we accept
 	// from lighthouses.
-	remoteAllowList *AllowList
+	remoteAllowList *RemoteAllowList
 
 	// filters local addresses that we advertise to lighthouses
-	localAllowList *AllowList
+	localAllowList *LocalAllowList
 
 	// used to trigger the HandshakeManager when we receive HostQueryReply
-	handshakeTrigger chan<- uint32
+	handshakeTrigger chan<- iputil.VpnIp
 
 	// staticList exists to avoid having a bool in each addrMap entry
 	// since static should be rare
-	staticList  map[uint32]struct{}
-	lighthouses map[uint32]struct{}
+	staticList  map[iputil.VpnIp]struct{}
+	lighthouses map[iputil.VpnIp]struct{}
 	interval    int
-	nebulaPort  uint32
+	nebulaPort  uint32 // 32 bits because protobuf does not have a uint16
 	punchBack   bool
 	punchDelay  time.Duration
 
 	metrics           *MessageMetrics
 	metricHolepunchTx metrics.Counter
+	l                 *logrus.Logger
 }
 
-type EncWriter interface {
-	SendMessageToVpnIp(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte)
-	SendMessageToAll(t NebulaMessageType, st NebulaMessageSubType, vpnIp uint32, p, nb, out []byte)
-}
-
-func NewLightHouse(amLighthouse bool, myIp uint32, ips []uint32, interval int, nebulaPort uint32, pc *udpConn, punchBack bool, punchDelay time.Duration, metricsEnabled bool) *LightHouse {
+func NewLightHouse(l *logrus.Logger, amLighthouse bool, myVpnIpNet *net.IPNet, ips []iputil.VpnIp, interval int, nebulaPort uint32, pc *udp.Conn, punchBack bool, punchDelay time.Duration, metricsEnabled bool) *LightHouse {
+	ones, _ := myVpnIpNet.Mask.Size()
 	h := LightHouse{
 		amLighthouse: amLighthouse,
-		myIp:         myIp,
-		addrMap:      make(map[uint32][]udpAddr),
+		myVpnIp:      iputil.Ip2VpnIp(myVpnIpNet.IP),
+		myVpnZeros:   iputil.VpnIp(32 - ones),
+		addrMap:      make(map[iputil.VpnIp]*RemoteList),
 		nebulaPort:   nebulaPort,
-		lighthouses:  make(map[uint32]struct{}),
-		staticList:   make(map[uint32]struct{}),
+		lighthouses:  make(map[iputil.VpnIp]struct{}),
+		staticList:   make(map[iputil.VpnIp]struct{}),
 		interval:     interval,
 		punchConn:    pc,
 		punchBack:    punchBack,
 		punchDelay:   punchDelay,
+		l:            l,
 	}
 
 	if metricsEnabled {
@@ -83,14 +93,14 @@ func NewLightHouse(amLighthouse bool, myIp uint32, ips []uint32, interval int, n
 	return &h
 }
 
-func (lh *LightHouse) SetRemoteAllowList(allowList *AllowList) {
+func (lh *LightHouse) SetRemoteAllowList(allowList *RemoteAllowList) {
 	lh.Lock()
 	defer lh.Unlock()
 
 	lh.remoteAllowList = allowList
 }
 
-func (lh *LightHouse) SetLocalAllowList(allowList *AllowList) {
+func (lh *LightHouse) SetLocalAllowList(allowList *LocalAllowList) {
 	lh.Lock()
 	defer lh.Unlock()
 
@@ -100,46 +110,16 @@ func (lh *LightHouse) SetLocalAllowList(allowList *AllowList) {
 func (lh *LightHouse) ValidateLHStaticEntries() error {
 	for lhIP, _ := range lh.lighthouses {
 		if _, ok := lh.staticList[lhIP]; !ok {
-			return fmt.Errorf("Lighthouse %s does not have a static_host_map entry", IntIp(lhIP))
+			return fmt.Errorf("Lighthouse %s does not have a static_host_map entry", lhIP)
 		}
 	}
 	return nil
 }
 
-func (lh *LightHouse) Query(ip uint32, f EncWriter) ([]udpAddr, error) {
+func (lh *LightHouse) Query(ip iputil.VpnIp, f udp.EncWriter) *RemoteList {
 	if !lh.IsLighthouseIP(ip) {
 		lh.QueryServer(ip, f)
 	}
-	lh.RLock()
-	if v, ok := lh.addrMap[ip]; ok {
-		lh.RUnlock()
-		return v, nil
-	}
-	lh.RUnlock()
-	return nil, ErrHostNotKnown
-}
-
-// This is asynchronous so no reply should be expected
-func (lh *LightHouse) QueryServer(ip uint32, f EncWriter) {
-	if !lh.amLighthouse {
-		// Send a query to the lighthouses and hope for the best next time
-		query, err := proto.Marshal(NewLhQueryByInt(ip))
-		if err != nil {
-			l.WithError(err).WithField("vpnIp", IntIp(ip)).Error("Failed to marshal lighthouse query payload")
-			return
-		}
-
-		lh.metricTx(NebulaMeta_HostQuery, int64(len(lh.lighthouses)))
-		nb := make([]byte, 12, 12)
-		out := make([]byte, mtu)
-		for n := range lh.lighthouses {
-			f.SendMessageToVpnIp(lightHouse, 0, n, query, nb, out)
-		}
-	}
-}
-
-// Query our local lighthouse cached results
-func (lh *LightHouse) QueryCache(ip uint32) []udpAddr {
 	lh.RLock()
 	if v, ok := lh.addrMap[ip]; ok {
 		lh.RUnlock()
@@ -149,121 +129,262 @@ func (lh *LightHouse) QueryCache(ip uint32) []udpAddr {
 	return nil
 }
 
-func (lh *LightHouse) DeleteVpnIP(vpnIP uint32) {
+// This is asynchronous so no reply should be expected
+func (lh *LightHouse) QueryServer(ip iputil.VpnIp, f udp.EncWriter) {
+	if lh.amLighthouse {
+		return
+	}
+
+	if lh.IsLighthouseIP(ip) {
+		return
+	}
+
+	// Send a query to the lighthouses and hope for the best next time
+	query, err := proto.Marshal(NewLhQueryByInt(ip))
+	if err != nil {
+		lh.l.WithError(err).WithField("vpnIp", ip).Error("Failed to marshal lighthouse query payload")
+		return
+	}
+
+	lh.metricTx(NebulaMeta_HostQuery, int64(len(lh.lighthouses)))
+	nb := make([]byte, 12, 12)
+	out := make([]byte, mtu)
+	for n := range lh.lighthouses {
+		f.SendMessageToVpnIp(header.LightHouse, 0, n, query, nb, out)
+	}
+}
+
+func (lh *LightHouse) QueryCache(ip iputil.VpnIp) *RemoteList {
+	lh.RLock()
+	if v, ok := lh.addrMap[ip]; ok {
+		lh.RUnlock()
+		return v
+	}
+	lh.RUnlock()
+
+	lh.Lock()
+	defer lh.Unlock()
+	// Add an entry if we don't already have one
+	return lh.unlockedGetRemoteList(ip)
+}
+
+// queryAndPrepMessage is a lock helper on RemoteList, assisting the caller to build a lighthouse message containing
+// details from the remote list. It looks for a hit in the addrMap and a hit in the RemoteList under the owner vpnIp
+// If one is found then f() is called with proper locking, f() must return result of n.MarshalTo()
+func (lh *LightHouse) queryAndPrepMessage(vpnIp iputil.VpnIp, f func(*cache) (int, error)) (bool, int, error) {
+	lh.RLock()
+	// Do we have an entry in the main cache?
+	if v, ok := lh.addrMap[vpnIp]; ok {
+		// Swap lh lock for remote list lock
+		v.RLock()
+		defer v.RUnlock()
+
+		lh.RUnlock()
+
+		// vpnIp should also be the owner here since we are a lighthouse.
+		c := v.cache[vpnIp]
+		// Make sure we have
+		if c != nil {
+			n, err := f(c)
+			return true, n, err
+		}
+		return false, 0, nil
+	}
+	lh.RUnlock()
+	return false, 0, nil
+}
+
+func (lh *LightHouse) DeleteVpnIp(vpnIp iputil.VpnIp) {
 	// First we check the static mapping
 	// and do nothing if it is there
-	if _, ok := lh.staticList[vpnIP]; ok {
+	if _, ok := lh.staticList[vpnIp]; ok {
 		return
 	}
 	lh.Lock()
 	//l.Debugln(lh.addrMap)
-	delete(lh.addrMap, vpnIP)
-	l.Debugf("deleting %s from lighthouse.", IntIp(vpnIP))
+	delete(lh.addrMap, vpnIp)
+
+	if lh.l.Level >= logrus.DebugLevel {
+		lh.l.Debugf("deleting %s from lighthouse.", vpnIp)
+	}
+
 	lh.Unlock()
 }
 
-func (lh *LightHouse) AddRemote(vpnIP uint32, toIp *udpAddr, static bool) {
-	// First we check if the sender thinks this is a static entry
-	// and do nothing if it is not, but should be considered static
-	if static == false {
-		if _, ok := lh.staticList[vpnIP]; ok {
-			return
-		}
-	}
-
+// AddStaticRemote adds a static host entry for vpnIp as ourselves as the owner
+// We are the owner because we don't want a lighthouse server to advertise for static hosts it was configured with
+// And we don't want a lighthouse query reply to interfere with our learned cache if we are a client
+func (lh *LightHouse) AddStaticRemote(vpnIp iputil.VpnIp, toAddr *udp.Addr) {
 	lh.Lock()
-	defer lh.Unlock()
-	for _, v := range lh.addrMap[vpnIP] {
-		if v.Equals(toIp) {
+	am := lh.unlockedGetRemoteList(vpnIp)
+	am.Lock()
+	defer am.Unlock()
+	lh.Unlock()
+
+	if ipv4 := toAddr.IP.To4(); ipv4 != nil {
+		to := NewIp4AndPort(ipv4, uint32(toAddr.Port))
+		if !lh.unlockedShouldAddV4(vpnIp, to) {
 			return
 		}
+		am.unlockedPrependV4(lh.myVpnIp, to)
+
+	} else {
+		to := NewIp6AndPort(toAddr.IP, uint32(toAddr.Port))
+		if !lh.unlockedShouldAddV6(vpnIp, to) {
+			return
+		}
+		am.unlockedPrependV6(lh.myVpnIp, to)
 	}
 
-	allow := lh.remoteAllowList.Allow(udp2ipInt(toIp))
-	l.WithField("remoteIp", toIp).WithField("allow", allow).Debug("remoteAllowList.Allow")
+	// Mark it as static
+	lh.staticList[vpnIp] = struct{}{}
+}
+
+// unlockedGetRemoteList assumes you have the lh lock
+func (lh *LightHouse) unlockedGetRemoteList(vpnIp iputil.VpnIp) *RemoteList {
+	am, ok := lh.addrMap[vpnIp]
+	if !ok {
+		am = NewRemoteList()
+		lh.addrMap[vpnIp] = am
+	}
+	return am
+}
+
+// unlockedShouldAddV4 checks if to is allowed by our allow list
+func (lh *LightHouse) unlockedShouldAddV4(vpnIp iputil.VpnIp, to *Ip4AndPort) bool {
+	allow := lh.remoteAllowList.AllowIpV4(vpnIp, iputil.VpnIp(to.Ip))
+	if lh.l.Level >= logrus.TraceLevel {
+		lh.l.WithField("remoteIp", vpnIp).WithField("allow", allow).Trace("remoteAllowList.Allow")
+	}
+
+	if !allow || ipMaskContains(lh.myVpnIp, lh.myVpnZeros, iputil.VpnIp(to.Ip)) {
+		return false
+	}
+
+	return true
+}
+
+// unlockedShouldAddV6 checks if to is allowed by our allow list
+func (lh *LightHouse) unlockedShouldAddV6(vpnIp iputil.VpnIp, to *Ip6AndPort) bool {
+	allow := lh.remoteAllowList.AllowIpV6(vpnIp, to.Hi, to.Lo)
+	if lh.l.Level >= logrus.TraceLevel {
+		lh.l.WithField("remoteIp", lhIp6ToIp(to)).WithField("allow", allow).Trace("remoteAllowList.Allow")
+	}
+
+	// We don't check our vpn network here because nebula does not support ipv6 on the inside
 	if !allow {
-		return
+		return false
 	}
 
-	//l.Debugf("Adding reply of %s as %s\n", IntIp(vpnIP), toIp)
-	if static {
-		lh.staticList[vpnIP] = struct{}{}
-	}
-	lh.addrMap[vpnIP] = append(lh.addrMap[vpnIP], *toIp)
+	return true
 }
 
-func (lh *LightHouse) AddRemoteAndReset(vpnIP uint32, toIp *udpAddr) {
-	if lh.amLighthouse {
-		lh.DeleteVpnIP(vpnIP)
-		lh.AddRemote(vpnIP, toIp, false)
-	}
-
+func lhIp6ToIp(v *Ip6AndPort) net.IP {
+	ip := make(net.IP, 16)
+	binary.BigEndian.PutUint64(ip[:8], v.Hi)
+	binary.BigEndian.PutUint64(ip[8:], v.Lo)
+	return ip
 }
 
-func (lh *LightHouse) IsLighthouseIP(vpnIP uint32) bool {
-	if _, ok := lh.lighthouses[vpnIP]; ok {
+func (lh *LightHouse) IsLighthouseIP(vpnIp iputil.VpnIp) bool {
+	if _, ok := lh.lighthouses[vpnIp]; ok {
 		return true
 	}
 	return false
 }
 
-func NewLhQueryByInt(VpnIp uint32) *NebulaMeta {
+func NewLhQueryByInt(VpnIp iputil.VpnIp) *NebulaMeta {
 	return &NebulaMeta{
 		Type: NebulaMeta_HostQuery,
 		Details: &NebulaMetaDetails{
-			VpnIp: VpnIp,
+			VpnIp: uint32(VpnIp),
 		},
 	}
 }
 
-func NewIpAndPort(ip net.IP, port uint32) IpAndPort {
-	return IpAndPort{Ip: ip2int(ip), Port: port}
+func NewIp4AndPort(ip net.IP, port uint32) *Ip4AndPort {
+	ipp := Ip4AndPort{Port: port}
+	ipp.Ip = uint32(iputil.Ip2VpnIp(ip))
+	return &ipp
 }
 
-func NewIpAndPortFromUDPAddr(addr udpAddr) IpAndPort {
-	return IpAndPort{Ip: udp2ipInt(&addr), Port: uint32(addr.Port)}
+func NewIp6AndPort(ip net.IP, port uint32) *Ip6AndPort {
+	return &Ip6AndPort{
+		Hi:   binary.BigEndian.Uint64(ip[:8]),
+		Lo:   binary.BigEndian.Uint64(ip[8:]),
+		Port: port,
+	}
 }
 
-func (lh *LightHouse) LhUpdateWorker(f EncWriter) {
+func NewUDPAddrFromLH4(ipp *Ip4AndPort) *udp.Addr {
+	ip := ipp.Ip
+	return udp.NewAddr(
+		net.IPv4(byte(ip&0xff000000>>24), byte(ip&0x00ff0000>>16), byte(ip&0x0000ff00>>8), byte(ip&0x000000ff)),
+		uint16(ipp.Port),
+	)
+}
+
+func NewUDPAddrFromLH6(ipp *Ip6AndPort) *udp.Addr {
+	return udp.NewAddr(lhIp6ToIp(ipp), uint16(ipp.Port))
+}
+
+func (lh *LightHouse) LhUpdateWorker(ctx context.Context, f udp.EncWriter) {
 	if lh.amLighthouse || lh.interval == 0 {
 		return
 	}
 
+	clockSource := time.NewTicker(time.Second * time.Duration(lh.interval))
+	defer clockSource.Stop()
+
 	for {
 		lh.SendUpdate(f)
-		time.Sleep(time.Second * time.Duration(lh.interval))
+
+		select {
+		case <-ctx.Done():
+			return
+		case <-clockSource.C:
+			continue
+		}
 	}
 }
 
-func (lh *LightHouse) SendUpdate(f EncWriter) {
-	var ipps []*IpAndPort
+func (lh *LightHouse) SendUpdate(f udp.EncWriter) {
+	var v4 []*Ip4AndPort
+	var v6 []*Ip6AndPort
+
+	for _, e := range *localIps(lh.l, lh.localAllowList) {
+		if ip4 := e.To4(); ip4 != nil && ipMaskContains(lh.myVpnIp, lh.myVpnZeros, iputil.Ip2VpnIp(ip4)) {
+			continue
+		}
 
-	for _, e := range *localIps(lh.localAllowList) {
 		// Only add IPs that aren't my VPN/tun IP
-		if ip2int(e) != lh.myIp {
-			ipp := NewIpAndPort(e, lh.nebulaPort)
-			ipps = append(ipps, &ipp)
+		if ip := e.To4(); ip != nil {
+			v4 = append(v4, NewIp4AndPort(e, lh.nebulaPort))
+		} else {
+			v6 = append(v6, NewIp6AndPort(e, lh.nebulaPort))
 		}
 	}
 	m := &NebulaMeta{
 		Type: NebulaMeta_HostUpdateNotification,
 		Details: &NebulaMetaDetails{
-			VpnIp:      lh.myIp,
-			IpAndPorts: ipps,
+			VpnIp:       uint32(lh.myVpnIp),
+			Ip4AndPorts: v4,
+			Ip6AndPorts: v6,
 		},
 	}
 
 	lh.metricTx(NebulaMeta_HostUpdateNotification, int64(len(lh.lighthouses)))
 	nb := make([]byte, 12, 12)
 	out := make([]byte, mtu)
-	for vpnIp := range lh.lighthouses {
-		mm, err := proto.Marshal(m)
-		if err != nil {
-			l.Debugf("Invalid marshal to update")
-		}
-		//l.Error("LIGHTHOUSE PACKET SEND", mm)
-		f.SendMessageToVpnIp(lightHouse, 0, vpnIp, mm, nb, out)
 
+	mm, err := proto.Marshal(m)
+	if err != nil {
+		lh.l.WithError(err).Error("Error while marshaling for lighthouse update")
+		return
+	}
+
+	for vpnIp := range lh.lighthouses {
+		f.SendMessageToVpnIp(header.LightHouse, 0, vpnIp, mm, nb, out)
 	}
 }
 
@@ -271,9 +392,9 @@ type LightHouseHandler struct {
 	lh   *LightHouse
 	nb   []byte
 	out  []byte
+	pb   []byte
 	meta *NebulaMeta
-	iap  []IpAndPort
-	iapp []*IpAndPort
+	l    *logrus.Logger
 }
 
 func (lh *LightHouse) NewRequestHandler() *LightHouseHandler {
@@ -281,203 +402,252 @@ func (lh *LightHouse) NewRequestHandler() *LightHouseHandler {
 		lh:  lh,
 		nb:  make([]byte, 12, 12),
 		out: make([]byte, mtu),
+		l:   lh.l,
+		pb:  make([]byte, mtu),
 
 		meta: &NebulaMeta{
 			Details: &NebulaMetaDetails{},
 		},
 	}
 
-	lhh.resizeIpAndPorts(10)
-
 	return lhh
 }
 
+func (lh *LightHouse) metricRx(t NebulaMeta_MessageType, i int64) {
+	lh.metrics.Rx(header.MessageType(t), 0, i)
+}
+
+func (lh *LightHouse) metricTx(t NebulaMeta_MessageType, i int64) {
+	lh.metrics.Tx(header.MessageType(t), 0, i)
+}
+
 // This method is similar to Reset(), but it re-uses the pointer structs
 // so that we don't have to re-allocate them
 func (lhh *LightHouseHandler) resetMeta() *NebulaMeta {
 	details := lhh.meta.Details
-
-	details.Reset()
 	lhh.meta.Reset()
+
+	// Keep the array memory around
+	details.Ip4AndPorts = details.Ip4AndPorts[:0]
+	details.Ip6AndPorts = details.Ip6AndPorts[:0]
 	lhh.meta.Details = details
 
 	return lhh.meta
 }
 
-func (lhh *LightHouseHandler) resizeIpAndPorts(n int) {
-	if cap(lhh.iap) < n {
-		lhh.iap = make([]IpAndPort, n)
-		lhh.iapp = make([]*IpAndPort, n)
-
-		for i := range lhh.iap {
-			lhh.iapp[i] = &lhh.iap[i]
-		}
-	}
-	lhh.iap = lhh.iap[:n]
-	lhh.iapp = lhh.iapp[:n]
-}
-
-func (lhh *LightHouseHandler) setIpAndPortsFromNetIps(ips []udpAddr) []*IpAndPort {
-	lhh.resizeIpAndPorts(len(ips))
-	for i, e := range ips {
-		lhh.iap[i] = NewIpAndPortFromUDPAddr(e)
-	}
-	return lhh.iapp
-}
-
-func (lhh *LightHouseHandler) HandleRequest(rAddr *udpAddr, vpnIp uint32, p []byte, c *cert.NebulaCertificate, f EncWriter) {
-	lh := lhh.lh
+func (lhh *LightHouseHandler) HandleRequest(rAddr *udp.Addr, vpnIp iputil.VpnIp, p []byte, w udp.EncWriter) {
 	n := lhh.resetMeta()
-	err := proto.UnmarshalMerge(p, n)
+	err := n.Unmarshal(p)
 	if err != nil {
-		l.WithError(err).WithField("vpnIp", IntIp(vpnIp)).WithField("udpAddr", rAddr).
+		lhh.l.WithError(err).WithField("vpnIp", vpnIp).WithField("udpAddr", rAddr).
 			Error("Failed to unmarshal lighthouse packet")
 		//TODO: send recv_error?
 		return
 	}
 
 	if n.Details == nil {
-		l.WithField("vpnIp", IntIp(vpnIp)).WithField("udpAddr", rAddr).
+		lhh.l.WithField("vpnIp", vpnIp).WithField("udpAddr", rAddr).
 			Error("Invalid lighthouse update")
 		//TODO: send recv_error?
 		return
 	}
 
-	lh.metricRx(n.Type, 1)
+	lhh.lh.metricRx(n.Type, 1)
 
 	switch n.Type {
 	case NebulaMeta_HostQuery:
-		// Exit if we don't answer queries
-		if !lh.amLighthouse {
-			l.Debugln("I don't answer queries, but received from: ", rAddr)
-			return
-		}
-
-		//l.Debugln("Got Query")
-		ips, err := lh.Query(n.Details.VpnIp, f)
-		if err != nil {
-			//l.Debugf("Can't answer query %s from %s because error: %s", IntIp(n.Details.VpnIp), rAddr, err)
-			return
-		} else {
-			reqVpnIP := n.Details.VpnIp
-			n = lhh.resetMeta()
-			n.Type = NebulaMeta_HostQueryReply
-			n.Details.VpnIp = reqVpnIP
-			n.Details.IpAndPorts = lhh.setIpAndPortsFromNetIps(ips)
-			reply, err := proto.Marshal(n)
-			if err != nil {
-				l.WithError(err).WithField("vpnIp", IntIp(vpnIp)).Error("Failed to marshal lighthouse host query reply")
-				return
-			}
-			lh.metricTx(NebulaMeta_HostQueryReply, 1)
-			f.SendMessageToVpnIp(lightHouse, 0, vpnIp, reply, lhh.nb, lhh.out[:0])
-
-			// This signals the other side to punch some zero byte udp packets
-			ips, err = lh.Query(vpnIp, f)
-			if err != nil {
-				l.WithField("vpnIp", IntIp(vpnIp)).Debugln("Can't notify host to punch")
-				return
-			} else {
-				//l.Debugln("Notify host to punch", iap)
-				n = lhh.resetMeta()
-				n.Type = NebulaMeta_HostPunchNotification
-				n.Details.VpnIp = vpnIp
-				n.Details.IpAndPorts = lhh.setIpAndPortsFromNetIps(ips)
-				reply, _ := proto.Marshal(n)
-				lh.metricTx(NebulaMeta_HostPunchNotification, 1)
-				f.SendMessageToVpnIp(lightHouse, 0, reqVpnIP, reply, lhh.nb, lhh.out[:0])
-			}
-			//fmt.Println(reply, remoteaddr)
-		}
+		lhh.handleHostQuery(n, vpnIp, rAddr, w)
 
 	case NebulaMeta_HostQueryReply:
-		if !lh.IsLighthouseIP(vpnIp) {
-			return
-		}
-		for _, a := range n.Details.IpAndPorts {
-			//first := n.Details.IpAndPorts[0]
-			ans := NewUDPAddr(a.Ip, uint16(a.Port))
-			lh.AddRemote(n.Details.VpnIp, ans, false)
-		}
-		// Non-blocking attempt to trigger, skip if it would block
-		select {
-		case lh.handshakeTrigger <- n.Details.VpnIp:
-		default:
-		}
+		lhh.handleHostQueryReply(n, vpnIp)
 
 	case NebulaMeta_HostUpdateNotification:
-		//Simple check that the host sent this not someone else
-		if n.Details.VpnIp != vpnIp {
-			l.WithField("vpnIp", IntIp(vpnIp)).WithField("answer", IntIp(n.Details.VpnIp)).Debugln("Host sent invalid update")
-			return
-		}
-		for _, a := range n.Details.IpAndPorts {
-			ans := NewUDPAddr(a.Ip, uint16(a.Port))
-			lh.AddRemote(n.Details.VpnIp, ans, false)
-		}
+		lhh.handleHostUpdateNotification(n, vpnIp)
+
 	case NebulaMeta_HostMovedNotification:
 	case NebulaMeta_HostPunchNotification:
-		if !lh.IsLighthouseIP(vpnIp) {
+		lhh.handleHostPunchNotification(n, vpnIp, w)
+	}
+}
+
+func (lhh *LightHouseHandler) handleHostQuery(n *NebulaMeta, vpnIp iputil.VpnIp, addr *udp.Addr, w udp.EncWriter) {
+	// Exit if we don't answer queries
+	if !lhh.lh.amLighthouse {
+		if lhh.l.Level >= logrus.DebugLevel {
+			lhh.l.Debugln("I don't answer queries, but received from: ", addr)
+		}
+		return
+	}
+
+	//TODO: we can DRY this further
+	reqVpnIp := n.Details.VpnIp
+	//TODO: Maybe instead of marshalling into n we marshal into a new `r` to not nuke our current request data
+	found, ln, err := lhh.lh.queryAndPrepMessage(iputil.VpnIp(n.Details.VpnIp), func(c *cache) (int, error) {
+		n = lhh.resetMeta()
+		n.Type = NebulaMeta_HostQueryReply
+		n.Details.VpnIp = reqVpnIp
+
+		lhh.coalesceAnswers(c, n)
+
+		return n.MarshalTo(lhh.pb)
+	})
+
+	if !found {
+		return
+	}
+
+	if err != nil {
+		lhh.l.WithError(err).WithField("vpnIp", vpnIp).Error("Failed to marshal lighthouse host query reply")
+		return
+	}
+
+	lhh.lh.metricTx(NebulaMeta_HostQueryReply, 1)
+	w.SendMessageToVpnIp(header.LightHouse, 0, vpnIp, lhh.pb[:ln], lhh.nb, lhh.out[:0])
+
+	// This signals the other side to punch some zero byte udp packets
+	found, ln, err = lhh.lh.queryAndPrepMessage(vpnIp, func(c *cache) (int, error) {
+		n = lhh.resetMeta()
+		n.Type = NebulaMeta_HostPunchNotification
+		n.Details.VpnIp = uint32(vpnIp)
+
+		lhh.coalesceAnswers(c, n)
+
+		return n.MarshalTo(lhh.pb)
+	})
+
+	if !found {
+		return
+	}
+
+	if err != nil {
+		lhh.l.WithError(err).WithField("vpnIp", vpnIp).Error("Failed to marshal lighthouse host was queried for")
+		return
+	}
+
+	lhh.lh.metricTx(NebulaMeta_HostPunchNotification, 1)
+	w.SendMessageToVpnIp(header.LightHouse, 0, iputil.VpnIp(reqVpnIp), lhh.pb[:ln], lhh.nb, lhh.out[:0])
+}
+
+func (lhh *LightHouseHandler) coalesceAnswers(c *cache, n *NebulaMeta) {
+	if c.v4 != nil {
+		if c.v4.learned != nil {
+			n.Details.Ip4AndPorts = append(n.Details.Ip4AndPorts, c.v4.learned)
+		}
+		if c.v4.reported != nil && len(c.v4.reported) > 0 {
+			n.Details.Ip4AndPorts = append(n.Details.Ip4AndPorts, c.v4.reported...)
+		}
+	}
+
+	if c.v6 != nil {
+		if c.v6.learned != nil {
+			n.Details.Ip6AndPorts = append(n.Details.Ip6AndPorts, c.v6.learned)
+		}
+		if c.v6.reported != nil && len(c.v6.reported) > 0 {
+			n.Details.Ip6AndPorts = append(n.Details.Ip6AndPorts, c.v6.reported...)
+		}
+	}
+}
+
+func (lhh *LightHouseHandler) handleHostQueryReply(n *NebulaMeta, vpnIp iputil.VpnIp) {
+	if !lhh.lh.IsLighthouseIP(vpnIp) {
+		return
+	}
+
+	lhh.lh.Lock()
+	am := lhh.lh.unlockedGetRemoteList(iputil.VpnIp(n.Details.VpnIp))
+	am.Lock()
+	lhh.lh.Unlock()
+
+	certVpnIp := iputil.VpnIp(n.Details.VpnIp)
+	am.unlockedSetV4(vpnIp, certVpnIp, n.Details.Ip4AndPorts, lhh.lh.unlockedShouldAddV4)
+	am.unlockedSetV6(vpnIp, certVpnIp, n.Details.Ip6AndPorts, lhh.lh.unlockedShouldAddV6)
+	am.Unlock()
+
+	// Non-blocking attempt to trigger, skip if it would block
+	select {
+	case lhh.lh.handshakeTrigger <- iputil.VpnIp(n.Details.VpnIp):
+	default:
+	}
+}
+
+func (lhh *LightHouseHandler) handleHostUpdateNotification(n *NebulaMeta, vpnIp iputil.VpnIp) {
+	if !lhh.lh.amLighthouse {
+		if lhh.l.Level >= logrus.DebugLevel {
+			lhh.l.Debugln("I am not a lighthouse, do not take host updates: ", vpnIp)
+		}
+		return
+	}
+
+	//Simple check that the host sent this not someone else
+	if n.Details.VpnIp != uint32(vpnIp) {
+		if lhh.l.Level >= logrus.DebugLevel {
+			lhh.l.WithField("vpnIp", vpnIp).WithField("answer", iputil.VpnIp(n.Details.VpnIp)).Debugln("Host sent invalid update")
+		}
+		return
+	}
+
+	lhh.lh.Lock()
+	am := lhh.lh.unlockedGetRemoteList(vpnIp)
+	am.Lock()
+	lhh.lh.Unlock()
+
+	certVpnIp := iputil.VpnIp(n.Details.VpnIp)
+	am.unlockedSetV4(vpnIp, certVpnIp, n.Details.Ip4AndPorts, lhh.lh.unlockedShouldAddV4)
+	am.unlockedSetV6(vpnIp, certVpnIp, n.Details.Ip6AndPorts, lhh.lh.unlockedShouldAddV6)
+	am.Unlock()
+}
+
+func (lhh *LightHouseHandler) handleHostPunchNotification(n *NebulaMeta, vpnIp iputil.VpnIp, w udp.EncWriter) {
+	if !lhh.lh.IsLighthouseIP(vpnIp) {
+		return
+	}
+
+	empty := []byte{0}
+	punch := func(vpnPeer *udp.Addr) {
+		if vpnPeer == nil {
 			return
 		}
 
-		empty := []byte{0}
-		for _, a := range n.Details.IpAndPorts {
-			vpnPeer := NewUDPAddr(a.Ip, uint16(a.Port))
-			go func() {
-				time.Sleep(lh.punchDelay)
-				lh.metricHolepunchTx.Inc(1)
-				lh.punchConn.WriteTo(empty, vpnPeer)
+		go func() {
+			time.Sleep(lhh.lh.punchDelay)
+			lhh.lh.metricHolepunchTx.Inc(1)
+			lhh.lh.punchConn.WriteTo(empty, vpnPeer)
+		}()
 
-			}()
-			l.Debugf("Punching %s on %d for %s", IntIp(a.Ip), a.Port, IntIp(n.Details.VpnIp))
-		}
-		// This sends a nebula test packet to the host trying to contact us. In the case
-		// of a double nat or other difficult scenario, this may help establish
-		// a tunnel.
-		if lh.punchBack {
-			go func() {
-				time.Sleep(time.Second * 5)
-				l.Debugf("Sending a nebula test packet to vpn ip %s", IntIp(n.Details.VpnIp))
-				// TODO we have to allocate a new output buffer here since we are spawning a new goroutine
-				// for each punchBack packet. We should move this into a timerwheel or a single goroutine
-				// managed by a channel.
-				f.SendMessageToVpnIp(test, testRequest, n.Details.VpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
-			}()
+		if lhh.l.Level >= logrus.DebugLevel {
+			//TODO: lacking the ip we are actually punching on, old: l.Debugf("Punching %s on %d for %s", IntIp(a.Ip), a.Port, IntIp(n.Details.VpnIp))
+			lhh.l.Debugf("Punching on %d for %s", vpnPeer.Port, iputil.VpnIp(n.Details.VpnIp))
 		}
 	}
-}
 
-func (lh *LightHouse) metricRx(t NebulaMeta_MessageType, i int64) {
-	lh.metrics.Rx(NebulaMessageType(t), 0, i)
-}
-func (lh *LightHouse) metricTx(t NebulaMeta_MessageType, i int64) {
-	lh.metrics.Tx(NebulaMessageType(t), 0, i)
-}
+	for _, a := range n.Details.Ip4AndPorts {
+		punch(NewUDPAddrFromLH4(a))
+	}
 
-/*
-func (f *Interface) sendPathCheck(ci *ConnectionState, endpoint *net.UDPAddr, counter int) {
-	c := ci.messageCounter
-    b := HeaderEncode(nil, Version, uint8(path_check), 0, ci.remoteIndex, c)
-	ci.messageCounter++
+	for _, a := range n.Details.Ip6AndPorts {
+		punch(NewUDPAddrFromLH6(a))
+	}
 
-	if ci.eKey != nil {
-		msg := ci.eKey.EncryptDanger(b, nil, []byte(strconv.Itoa(counter)), c)
-		//msg := ci.eKey.EncryptDanger(b, nil, []byte(fmt.Sprintf("%d", counter)), c)
-		f.outside.WriteTo(msg, endpoint)
-		l.Debugf("path_check sent, remote index: %d, pathCounter %d", ci.remoteIndex, counter)
+	// This sends a nebula test packet to the host trying to contact us. In the case
+	// of a double nat or other difficult scenario, this may help establish
+	// a tunnel.
+	if lhh.lh.punchBack {
+		queryVpnIp := iputil.VpnIp(n.Details.VpnIp)
+		go func() {
+			time.Sleep(time.Second * 5)
+			if lhh.l.Level >= logrus.DebugLevel {
+				lhh.l.Debugf("Sending a nebula test packet to vpn ip %s", queryVpnIp)
+			}
+			//NOTE: we have to allocate a new output buffer here since we are spawning a new goroutine
+			// for each punchBack packet. We should move this into a timerwheel or a single goroutine
+			// managed by a channel.
+			w.SendMessageToVpnIp(header.Test, header.TestRequest, queryVpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
+		}()
 	}
 }
 
-func (f *Interface) sendPathCheckReply(ci *ConnectionState, endpoint *net.UDPAddr, counter []byte) {
-	c := ci.messageCounter
-    b := HeaderEncode(nil, Version, uint8(path_check_reply), 0, ci.remoteIndex, c)
-	ci.messageCounter++
-
-	if ci.eKey != nil {
-		msg := ci.eKey.EncryptDanger(b, nil, counter, c)
-		f.outside.WriteTo(msg, endpoint)
-		l.Debugln("path_check sent, remote index: ", ci.remoteIndex)
-	}
+// ipMaskContains checks if testIp is contained by ip after applying a cidr
+// zeros is 32 - bits from net.IPMask.Size()
+func ipMaskContains(ip iputil.VpnIp, zeros iputil.VpnIp, testIp iputil.VpnIp) bool {
+	return (testIp^ip)>>zeros == 0
 }
-*/

lighthouse_test.go

@@ -1,16 +1,32 @@
 package nebula
 
 import (
+	"fmt"
 	"net"
 	"testing"
 
 	"github.com/golang/protobuf/proto"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
+//TODO: Add a test to ensure udpAddr is copied and not reused
+
+func TestOldIPv4Only(t *testing.T) {
+	// This test ensures our new ipv6 enabled LH protobuf IpAndPorts works with the old style to enable backwards compatibility
+	b := []byte{8, 129, 130, 132, 80, 16, 10}
+	var m Ip4AndPort
+	err := proto.Unmarshal(b, &m)
+	assert.NoError(t, err)
+	assert.Equal(t, "10.1.1.1", iputil.VpnIp(m.GetIp()).String())
+}
+
 func TestNewLhQuery(t *testing.T) {
 	myIp := net.ParseIP("192.1.1.1")
-	myIpint := ip2int(myIp)
+	myIpint := iputil.Ip2VpnIp(myIp)
 
 	// Generating a new lh query should work
 	a := NewLhQueryByInt(myIpint)
@@ -29,65 +45,61 @@ func TestNewLhQuery(t *testing.T) {
 
 }
 
-func TestNewipandportfromudpaddr(t *testing.T) {
-	blah := NewUDPAddrFromString("1.2.2.3:12345")
-	meh := NewIpAndPortFromUDPAddr(*blah)
-	assert.Equal(t, uint32(16908803), meh.Ip)
-	assert.Equal(t, uint32(12345), meh.Port)
-}
-
-func TestSetipandportsfromudpaddrs(t *testing.T) {
-	blah := NewUDPAddrFromString("1.2.2.3:12345")
-	blah2 := NewUDPAddrFromString("9.9.9.9:47828")
-	group := []udpAddr{*blah, *blah2}
-	var lh *LightHouse
-	lhh := lh.NewRequestHandler()
-	result := lhh.setIpAndPortsFromNetIps(group)
-	assert.IsType(t, []*IpAndPort{}, result)
-	assert.Len(t, result, 2)
-	assert.Equal(t, uint32(0x01020203), result[0].Ip)
-	assert.Equal(t, uint32(12345), result[0].Port)
-	assert.Equal(t, uint32(0x09090909), result[1].Ip)
-	assert.Equal(t, uint32(47828), result[1].Port)
-	//t.Error(reflect.TypeOf(hah))
-
-}
-
 func Test_lhStaticMapping(t *testing.T) {
+	l := util.NewTestLogger()
 	lh1 := "10.128.0.2"
 	lh1IP := net.ParseIP(lh1)
 
-	udpServer, _ := NewListener("0.0.0.0", 0, true)
+	udpServer, _ := udp.NewListener(l, "0.0.0.0", 0, true, 2)
 
-	meh := NewLightHouse(true, 1, []uint32{ip2int(lh1IP)}, 10, 10003, udpServer, false, 1, false)
-	meh.AddRemote(ip2int(lh1IP), NewUDPAddr(ip2int(lh1IP), uint16(4242)), true)
+	meh := NewLightHouse(l, true, &net.IPNet{IP: net.IP{0, 0, 0, 1}, Mask: net.IPMask{255, 255, 255, 255}}, []iputil.VpnIp{iputil.Ip2VpnIp(lh1IP)}, 10, 10003, udpServer, false, 1, false)
+	meh.AddStaticRemote(iputil.Ip2VpnIp(lh1IP), udp.NewAddr(lh1IP, uint16(4242)))
 	err := meh.ValidateLHStaticEntries()
 	assert.Nil(t, err)
 
 	lh2 := "10.128.0.3"
 	lh2IP := net.ParseIP(lh2)
 
-	meh = NewLightHouse(true, 1, []uint32{ip2int(lh1IP), ip2int(lh2IP)}, 10, 10003, udpServer, false, 1, false)
-	meh.AddRemote(ip2int(lh1IP), NewUDPAddr(ip2int(lh1IP), uint16(4242)), true)
+	meh = NewLightHouse(l, true, &net.IPNet{IP: net.IP{0, 0, 0, 1}, Mask: net.IPMask{255, 255, 255, 255}}, []iputil.VpnIp{iputil.Ip2VpnIp(lh1IP), iputil.Ip2VpnIp(lh2IP)}, 10, 10003, udpServer, false, 1, false)
+	meh.AddStaticRemote(iputil.Ip2VpnIp(lh1IP), udp.NewAddr(lh1IP, uint16(4242)))
 	err = meh.ValidateLHStaticEntries()
 	assert.EqualError(t, err, "Lighthouse 10.128.0.3 does not have a static_host_map entry")
 }
 
 func BenchmarkLighthouseHandleRequest(b *testing.B) {
+	l := util.NewTestLogger()
 	lh1 := "10.128.0.2"
 	lh1IP := net.ParseIP(lh1)
 
-	udpServer, _ := NewListener("0.0.0.0", 0, true)
+	udpServer, _ := udp.NewListener(l, "0.0.0.0", 0, true, 2)
 
-	lh := NewLightHouse(true, 1, []uint32{ip2int(lh1IP)}, 10, 10003, udpServer, false, 1, false)
+	lh := NewLightHouse(l, true, &net.IPNet{IP: net.IP{0, 0, 0, 1}, Mask: net.IPMask{0, 0, 0, 0}}, []iputil.VpnIp{iputil.Ip2VpnIp(lh1IP)}, 10, 10003, udpServer, false, 1, false)
 
-	hAddr := NewUDPAddrFromString("4.5.6.7:12345")
-	hAddr2 := NewUDPAddrFromString("4.5.6.7:12346")
-	lh.addrMap[3] = []udpAddr{*hAddr, *hAddr2}
+	hAddr := udp.NewAddrFromString("4.5.6.7:12345")
+	hAddr2 := udp.NewAddrFromString("4.5.6.7:12346")
+	lh.addrMap[3] = NewRemoteList()
+	lh.addrMap[3].unlockedSetV4(
+		3,
+		3,
+		[]*Ip4AndPort{
+			NewIp4AndPort(hAddr.IP, uint32(hAddr.Port)),
+			NewIp4AndPort(hAddr2.IP, uint32(hAddr2.Port)),
+		},
+		func(iputil.VpnIp, *Ip4AndPort) bool { return true },
+	)
 
-	rAddr := NewUDPAddrFromString("1.2.2.3:12345")
-	rAddr2 := NewUDPAddrFromString("1.2.2.3:12346")
-	lh.addrMap[2] = []udpAddr{*rAddr, *rAddr2}
+	rAddr := udp.NewAddrFromString("1.2.2.3:12345")
+	rAddr2 := udp.NewAddrFromString("1.2.2.3:12346")
+	lh.addrMap[2] = NewRemoteList()
+	lh.addrMap[2].unlockedSetV4(
+		3,
+		3,
+		[]*Ip4AndPort{
+			NewIp4AndPort(rAddr.IP, uint32(rAddr.Port)),
+			NewIp4AndPort(rAddr2.IP, uint32(rAddr2.Port)),
+		},
+		func(iputil.VpnIp, *Ip4AndPort) bool { return true },
+	)
 
 	mw := &mockEncWriter{}
 
@@ -96,14 +108,14 @@ func BenchmarkLighthouseHandleRequest(b *testing.B) {
 		req := &NebulaMeta{
 			Type: NebulaMeta_HostQuery,
 			Details: &NebulaMetaDetails{
-				VpnIp:      4,
-				IpAndPorts: nil,
+				VpnIp:       4,
+				Ip4AndPorts: nil,
 			},
 		}
 		p, err := proto.Marshal(req)
 		assert.NoError(b, err)
 		for n := 0; n < b.N; n++ {
-			lhh.HandleRequest(rAddr, 2, p, nil, mw)
+			lhh.HandleRequest(rAddr, 2, p, mw)
 		}
 	})
 	b.Run("found", func(b *testing.B) {
@@ -111,75 +123,268 @@ func BenchmarkLighthouseHandleRequest(b *testing.B) {
 		req := &NebulaMeta{
 			Type: NebulaMeta_HostQuery,
 			Details: &NebulaMetaDetails{
-				VpnIp:      3,
-				IpAndPorts: nil,
+				VpnIp:       3,
+				Ip4AndPorts: nil,
 			},
 		}
 		p, err := proto.Marshal(req)
 		assert.NoError(b, err)
 
 		for n := 0; n < b.N; n++ {
-			lhh.HandleRequest(rAddr, 2, p, nil, mw)
+			lhh.HandleRequest(rAddr, 2, p, mw)
 		}
 	})
 }
 
-func Test_lhRemoteAllowList(t *testing.T) {
-	c := NewConfig()
-	c.Settings["remoteallowlist"] = map[interface{}]interface{}{
-		"10.20.0.0/12": false,
-	}
-	allowList, err := c.GetAllowList("remoteallowlist", false)
-	assert.Nil(t, err)
+func TestLighthouse_Memory(t *testing.T) {
+	l := util.NewTestLogger()
 
-	lh1 := "10.128.0.2"
-	lh1IP := net.ParseIP(lh1)
+	myUdpAddr0 := &udp.Addr{IP: net.ParseIP("10.0.0.2"), Port: 4242}
+	myUdpAddr1 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4242}
+	myUdpAddr2 := &udp.Addr{IP: net.ParseIP("172.16.0.2"), Port: 4242}
+	myUdpAddr3 := &udp.Addr{IP: net.ParseIP("100.152.0.2"), Port: 4242}
+	myUdpAddr4 := &udp.Addr{IP: net.ParseIP("24.15.0.2"), Port: 4242}
+	myUdpAddr5 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4243}
+	myUdpAddr6 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4244}
+	myUdpAddr7 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4245}
+	myUdpAddr8 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4246}
+	myUdpAddr9 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4247}
+	myUdpAddr10 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4248}
+	myUdpAddr11 := &udp.Addr{IP: net.ParseIP("192.168.0.2"), Port: 4249}
+	myVpnIp := iputil.Ip2VpnIp(net.ParseIP("10.128.0.2"))
 
-	udpServer, _ := NewListener("0.0.0.0", 0, true)
+	theirUdpAddr0 := &udp.Addr{IP: net.ParseIP("10.0.0.3"), Port: 4242}
+	theirUdpAddr1 := &udp.Addr{IP: net.ParseIP("192.168.0.3"), Port: 4242}
+	theirUdpAddr2 := &udp.Addr{IP: net.ParseIP("172.16.0.3"), Port: 4242}
+	theirUdpAddr3 := &udp.Addr{IP: net.ParseIP("100.152.0.3"), Port: 4242}
+	theirUdpAddr4 := &udp.Addr{IP: net.ParseIP("24.15.0.3"), Port: 4242}
+	theirVpnIp := iputil.Ip2VpnIp(net.ParseIP("10.128.0.3"))
 
-	lh := NewLightHouse(true, 1, []uint32{ip2int(lh1IP)}, 10, 10003, udpServer, false, 1, false)
-	lh.SetRemoteAllowList(allowList)
+	udpServer, _ := udp.NewListener(l, "0.0.0.0", 0, true, 2)
+	lh := NewLightHouse(l, true, &net.IPNet{IP: net.IP{10, 128, 0, 1}, Mask: net.IPMask{255, 255, 255, 0}}, []iputil.VpnIp{}, 10, 10003, udpServer, false, 1, false)
+	lhh := lh.NewRequestHandler()
 
-	remote1 := "10.20.0.3"
-	remote1IP := net.ParseIP(remote1)
-	lh.AddRemote(ip2int(remote1IP), NewUDPAddr(ip2int(remote1IP), uint16(4242)), true)
-	assert.Nil(t, lh.addrMap[ip2int(remote1IP)])
+	// Test that my first update responds with just that
+	newLHHostUpdate(myUdpAddr0, myVpnIp, []*udp.Addr{myUdpAddr1, myUdpAddr2}, lhh)
+	r := newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, myUdpAddr1, myUdpAddr2)
 
-	remote2 := "10.128.0.3"
-	remote2IP := net.ParseIP(remote2)
-	remote2UDPAddr := NewUDPAddr(ip2int(remote2IP), uint16(4242))
+	// Ensure we don't accumulate addresses
+	newLHHostUpdate(myUdpAddr0, myVpnIp, []*udp.Addr{myUdpAddr3}, lhh)
+	r = newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, myUdpAddr3)
 
-	lh.AddRemote(ip2int(remote2IP), remote2UDPAddr, true)
-	assert.Equal(t, remote2UDPAddr, &lh.addrMap[ip2int(remote2IP)][0])
+	// Grow it back to 2
+	newLHHostUpdate(myUdpAddr0, myVpnIp, []*udp.Addr{myUdpAddr1, myUdpAddr4}, lhh)
+	r = newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, myUdpAddr1, myUdpAddr4)
+
+	// Update a different host
+	newLHHostUpdate(theirUdpAddr0, theirVpnIp, []*udp.Addr{theirUdpAddr1, theirUdpAddr2, theirUdpAddr3, theirUdpAddr4}, lhh)
+	r = newLHHostRequest(theirUdpAddr0, theirVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, theirUdpAddr1, theirUdpAddr2, theirUdpAddr3, theirUdpAddr4)
+
+	// Make sure we didn't get changed
+	r = newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, myUdpAddr1, myUdpAddr4)
+
+	// Ensure proper ordering and limiting
+	// Send 12 addrs, get 10 back, the last 2 removed, allowing the duplicate to remain (clients dedupe)
+	newLHHostUpdate(
+		myUdpAddr0,
+		myVpnIp,
+		[]*udp.Addr{
+			myUdpAddr1,
+			myUdpAddr2,
+			myUdpAddr3,
+			myUdpAddr4,
+			myUdpAddr5,
+			myUdpAddr5, //Duplicated on purpose
+			myUdpAddr6,
+			myUdpAddr7,
+			myUdpAddr8,
+			myUdpAddr9,
+			myUdpAddr10,
+			myUdpAddr11, // This should get cut
+		}, lhh)
+
+	r = newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(
+		t,
+		r.msg.Details.Ip4AndPorts,
+		myUdpAddr1, myUdpAddr2, myUdpAddr3, myUdpAddr4, myUdpAddr5, myUdpAddr5, myUdpAddr6, myUdpAddr7, myUdpAddr8, myUdpAddr9,
+	)
+
+	// Make sure we won't add ips in our vpn network
+	bad1 := &udp.Addr{IP: net.ParseIP("10.128.0.99"), Port: 4242}
+	bad2 := &udp.Addr{IP: net.ParseIP("10.128.0.100"), Port: 4242}
+	good := &udp.Addr{IP: net.ParseIP("1.128.0.99"), Port: 4242}
+	newLHHostUpdate(myUdpAddr0, myVpnIp, []*udp.Addr{bad1, bad2, good}, lhh)
+	r = newLHHostRequest(myUdpAddr0, myVpnIp, myVpnIp, lhh)
+	assertIp4InArray(t, r.msg.Details.Ip4AndPorts, good)
 }
 
-//func NewLightHouse(amLighthouse bool, myIp uint32, ips []string, interval int, nebulaPort int, pc *udpConn, punchBack bool) *LightHouse {
-
-/*
-func TestLHQuery(t *testing.T) {
-	//n := NewLhQueryByIpString("10.128.0.3")
-	_, myNet, _ := net.ParseCIDR("10.128.0.0/16")
-	m := NewHostMap(myNet)
-	y, _ := net.ResolveUDPAddr("udp", "10.128.0.3:11111")
-	m.Add(ip2int(net.ParseIP("127.0.0.1")), y)
-	//t.Errorf("%s", m)
-	_ = m
-
-	_, n, _ := net.ParseCIDR("127.0.0.1/8")
-
-	/*udpServer, err := net.ListenUDP("udp", &net.UDPAddr{Port: 10009})
-	if err != nil {
-		t.Errorf("%s", err)
+func newLHHostRequest(fromAddr *udp.Addr, myVpnIp, queryVpnIp iputil.VpnIp, lhh *LightHouseHandler) testLhReply {
+	req := &NebulaMeta{
+		Type: NebulaMeta_HostQuery,
+		Details: &NebulaMetaDetails{
+			VpnIp: uint32(queryVpnIp),
+		},
 	}
 
-	meh := NewLightHouse(n, m, []string{"10.128.0.2"}, false, 10, 10003, 10004)
-	//t.Error(m.Hosts)
-	meh2, err := meh.Query(ip2int(net.ParseIP("10.128.0.3")))
-	t.Error(err)
+	b, err := req.Marshal()
 	if err != nil {
-		return
+		panic(err)
 	}
-	t.Errorf("%s", meh2)
-	t.Errorf("%s", n)
+
+	w := &testEncWriter{}
+	lhh.HandleRequest(fromAddr, myVpnIp, b, w)
+	return w.lastReply
+}
+
+func newLHHostUpdate(fromAddr *udp.Addr, vpnIp iputil.VpnIp, addrs []*udp.Addr, lhh *LightHouseHandler) {
+	req := &NebulaMeta{
+		Type: NebulaMeta_HostUpdateNotification,
+		Details: &NebulaMetaDetails{
+			VpnIp:       uint32(vpnIp),
+			Ip4AndPorts: make([]*Ip4AndPort, len(addrs)),
+		},
+	}
+
+	for k, v := range addrs {
+		req.Details.Ip4AndPorts[k] = &Ip4AndPort{Ip: uint32(iputil.Ip2VpnIp(v.IP)), Port: uint32(v.Port)}
+	}
+
+	b, err := req.Marshal()
+	if err != nil {
+		panic(err)
+	}
+
+	w := &testEncWriter{}
+	lhh.HandleRequest(fromAddr, vpnIp, b, w)
+}
+
+//TODO: this is a RemoteList test
+//func Test_lhRemoteAllowList(t *testing.T) {
+//	l := NewTestLogger()
+//	c := NewConfig(l)
+//	c.Settings["remoteallowlist"] = map[interface{}]interface{}{
+//		"10.20.0.0/12": false,
+//	}
+//	allowList, err := c.GetAllowList("remoteallowlist", false)
+//	assert.Nil(t, err)
+//
+//	lh1 := "10.128.0.2"
+//	lh1IP := net.ParseIP(lh1)
+//
+//	udpServer, _ := NewListener(l, "0.0.0.0", 0, true)
+//
+//	lh := NewLightHouse(l, true, &net.IPNet{IP: net.IP{0, 0, 0, 1}, Mask: net.IPMask{255, 255, 255, 0}}, []uint32{ip2int(lh1IP)}, 10, 10003, udpServer, false, 1, false)
+//	lh.SetRemoteAllowList(allowList)
+//
+//	// A disallowed ip should not enter the cache but we should end up with an empty entry in the addrMap
+//	remote1IP := net.ParseIP("10.20.0.3")
+//	remotes := lh.unlockedGetRemoteList(ip2int(remote1IP))
+//	remotes.unlockedPrependV4(ip2int(remote1IP), NewIp4AndPort(remote1IP, 4242))
+//	assert.NotNil(t, lh.addrMap[ip2int(remote1IP)])
+//	assert.Empty(t, lh.addrMap[ip2int(remote1IP)].CopyAddrs([]*net.IPNet{}))
+//
+//	// Make sure a good ip enters the cache and addrMap
+//	remote2IP := net.ParseIP("10.128.0.3")
+//	remote2UDPAddr := NewUDPAddr(remote2IP, uint16(4242))
+//	lh.addRemoteV4(ip2int(remote2IP), ip2int(remote2IP), NewIp4AndPort(remote2UDPAddr.IP, uint32(remote2UDPAddr.Port)), false, false)
+//	assertUdpAddrInArray(t, lh.addrMap[ip2int(remote2IP)].CopyAddrs([]*net.IPNet{}), remote2UDPAddr)
+//
+//	// Another good ip gets into the cache, ordering is inverted
+//	remote3IP := net.ParseIP("10.128.0.4")
+//	remote3UDPAddr := NewUDPAddr(remote3IP, uint16(4243))
+//	lh.addRemoteV4(ip2int(remote2IP), ip2int(remote2IP), NewIp4AndPort(remote3UDPAddr.IP, uint32(remote3UDPAddr.Port)), false, false)
+//	assertUdpAddrInArray(t, lh.addrMap[ip2int(remote2IP)].CopyAddrs([]*net.IPNet{}), remote2UDPAddr, remote3UDPAddr)
+//
+//	// If we exceed the length limit we should only have the most recent addresses
+//	addedAddrs := []*udpAddr{}
+//	for i := 0; i < 11; i++ {
+//		remoteUDPAddr := NewUDPAddr(net.IP{10, 128, 0, 4}, uint16(4243+i))
+//		lh.addRemoteV4(ip2int(remote2IP), ip2int(remote2IP), NewIp4AndPort(remoteUDPAddr.IP, uint32(remoteUDPAddr.Port)), false, false)
+//		// The first entry here is a duplicate, don't add it to the assert list
+//		if i != 0 {
+//			addedAddrs = append(addedAddrs, remoteUDPAddr)
+//		}
+//	}
+//
+//	// We should only have the last 10 of what we tried to add
+//	assert.True(t, len(addedAddrs) >= 10, "We should have tried to add at least 10 addresses")
+//	assertUdpAddrInArray(
+//		t,
+//		lh.addrMap[ip2int(remote2IP)].CopyAddrs([]*net.IPNet{}),
+//		addedAddrs[0],
+//		addedAddrs[1],
+//		addedAddrs[2],
+//		addedAddrs[3],
+//		addedAddrs[4],
+//		addedAddrs[5],
+//		addedAddrs[6],
+//		addedAddrs[7],
+//		addedAddrs[8],
+//		addedAddrs[9],
+//	)
+//}
+
+func Test_ipMaskContains(t *testing.T) {
+	assert.True(t, ipMaskContains(iputil.Ip2VpnIp(net.ParseIP("10.0.0.1")), 32-24, iputil.Ip2VpnIp(net.ParseIP("10.0.0.255"))))
+	assert.False(t, ipMaskContains(iputil.Ip2VpnIp(net.ParseIP("10.0.0.1")), 32-24, iputil.Ip2VpnIp(net.ParseIP("10.0.1.1"))))
+	assert.True(t, ipMaskContains(iputil.Ip2VpnIp(net.ParseIP("10.0.0.1")), 32, iputil.Ip2VpnIp(net.ParseIP("10.0.1.1"))))
+}
+
+type testLhReply struct {
+	nebType    header.MessageType
+	nebSubType header.MessageSubType
+	vpnIp      iputil.VpnIp
+	msg        *NebulaMeta
+}
+
+type testEncWriter struct {
+	lastReply testLhReply
+}
+
+func (tw *testEncWriter) SendMessageToVpnIp(t header.MessageType, st header.MessageSubType, vpnIp iputil.VpnIp, p, _, _ []byte) {
+	tw.lastReply = testLhReply{
+		nebType:    t,
+		nebSubType: st,
+		vpnIp:      vpnIp,
+		msg:        &NebulaMeta{},
+	}
+
+	err := proto.Unmarshal(p, tw.lastReply.msg)
+	if err != nil {
+		panic(err)
+	}
+}
+
+// assertIp4InArray asserts every address in want is at the same position in have and that the lengths match
+func assertIp4InArray(t *testing.T, have []*Ip4AndPort, want ...*udp.Addr) {
+	assert.Len(t, have, len(want))
+	for k, w := range want {
+		if !(have[k].Ip == uint32(iputil.Ip2VpnIp(w.IP)) && have[k].Port == uint32(w.Port)) {
+			assert.Fail(t, fmt.Sprintf("Response did not contain: %v:%v at %v; %v", w.IP, w.Port, k, translateV4toUdpAddr(have)))
+		}
+	}
+}
+
+// assertUdpAddrInArray asserts every address in want is at the same position in have and that the lengths match
+func assertUdpAddrInArray(t *testing.T, have []*udp.Addr, want ...*udp.Addr) {
+	assert.Len(t, have, len(want))
+	for k, w := range want {
+		if !(have[k].IP.Equal(w.IP) && have[k].Port == w.Port) {
+			assert.Fail(t, fmt.Sprintf("Response did not contain: %v at %v; %v", w, k, have))
+		}
+	}
+}
+
+func translateV4toUdpAddr(ips []*Ip4AndPort) []*udp.Addr {
+	addrs := make([]*udp.Addr, len(ips))
+	for k, v := range ips {
+		addrs[k] = NewUDPAddrFromLH4(v)
+	}
+	return addrs
 }
-*/

logger.go

@@ -2,8 +2,12 @@ package nebula
 
 import (
 	"errors"
+	"fmt"
+	"strings"
+	"time"
 
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
 )
 
 type ContextualError struct {
@@ -37,3 +41,38 @@ func (ce *ContextualError) Log(lr *logrus.Logger) {
 		lr.WithFields(ce.Fields).Error(ce.Context)
 	}
 }
+
+func configLogger(l *logrus.Logger, c *config.C) 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)
+
+	disableTimestamp := c.GetBool("logging.disable_timestamp", false)
+	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,
+			DisableTimestamp: disableTimestamp,
+		}
+	case "json":
+		l.Formatter = &logrus.JSONFormatter{
+			TimestampFormat:  timestampFormat,
+			DisableTimestamp: disableTimestamp,
+		}
+	default:
+		return fmt.Errorf("unknown log format `%s`. possible formats: %s", logFormat, []string{"text", "json"})
+	}
+
+	return nil
+}

main.go

@@ -1,32 +1,39 @@
 package nebula
 
 import (
+	"context"
 	"encoding/binary"
 	"fmt"
 	"net"
-	"strconv"
-	"strings"
 	"time"
 
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/iputil"
 	"github.com/slackhq/nebula/sshd"
+	"github.com/slackhq/nebula/udp"
 	"gopkg.in/yaml.v2"
 )
 
-// The caller should provide a real logger, we have one just in case
-var l = logrus.New()
-
 type m map[string]interface{}
 
-func Main(config *Config, configTest bool, buildVersion string, logger *logrus.Logger, tunFd *int) (*Control, error) {
-	l = logger
+func Main(c *config.C, configTest bool, buildVersion string, logger *logrus.Logger, tunFd *int) (retcon *Control, reterr error) {
+	ctx, cancel := context.WithCancel(context.Background())
+	// Automatically cancel the context if Main returns an error, to signal all created goroutines to quit.
+	defer func() {
+		if reterr != nil {
+			cancel()
+		}
+	}()
+
+	l := logger
 	l.Formatter = &logrus.TextFormatter{
 		FullTimestamp: true,
 	}
 
 	// Print the config if in test, the exit comes later
 	if configTest {
-		b, err := yaml.Marshal(config.Settings)
+		b, err := yaml.Marshal(c.Settings)
 		if err != nil {
 			return nil, err
 		}
@@ -35,34 +42,33 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		l.Println(string(b))
 	}
 
-	err := configLogger(config)
+	err := configLogger(l, c)
 	if err != nil {
 		return nil, NewContextualError("Failed to configure the logger", nil, err)
 	}
 
-	config.RegisterReloadCallback(func(c *Config) {
-		err := configLogger(c)
+	c.RegisterReloadCallback(func(c *config.C) {
+		err := configLogger(l, c)
 		if err != nil {
 			l.WithError(err).Error("Failed to configure the logger")
 		}
 	})
 
-	// trustedCAs is currently a global, so loadCA operates on that global directly
-	trustedCAs, err = loadCAFromConfig(config)
+	caPool, err := loadCAFromConfig(l, c)
 	if err != nil {
 		//The errors coming out of loadCA are already nicely formatted
 		return nil, NewContextualError("Failed to load ca from config", nil, err)
 	}
-	l.WithField("fingerprints", trustedCAs.GetFingerprints()).Debug("Trusted CA fingerprints")
+	l.WithField("fingerprints", caPool.GetFingerprints()).Debug("Trusted CA fingerprints")
 
-	cs, err := NewCertStateFromConfig(config)
+	cs, err := NewCertStateFromConfig(c)
 	if err != nil {
 		//The errors coming out of NewCertStateFromConfig are already nicely formatted
 		return nil, NewContextualError("Failed to load certificate from config", nil, err)
 	}
 	l.WithField("cert", cs.certificate).Debug("Client nebula certificate")
 
-	fw, err := NewFirewallFromConfig(cs.certificate, config)
+	fw, err := NewFirewallFromConfig(l, cs.certificate, c)
 	if err != nil {
 		return nil, NewContextualError("Error while loading firewall rules", nil, err)
 	}
@@ -70,19 +76,20 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 
 	// TODO: make sure mask is 4 bytes
 	tunCidr := cs.certificate.Details.Ips[0]
-	routes, err := parseRoutes(config, tunCidr)
+	routes, err := parseRoutes(c, tunCidr)
 	if err != nil {
 		return nil, NewContextualError("Could not parse tun.routes", nil, err)
 	}
-	unsafeRoutes, err := parseUnsafeRoutes(config, tunCidr)
+	unsafeRoutes, err := parseUnsafeRoutes(c, tunCidr)
 	if err != nil {
 		return nil, NewContextualError("Could not parse tun.unsafe_routes", nil, err)
 	}
 
 	ssh, err := sshd.NewSSHServer(l.WithField("subsystem", "sshd"))
-	wireSSHReload(ssh, config)
-	if config.GetBool("sshd.enabled", false) {
-		err = configSSH(ssh, config)
+	wireSSHReload(l, ssh, c)
+	var sshStart func()
+	if c.GetBool("sshd.enabled", false) {
+		sshStart, err = configSSH(l, ssh, c)
 		if err != nil {
 			return nil, NewContextualError("Error while configuring the sshd", nil, err)
 		}
@@ -96,7 +103,7 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 	var routines int
 
 	// If `routines` is set, use that and ignore the specific values
-	if routines = config.GetInt("routines", 0); routines != 0 {
+	if routines = c.GetInt("routines", 0); routines != 0 {
 		if routines < 1 {
 			routines = 1
 		}
@@ -105,8 +112,8 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		}
 	} else {
 		// deprecated and undocumented
-		tunQueues := config.GetInt("tun.routines", 1)
-		udpQueues := config.GetInt("listen.routines", 1)
+		tunQueues := c.GetInt("tun.routines", 1)
+		udpQueues := c.GetInt("listen.routines", 1)
 		if tunQueues > udpQueues {
 			routines = tunQueues
 		} else {
@@ -120,8 +127,8 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 	// EXPERIMENTAL
 	// Intentionally not documented yet while we do more testing and determine
 	// a good default value.
-	conntrackCacheTimeout := config.GetDuration("firewall.conntrack.routine_cache_timeout", 0)
-	if routines > 1 && !config.IsSet("firewall.conntrack.routine_cache_timeout") {
+	conntrackCacheTimeout := c.GetDuration("firewall.conntrack.routine_cache_timeout", 0)
+	if routines > 1 && !c.IsSet("firewall.conntrack.routine_cache_timeout") {
 		// Use a different default if we are running with multiple routines
 		conntrackCacheTimeout = 1 * time.Second
 	}
@@ -131,28 +138,30 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 
 	var tun Inside
 	if !configTest {
-		config.CatchHUP()
+		c.CatchHUP(ctx)
 
 		switch {
-		case config.GetBool("tun.disabled", false):
-			tun = newDisabledTun(tunCidr, config.GetInt("tun.tx_queue", 500), config.GetBool("stats.message_metrics", false), l)
+		case c.GetBool("tun.disabled", false):
+			tun = newDisabledTun(tunCidr, c.GetInt("tun.tx_queue", 500), c.GetBool("stats.message_metrics", false), l)
 		case tunFd != nil:
 			tun, err = newTunFromFd(
+				l,
 				*tunFd,
 				tunCidr,
-				config.GetInt("tun.mtu", DEFAULT_MTU),
+				c.GetInt("tun.mtu", DEFAULT_MTU),
 				routes,
 				unsafeRoutes,
-				config.GetInt("tun.tx_queue", 500),
+				c.GetInt("tun.tx_queue", 500),
 			)
 		default:
 			tun, err = newTun(
-				config.GetString("tun.dev", ""),
+				l,
+				c.GetString("tun.dev", ""),
 				tunCidr,
-				config.GetInt("tun.mtu", DEFAULT_MTU),
+				c.GetInt("tun.mtu", DEFAULT_MTU),
 				routes,
 				unsafeRoutes,
-				config.GetInt("tun.tx_queue", 500),
+				c.GetInt("tun.tx_queue", 500),
 				routines > 1,
 			)
 		}
@@ -162,17 +171,23 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		}
 	}
 
+	defer func() {
+		if reterr != nil {
+			tun.Close()
+		}
+	}()
+
 	// set up our UDP listener
-	udpConns := make([]*udpConn, routines)
-	port := config.GetInt("listen.port", 0)
+	udpConns := make([]*udp.Conn, routines)
+	port := c.GetInt("listen.port", 0)
 
 	if !configTest {
 		for i := 0; i < routines; i++ {
-			udpServer, err := NewListener(config.GetString("listen.host", "0.0.0.0"), port, routines > 1)
+			udpServer, err := udp.NewListener(l, c.GetString("listen.host", "0.0.0.0"), port, routines > 1, c.GetInt("listen.batch", 64))
 			if err != nil {
 				return nil, NewContextualError("Failed to open udp listener", m{"queue": i}, err)
 			}
-			udpServer.reloadConfig(config)
+			udpServer.ReloadConfig(c)
 			udpConns[i] = udpServer
 
 			// If port is dynamic, discover it
@@ -188,7 +203,7 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 
 	// Set up my internal host map
 	var preferredRanges []*net.IPNet
-	rawPreferredRanges := config.GetStringSlice("preferred_ranges", []string{})
+	rawPreferredRanges := c.GetStringSlice("preferred_ranges", []string{})
 	// First, check if 'preferred_ranges' is set and fallback to 'local_range'
 	if len(rawPreferredRanges) > 0 {
 		for _, rawPreferredRange := range rawPreferredRanges {
@@ -203,7 +218,7 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 	// local_range was superseded by preferred_ranges. If it is still present,
 	// merge the local_range setting into preferred_ranges. We will probably
 	// deprecate local_range and remove in the future.
-	rawLocalRange := config.GetString("local_range", "")
+	rawLocalRange := c.GetString("local_range", "")
 	if rawLocalRange != "" {
 		_, localRange, err := net.ParseCIDR(rawLocalRange)
 		if err != nil {
@@ -224,10 +239,10 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		}
 	}
 
-	hostMap := NewHostMap("main", tunCidr, preferredRanges)
-	hostMap.SetDefaultRoute(ip2int(net.ParseIP(config.GetString("default_route", "0.0.0.0"))))
+	hostMap := NewHostMap(l, "main", tunCidr, preferredRanges)
+
 	hostMap.addUnsafeRoutes(&unsafeRoutes)
-	hostMap.metricsEnabled = config.GetBool("stats.message_metrics", false)
+	hostMap.metricsEnabled = c.GetBool("stats.message_metrics", false)
 
 	l.WithField("network", hostMap.vpnCIDR).WithField("preferredRanges", hostMap.preferredRanges).Info("Main HostMap created")
 
@@ -236,26 +251,26 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		go hostMap.Promoter(config.GetInt("promoter.interval"))
 	*/
 
-	punchy := NewPunchyFromConfig(config)
+	punchy := NewPunchyFromConfig(c)
 	if punchy.Punch && !configTest {
 		l.Info("UDP hole punching enabled")
-		go hostMap.Punchy(udpConns[0])
+		go hostMap.Punchy(ctx, udpConns[0])
 	}
 
-	amLighthouse := config.GetBool("lighthouse.am_lighthouse", false)
+	amLighthouse := c.GetBool("lighthouse.am_lighthouse", false)
 
 	// fatal if am_lighthouse is enabled but we are using an ephemeral port
-	if amLighthouse && (config.GetInt("listen.port", 0) == 0) {
+	if amLighthouse && (c.GetInt("listen.port", 0) == 0) {
 		return nil, NewContextualError("lighthouse.am_lighthouse enabled on node but no port number is set in config", nil, nil)
 	}
 
 	// warn if am_lighthouse is enabled but upstream lighthouses exists
-	rawLighthouseHosts := config.GetStringSlice("lighthouse.hosts", []string{})
+	rawLighthouseHosts := c.GetStringSlice("lighthouse.hosts", []string{})
 	if amLighthouse && len(rawLighthouseHosts) != 0 {
 		l.Warn("lighthouse.am_lighthouse enabled on node but upstream lighthouses exist in config")
 	}
 
-	lighthouseHosts := make([]uint32, len(rawLighthouseHosts))
+	lighthouseHosts := make([]iputil.VpnIp, len(rawLighthouseHosts))
 	for i, host := range rawLighthouseHosts {
 		ip := net.ParseIP(host)
 		if ip == nil {
@@ -264,66 +279,57 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		if !tunCidr.Contains(ip) {
 			return nil, NewContextualError("lighthouse host is not in our subnet, invalid", m{"vpnIp": ip, "network": tunCidr.String()}, nil)
 		}
-		lighthouseHosts[i] = ip2int(ip)
+		lighthouseHosts[i] = iputil.Ip2VpnIp(ip)
 	}
 
 	lightHouse := NewLightHouse(
+		l,
 		amLighthouse,
-		ip2int(tunCidr.IP),
+		tunCidr,
 		lighthouseHosts,
 		//TODO: change to a duration
-		config.GetInt("lighthouse.interval", 10),
+		c.GetInt("lighthouse.interval", 10),
 		uint32(port),
 		udpConns[0],
 		punchy.Respond,
 		punchy.Delay,
-		config.GetBool("stats.lighthouse_metrics", false),
+		c.GetBool("stats.lighthouse_metrics", false),
 	)
 
-	remoteAllowList, err := config.GetAllowList("lighthouse.remote_allow_list", false)
+	remoteAllowList, err := NewRemoteAllowListFromConfig(c, "lighthouse.remote_allow_list", "lighthouse.remote_allow_ranges")
 	if err != nil {
 		return nil, NewContextualError("Invalid lighthouse.remote_allow_list", nil, err)
 	}
 	lightHouse.SetRemoteAllowList(remoteAllowList)
 
-	localAllowList, err := config.GetAllowList("lighthouse.local_allow_list", true)
+	localAllowList, err := NewLocalAllowListFromConfig(c, "lighthouse.local_allow_list")
 	if err != nil {
 		return nil, NewContextualError("Invalid lighthouse.local_allow_list", nil, err)
 	}
 	lightHouse.SetLocalAllowList(localAllowList)
 
 	//TODO: Move all of this inside functions in lighthouse.go
-	for k, v := range config.GetMap("static_host_map", map[interface{}]interface{}{}) {
-		vpnIp := net.ParseIP(fmt.Sprintf("%v", k))
-		if !tunCidr.Contains(vpnIp) {
+	for k, v := range c.GetMap("static_host_map", map[interface{}]interface{}{}) {
+		ip := net.ParseIP(fmt.Sprintf("%v", k))
+		vpnIp := iputil.Ip2VpnIp(ip)
+		if !tunCidr.Contains(ip) {
 			return nil, NewContextualError("static_host_map key is not in our subnet, invalid", m{"vpnIp": vpnIp, "network": tunCidr.String()}, nil)
 		}
 		vals, ok := v.([]interface{})
 		if ok {
 			for _, v := range vals {
-				parts := strings.Split(fmt.Sprintf("%v", v), ":")
-				addr, err := net.ResolveIPAddr("ip", parts[0])
-				if err == nil {
-					ip := addr.IP
-					port, err := strconv.Atoi(parts[1])
-					if err != nil {
-						return nil, NewContextualError("Static host address could not be parsed", m{"vpnIp": vpnIp}, err)
-					}
-					lightHouse.AddRemote(ip2int(vpnIp), NewUDPAddr(ip2int(ip), uint16(port)), true)
-				}
-			}
-		} else {
-			//TODO: make this all a helper
-			parts := strings.Split(fmt.Sprintf("%v", v), ":")
-			addr, err := net.ResolveIPAddr("ip", parts[0])
-			if err == nil {
-				ip := addr.IP
-				port, err := strconv.Atoi(parts[1])
+				ip, port, err := udp.ParseIPAndPort(fmt.Sprintf("%v", v))
 				if err != nil {
 					return nil, NewContextualError("Static host address could not be parsed", m{"vpnIp": vpnIp}, err)
 				}
-				lightHouse.AddRemote(ip2int(vpnIp), NewUDPAddr(ip2int(ip), uint16(port)), true)
+				lightHouse.AddStaticRemote(vpnIp, udp.NewAddr(ip, port))
 			}
+		} else {
+			ip, port, err := udp.ParseIPAndPort(fmt.Sprintf("%v", v))
+			if err != nil {
+				return nil, NewContextualError("Static host address could not be parsed", m{"vpnIp": vpnIp}, err)
+			}
+			lightHouse.AddStaticRemote(vpnIp, udp.NewAddr(ip, port))
 		}
 	}
 
@@ -333,51 +339,60 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 	}
 
 	var messageMetrics *MessageMetrics
-	if config.GetBool("stats.message_metrics", false) {
+	if c.GetBool("stats.message_metrics", false) {
 		messageMetrics = newMessageMetrics()
 	} else {
 		messageMetrics = newMessageMetricsOnlyRecvError()
 	}
 
 	handshakeConfig := HandshakeConfig{
-		tryInterval:   config.GetDuration("handshakes.try_interval", DefaultHandshakeTryInterval),
-		retries:       config.GetInt("handshakes.retries", DefaultHandshakeRetries),
-		waitRotation:  config.GetInt("handshakes.wait_rotation", DefaultHandshakeWaitRotation),
-		triggerBuffer: config.GetInt("handshakes.trigger_buffer", DefaultHandshakeTriggerBuffer),
+		tryInterval:   c.GetDuration("handshakes.try_interval", DefaultHandshakeTryInterval),
+		retries:       c.GetInt("handshakes.retries", DefaultHandshakeRetries),
+		triggerBuffer: c.GetInt("handshakes.trigger_buffer", DefaultHandshakeTriggerBuffer),
 
 		messageMetrics: messageMetrics,
 	}
 
-	handshakeManager := NewHandshakeManager(tunCidr, preferredRanges, hostMap, lightHouse, udpConns[0], handshakeConfig)
+	handshakeManager := NewHandshakeManager(l, tunCidr, preferredRanges, hostMap, lightHouse, udpConns[0], handshakeConfig)
 	lightHouse.handshakeTrigger = handshakeManager.trigger
 
 	//TODO: These will be reused for psk
 	//handshakeMACKey := config.GetString("handshake_mac.key", "")
 	//handshakeAcceptedMACKeys := config.GetStringSlice("handshake_mac.accepted_keys", []string{})
 
-	serveDns := config.GetBool("lighthouse.serve_dns", false)
-	checkInterval := config.GetInt("timers.connection_alive_interval", 5)
-	pendingDeletionInterval := config.GetInt("timers.pending_deletion_interval", 10)
+	serveDns := false
+	if c.GetBool("lighthouse.serve_dns", false) {
+		if c.GetBool("lighthouse.am_lighthouse", false) {
+			serveDns = true
+		} else {
+			l.Warn("DNS server refusing to run because this host is not a lighthouse.")
+		}
+	}
+
+	checkInterval := c.GetInt("timers.connection_alive_interval", 5)
+	pendingDeletionInterval := c.GetInt("timers.pending_deletion_interval", 10)
 	ifConfig := &InterfaceConfig{
 		HostMap:                 hostMap,
 		Inside:                  tun,
 		Outside:                 udpConns[0],
 		certState:               cs,
-		Cipher:                  config.GetString("cipher", "aes"),
+		Cipher:                  c.GetString("cipher", "aes"),
 		Firewall:                fw,
 		ServeDns:                serveDns,
 		HandshakeManager:        handshakeManager,
 		lightHouse:              lightHouse,
 		checkInterval:           checkInterval,
 		pendingDeletionInterval: pendingDeletionInterval,
-		DropLocalBroadcast:      config.GetBool("tun.drop_local_broadcast", false),
-		DropMulticast:           config.GetBool("tun.drop_multicast", false),
-		UDPBatchSize:            config.GetInt("listen.batch", 64),
+		DropLocalBroadcast:      c.GetBool("tun.drop_local_broadcast", false),
+		DropMulticast:           c.GetBool("tun.drop_multicast", false),
 		routines:                routines,
 		MessageMetrics:          messageMetrics,
 		version:                 buildVersion,
+		caPool:                  caPool,
+		disconnectInvalid:       c.GetBool("pki.disconnect_invalid", false),
 
 		ConntrackCacheTimeout: conntrackCacheTimeout,
+		l:                     l,
 	}
 
 	switch ifConfig.Cipher {
@@ -391,7 +406,7 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 
 	var ifce *Interface
 	if !configTest {
-		ifce, err = NewInterface(ifConfig)
+		ifce, err = NewInterface(ctx, ifConfig)
 		if err != nil {
 			return nil, fmt.Errorf("failed to initialize interface: %s", err)
 		}
@@ -400,13 +415,16 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 		// I don't want to make this initial commit too far-reaching though
 		ifce.writers = udpConns
 
-		ifce.RegisterConfigChangeCallbacks(config)
+		ifce.RegisterConfigChangeCallbacks(c)
 
-		go handshakeManager.Run(ifce)
-		go lightHouse.LhUpdateWorker(ifce)
+		go handshakeManager.Run(ctx, ifce)
+		go lightHouse.LhUpdateWorker(ctx, ifce)
 	}
 
-	err = startStats(config, configTest)
+	// TODO - stats third-party modules start uncancellable goroutines. Update those libs to accept
+	// a context so that they can exit when the context is Done.
+	statsStart, err := startStats(l, c, buildVersion, configTest)
+
 	if err != nil {
 		return nil, NewContextualError("Failed to start stats emitter", nil, err)
 	}
@@ -416,15 +434,16 @@ func Main(config *Config, configTest bool, buildVersion string, logger *logrus.L
 	}
 
 	//TODO: check if we _should_ be emitting stats
-	go ifce.emitStats(config.GetDuration("stats.interval", time.Second*10))
+	go ifce.emitStats(ctx, c.GetDuration("stats.interval", time.Second*10))
 
-	attachCommands(ssh, hostMap, handshakeManager.pendingHostMap, lightHouse, ifce)
+	attachCommands(l, ssh, hostMap, handshakeManager.pendingHostMap, lightHouse, ifce)
 
 	// Start DNS server last to allow using the nebula IP as lighthouse.dns.host
+	var dnsStart func()
 	if amLighthouse && serveDns {
 		l.Debugln("Starting dns server")
-		go dnsMain(hostMap, config)
+		dnsStart = dnsMain(l, hostMap, c)
 	}
 
-	return &Control{ifce, l}, nil
+	return &Control{ifce, l, cancel, sshStart, statsStart, dnsStart}, nil
 }

main_test.go

@@ -1 +0,0 @@
-package nebula

message_metrics.go

@@ -4,8 +4,11 @@ import (
 	"fmt"
 
 	"github.com/rcrowley/go-metrics"
+	"github.com/slackhq/nebula/header"
 )
 
+//TODO: this can probably move into the header package
+
 type MessageMetrics struct {
 	rx [][]metrics.Counter
 	tx [][]metrics.Counter
@@ -14,7 +17,7 @@ type MessageMetrics struct {
 	txUnknown metrics.Counter
 }
 
-func (m *MessageMetrics) Rx(t NebulaMessageType, s NebulaMessageSubType, i int64) {
+func (m *MessageMetrics) Rx(t header.MessageType, s header.MessageSubType, i int64) {
 	if m != nil {
 		if t >= 0 && int(t) < len(m.rx) && s >= 0 && int(s) < len(m.rx[t]) {
 			m.rx[t][s].Inc(i)
@@ -23,7 +26,7 @@ func (m *MessageMetrics) Rx(t NebulaMessageType, s NebulaMessageSubType, i int64
 		}
 	}
 }
-func (m *MessageMetrics) Tx(t NebulaMessageType, s NebulaMessageSubType, i int64) {
+func (m *MessageMetrics) Tx(t header.MessageType, s header.MessageSubType, i int64) {
 	if m != nil {
 		if t >= 0 && int(t) < len(m.tx) && s >= 0 && int(s) < len(m.tx[t]) {
 			m.tx[t][s].Inc(i)

nebula.proto

@@ -1,6 +1,8 @@
 syntax = "proto3";
 package nebula;
 
+option go_package = "github.com/slackhq/nebula";
+
 message NebulaMeta {
   enum MessageType {
     None = 0;
@@ -20,19 +22,23 @@ message NebulaMeta {
   NebulaMetaDetails Details = 2;
 }
 
-
 message NebulaMetaDetails {
-
   uint32 VpnIp = 1;
-  repeated IpAndPort IpAndPorts = 2;
+  repeated Ip4AndPort Ip4AndPorts = 2;
+  repeated Ip6AndPort Ip6AndPorts = 4;
   uint32 counter = 3;
 }
 
-message IpAndPort {
+message Ip4AndPort {
   uint32 Ip = 1;
   uint32 Port = 2;
 }
 
+message Ip6AndPort {
+  uint64 Hi = 1;
+  uint64 Lo = 2;
+  uint32 Port = 3;
+}
 
 message NebulaPing {
   enum MessageType {

outside.go

@@ -10,6 +10,10 @@ import (
 	"github.com/golang/protobuf/proto"
 	"github.com/sirupsen/logrus"
 	"github.com/slackhq/nebula/cert"
+	"github.com/slackhq/nebula/firewall"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
 	"golang.org/x/net/ipv4"
 )
 
@@ -17,14 +21,14 @@ const (
 	minFwPacketLen = 4
 )
 
-func (f *Interface) readOutsidePackets(addr *udpAddr, out []byte, packet []byte, header *Header, fwPacket *FirewallPacket, lhh *LightHouseHandler, nb []byte, q int, localCache ConntrackCache) {
-	err := header.Parse(packet)
+func (f *Interface) readOutsidePackets(addr *udp.Addr, out []byte, packet []byte, h *header.H, fwPacket *firewall.Packet, lhf udp.LightHouseHandlerFunc, nb []byte, q int, localCache firewall.ConntrackCache) {
+	err := h.Parse(packet)
 	if err != nil {
 		// TODO: best if we return this and let caller log
 		// TODO: Might be better to send the literal []byte("holepunch") packet and ignore that?
 		// Hole punch packets are 0 or 1 byte big, so lets ignore printing those errors
 		if len(packet) > 1 {
-			l.WithField("packet", packet).Infof("Error while parsing inbound packet from %s: %s", addr, err)
+			f.l.WithField("packet", packet).Infof("Error while parsing inbound packet from %s: %s", addr, err)
 		}
 		return
 	}
@@ -32,32 +36,32 @@ func (f *Interface) readOutsidePackets(addr *udpAddr, out []byte, packet []byte,
 	//l.Error("in packet ", header, packet[HeaderLen:])
 
 	// verify if we've seen this index before, otherwise respond to the handshake initiation
-	hostinfo, err := f.hostMap.QueryIndex(header.RemoteIndex)
+	hostinfo, err := f.hostMap.QueryIndex(h.RemoteIndex)
 
 	var ci *ConnectionState
 	if err == nil {
 		ci = hostinfo.ConnectionState
 	}
 
-	switch header.Type {
-	case message:
-		if !f.handleEncrypted(ci, addr, header) {
+	switch h.Type {
+	case header.Message:
+		if !f.handleEncrypted(ci, addr, h) {
 			return
 		}
 
-		f.decryptToTun(hostinfo, header.MessageCounter, out, packet, fwPacket, nb, q, localCache)
+		f.decryptToTun(hostinfo, h.MessageCounter, out, packet, fwPacket, nb, q, localCache)
 
 		// Fallthrough to the bottom to record incoming traffic
 
-	case lightHouse:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		if !f.handleEncrypted(ci, addr, header) {
+	case header.LightHouse:
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		if !f.handleEncrypted(ci, addr, h) {
 			return
 		}
 
-		d, err := f.decrypt(hostinfo, header.MessageCounter, out, packet, header, nb)
+		d, err := f.decrypt(hostinfo, h.MessageCounter, out, packet, h, nb)
 		if err != nil {
-			hostinfo.logger().WithError(err).WithField("udpAddr", addr).
+			hostinfo.logger(f.l).WithError(err).WithField("udpAddr", addr).
 				WithField("packet", packet).
 				Error("Failed to decrypt lighthouse packet")
 
@@ -66,19 +70,19 @@ func (f *Interface) readOutsidePackets(addr *udpAddr, out []byte, packet []byte,
 			return
 		}
 
-		lhh.HandleRequest(addr, hostinfo.hostId, d, hostinfo.GetCert(), f)
+		lhf(addr, hostinfo.vpnIp, d, f)
 
 		// Fallthrough to the bottom to record incoming traffic
 
-	case test:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		if !f.handleEncrypted(ci, addr, header) {
+	case header.Test:
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		if !f.handleEncrypted(ci, addr, h) {
 			return
 		}
 
-		d, err := f.decrypt(hostinfo, header.MessageCounter, out, packet, header, nb)
+		d, err := f.decrypt(hostinfo, h.MessageCounter, out, packet, h, nb)
 		if err != nil {
-			hostinfo.logger().WithError(err).WithField("udpAddr", addr).
+			hostinfo.logger(f.l).WithError(err).WithField("udpAddr", addr).
 				WithField("packet", packet).
 				Error("Failed to decrypt test packet")
 
@@ -87,11 +91,11 @@ func (f *Interface) readOutsidePackets(addr *udpAddr, out []byte, packet []byte,
 			return
 		}
 
-		if header.Subtype == testRequest {
+		if h.Subtype == header.TestRequest {
 			// This testRequest might be from TryPromoteBest, so we should roam
 			// to the new IP address before responding
 			f.handleHostRoaming(hostinfo, addr)
-			f.send(test, testReply, ci, hostinfo, hostinfo.remote, d, nb, out)
+			f.send(header.Test, header.TestReply, ci, hostinfo, hostinfo.remote, d, nb, out)
 		}
 
 		// Fallthrough to the bottom to record incoming traffic
@@ -99,79 +103,87 @@ func (f *Interface) readOutsidePackets(addr *udpAddr, out []byte, packet []byte,
 		// Non encrypted messages below here, they should not fall through to avoid tracking incoming traffic since they
 		// are unauthenticated
 
-	case handshake:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		HandleIncomingHandshake(f, addr, packet, header, hostinfo)
+	case header.Handshake:
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		HandleIncomingHandshake(f, addr, packet, h, hostinfo)
 		return
 
-	case recvError:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		f.handleRecvError(addr, header)
+	case header.RecvError:
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		f.handleRecvError(addr, h)
 		return
 
-	case closeTunnel:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		if !f.handleEncrypted(ci, addr, header) {
+	case header.CloseTunnel:
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		if !f.handleEncrypted(ci, addr, h) {
 			return
 		}
 
-		hostinfo.logger().WithField("udpAddr", addr).
+		hostinfo.logger(f.l).WithField("udpAddr", addr).
 			Info("Close tunnel received, tearing down.")
 
-		f.closeTunnel(hostinfo)
+		f.closeTunnel(hostinfo, false)
 		return
 
 	default:
-		f.messageMetrics.Rx(header.Type, header.Subtype, 1)
-		hostinfo.logger().Debugf("Unexpected packet received from %s", addr)
+		f.messageMetrics.Rx(h.Type, h.Subtype, 1)
+		hostinfo.logger(f.l).Debugf("Unexpected packet received from %s", addr)
 		return
 	}
 
 	f.handleHostRoaming(hostinfo, addr)
 
-	f.connectionManager.In(hostinfo.hostId)
+	f.connectionManager.In(hostinfo.vpnIp)
 }
 
-func (f *Interface) closeTunnel(hostInfo *HostInfo) {
+// closeTunnel closes a tunnel locally, it does not send a closeTunnel packet to the remote
+func (f *Interface) closeTunnel(hostInfo *HostInfo, hasHostMapLock bool) {
 	//TODO: this would be better as a single function in ConnectionManager that handled locks appropriately
-	f.connectionManager.ClearIP(hostInfo.hostId)
-	f.connectionManager.ClearPendingDeletion(hostInfo.hostId)
-	f.lightHouse.DeleteVpnIP(hostInfo.hostId)
-	f.hostMap.DeleteHostInfo(hostInfo)
+	f.connectionManager.ClearIP(hostInfo.vpnIp)
+	f.connectionManager.ClearPendingDeletion(hostInfo.vpnIp)
+	f.lightHouse.DeleteVpnIp(hostInfo.vpnIp)
+
+	if hasHostMapLock {
+		f.hostMap.unlockedDeleteHostInfo(hostInfo)
+	} else {
+		f.hostMap.DeleteHostInfo(hostInfo)
+	}
 }
 
-func (f *Interface) handleHostRoaming(hostinfo *HostInfo, addr *udpAddr) {
-	if hostDidRoam(hostinfo.remote, addr) {
-		if !f.lightHouse.remoteAllowList.Allow(udp2ipInt(addr)) {
-			hostinfo.logger().WithField("newAddr", addr).Debug("lighthouse.remote_allow_list denied roaming")
+// sendCloseTunnel is a helper function to send a proper close tunnel packet to a remote
+func (f *Interface) sendCloseTunnel(h *HostInfo) {
+	f.send(header.CloseTunnel, 0, h.ConnectionState, h, h.remote, []byte{}, make([]byte, 12, 12), make([]byte, mtu))
+}
+
+func (f *Interface) handleHostRoaming(hostinfo *HostInfo, addr *udp.Addr) {
+	if !hostinfo.remote.Equals(addr) {
+		if !f.lightHouse.remoteAllowList.Allow(hostinfo.vpnIp, addr.IP) {
+			hostinfo.logger(f.l).WithField("newAddr", addr).Debug("lighthouse.remote_allow_list denied roaming")
 			return
 		}
 		if !hostinfo.lastRoam.IsZero() && addr.Equals(hostinfo.lastRoamRemote) && time.Since(hostinfo.lastRoam) < RoamingSuppressSeconds*time.Second {
-			if l.Level >= logrus.DebugLevel {
-				hostinfo.logger().WithField("udpAddr", hostinfo.remote).WithField("newAddr", addr).
+			if f.l.Level >= logrus.DebugLevel {
+				hostinfo.logger(f.l).WithField("udpAddr", hostinfo.remote).WithField("newAddr", addr).
 					Debugf("Suppressing roam back to previous remote for %d seconds", RoamingSuppressSeconds)
 			}
 			return
 		}
 
-		hostinfo.logger().WithField("udpAddr", hostinfo.remote).WithField("newAddr", addr).
+		hostinfo.logger(f.l).WithField("udpAddr", hostinfo.remote).WithField("newAddr", addr).
 			Info("Host roamed to new udp ip/port.")
 		hostinfo.lastRoam = time.Now()
 		remoteCopy := *hostinfo.remote
 		hostinfo.lastRoamRemote = &remoteCopy
-		hostinfo.SetRemote(*addr)
-		if f.lightHouse.amLighthouse {
-			f.lightHouse.AddRemote(hostinfo.hostId, addr, false)
-		}
+		hostinfo.SetRemote(addr)
 	}
 
 }
 
-func (f *Interface) handleEncrypted(ci *ConnectionState, addr *udpAddr, header *Header) bool {
+func (f *Interface) handleEncrypted(ci *ConnectionState, addr *udp.Addr, h *header.H) bool {
 	// If connectionstate exists and the replay protector allows, process packet
 	// Else, send recv errors for 300 seconds after a restart to allow fast reconnection.
-	if ci == nil || !ci.window.Check(header.MessageCounter) {
-		f.sendRecvError(addr, header.RemoteIndex)
+	if ci == nil || !ci.window.Check(f.l, h.MessageCounter) {
+		f.sendRecvError(addr, h.RemoteIndex)
 		return false
 	}
 
@@ -179,7 +191,7 @@ func (f *Interface) handleEncrypted(ci *ConnectionState, addr *udpAddr, header *
 }
 
 // newPacket validates and parses the interesting bits for the firewall out of the ip and sub protocol headers
-func newPacket(data []byte, incoming bool, fp *FirewallPacket) error {
+func newPacket(data []byte, incoming bool, fp *firewall.Packet) error {
 	// Do we at least have an ipv4 header worth of data?
 	if len(data) < ipv4.HeaderLen {
 		return fmt.Errorf("packet is less than %v bytes", ipv4.HeaderLen)
@@ -207,7 +219,7 @@ func newPacket(data []byte, incoming bool, fp *FirewallPacket) error {
 
 	// Accounting for a variable header length, do we have enough data for our src/dst tuples?
 	minLen := ihl
-	if !fp.Fragment && fp.Protocol != fwProtoICMP {
+	if !fp.Fragment && fp.Protocol != firewall.ProtoICMP {
 		minLen += minFwPacketLen
 	}
 	if len(data) < minLen {
@@ -216,9 +228,9 @@ func newPacket(data []byte, incoming bool, fp *FirewallPacket) error {
 
 	// Firewall packets are locally oriented
 	if incoming {
-		fp.RemoteIP = binary.BigEndian.Uint32(data[12:16])
-		fp.LocalIP = binary.BigEndian.Uint32(data[16:20])
-		if fp.Fragment || fp.Protocol == fwProtoICMP {
+		fp.RemoteIP = iputil.Ip2VpnIp(data[12:16])
+		fp.LocalIP = iputil.Ip2VpnIp(data[16:20])
+		if fp.Fragment || fp.Protocol == firewall.ProtoICMP {
 			fp.RemotePort = 0
 			fp.LocalPort = 0
 		} else {
@@ -226,9 +238,9 @@ func newPacket(data []byte, incoming bool, fp *FirewallPacket) error {
 			fp.LocalPort = binary.BigEndian.Uint16(data[ihl+2 : ihl+4])
 		}
 	} else {
-		fp.LocalIP = binary.BigEndian.Uint32(data[12:16])
-		fp.RemoteIP = binary.BigEndian.Uint32(data[16:20])
-		if fp.Fragment || fp.Protocol == fwProtoICMP {
+		fp.LocalIP = iputil.Ip2VpnIp(data[12:16])
+		fp.RemoteIP = iputil.Ip2VpnIp(data[16:20])
+		if fp.Fragment || fp.Protocol == firewall.ProtoICMP {
 			fp.RemotePort = 0
 			fp.LocalPort = 0
 		} else {
@@ -240,15 +252,15 @@ func newPacket(data []byte, incoming bool, fp *FirewallPacket) error {
 	return nil
 }
 
-func (f *Interface) decrypt(hostinfo *HostInfo, mc uint64, out []byte, packet []byte, header *Header, nb []byte) ([]byte, error) {
+func (f *Interface) decrypt(hostinfo *HostInfo, mc uint64, out []byte, packet []byte, h *header.H, nb []byte) ([]byte, error) {
 	var err error
-	out, err = hostinfo.ConnectionState.dKey.DecryptDanger(out, packet[:HeaderLen], packet[HeaderLen:], mc, nb)
+	out, err = hostinfo.ConnectionState.dKey.DecryptDanger(out, packet[:header.Len], packet[header.Len:], mc, nb)
 	if err != nil {
 		return nil, err
 	}
 
-	if !hostinfo.ConnectionState.window.Update(mc) {
-		hostinfo.logger().WithField("header", header).
+	if !hostinfo.ConnectionState.window.Update(f.l, mc) {
+		hostinfo.logger(f.l).WithField("header", h).
 			Debugln("dropping out of window packet")
 		return nil, errors.New("out of window packet")
 	}
@@ -256,12 +268,12 @@ func (f *Interface) decrypt(hostinfo *HostInfo, mc uint64, out []byte, packet []
 	return out, nil
 }
 
-func (f *Interface) decryptToTun(hostinfo *HostInfo, messageCounter uint64, out []byte, packet []byte, fwPacket *FirewallPacket, nb []byte, q int, localCache ConntrackCache) {
+func (f *Interface) decryptToTun(hostinfo *HostInfo, messageCounter uint64, out []byte, packet []byte, fwPacket *firewall.Packet, nb []byte, q int, localCache firewall.ConntrackCache) {
 	var err error
 
-	out, err = hostinfo.ConnectionState.dKey.DecryptDanger(out, packet[:HeaderLen], packet[HeaderLen:], messageCounter, nb)
+	out, err = hostinfo.ConnectionState.dKey.DecryptDanger(out, packet[:header.Len], packet[header.Len:], messageCounter, nb)
 	if err != nil {
-		hostinfo.logger().WithError(err).Error("Failed to decrypt packet")
+		hostinfo.logger(f.l).WithError(err).Error("Failed to decrypt packet")
 		//TODO: maybe after build 64 is out? 06/14/2018 - NB
 		//f.sendRecvError(hostinfo.remote, header.RemoteIndex)
 		return
@@ -269,50 +281,50 @@ func (f *Interface) decryptToTun(hostinfo *HostInfo, messageCounter uint64, out
 
 	err = newPacket(out, true, fwPacket)
 	if err != nil {
-		hostinfo.logger().WithError(err).WithField("packet", out).
+		hostinfo.logger(f.l).WithError(err).WithField("packet", out).
 			Warnf("Error while validating inbound packet")
 		return
 	}
 
-	if !hostinfo.ConnectionState.window.Update(messageCounter) {
-		hostinfo.logger().WithField("fwPacket", fwPacket).
+	if !hostinfo.ConnectionState.window.Update(f.l, messageCounter) {
+		hostinfo.logger(f.l).WithField("fwPacket", fwPacket).
 			Debugln("dropping out of window packet")
 		return
 	}
 
-	dropReason := f.firewall.Drop(out, *fwPacket, true, hostinfo, trustedCAs, localCache)
+	dropReason := f.firewall.Drop(out, *fwPacket, true, hostinfo, f.caPool, localCache)
 	if dropReason != nil {
-		if l.Level >= logrus.DebugLevel {
-			hostinfo.logger().WithField("fwPacket", fwPacket).
+		if f.l.Level >= logrus.DebugLevel {
+			hostinfo.logger(f.l).WithField("fwPacket", fwPacket).
 				WithField("reason", dropReason).
 				Debugln("dropping inbound packet")
 		}
 		return
 	}
 
-	f.connectionManager.In(hostinfo.hostId)
+	f.connectionManager.In(hostinfo.vpnIp)
 	_, err = f.readers[q].Write(out)
 	if err != nil {
-		l.WithError(err).Error("Failed to write to tun")
+		f.l.WithError(err).Error("Failed to write to tun")
 	}
 }
 
-func (f *Interface) sendRecvError(endpoint *udpAddr, index uint32) {
-	f.messageMetrics.Tx(recvError, 0, 1)
+func (f *Interface) sendRecvError(endpoint *udp.Addr, index uint32) {
+	f.messageMetrics.Tx(header.RecvError, 0, 1)
 
 	//TODO: this should be a signed message so we can trust that we should drop the index
-	b := HeaderEncode(make([]byte, HeaderLen), Version, uint8(recvError), 0, index, 0)
+	b := header.Encode(make([]byte, header.Len), header.Version, header.RecvError, 0, index, 0)
 	f.outside.WriteTo(b, endpoint)
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("index", index).
+	if f.l.Level >= logrus.DebugLevel {
+		f.l.WithField("index", index).
 			WithField("udpAddr", endpoint).
 			Debug("Recv error sent")
 	}
 }
 
-func (f *Interface) handleRecvError(addr *udpAddr, h *Header) {
-	if l.Level >= logrus.DebugLevel {
-		l.WithField("index", h.RemoteIndex).
+func (f *Interface) handleRecvError(addr *udp.Addr, h *header.H) {
+	if f.l.Level >= logrus.DebugLevel {
+		f.l.WithField("index", h.RemoteIndex).
 			WithField("udpAddr", addr).
 			Debug("Recv error received")
 	}
@@ -322,7 +334,7 @@ func (f *Interface) handleRecvError(addr *udpAddr, h *Header) {
 
 	hostinfo, err := f.hostMap.QueryReverseIndex(h.RemoteIndex)
 	if err != nil {
-		l.Debugln(err, ": ", h.RemoteIndex)
+		f.l.Debugln(err, ": ", h.RemoteIndex)
 		return
 	}
 
@@ -332,8 +344,8 @@ func (f *Interface) handleRecvError(addr *udpAddr, h *Header) {
 	if !hostinfo.RecvErrorExceeded() {
 		return
 	}
-	if hostinfo.remote != nil && hostinfo.remote.String() != addr.String() {
-		l.Infoln("Someone spoofing recv_errors? ", addr, hostinfo.remote)
+	if hostinfo.remote != nil && !hostinfo.remote.Equals(addr) {
+		f.l.Infoln("Someone spoofing recv_errors? ", addr, hostinfo.remote)
 		return
 	}
 
@@ -368,7 +380,7 @@ func (f *Interface) sendMeta(ci *ConnectionState, endpoint *net.UDPAddr, meta *N
 }
 */
 
-func RecombineCertAndValidate(h *noise.HandshakeState, rawCertBytes []byte) (*cert.NebulaCertificate, error) {
+func RecombineCertAndValidate(h *noise.HandshakeState, rawCertBytes []byte, caPool *cert.NebulaCAPool) (*cert.NebulaCertificate, error) {
 	pk := h.PeerStatic()
 
 	if pk == nil {
@@ -397,7 +409,7 @@ func RecombineCertAndValidate(h *noise.HandshakeState, rawCertBytes []byte) (*ce
 	}
 
 	c, _ := cert.UnmarshalNebulaCertificate(recombined)
-	isValid, err := c.Verify(time.Now(), trustedCAs)
+	isValid, err := c.Verify(time.Now(), caPool)
 	if err != nil {
 		return c, fmt.Errorf("certificate validation failed: %s", err)
 	} else if !isValid {

outside_test.go

@@ -4,12 +4,14 @@ import (
 	"net"
 	"testing"
 
+	"github.com/slackhq/nebula/firewall"
+	"github.com/slackhq/nebula/iputil"
 	"github.com/stretchr/testify/assert"
 	"golang.org/x/net/ipv4"
 )
 
 func Test_newPacket(t *testing.T) {
-	p := &FirewallPacket{}
+	p := &firewall.Packet{}
 
 	// length fail
 	err := newPacket([]byte{0, 1}, true, p)
@@ -44,7 +46,7 @@ func Test_newPacket(t *testing.T) {
 		Src:      net.IPv4(10, 0, 0, 1),
 		Dst:      net.IPv4(10, 0, 0, 2),
 		Options:  []byte{0, 1, 0, 2},
-		Protocol: fwProtoTCP,
+		Protocol: firewall.ProtoTCP,
 	}
 
 	b, _ = h.Marshal()
@@ -52,9 +54,9 @@ func Test_newPacket(t *testing.T) {
 	err = newPacket(b, true, p)
 
 	assert.Nil(t, err)
-	assert.Equal(t, p.Protocol, uint8(fwProtoTCP))
-	assert.Equal(t, p.LocalIP, ip2int(net.IPv4(10, 0, 0, 2)))
-	assert.Equal(t, p.RemoteIP, ip2int(net.IPv4(10, 0, 0, 1)))
+	assert.Equal(t, p.Protocol, uint8(firewall.ProtoTCP))
+	assert.Equal(t, p.LocalIP, iputil.Ip2VpnIp(net.IPv4(10, 0, 0, 2)))
+	assert.Equal(t, p.RemoteIP, iputil.Ip2VpnIp(net.IPv4(10, 0, 0, 1)))
 	assert.Equal(t, p.RemotePort, uint16(3))
 	assert.Equal(t, p.LocalPort, uint16(4))
 
@@ -74,8 +76,8 @@ func Test_newPacket(t *testing.T) {
 
 	assert.Nil(t, err)
 	assert.Equal(t, p.Protocol, uint8(2))
-	assert.Equal(t, p.LocalIP, ip2int(net.IPv4(10, 0, 0, 1)))
-	assert.Equal(t, p.RemoteIP, ip2int(net.IPv4(10, 0, 0, 2)))
+	assert.Equal(t, p.LocalIP, iputil.Ip2VpnIp(net.IPv4(10, 0, 0, 1)))
+	assert.Equal(t, p.RemoteIP, iputil.Ip2VpnIp(net.IPv4(10, 0, 0, 2)))
 	assert.Equal(t, p.RemotePort, uint16(6))
 	assert.Equal(t, p.LocalPort, uint16(5))
 }

punchy.go

@@ -1,6 +1,10 @@
 package nebula
 
-import "time"
+import (
+	"time"
+
+	"github.com/slackhq/nebula/config"
+)
 
 type Punchy struct {
 	Punch   bool
@@ -8,7 +12,7 @@ type Punchy struct {
 	Delay   time.Duration
 }
 
-func NewPunchyFromConfig(c *Config) *Punchy {
+func NewPunchyFromConfig(c *config.C) *Punchy {
 	p := &Punchy{}
 
 	if c.IsSet("punchy.punch") {

punchy_test.go

@@ -4,11 +4,14 @@ import (
 	"testing"
 	"time"
 
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/util"
 	"github.com/stretchr/testify/assert"
 )
 
 func TestNewPunchyFromConfig(t *testing.T) {
-	c := NewConfig()
+	l := util.NewTestLogger()
+	c := config.NewC(l)
 
 	// Test defaults
 	p := NewPunchyFromConfig(c)

remote_list.go

@@ -0,0 +1,507 @@
+package nebula
+
+import (
+	"bytes"
+	"net"
+	"sort"
+	"sync"
+
+	"github.com/slackhq/nebula/iputil"
+	"github.com/slackhq/nebula/udp"
+)
+
+// forEachFunc is used to benefit folks that want to do work inside the lock
+type forEachFunc func(addr *udp.Addr, preferred bool)
+
+// The checkFuncs here are to simplify bulk importing LH query response logic into a single function (reset slice and iterate)
+type checkFuncV4 func(vpnIp iputil.VpnIp, to *Ip4AndPort) bool
+type checkFuncV6 func(vpnIp iputil.VpnIp, to *Ip6AndPort) bool
+
+// CacheMap is a struct that better represents the lighthouse cache for humans
+// The string key is the owners vpnIp
+type CacheMap map[string]*Cache
+
+// Cache is the other part of CacheMap to better represent the lighthouse cache for humans
+// We don't reason about ipv4 vs ipv6 here
+type Cache struct {
+	Learned  []*udp.Addr `json:"learned,omitempty"`
+	Reported []*udp.Addr `json:"reported,omitempty"`
+}
+
+//TODO: Seems like we should plop static host entries in here too since the are protected by the lighthouse from deletion
+// We will never clean learned/reported information for them as it stands today
+
+// cache is an internal struct that splits v4 and v6 addresses inside the cache map
+type cache struct {
+	v4 *cacheV4
+	v6 *cacheV6
+}
+
+// cacheV4 stores learned and reported ipv4 records under cache
+type cacheV4 struct {
+	learned  *Ip4AndPort
+	reported []*Ip4AndPort
+}
+
+// cacheV4 stores learned and reported ipv6 records under cache
+type cacheV6 struct {
+	learned  *Ip6AndPort
+	reported []*Ip6AndPort
+}
+
+// RemoteList is a unifying concept for lighthouse servers and clients as well as hostinfos.
+// It serves as a local cache of query replies, host update notifications, and locally learned addresses
+type RemoteList struct {
+	// Every interaction with internals requires a lock!
+	sync.RWMutex
+
+	// A deduplicated set of addresses. Any accessor should lock beforehand.
+	addrs []*udp.Addr
+
+	// These are maps to store v4 and v6 addresses per lighthouse
+	// Map key is the vpnIp of the person that told us about this the cached entries underneath.
+	// For learned addresses, this is the vpnIp that sent the packet
+	cache map[iputil.VpnIp]*cache
+
+	// This is a list of remotes that we have tried to handshake with and have returned from the wrong vpn ip.
+	// They should not be tried again during a handshake
+	badRemotes []*udp.Addr
+
+	// A flag that the cache may have changed and addrs needs to be rebuilt
+	shouldRebuild bool
+}
+
+// NewRemoteList creates a new empty RemoteList
+func NewRemoteList() *RemoteList {
+	return &RemoteList{
+		addrs: make([]*udp.Addr, 0),
+		cache: make(map[iputil.VpnIp]*cache),
+	}
+}
+
+// Len locks and reports the size of the deduplicated address list
+// The deduplication work may need to occur here, so you must pass preferredRanges
+func (r *RemoteList) Len(preferredRanges []*net.IPNet) int {
+	r.Rebuild(preferredRanges)
+	r.RLock()
+	defer r.RUnlock()
+	return len(r.addrs)
+}
+
+// ForEach locks and will call the forEachFunc for every deduplicated address in the list
+// The deduplication work may need to occur here, so you must pass preferredRanges
+func (r *RemoteList) ForEach(preferredRanges []*net.IPNet, forEach forEachFunc) {
+	r.Rebuild(preferredRanges)
+	r.RLock()
+	for _, v := range r.addrs {
+		forEach(v, isPreferred(v.IP, preferredRanges))
+	}
+	r.RUnlock()
+}
+
+// CopyAddrs locks and makes a deep copy of the deduplicated address list
+// The deduplication work may need to occur here, so you must pass preferredRanges
+func (r *RemoteList) CopyAddrs(preferredRanges []*net.IPNet) []*udp.Addr {
+	if r == nil {
+		return nil
+	}
+
+	r.Rebuild(preferredRanges)
+
+	r.RLock()
+	defer r.RUnlock()
+	c := make([]*udp.Addr, len(r.addrs))
+	for i, v := range r.addrs {
+		c[i] = v.Copy()
+	}
+	return c
+}
+
+// LearnRemote locks and sets the learned slot for the owner vpn ip to the provided addr
+// Currently this is only needed when HostInfo.SetRemote is called as that should cover both handshaking and roaming.
+// It will mark the deduplicated address list as dirty, so do not call it unless new information is available
+//TODO: this needs to support the allow list list
+func (r *RemoteList) LearnRemote(ownerVpnIp iputil.VpnIp, addr *udp.Addr) {
+	r.Lock()
+	defer r.Unlock()
+	if v4 := addr.IP.To4(); v4 != nil {
+		r.unlockedSetLearnedV4(ownerVpnIp, NewIp4AndPort(v4, uint32(addr.Port)))
+	} else {
+		r.unlockedSetLearnedV6(ownerVpnIp, NewIp6AndPort(addr.IP, uint32(addr.Port)))
+	}
+}
+
+// CopyCache locks and creates a more human friendly form of the internal address cache.
+// This may contain duplicates and blocked addresses
+func (r *RemoteList) CopyCache() *CacheMap {
+	r.RLock()
+	defer r.RUnlock()
+
+	cm := make(CacheMap)
+	getOrMake := func(vpnIp string) *Cache {
+		c := cm[vpnIp]
+		if c == nil {
+			c = &Cache{
+				Learned:  make([]*udp.Addr, 0),
+				Reported: make([]*udp.Addr, 0),
+			}
+			cm[vpnIp] = c
+		}
+		return c
+	}
+
+	for owner, mc := range r.cache {
+		c := getOrMake(owner.String())
+
+		if mc.v4 != nil {
+			if mc.v4.learned != nil {
+				c.Learned = append(c.Learned, NewUDPAddrFromLH4(mc.v4.learned))
+			}
+
+			for _, a := range mc.v4.reported {
+				c.Reported = append(c.Reported, NewUDPAddrFromLH4(a))
+			}
+		}
+
+		if mc.v6 != nil {
+			if mc.v6.learned != nil {
+				c.Learned = append(c.Learned, NewUDPAddrFromLH6(mc.v6.learned))
+			}
+
+			for _, a := range mc.v6.reported {
+				c.Reported = append(c.Reported, NewUDPAddrFromLH6(a))
+			}
+		}
+	}
+
+	return &cm
+}
+
+// BlockRemote locks and records the address as bad, it will be excluded from the deduplicated address list
+func (r *RemoteList) BlockRemote(bad *udp.Addr) {
+	r.Lock()
+	defer r.Unlock()
+
+	// Check if we already blocked this addr
+	if r.unlockedIsBad(bad) {
+		return
+	}
+
+	// We copy here because we are taking something else's memory and we can't trust everything
+	r.badRemotes = append(r.badRemotes, bad.Copy())
+
+	// Mark the next interaction must recollect/dedupe
+	r.shouldRebuild = true
+}
+
+// CopyBlockedRemotes locks and makes a deep copy of the blocked remotes list
+func (r *RemoteList) CopyBlockedRemotes() []*udp.Addr {
+	r.RLock()
+	defer r.RUnlock()
+
+	c := make([]*udp.Addr, len(r.badRemotes))
+	for i, v := range r.badRemotes {
+		c[i] = v.Copy()
+	}
+	return c
+}
+
+// ResetBlockedRemotes locks and clears the blocked remotes list
+func (r *RemoteList) ResetBlockedRemotes() {
+	r.Lock()
+	r.badRemotes = nil
+	r.Unlock()
+}
+
+// Rebuild locks and generates the deduplicated address list only if there is work to be done
+// There is generally no reason to call this directly but it is safe to do so
+func (r *RemoteList) Rebuild(preferredRanges []*net.IPNet) {
+	r.Lock()
+	defer r.Unlock()
+
+	// Only rebuild if the cache changed
+	//TODO: shouldRebuild is probably pointless as we don't check for actual change when lighthouse updates come in
+	if r.shouldRebuild {
+		r.unlockedCollect()
+		r.shouldRebuild = false
+	}
+
+	// Always re-sort, preferredRanges can change via HUP
+	r.unlockedSort(preferredRanges)
+}
+
+// unlockedIsBad assumes you have the write lock and checks if the remote matches any entry in the blocked address list
+func (r *RemoteList) unlockedIsBad(remote *udp.Addr) bool {
+	for _, v := range r.badRemotes {
+		if v.Equals(remote) {
+			return true
+		}
+	}
+	return false
+}
+
+// unlockedSetLearnedV4 assumes you have the write lock and sets the current learned address for this owner and marks the
+// deduplicated address list as dirty
+func (r *RemoteList) unlockedSetLearnedV4(ownerVpnIp iputil.VpnIp, to *Ip4AndPort) {
+	r.shouldRebuild = true
+	r.unlockedGetOrMakeV4(ownerVpnIp).learned = to
+}
+
+// unlockedSetV4 assumes you have the write lock and resets the reported list of ips for this owner to the list provided
+// and marks the deduplicated address list as dirty
+func (r *RemoteList) unlockedSetV4(ownerVpnIp iputil.VpnIp, vpnIp iputil.VpnIp, to []*Ip4AndPort, check checkFuncV4) {
+	r.shouldRebuild = true
+	c := r.unlockedGetOrMakeV4(ownerVpnIp)
+
+	// Reset the slice
+	c.reported = c.reported[:0]
+
+	// We can't take their array but we can take their pointers
+	for _, v := range to[:minInt(len(to), MaxRemotes)] {
+		if check(vpnIp, v) {
+			c.reported = append(c.reported, v)
+		}
+	}
+}
+
+// unlockedPrependV4 assumes you have the write lock and prepends the address in the reported list for this owner
+// This is only useful for establishing static hosts
+func (r *RemoteList) unlockedPrependV4(ownerVpnIp iputil.VpnIp, to *Ip4AndPort) {
+	r.shouldRebuild = true
+	c := r.unlockedGetOrMakeV4(ownerVpnIp)
+
+	// We are doing the easy append because this is rarely called
+	c.reported = append([]*Ip4AndPort{to}, c.reported...)
+	if len(c.reported) > MaxRemotes {
+		c.reported = c.reported[:MaxRemotes]
+	}
+}
+
+// unlockedSetLearnedV6 assumes you have the write lock and sets the current learned address for this owner and marks the
+// deduplicated address list as dirty
+func (r *RemoteList) unlockedSetLearnedV6(ownerVpnIp iputil.VpnIp, to *Ip6AndPort) {
+	r.shouldRebuild = true
+	r.unlockedGetOrMakeV6(ownerVpnIp).learned = to
+}
+
+// unlockedSetV6 assumes you have the write lock and resets the reported list of ips for this owner to the list provided
+// and marks the deduplicated address list as dirty
+func (r *RemoteList) unlockedSetV6(ownerVpnIp iputil.VpnIp, vpnIp iputil.VpnIp, to []*Ip6AndPort, check checkFuncV6) {
+	r.shouldRebuild = true
+	c := r.unlockedGetOrMakeV6(ownerVpnIp)
+
+	// Reset the slice
+	c.reported = c.reported[:0]
+
+	// We can't take their array but we can take their pointers
+	for _, v := range to[:minInt(len(to), MaxRemotes)] {
+		if check(vpnIp, v) {
+			c.reported = append(c.reported, v)
+		}
+	}
+}
+
+// unlockedPrependV6 assumes you have the write lock and prepends the address in the reported list for this owner
+// This is only useful for establishing static hosts
+func (r *RemoteList) unlockedPrependV6(ownerVpnIp iputil.VpnIp, to *Ip6AndPort) {
+	r.shouldRebuild = true
+	c := r.unlockedGetOrMakeV6(ownerVpnIp)
+
+	// We are doing the easy append because this is rarely called
+	c.reported = append([]*Ip6AndPort{to}, c.reported...)
+	if len(c.reported) > MaxRemotes {
+		c.reported = c.reported[:MaxRemotes]
+	}
+}
+
+// unlockedGetOrMakeV4 assumes you have the write lock and builds the cache and owner entry. Only the v4 pointer is established.
+// The caller must dirty the learned address cache if required
+func (r *RemoteList) unlockedGetOrMakeV4(ownerVpnIp iputil.VpnIp) *cacheV4 {
+	am := r.cache[ownerVpnIp]
+	if am == nil {
+		am = &cache{}
+		r.cache[ownerVpnIp] = am
+	}
+	// Avoid occupying memory for v6 addresses if we never have any
+	if am.v4 == nil {
+		am.v4 = &cacheV4{}
+	}
+	return am.v4
+}
+
+// unlockedGetOrMakeV6 assumes you have the write lock and builds the cache and owner entry. Only the v6 pointer is established.
+// The caller must dirty the learned address cache if required
+func (r *RemoteList) unlockedGetOrMakeV6(ownerVpnIp iputil.VpnIp) *cacheV6 {
+	am := r.cache[ownerVpnIp]
+	if am == nil {
+		am = &cache{}
+		r.cache[ownerVpnIp] = am
+	}
+	// Avoid occupying memory for v4 addresses if we never have any
+	if am.v6 == nil {
+		am.v6 = &cacheV6{}
+	}
+	return am.v6
+}
+
+// unlockedCollect assumes you have the write lock and collects/transforms the cache into the deduped address list.
+// The result of this function can contain duplicates. unlockedSort handles cleaning it.
+func (r *RemoteList) unlockedCollect() {
+	addrs := r.addrs[:0]
+
+	for _, c := range r.cache {
+		if c.v4 != nil {
+			if c.v4.learned != nil {
+				u := NewUDPAddrFromLH4(c.v4.learned)
+				if !r.unlockedIsBad(u) {
+					addrs = append(addrs, u)
+				}
+			}
+
+			for _, v := range c.v4.reported {
+				u := NewUDPAddrFromLH4(v)
+				if !r.unlockedIsBad(u) {
+					addrs = append(addrs, u)
+				}
+			}
+		}
+
+		if c.v6 != nil {
+			if c.v6.learned != nil {
+				u := NewUDPAddrFromLH6(c.v6.learned)
+				if !r.unlockedIsBad(u) {
+					addrs = append(addrs, u)
+				}
+			}
+
+			for _, v := range c.v6.reported {
+				u := NewUDPAddrFromLH6(v)
+				if !r.unlockedIsBad(u) {
+					addrs = append(addrs, u)
+				}
+			}
+		}
+	}
+
+	r.addrs = addrs
+}
+
+// unlockedSort assumes you have the write lock and performs the deduping and sorting of the address list
+func (r *RemoteList) unlockedSort(preferredRanges []*net.IPNet) {
+	n := len(r.addrs)
+	if n < 2 {
+		return
+	}
+
+	lessFunc := func(i, j int) bool {
+		a := r.addrs[i]
+		b := r.addrs[j]
+		// Preferred addresses first
+
+		aPref := isPreferred(a.IP, preferredRanges)
+		bPref := isPreferred(b.IP, preferredRanges)
+		switch {
+		case aPref && !bPref:
+			// If i is preferred and j is not, i is less than j
+			return true
+
+		case !aPref && bPref:
+			// If j is preferred then i is not due to the else, i is not less than j
+			return false
+
+		default:
+			// Both i an j are either preferred or not, sort within that
+		}
+
+		// ipv6 addresses 2nd
+		a4 := a.IP.To4()
+		b4 := b.IP.To4()
+		switch {
+		case a4 == nil && b4 != nil:
+			// If i is v6 and j is v4, i is less than j
+			return true
+
+		case a4 != nil && b4 == nil:
+			// If j is v6 and i is v4, i is not less than j
+			return false
+
+		case a4 != nil && b4 != nil:
+			// Special case for ipv4, a4 and b4 are not nil
+			aPrivate := isPrivateIP(a4)
+			bPrivate := isPrivateIP(b4)
+			switch {
+			case !aPrivate && bPrivate:
+				// If i is a public ip (not private) and j is a private ip, i is less then j
+				return true
+
+			case aPrivate && !bPrivate:
+				// If j is public (not private) then i is private due to the else, i is not less than j
+				return false
+
+			default:
+				// Both i an j are either public or private, sort within that
+			}
+
+		default:
+			// Both i an j are either ipv4 or ipv6, sort within that
+		}
+
+		// lexical order of ips 3rd
+		c := bytes.Compare(a.IP, b.IP)
+		if c == 0 {
+			// Ips are the same, Lexical order of ports 4th
+			return a.Port < b.Port
+		}
+
+		// Ip wasn't the same
+		return c < 0
+	}
+
+	// Sort it
+	sort.Slice(r.addrs, lessFunc)
+
+	// Deduplicate
+	a, b := 0, 1
+	for b < n {
+		if !r.addrs[a].Equals(r.addrs[b]) {
+			a++
+			if a != b {
+				r.addrs[a], r.addrs[b] = r.addrs[b], r.addrs[a]
+			}
+		}
+		b++
+	}
+
+	r.addrs = r.addrs[:a+1]
+	return
+}
+
+// minInt returns the minimum integer of a or b
+func minInt(a, b int) int {
+	if a < b {
+		return a
+	}
+	return b
+}
+
+// isPreferred returns true of the ip is contained in the preferredRanges list
+func isPreferred(ip net.IP, preferredRanges []*net.IPNet) bool {
+	//TODO: this would be better in a CIDR6Tree
+	for _, p := range preferredRanges {
+		if p.Contains(ip) {
+			return true
+		}
+	}
+	return false
+}
+
+var _, private24BitBlock, _ = net.ParseCIDR("10.0.0.0/8")
+var _, private20BitBlock, _ = net.ParseCIDR("172.16.0.0/12")
+var _, private16BitBlock, _ = net.ParseCIDR("192.168.0.0/16")
+
+// isPrivateIP returns true if the ip is contained by a rfc 1918 private range
+func isPrivateIP(ip net.IP) bool {
+	//TODO: another great cidrtree option
+	//TODO: Private for ipv6 or just let it ride?
+	return private24BitBlock.Contains(ip) || private20BitBlock.Contains(ip) || private16BitBlock.Contains(ip)
+}

remote_list_test.go

@@ -0,0 +1,235 @@
+package nebula
+
+import (
+	"net"
+	"testing"
+
+	"github.com/slackhq/nebula/iputil"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestRemoteList_Rebuild(t *testing.T) {
+	rl := NewRemoteList()
+	rl.unlockedSetV4(
+		0,
+		0,
+		[]*Ip4AndPort{
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1475}, // this is duped
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.0.182"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101}, // this is duped
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.18.0.1"))), Port: 10101}, // this is duped
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.18.0.1"))), Port: 10101}, // this is a dupe
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.19.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.31.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101},   // this is a dupe
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1476}, // almost dupe of 0 with a diff port
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1475}, // this is a dupe
+		},
+		func(iputil.VpnIp, *Ip4AndPort) bool { return true },
+	)
+
+	rl.unlockedSetV6(
+		1,
+		1,
+		[]*Ip6AndPort{
+			NewIp6AndPort(net.ParseIP("1::1"), 1), // this is duped
+			NewIp6AndPort(net.ParseIP("1::1"), 2), // almost dupe of 0 with a diff port, also gets duped
+			NewIp6AndPort(net.ParseIP("1:100::1"), 1),
+			NewIp6AndPort(net.ParseIP("1::1"), 1), // this is a dupe
+			NewIp6AndPort(net.ParseIP("1::1"), 2), // this is a dupe
+		},
+		func(iputil.VpnIp, *Ip6AndPort) bool { return true },
+	)
+
+	rl.Rebuild([]*net.IPNet{})
+	assert.Len(t, rl.addrs, 10, "addrs contains too many entries")
+
+	// ipv6 first, sorted lexically within
+	assert.Equal(t, "[1::1]:1", rl.addrs[0].String())
+	assert.Equal(t, "[1::1]:2", rl.addrs[1].String())
+	assert.Equal(t, "[1:100::1]:1", rl.addrs[2].String())
+
+	// ipv4 last, sorted by public first, then private, lexically within them
+	assert.Equal(t, "70.199.182.92:1475", rl.addrs[3].String())
+	assert.Equal(t, "70.199.182.92:1476", rl.addrs[4].String())
+	assert.Equal(t, "172.17.0.182:10101", rl.addrs[5].String())
+	assert.Equal(t, "172.17.1.1:10101", rl.addrs[6].String())
+	assert.Equal(t, "172.18.0.1:10101", rl.addrs[7].String())
+	assert.Equal(t, "172.19.0.1:10101", rl.addrs[8].String())
+	assert.Equal(t, "172.31.0.1:10101", rl.addrs[9].String())
+
+	// Now ensure we can hoist ipv4 up
+	_, ipNet, err := net.ParseCIDR("0.0.0.0/0")
+	assert.NoError(t, err)
+	rl.Rebuild([]*net.IPNet{ipNet})
+	assert.Len(t, rl.addrs, 10, "addrs contains too many entries")
+
+	// ipv4 first, public then private, lexically within them
+	assert.Equal(t, "70.199.182.92:1475", rl.addrs[0].String())
+	assert.Equal(t, "70.199.182.92:1476", rl.addrs[1].String())
+	assert.Equal(t, "172.17.0.182:10101", rl.addrs[2].String())
+	assert.Equal(t, "172.17.1.1:10101", rl.addrs[3].String())
+	assert.Equal(t, "172.18.0.1:10101", rl.addrs[4].String())
+	assert.Equal(t, "172.19.0.1:10101", rl.addrs[5].String())
+	assert.Equal(t, "172.31.0.1:10101", rl.addrs[6].String())
+
+	// ipv6 last, sorted by public first, then private, lexically within them
+	assert.Equal(t, "[1::1]:1", rl.addrs[7].String())
+	assert.Equal(t, "[1::1]:2", rl.addrs[8].String())
+	assert.Equal(t, "[1:100::1]:1", rl.addrs[9].String())
+
+	// Ensure we can hoist a specific ipv4 range over anything else
+	_, ipNet, err = net.ParseCIDR("172.17.0.0/16")
+	assert.NoError(t, err)
+	rl.Rebuild([]*net.IPNet{ipNet})
+	assert.Len(t, rl.addrs, 10, "addrs contains too many entries")
+
+	// Preferred ipv4 first
+	assert.Equal(t, "172.17.0.182:10101", rl.addrs[0].String())
+	assert.Equal(t, "172.17.1.1:10101", rl.addrs[1].String())
+
+	// ipv6 next
+	assert.Equal(t, "[1::1]:1", rl.addrs[2].String())
+	assert.Equal(t, "[1::1]:2", rl.addrs[3].String())
+	assert.Equal(t, "[1:100::1]:1", rl.addrs[4].String())
+
+	// the remaining ipv4 last
+	assert.Equal(t, "70.199.182.92:1475", rl.addrs[5].String())
+	assert.Equal(t, "70.199.182.92:1476", rl.addrs[6].String())
+	assert.Equal(t, "172.18.0.1:10101", rl.addrs[7].String())
+	assert.Equal(t, "172.19.0.1:10101", rl.addrs[8].String())
+	assert.Equal(t, "172.31.0.1:10101", rl.addrs[9].String())
+}
+
+func BenchmarkFullRebuild(b *testing.B) {
+	rl := NewRemoteList()
+	rl.unlockedSetV4(
+		0,
+		0,
+		[]*Ip4AndPort{
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1475},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.0.182"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.18.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.19.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.31.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101},   // this is a dupe
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1476}, // dupe of 0 with a diff port
+		},
+		func(iputil.VpnIp, *Ip4AndPort) bool { return true },
+	)
+
+	rl.unlockedSetV6(
+		0,
+		0,
+		[]*Ip6AndPort{
+			NewIp6AndPort(net.ParseIP("1::1"), 1),
+			NewIp6AndPort(net.ParseIP("1::1"), 2), // dupe of 0 with a diff port
+			NewIp6AndPort(net.ParseIP("1:100::1"), 1),
+			NewIp6AndPort(net.ParseIP("1::1"), 1), // this is a dupe
+		},
+		func(iputil.VpnIp, *Ip6AndPort) bool { return true },
+	)
+
+	b.Run("no preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.shouldRebuild = true
+			rl.Rebuild([]*net.IPNet{})
+		}
+	})
+
+	_, ipNet, err := net.ParseCIDR("172.17.0.0/16")
+	assert.NoError(b, err)
+	b.Run("1 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.shouldRebuild = true
+			rl.Rebuild([]*net.IPNet{ipNet})
+		}
+	})
+
+	_, ipNet2, err := net.ParseCIDR("70.0.0.0/8")
+	assert.NoError(b, err)
+	b.Run("2 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.shouldRebuild = true
+			rl.Rebuild([]*net.IPNet{ipNet, ipNet2})
+		}
+	})
+
+	_, ipNet3, err := net.ParseCIDR("0.0.0.0/0")
+	assert.NoError(b, err)
+	b.Run("3 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.shouldRebuild = true
+			rl.Rebuild([]*net.IPNet{ipNet, ipNet2, ipNet3})
+		}
+	})
+}
+
+func BenchmarkSortRebuild(b *testing.B) {
+	rl := NewRemoteList()
+	rl.unlockedSetV4(
+		0,
+		0,
+		[]*Ip4AndPort{
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1475},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.0.182"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.18.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.19.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.31.0.1"))), Port: 10101},
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("172.17.1.1"))), Port: 10101},   // this is a dupe
+			{Ip: uint32(iputil.Ip2VpnIp(net.ParseIP("70.199.182.92"))), Port: 1476}, // dupe of 0 with a diff port
+		},
+		func(iputil.VpnIp, *Ip4AndPort) bool { return true },
+	)
+
+	rl.unlockedSetV6(
+		0,
+		0,
+		[]*Ip6AndPort{
+			NewIp6AndPort(net.ParseIP("1::1"), 1),
+			NewIp6AndPort(net.ParseIP("1::1"), 2), // dupe of 0 with a diff port
+			NewIp6AndPort(net.ParseIP("1:100::1"), 1),
+			NewIp6AndPort(net.ParseIP("1::1"), 1), // this is a dupe
+		},
+		func(iputil.VpnIp, *Ip6AndPort) bool { return true },
+	)
+
+	b.Run("no preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.shouldRebuild = true
+			rl.Rebuild([]*net.IPNet{})
+		}
+	})
+
+	_, ipNet, err := net.ParseCIDR("172.17.0.0/16")
+	rl.Rebuild([]*net.IPNet{ipNet})
+
+	assert.NoError(b, err)
+	b.Run("1 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.Rebuild([]*net.IPNet{ipNet})
+		}
+	})
+
+	_, ipNet2, err := net.ParseCIDR("70.0.0.0/8")
+	rl.Rebuild([]*net.IPNet{ipNet, ipNet2})
+
+	assert.NoError(b, err)
+	b.Run("2 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.Rebuild([]*net.IPNet{ipNet, ipNet2})
+		}
+	})
+
+	_, ipNet3, err := net.ParseCIDR("0.0.0.0/0")
+	rl.Rebuild([]*net.IPNet{ipNet, ipNet2, ipNet3})
+
+	assert.NoError(b, err)
+	b.Run("3 preferred", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			rl.Rebuild([]*net.IPNet{ipNet, ipNet2, ipNet3})
+		}
+	})
+}

ssh.go

@@ -10,12 +10,16 @@ import (
 	"os"
 	"reflect"
 	"runtime/pprof"
+	"sort"
 	"strings"
-	"sync/atomic"
 	"syscall"
 
 	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
+	"github.com/slackhq/nebula/header"
+	"github.com/slackhq/nebula/iputil"
 	"github.com/slackhq/nebula/sshd"
+	"github.com/slackhq/nebula/udp"
 )
 
 type sshListHostMapFlags struct {
@@ -26,6 +30,7 @@ type sshListHostMapFlags struct {
 type sshPrintCertFlags struct {
 	Json   bool
 	Pretty bool
+	Raw    bool
 }
 
 type sshPrintTunnelFlags struct {
@@ -44,51 +49,58 @@ type sshCreateTunnelFlags struct {
 	Address string
 }
 
-func wireSSHReload(ssh *sshd.SSHServer, c *Config) {
-	c.RegisterReloadCallback(func(c *Config) {
+func wireSSHReload(l *logrus.Logger, ssh *sshd.SSHServer, c *config.C) {
+	c.RegisterReloadCallback(func(c *config.C) {
 		if c.GetBool("sshd.enabled", false) {
-			err := configSSH(ssh, c)
+			sshRun, err := configSSH(l, ssh, c)
 			if err != nil {
 				l.WithError(err).Error("Failed to reconfigure the sshd")
 				ssh.Stop()
 			}
+			if sshRun != nil {
+				go sshRun()
+			}
 		} else {
 			ssh.Stop()
 		}
 	})
 }
 
-func configSSH(ssh *sshd.SSHServer, c *Config) error {
+// configSSH reads the ssh info out of the passed-in Config and
+// updates the passed-in SSHServer. On success, it returns a function
+// that callers may invoke to run the configured ssh server. On
+// failure, it returns nil, error.
+func configSSH(l *logrus.Logger, ssh *sshd.SSHServer, c *config.C) (func(), error) {
 	//TODO conntrack list
 	//TODO print firewall rules or hash?
 
 	listen := c.GetString("sshd.listen", "")
 	if listen == "" {
-		return fmt.Errorf("sshd.listen must be provided")
+		return nil, fmt.Errorf("sshd.listen must be provided")
 	}
 
 	_, port, err := net.SplitHostPort(listen)
 	if err != nil {
-		return fmt.Errorf("invalid sshd.listen address: %s", err)
+		return nil, fmt.Errorf("invalid sshd.listen address: %s", err)
 	}
 	if port == "22" {
-		return fmt.Errorf("sshd.listen can not use port 22")
+		return nil, fmt.Errorf("sshd.listen can not use port 22")
 	}
 
 	//TODO: no good way to reload this right now
 	hostKeyFile := c.GetString("sshd.host_key", "")
 	if hostKeyFile == "" {
-		return fmt.Errorf("sshd.host_key must be provided")
+		return nil, fmt.Errorf("sshd.host_key must be provided")
 	}
 
 	hostKeyBytes, err := ioutil.ReadFile(hostKeyFile)
 	if err != nil {
-		return fmt.Errorf("error while loading sshd.host_key file: %s", err)
+		return nil, fmt.Errorf("error while loading sshd.host_key file: %s", err)
 	}
 
 	err = ssh.SetHostKey(hostKeyBytes)
 	if err != nil {
-		return fmt.Errorf("error while adding sshd.host_key: %s", err)
+		return nil, fmt.Errorf("error while adding sshd.host_key: %s", err)
 	}
 
 	rawKeys := c.Get("sshd.authorized_users")
@@ -139,17 +151,22 @@ func configSSH(ssh *sshd.SSHServer, c *Config) error {
 		l.Info("no ssh users to authorize")
 	}
 
+	var runner func()
 	if c.GetBool("sshd.enabled", false) {
 		ssh.Stop()
-		go ssh.Run(listen)
+		runner = func() {
+			if err := ssh.Run(listen); err != nil {
+				l.WithField("err", err).Warn("Failed to run the SSH server")
+			}
+		}
 	} else {
 		ssh.Stop()
 	}
 
-	return nil
+	return runner, nil
 }
 
-func attachCommands(ssh *sshd.SSHServer, hostMap *HostMap, pendingHostMap *HostMap, lightHouse *LightHouse, ifce *Interface) {
+func attachCommands(l *logrus.Logger, ssh *sshd.SSHServer, hostMap *HostMap, pendingHostMap *HostMap, lightHouse *LightHouse, ifce *Interface) {
 	ssh.RegisterCommand(&sshd.Command{
 		Name:             "list-hostmap",
 		ShortDescription: "List all known previously connected hosts",
@@ -225,13 +242,17 @@ func attachCommands(ssh *sshd.SSHServer, hostMap *HostMap, pendingHostMap *HostM
 	ssh.RegisterCommand(&sshd.Command{
 		Name:             "log-level",
 		ShortDescription: "Gets or sets the current log level",
-		Callback:         sshLogLevel,
+		Callback: func(fs interface{}, a []string, w sshd.StringWriter) error {
+			return sshLogLevel(l, fs, a, w)
+		},
 	})
 
 	ssh.RegisterCommand(&sshd.Command{
 		Name:             "log-format",
 		ShortDescription: "Gets or sets the current log format",
-		Callback:         sshLogFormat,
+		Callback: func(fs interface{}, a []string, w sshd.StringWriter) error {
+			return sshLogFormat(l, fs, a, w)
+		},
 	})
 
 	ssh.RegisterCommand(&sshd.Command{
@@ -250,6 +271,7 @@ func attachCommands(ssh *sshd.SSHServer, hostMap *HostMap, pendingHostMap *HostM
 			s := sshPrintCertFlags{}
 			fl.BoolVar(&s.Json, "json", false, "outputs as json")
 			fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json, assumes -json")
+			fl.BoolVar(&s.Raw, "raw", false, "raw prints the PEM encoded certificate, not compatible with -json or -pretty")
 			return fl, &s
 		},
 		Callback: func(fs interface{}, a []string, w sshd.StringWriter) error {
@@ -331,8 +353,10 @@ func sshListHostMap(hostMap *HostMap, a interface{}, w sshd.StringWriter) error
 		return nil
 	}
 
-	hostMap.RLock()
-	defer hostMap.RUnlock()
+	hm := listHostMap(hostMap)
+	sort.Slice(hm, func(i, j int) bool {
+		return bytes.Compare(hm[i].VpnIp, hm[j].VpnIp) < 0
+	})
 
 	if fs.Json || fs.Pretty {
 		js := json.NewEncoder(w.GetWriter())
@@ -340,35 +364,15 @@ func sshListHostMap(hostMap *HostMap, a interface{}, w sshd.StringWriter) error
 			js.SetIndent("", "    ")
 		}
 
-		d := make([]m, len(hostMap.Hosts))
-		x := 0
-		var h m
-		for _, v := range hostMap.Hosts {
-			h = m{
-				"vpnIp":         int2ip(v.hostId),
-				"localIndex":    v.localIndexId,
-				"remoteIndex":   v.remoteIndexId,
-				"remoteAddrs":   v.RemoteUDPAddrs(),
-				"cachedPackets": len(v.packetStore),
-				"cert":          v.GetCert(),
-			}
-
-			if v.ConnectionState != nil {
-				h["messageCounter"] = atomic.LoadUint64(&v.ConnectionState.atomicMessageCounter)
-			}
-
-			d[x] = h
-			x++
-		}
-
-		err := js.Encode(d)
+		err := js.Encode(hm)
 		if err != nil {
 			//TODO
 			return nil
 		}
+
 	} else {
-		for i, v := range hostMap.Hosts {
-			err := w.WriteLine(fmt.Sprintf("%s: %s", int2ip(i), v.RemoteUDPAddrs()))
+		for _, v := range hm {
+			err := w.WriteLine(fmt.Sprintf("%s: %s", v.VpnIp, v.RemoteAddrs))
 			if err != nil {
 				return err
 			}
@@ -385,8 +389,26 @@ func sshListLighthouseMap(lightHouse *LightHouse, a interface{}, w sshd.StringWr
 		return nil
 	}
 
+	type lighthouseInfo struct {
+		VpnIp string    `json:"vpnIp"`
+		Addrs *CacheMap `json:"addrs"`
+	}
+
 	lightHouse.RLock()
-	defer lightHouse.RUnlock()
+	addrMap := make([]lighthouseInfo, len(lightHouse.addrMap))
+	x := 0
+	for k, v := range lightHouse.addrMap {
+		addrMap[x] = lighthouseInfo{
+			VpnIp: k.String(),
+			Addrs: v.CopyCache(),
+		}
+		x++
+	}
+	lightHouse.RUnlock()
+
+	sort.Slice(addrMap, func(i, j int) bool {
+		return strings.Compare(addrMap[i].VpnIp, addrMap[j].VpnIp) < 0
+	})
 
 	if fs.Json || fs.Pretty {
 		js := json.NewEncoder(w.GetWriter())
@@ -394,36 +416,19 @@ func sshListLighthouseMap(lightHouse *LightHouse, a interface{}, w sshd.StringWr
 			js.SetIndent("", "    ")
 		}
 
-		d := make([]m, len(lightHouse.addrMap))
-		x := 0
-		var h m
-		for vpnIp, v := range lightHouse.addrMap {
-			ips := make([]string, len(v))
-			for i, ip := range v {
-				ips[i] = ip.String()
-			}
-
-			h = m{
-				"vpnIp": int2ip(vpnIp),
-				"addrs": ips,
-			}
-
-			d[x] = h
-			x++
-		}
-
-		err := js.Encode(d)
+		err := js.Encode(addrMap)
 		if err != nil {
 			//TODO
 			return nil
 		}
+
 	} else {
-		for vpnIp, v := range lightHouse.addrMap {
-			ips := make([]string, len(v))
-			for i, ip := range v {
-				ips[i] = ip.String()
+		for _, v := range addrMap {
+			b, err := json.Marshal(v.Addrs)
+			if err != nil {
+				return err
 			}
-			err := w.WriteLine(fmt.Sprintf("%s: %s", int2ip(vpnIp), ips))
+			err = w.WriteLine(fmt.Sprintf("%s: %s", v.VpnIp, string(b)))
 			if err != nil {
 				return err
 			}
@@ -469,13 +474,17 @@ func sshQueryLighthouse(ifce *Interface, fs interface{}, a []string, w sshd.Stri
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	vpnIp := ip2int(parsedIp)
+	vpnIp := iputil.Ip2VpnIp(parsedIp)
 	if vpnIp == 0 {
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	ips, _ := ifce.lightHouse.Query(vpnIp, ifce)
-	return json.NewEncoder(w.GetWriter()).Encode(ips)
+	var cm *CacheMap
+	rl := ifce.lightHouse.Query(vpnIp, ifce)
+	if rl != nil {
+		cm = rl.CopyCache()
+	}
+	return json.NewEncoder(w.GetWriter()).Encode(cm)
 }
 
 func sshCloseTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringWriter) error {
@@ -494,19 +503,19 @@ func sshCloseTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringWr
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	vpnIp := ip2int(parsedIp)
+	vpnIp := iputil.Ip2VpnIp(parsedIp)
 	if vpnIp == 0 {
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	hostInfo, err := ifce.hostMap.QueryVpnIP(uint32(vpnIp))
+	hostInfo, err := ifce.hostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return w.WriteLine(fmt.Sprintf("Could not find tunnel for vpn ip: %v", a[0]))
 	}
 
 	if !flags.LocalOnly {
 		ifce.send(
-			closeTunnel,
+			header.CloseTunnel,
 			0,
 			hostInfo.ConnectionState,
 			hostInfo,
@@ -517,7 +526,7 @@ func sshCloseTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringWr
 		)
 	}
 
-	ifce.closeTunnel(hostInfo)
+	ifce.closeTunnel(hostInfo, false)
 	return w.WriteLine("Closed")
 }
 
@@ -537,32 +546,32 @@ func sshCreateTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringW
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	vpnIp := ip2int(parsedIp)
+	vpnIp := iputil.Ip2VpnIp(parsedIp)
 	if vpnIp == 0 {
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	hostInfo, _ := ifce.hostMap.QueryVpnIP(uint32(vpnIp))
+	hostInfo, _ := ifce.hostMap.QueryVpnIp(vpnIp)
 	if hostInfo != nil {
 		return w.WriteLine(fmt.Sprintf("Tunnel already exists"))
 	}
 
-	hostInfo, _ = ifce.handshakeManager.pendingHostMap.QueryVpnIP(uint32(vpnIp))
+	hostInfo, _ = ifce.handshakeManager.pendingHostMap.QueryVpnIp(vpnIp)
 	if hostInfo != nil {
 		return w.WriteLine(fmt.Sprintf("Tunnel already handshaking"))
 	}
 
-	var addr *udpAddr
+	var addr *udp.Addr
 	if flags.Address != "" {
-		addr = NewUDPAddrFromString(flags.Address)
+		addr = udp.NewAddrFromString(flags.Address)
 		if addr == nil {
 			return w.WriteLine("Address could not be parsed")
 		}
 	}
 
-	hostInfo = ifce.handshakeManager.AddVpnIP(vpnIp)
+	hostInfo = ifce.handshakeManager.AddVpnIp(vpnIp, ifce.initHostInfo)
 	if addr != nil {
-		hostInfo.SetRemote(*addr)
+		hostInfo.SetRemote(addr)
 	}
 	ifce.getOrHandshake(vpnIp)
 
@@ -584,7 +593,7 @@ func sshChangeRemote(ifce *Interface, fs interface{}, a []string, w sshd.StringW
 		return w.WriteLine("No address was provided")
 	}
 
-	addr := NewUDPAddrFromString(flags.Address)
+	addr := udp.NewAddrFromString(flags.Address)
 	if addr == nil {
 		return w.WriteLine("Address could not be parsed")
 	}
@@ -594,17 +603,17 @@ func sshChangeRemote(ifce *Interface, fs interface{}, a []string, w sshd.StringW
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	vpnIp := ip2int(parsedIp)
+	vpnIp := iputil.Ip2VpnIp(parsedIp)
 	if vpnIp == 0 {
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	hostInfo, err := ifce.hostMap.QueryVpnIP(uint32(vpnIp))
+	hostInfo, err := ifce.hostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return w.WriteLine(fmt.Sprintf("Could not find tunnel for vpn ip: %v", a[0]))
 	}
 
-	hostInfo.SetRemote(*addr)
+	hostInfo.SetRemote(addr)
 	return w.WriteLine("Changed")
 }
 
@@ -629,7 +638,7 @@ func sshGetHeapProfile(fs interface{}, a []string, w sshd.StringWriter) error {
 	return err
 }
 
-func sshLogLevel(fs interface{}, a []string, w sshd.StringWriter) error {
+func sshLogLevel(l *logrus.Logger, fs interface{}, a []string, w sshd.StringWriter) error {
 	if len(a) == 0 {
 		return w.WriteLine(fmt.Sprintf("Log level is: %s", l.Level))
 	}
@@ -643,7 +652,7 @@ func sshLogLevel(fs interface{}, a []string, w sshd.StringWriter) error {
 	return w.WriteLine(fmt.Sprintf("Log level is: %s", l.Level))
 }
 
-func sshLogFormat(fs interface{}, a []string, w sshd.StringWriter) error {
+func sshLogFormat(l *logrus.Logger, fs interface{}, a []string, w sshd.StringWriter) error {
 	if len(a) == 0 {
 		return w.WriteLine(fmt.Sprintf("Log format is: %s", reflect.TypeOf(l.Formatter)))
 	}
@@ -675,12 +684,12 @@ func sshPrintCert(ifce *Interface, fs interface{}, a []string, w sshd.StringWrit
 			return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 		}
 
-		vpnIp := ip2int(parsedIp)
+		vpnIp := iputil.Ip2VpnIp(parsedIp)
 		if vpnIp == 0 {
 			return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 		}
 
-		hostInfo, err := ifce.hostMap.QueryVpnIP(uint32(vpnIp))
+		hostInfo, err := ifce.hostMap.QueryVpnIp(vpnIp)
 		if err != nil {
 			return w.WriteLine(fmt.Sprintf("Could not find tunnel for vpn ip: %v", a[0]))
 		}
@@ -708,6 +717,16 @@ func sshPrintCert(ifce *Interface, fs interface{}, a []string, w sshd.StringWrit
 		return w.WriteBytes(b)
 	}
 
+	if args.Raw {
+		b, err := cert.MarshalToPEM()
+		if err != nil {
+			//TODO: handle it
+			return nil
+		}
+
+		return w.WriteBytes(b)
+	}
+
 	return w.WriteLine(cert.String())
 }
 
@@ -727,12 +746,12 @@ func sshPrintTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringWr
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	vpnIp := ip2int(parsedIp)
+	vpnIp := iputil.Ip2VpnIp(parsedIp)
 	if vpnIp == 0 {
 		return w.WriteLine(fmt.Sprintf("The provided vpn ip could not be parsed: %s", a[0]))
 	}
 
-	hostInfo, err := ifce.hostMap.QueryVpnIP(uint32(vpnIp))
+	hostInfo, err := ifce.hostMap.QueryVpnIp(vpnIp)
 	if err != nil {
 		return w.WriteLine(fmt.Sprintf("Could not find tunnel for vpn ip: %v", a[0]))
 	}
@@ -742,7 +761,7 @@ func sshPrintTunnel(ifce *Interface, fs interface{}, a []string, w sshd.StringWr
 		enc.SetIndent("", "    ")
 	}
 
-	return enc.Encode(hostInfo)
+	return enc.Encode(copyHostInfo(hostInfo, ifce.hostMap.preferredRanges))
 }
 
 func sshReload(fs interface{}, a []string, w sshd.StringWriter) error {

sshd/server.go

@@ -1,8 +1,10 @@
 package sshd
 
 import (
+	"errors"
 	"fmt"
 	"net"
+	"sync"
 
 	"github.com/armon/go-radix"
 	"github.com/sirupsen/logrus"
@@ -20,8 +22,11 @@ type SSHServer struct {
 	helpCommand *Command
 	commands    *radix.Tree
 	listener    net.Listener
-	conns       map[int]*session
-	counter     int
+
+	// Locks the conns/counter to avoid concurrent map access
+	connsLock sync.Mutex
+	conns     map[int]*session
+	counter   int
 }
 
 // NewSSHServer creates a new ssh server rigged with default commands and prepares to listen
@@ -97,11 +102,24 @@ func (s *SSHServer) Run(addr string) error {
 	}
 
 	s.l.WithField("sshListener", addr).Info("SSH server is listening")
+
+	// Run loops until there is an error
+	s.run()
+	s.closeSessions()
+
+	s.l.Info("SSH server stopped listening")
+	// We don't return an error because run logs for us
+	return nil
+}
+
+func (s *SSHServer) run() {
 	for {
 		c, err := s.listener.Accept()
 		if err != nil {
-			s.l.WithError(err).Warn("Error in listener, shutting down")
-			return nil
+			if !errors.Is(err, net.ErrClosed) {
+				s.l.WithError(err).Warn("Error in listener, shutting down")
+			}
+			return
 		}
 
 		conn, chans, reqs, err := ssh.NewServerConn(c, s.config)
@@ -127,36 +145,38 @@ func (s *SSHServer) Run(addr string) error {
 		l.WithField("remoteAddress", c.RemoteAddr()).WithField("sshFingerprint", fp).Info("ssh user logged in")
 
 		session := NewSession(s.commands, conn, chans, l.WithField("subsystem", "sshd.session"))
+		s.connsLock.Lock()
 		s.counter++
 		counter := s.counter
 		s.conns[counter] = session
+		s.connsLock.Unlock()
 
 		go ssh.DiscardRequests(reqs)
 		go func() {
 			<-session.exitChan
 			s.l.WithField("id", counter).Debug("closing conn")
+			s.connsLock.Lock()
 			delete(s.conns, counter)
+			s.connsLock.Unlock()
 		}()
 	}
 }
 
 func (s *SSHServer) Stop() {
+	// Close the listener, this will cause all session to terminate as well, see SSHServer.Run
+	if s.listener != nil {
+		if err := s.listener.Close(); err != nil {
+			s.l.WithError(err).Warn("Failed to close the sshd listener")
+		}
+	}
+}
+
+func (s *SSHServer) closeSessions() {
+	s.connsLock.Lock()
 	for _, c := range s.conns {
 		c.Close()
 	}
-
-	if s.listener == nil {
-		return
-	}
-
-	err := s.listener.Close()
-	if err != nil {
-		s.l.WithError(err).Warn("Failed to close the sshd listener")
-		return
-	}
-
-	s.l.Info("SSH server stopped listening")
-	return
+	s.connsLock.Unlock()
 }
 
 func (s *SSHServer) matchPubKey(c ssh.ConnMetadata, pubKey ssh.PublicKey) (*ssh.Permissions, error) {

sshd/session.go

@@ -81,11 +81,18 @@ func (s *session) handleRequests(in <-chan *ssh.Request, channel ssh.Channel) {
 		case "exec":
 			var payload = struct{ Value string }{}
 			cErr := ssh.Unmarshal(req.Payload, &payload)
-			if cErr == nil {
-				s.dispatchCommand(payload.Value, &stringWriter{channel})
-			} else {
-				//TODO: log it
+			if cErr != nil {
+				req.Reply(false, nil)
+				return
 			}
+
+			req.Reply(true, nil)
+			s.dispatchCommand(payload.Value, &stringWriter{channel})
+
+			//TODO: Fix error handling and report the proper status back
+			status := struct{ Status uint32 }{uint32(0)}
+			//TODO: I think this is how we shut down a shell as well?
+			channel.SendRequest("exit-status", false, ssh.Marshal(status))
 			channel.Close()
 			return
 

stats.go

@@ -6,6 +6,7 @@ import (
 	"log"
 	"net"
 	"net/http"
+	"runtime"
 	"time"
 
 	graphite "github.com/cyberdelia/go-metrics-graphite"
@@ -13,26 +14,39 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"github.com/rcrowley/go-metrics"
+	"github.com/sirupsen/logrus"
+	"github.com/slackhq/nebula/config"
 )
 
-func startStats(c *Config, configTest bool) error {
+// startStats initializes stats from config. On success, if any futher work
+// is needed to serve stats, it returns a func to handle that work. If no
+// work is needed, it'll return nil. On failure, it returns nil, error.
+func startStats(l *logrus.Logger, c *config.C, buildVersion string, configTest bool) (func(), error) {
 	mType := c.GetString("stats.type", "")
 	if mType == "" || mType == "none" {
-		return nil
+		return nil, nil
 	}
 
 	interval := c.GetDuration("stats.interval", 0)
 	if interval == 0 {
-		return fmt.Errorf("stats.interval was an invalid duration: %s", c.GetString("stats.interval", ""))
+		return nil, fmt.Errorf("stats.interval was an invalid duration: %s", c.GetString("stats.interval", ""))
 	}
 
+	var startFn func()
 	switch mType {
 	case "graphite":
-		startGraphiteStats(interval, c, configTest)
+		err := startGraphiteStats(l, interval, c, configTest)
+		if err != nil {
+			return nil, err
+		}
 	case "prometheus":
-		startPrometheusStats(interval, c, configTest)
+		var err error
+		startFn, err = startPrometheusStats(l, interval, c, buildVersion, configTest)
+		if err != nil {
+			return nil, err
+		}
 	default:
-		return fmt.Errorf("stats.type was not understood: %s", mType)
+		return nil, fmt.Errorf("stats.type was not understood: %s", mType)
 	}
 
 	metrics.RegisterDebugGCStats(metrics.DefaultRegistry)
@@ -41,10 +55,10 @@ func startStats(c *Config, configTest bool) error {
 	go metrics.CaptureDebugGCStats(metrics.DefaultRegistry, interval)
 	go metrics.CaptureRuntimeMemStats(metrics.DefaultRegistry, interval)
 
-	return nil
+	return startFn, nil
 }
 
-func startGraphiteStats(i time.Duration, c *Config, configTest bool) error {
+func startGraphiteStats(l *logrus.Logger, i time.Duration, c *config.C, configTest bool) error {
 	proto := c.GetString("stats.protocol", "tcp")
 	host := c.GetString("stats.host", "")
 	if host == "" {
@@ -57,38 +71,55 @@ func startGraphiteStats(i time.Duration, c *Config, configTest bool) error {
 		return fmt.Errorf("error while setting up graphite sink: %s", err)
 	}
 
-	l.Infof("Starting graphite. Interval: %s, prefix: %s, addr: %s", i, prefix, addr)
 	if !configTest {
+		l.Infof("Starting graphite. Interval: %s, prefix: %s, addr: %s", i, prefix, addr)
 		go graphite.Graphite(metrics.DefaultRegistry, i, prefix, addr)
 	}
 	return nil
 }
 
-func startPrometheusStats(i time.Duration, c *Config, configTest bool) error {
+func startPrometheusStats(l *logrus.Logger, i time.Duration, c *config.C, buildVersion string, configTest bool) (func(), error) {
 	namespace := c.GetString("stats.namespace", "")
 	subsystem := c.GetString("stats.subsystem", "")
 
 	listen := c.GetString("stats.listen", "")
 	if listen == "" {
-		return fmt.Errorf("stats.listen should not be empty")
+		return nil, fmt.Errorf("stats.listen should not be empty")
 	}
 
 	path := c.GetString("stats.path", "")
 	if path == "" {
-		return fmt.Errorf("stats.path should not be empty")
+		return nil, fmt.Errorf("stats.path should not be empty")
 	}
 
 	pr := prometheus.NewRegistry()
 	pClient := mp.NewPrometheusProvider(metrics.DefaultRegistry, namespace, subsystem, pr, i)
-	go pClient.UpdatePrometheusMetrics()
-
 	if !configTest {
-		go func() {
+		go pClient.UpdatePrometheusMetrics()
+	}
+
+	// Export our version information as labels on a static gauge
+	g := prometheus.NewGauge(prometheus.GaugeOpts{
+		Namespace: namespace,
+		Subsystem: subsystem,
+		Name:      "info",
+		Help:      "Version information for the Nebula binary",
+		ConstLabels: prometheus.Labels{
+			"version":   buildVersion,
+			"goversion": runtime.Version(),
+		},
+	})
+	pr.MustRegister(g)
+	g.Set(1)
+
+	var startFn func()
+	if !configTest {
+		startFn = func() {
 			l.Infof("Prometheus stats listening on %s at %s", listen, path)
 			http.Handle(path, promhttp.HandlerFor(pr, promhttp.HandlerOpts{ErrorLog: l}))
 			log.Fatal(http.ListenAndServe(listen, nil))
-		}()
+		}
 	}
 
-	return nil
+	return startFn, nil
 }

timeout.go

@@ -2,12 +2,14 @@ package nebula
 
 import (
 	"time"
+
+	"github.com/slackhq/nebula/firewall"
 )
 
 // How many timer objects should be cached
 const timerCacheMax = 50000
 
-var emptyFWPacket = FirewallPacket{}
+var emptyFWPacket = firewall.Packet{}
 
 type TimerWheel struct {
 	// Current tick
@@ -42,7 +44,7 @@ type TimeoutList struct {
 
 // Represents an item within a tick
 type TimeoutItem struct {
-	Packet FirewallPacket
+	Packet firewall.Packet
 	Next   *TimeoutItem
 }
 
@@ -73,8 +75,8 @@ func NewTimerWheel(min, max time.Duration) *TimerWheel {
 	return &tw
 }
 
-// Add will add a FirewallPacket to the wheel in it's proper timeout
-func (tw *TimerWheel) Add(v FirewallPacket, timeout time.Duration) *TimeoutItem {
+// Add will add a firewall.Packet to the wheel in it's proper timeout
+func (tw *TimerWheel) Add(v firewall.Packet, timeout time.Duration) *TimeoutItem {
 	// Check and see if we should progress the tick
 	tw.advance(time.Now())
 
@@ -103,7 +105,7 @@ func (tw *TimerWheel) Add(v FirewallPacket, timeout time.Duration) *TimeoutItem
 	return ti
 }
 
-func (tw *TimerWheel) Purge() (FirewallPacket, bool) {
+func (tw *TimerWheel) Purge() (firewall.Packet, bool) {
 	if tw.expired.Head == nil {
 		return emptyFWPacket, false
 	}

timeout_system.go

@@ -3,6 +3,8 @@ package nebula
 import (
 	"sync"
 	"time"
+
+	"github.com/slackhq/nebula/iputil"
 )
 
 // How many timer objects should be cached
@@ -43,7 +45,7 @@ type SystemTimeoutList struct {
 
 // Represents an item within a tick
 type SystemTimeoutItem struct {
-	Item uint32
+	Item iputil.VpnIp
 	Next *SystemTimeoutItem
 }
 
@@ -74,7 +76,7 @@ func NewSystemTimerWheel(min, max time.Duration) *SystemTimerWheel {
 	return &tw
 }
 
-func (tw *SystemTimerWheel) Add(v uint32, timeout time.Duration) *SystemTimeoutItem {
+func (tw *SystemTimerWheel) Add(v iputil.VpnIp, timeout time.Duration) *SystemTimeoutItem {
 	tw.lock.Lock()
 	defer tw.lock.Unlock()
 

timeout_system_test.go

@@ -5,6 +5,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/slackhq/nebula/iputil"
 	"github.com/stretchr/testify/assert"
 )
 
@@ -51,7 +52,7 @@ func TestSystemTimerWheel_findWheel(t *testing.T) {
 func TestSystemTimerWheel_Add(t *testing.T) {
 	tw := NewSystemTimerWheel(time.Second, time.Second*10)
 
-	fp1 := ip2int(net.ParseIP("1.2.3.4"))
+	fp1 := iputil.Ip2VpnIp(net.ParseIP("1.2.3.4"))
 	tw.Add(fp1, time.Second*1)
 
 	// Make sure we set head and tail properly
@@ -62,7 +63,7 @@ func TestSystemTimerWheel_Add(t *testing.T) {
 	assert.Nil(t, tw.wheel[2].Tail.Next)
 
 	// Make sure we only modify head
-	fp2 := ip2int(net.ParseIP("1.2.3.4"))
+	fp2 := iputil.Ip2VpnIp(net.ParseIP("1.2.3.4"))
 	tw.Add(fp2, time.Second*1)
 	assert.Equal(t, fp2, tw.wheel[2].Head.Item)
 	assert.Equal(t, fp1, tw.wheel[2].Head.Next.Item)
@@ -85,7 +86,7 @@ func TestSystemTimerWheel_Purge(t *testing.T) {
 	assert.NotNil(t, tw.lastTick)
 	assert.Equal(t, 0, tw.current)
 
-	fps := []uint32{9, 10, 11, 12}
+	fps := []iputil.VpnIp{9, 10, 11, 12}
 
 	//fp1 := ip2int(net.ParseIP("1.2.3.4"))
 

timeout_test.go

@@ -4,6 +4,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/slackhq/nebula/firewall"
 	"github.com/stretchr/testify/assert"
 )
 
@@ -50,7 +51,7 @@ func TestTimerWheel_findWheel(t *testing.T) {
 func TestTimerWheel_Add(t *testing.T) {
 	tw := NewTimerWheel(time.Second, time.Second*10)
 
-	fp1 := FirewallPacket{}
+	fp1 := firewall.Packet{}
 	tw.Add(fp1, time.Second*1)
 
 	// Make sure we set head and tail properly
@@ -61,7 +62,7 @@ func TestTimerWheel_Add(t *testing.T) {
 	assert.Nil(t, tw.wheel[2].Tail.Next)
 
 	// Make sure we only modify head
-	fp2 := FirewallPacket{}
+	fp2 := firewall.Packet{}
 	tw.Add(fp2, time.Second*1)
 	assert.Equal(t, fp2, tw.wheel[2].Head.Packet)
 	assert.Equal(t, fp1, tw.wheel[2].Head.Next.Packet)
@@ -84,7 +85,7 @@ func TestTimerWheel_Purge(t *testing.T) {
 	assert.NotNil(t, tw.lastTick)
 	assert.Equal(t, 0, tw.current)
 
-	fps := []FirewallPacket{
+	fps := []firewall.Packet{
 		{LocalIP: 1},
 		{LocalIP: 2},
 		{LocalIP: 3},

tun_android.go

@@ -1,3 +1,6 @@
+//go:build !e2e_testing
+// +build !e2e_testing
+
 package nebula
 
 import (
@@ -6,6 +9,7 @@ import (
 	"net"
 	"os"
 
+	"github.com/sirupsen/logrus"
 	"golang.org/x/sys/unix"
 )
 
@@ -19,9 +23,10 @@ type Tun struct {
 	TXQueueLen   int
 	Routes       []route
 	UnsafeRoutes []route
+	l            *logrus.Logger
 }
 
-func newTunFromFd(deviceFd int, cidr *net.IPNet, defaultMTU int, routes []route, unsafeRoutes []route, txQueueLen int) (ifce *Tun, err error) {
+func newTunFromFd(l *logrus.Logger, deviceFd int, cidr *net.IPNet, defaultMTU int, routes []route, unsafeRoutes []route, txQueueLen int) (ifce *Tun, err error) {
 	file := os.NewFile(uintptr(deviceFd), "/dev/net/tun")
 
 	ifce = &Tun{
@@ -33,11 +38,12 @@ func newTunFromFd(deviceFd int, cidr *net.IPNet, defaultMTU int, routes []route,
 		TXQueueLen:      txQueueLen,
 		Routes:          routes,
 		UnsafeRoutes:    unsafeRoutes,
+		l:               l,
 	}
 	return
 }
 
-func newTun(deviceName string, cidr *net.IPNet, defaultMTU int, routes []route, unsafeRoutes []route, txQueueLen int, multiqueue bool) (ifce *Tun, err error) {
+func newTun(l *logrus.Logger, deviceName string, cidr *net.IPNet, defaultMTU int, routes []route, unsafeRoutes []route, txQueueLen int, multiqueue bool) (ifce *Tun, err error) {
 	return nil, fmt.Errorf("newTun not supported in Android")
 }