Fix ebpf free call

* Fix error handling in iptables initialization

* Change log level

.github/workflows/golang-test-darwin.yml

@@ -17,7 +17,7 @@ jobs:
       - name: Install Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
       - name: Checkout code
         uses: actions/checkout@v2
 

.github/workflows/golang-test-linux.yml

@@ -20,7 +20,7 @@ jobs:
       - name: Install Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
 
 
       - name: Cache Go modules
@@ -49,7 +49,7 @@ jobs:
       - name: Install Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
 
 
       - name: Cache Go modules
@@ -70,27 +70,40 @@ jobs:
         run: go mod tidy
 
       - name: Generate Iface Test bin
-        run: go test -c -o iface-testing.bin ./iface/
+        run: CGO_ENABLED=0 go test -c -o iface-testing.bin ./iface/
+
+      - name: Generate Shared Sock Test bin
+        run: CGO_ENABLED=0 go test -c -o sharedsock-testing.bin ./sharedsock
 
       - name: Generate RouteManager Test bin
-        run: go test -c -o routemanager-testing.bin ./client/internal/routemanager/...
+        run: CGO_ENABLED=0 go test -c -o routemanager-testing.bin ./client/internal/routemanager/...
+
+      - name: Generate nftables Manager Test bin
+        run: CGO_ENABLED=0 go test -c -o nftablesmanager-testing.bin ./client/firewall/nftables/...
 
       - name: Generate Engine Test bin
-        run: go test -c -o engine-testing.bin ./client/internal
+        run: CGO_ENABLED=0 go test -c -o engine-testing.bin ./client/internal
 
       - name: Generate Peer Test bin
-        run: go test -c -o peer-testing.bin ./client/internal/peer/...
+        run: CGO_ENABLED=0 go test -c -o peer-testing.bin ./client/internal/peer/...
 
       - run: chmod +x *testing.bin
 
+      - name: Run Shared Sock tests in docker
+        run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/sharedsock --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/sharedsock-testing.bin -test.timeout 5m -test.parallel 1
+
       - name: Run Iface tests in docker
         run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/iface --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/iface-testing.bin -test.timeout 5m -test.parallel 1
 
+
       - name: Run RouteManager tests in docker
         run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/client/internal/routemanager --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/routemanager-testing.bin  -test.timeout 5m -test.parallel 1
 
+      - name: Run nftables Manager tests in docker
+        run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/client/firewall --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/nftablesmanager-testing.bin  -test.timeout 5m -test.parallel 1
+
       - name: Run Engine tests in docker
         run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/client/internal --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/engine-testing.bin  -test.timeout 5m -test.parallel 1
 
       - name: Run Peer tests in docker
-        run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/client/internal/peer  --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/peer-testing.bin  -test.timeout 5m -test.parallel 1
+        run: docker run -t --cap-add=NET_ADMIN --privileged --rm -v $PWD:/ci -w /ci/client/internal/peer  --entrypoint /busybox/sh gcr.io/distroless/base:debug -c /ci/peer-testing.bin  -test.timeout 5m -test.parallel 1

.github/workflows/golang-test-windows.yml

@@ -23,7 +23,7 @@ jobs:
         uses: actions/setup-go@v4
         id: go
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
 
       - name: Download wintun
         uses: carlosperate/download-file-action@v2

.github/workflows/golangci-lint.yml

@@ -12,7 +12,7 @@ jobs:
       - name: Install Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
       - name: Install dependencies
         run: sudo apt update && sudo apt install -y -q libgtk-3-dev libayatana-appindicator3-dev libgl1-mesa-dev xorg-dev
       - name: golangci-lint

.github/workflows/release.yml

@@ -7,9 +7,19 @@ on:
     branches:
       - main
   pull_request:
+    paths:
+      - 'go.mod'
+      - 'go.sum'
+      - '.goreleaser.yml'
+      - '.goreleaser_ui.yaml'
+      - '.goreleaser_ui_darwin.yaml'
+      - '.github/workflows/release.yml'
+      - 'release_files/**'
+      - '**/Dockerfile'
+      - '**/Dockerfile.*'
 
 env:
-  SIGN_PIPE_VER: "v0.0.6"
+  SIGN_PIPE_VER: "v0.0.8"
   GORELEASER_VER: "v1.14.1"
 
 concurrency:
@@ -29,7 +39,7 @@ jobs:
         name: Set up Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19
+          go-version: "1.20"
       -
         name: Cache Go modules
         uses: actions/cache@v1
@@ -100,7 +110,7 @@ jobs:
       - name: Set up Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19
+          go-version: "1.20"
       - name: Cache Go modules
         uses: actions/cache@v1
         with:
@@ -116,7 +126,7 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Install dependencies
-        run: sudo apt update && sudo apt install -y -q libgtk-3-dev libayatana-appindicator3-dev libgl1-mesa-dev xorg-dev gcc-mingw-w64-x86-64
+        run: sudo apt update && sudo apt install -y -q libappindicator3-dev gir1.2-appindicator3-0.1 libxxf86vm-dev gcc-mingw-w64-x86-64
       - name: Install rsrc
         run: go install github.com/akavel/rsrc@v0.10.2
       - name: Generate windows rsrc
@@ -150,7 +160,7 @@ jobs:
         name: Set up Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19
+          go-version: "1.20"
       -
         name: Cache Go modules
         uses: actions/cache@v1
@@ -195,7 +205,7 @@ jobs:
 
   trigger_darwin_signer:
     runs-on: ubuntu-latest
-    needs: release_ui_darwin
+    needs: [release,release_ui_darwin]
     if: startsWith(github.ref, 'refs/tags/')
     steps:
       - name: Trigger Darwin App binaries sign pipeline
@@ -205,4 +215,4 @@ jobs:
           repo: netbirdio/sign-pipelines
           ref: ${{ env.SIGN_PIPE_VER }}
           token: ${{ secrets.SIGN_GITHUB_TOKEN }}
-          inputs: '{ "tag": "${{ github.ref }}" }'
+          inputs: '{ "tag": "${{ github.ref }}" }'

.github/workflows/test-infrastructure-files.yml

@@ -1,17 +1,21 @@
-name: Test Docker Compose Linux
+name: Test Infrastructure files
 
 on:
   push:
     branches:
       - main
   pull_request:
+    paths:
+      - 'infrastructure_files/**'
+      - '.github/workflows/test-infrastructure-files.yml'
+
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.actor_id }}
   cancel-in-progress: true
 
 jobs:
-  test:
+  test-docker-compose:
     runs-on: ubuntu-latest
     steps:
       - name: Install jq
@@ -23,7 +27,7 @@ jobs:
       - name: Install Go
         uses: actions/setup-go@v2
         with:
-          go-version: 1.19.x
+          go-version: "1.20.x"
 
       - name: Cache Go modules
         uses: actions/cache@v2
@@ -34,7 +38,7 @@ jobs:
             ${{ runner.os }}-go-
 
       - name: Checkout code
-        uses: actions/checkout@v2
+        uses: actions/checkout@v3
 
       - name: cp setup.env
         run: cp infrastructure_files/tests/setup.env infrastructure_files/
@@ -45,15 +49,21 @@ jobs:
         env:
           CI_NETBIRD_DOMAIN: localhost
           CI_NETBIRD_AUTH_CLIENT_ID: testing.client.id
+          CI_NETBIRD_AUTH_CLIENT_SECRET: testing.client.secret
           CI_NETBIRD_AUTH_AUDIENCE: testing.ci
           CI_NETBIRD_AUTH_OIDC_CONFIGURATION_ENDPOINT: https://example.eu.auth0.com/.well-known/openid-configuration
           CI_NETBIRD_USE_AUTH0: true
+          CI_NETBIRD_MGMT_IDP: "none"
+          CI_NETBIRD_IDP_MGMT_CLIENT_ID: testing.client.id
+          CI_NETBIRD_IDP_MGMT_CLIENT_SECRET: testing.client.secret
+          CI_NETBIRD_AUTH_SUPPORTED_SCOPES: "openid profile email offline_access api email_verified"
 
       - name: check values
         working-directory: infrastructure_files
         env:
           CI_NETBIRD_DOMAIN: localhost
           CI_NETBIRD_AUTH_CLIENT_ID: testing.client.id
+          CI_NETBIRD_AUTH_CLIENT_SECRET: testing.client.secret
           CI_NETBIRD_AUTH_AUDIENCE: testing.ci
           CI_NETBIRD_AUTH_OIDC_CONFIGURATION_ENDPOINT: https://example.eu.auth0.com/.well-known/openid-configuration
           CI_NETBIRD_USE_AUTH0: true
@@ -62,14 +72,19 @@ jobs:
           CI_NETBIRD_AUTH_JWT_CERTS: https://example.eu.auth0.com/.well-known/jwks.json
           CI_NETBIRD_AUTH_TOKEN_ENDPOINT: https://example.eu.auth0.com/oauth/token
           CI_NETBIRD_AUTH_DEVICE_AUTH_ENDPOINT: https://example.eu.auth0.com/oauth/device/code
+          CI_NETBIRD_AUTH_PKCE_AUTHORIZATION_ENDPOINT: https://example.eu.auth0.com/authorize
           CI_NETBIRD_AUTH_REDIRECT_URI: "/peers"
           CI_NETBIRD_TOKEN_SOURCE: "idToken"
           CI_NETBIRD_AUTH_USER_ID_CLAIM: "email"
           CI_NETBIRD_AUTH_DEVICE_AUTH_AUDIENCE: "super"
           CI_NETBIRD_AUTH_DEVICE_AUTH_SCOPE: "openid email"
+          CI_NETBIRD_MGMT_IDP: "none"
+          CI_NETBIRD_IDP_MGMT_CLIENT_ID: testing.client.id
+          CI_NETBIRD_IDP_MGMT_CLIENT_SECRET: testing.client.secret
 
         run: |
           grep AUTH_CLIENT_ID docker-compose.yml | grep $CI_NETBIRD_AUTH_CLIENT_ID
+          grep AUTH_CLIENT_SECRET docker-compose.yml | grep $CI_NETBIRD_AUTH_CLIENT_SECRET
           grep AUTH_AUTHORITY docker-compose.yml | grep $CI_NETBIRD_AUTH_AUTHORITY
           grep AUTH_AUDIENCE docker-compose.yml | grep $CI_NETBIRD_AUTH_AUDIENCE
           grep AUTH_SUPPORTED_SCOPES docker-compose.yml | grep "$CI_NETBIRD_AUTH_SUPPORTED_SCOPES"
@@ -80,13 +95,25 @@ jobs:
           grep LETSENCRYPT_DOMAIN docker-compose.yml | egrep 'LETSENCRYPT_DOMAIN=$'
           grep NETBIRD_TOKEN_SOURCE docker-compose.yml | grep $CI_NETBIRD_TOKEN_SOURCE
           grep AuthUserIDClaim management.json | grep $CI_NETBIRD_AUTH_USER_ID_CLAIM
-          grep -A 1 ProviderConfig management.json | grep Audience | grep $CI_NETBIRD_AUTH_DEVICE_AUTH_AUDIENCE
-          grep Scope management.json | grep "$CI_NETBIRD_AUTH_DEVICE_AUTH_SCOPE"
+          grep -A 3 DeviceAuthorizationFlow management.json | grep -A 1 ProviderConfig | grep Audience | grep $CI_NETBIRD_AUTH_DEVICE_AUTH_AUDIENCE
+          grep -A 8 DeviceAuthorizationFlow management.json | grep -A 6 ProviderConfig | grep Scope | grep "$CI_NETBIRD_AUTH_DEVICE_AUTH_SCOPE"
           grep UseIDToken management.json | grep false
+          grep -A 1 IdpManagerConfig management.json | grep ManagerType | grep $CI_NETBIRD_MGMT_IDP 
+          grep -A 3 IdpManagerConfig management.json | grep -A 1 ClientConfig | grep Issuer | grep $CI_NETBIRD_AUTH_AUTHORITY
+          grep -A 4 IdpManagerConfig management.json | grep -A 2 ClientConfig | grep TokenEndpoint | grep $CI_NETBIRD_AUTH_TOKEN_ENDPOINT
+          grep -A 5 IdpManagerConfig management.json | grep -A 3 ClientConfig | grep ClientID | grep $CI_NETBIRD_IDP_MGMT_CLIENT_ID
+          grep -A 6 IdpManagerConfig management.json | grep -A 4 ClientConfig | grep ClientSecret | grep $CI_NETBIRD_IDP_MGMT_CLIENT_SECRET
+          grep -A 7 IdpManagerConfig management.json | grep -A 5 ClientConfig | grep GrantType | grep client_credentials
+          grep -A 2 PKCEAuthorizationFlow management.json | grep -A 1 ProviderConfig | grep Audience | grep $CI_NETBIRD_AUTH_AUDIENCE
+          grep -A 3 PKCEAuthorizationFlow management.json | grep -A 2 ProviderConfig | grep ClientID | grep $CI_NETBIRD_AUTH_CLIENT_ID
+          grep -A 4 PKCEAuthorizationFlow management.json | grep -A 3 ProviderConfig | grep ClientSecret | grep $CI_NETBIRD_AUTH_CLIENT_SECRET
+          grep -A 5 PKCEAuthorizationFlow management.json | grep -A 4 ProviderConfig | grep AuthorizationEndpoint | grep $CI_NETBIRD_AUTH_PKCE_AUTHORIZATION_ENDPOINT
+          grep -A 6 PKCEAuthorizationFlow management.json | grep -A 5 ProviderConfig | grep TokenEndpoint | grep $CI_NETBIRD_AUTH_TOKEN_ENDPOINT
+          grep -A 7 PKCEAuthorizationFlow management.json | grep -A 6 ProviderConfig | grep Scope | grep "$CI_NETBIRD_AUTH_SUPPORTED_SCOPES"
 
       - name: run docker compose up
         working-directory: infrastructure_files
-        run: | 
+        run: |
           docker-compose up -d
           sleep 5
           docker-compose ps
@@ -97,3 +124,28 @@ jobs:
           count=$(docker compose ps --format json | jq '.[] | select(.Project | contains("infrastructure_files")) | .State' | grep -c running)
           test $count -eq 4
         working-directory: infrastructure_files
+
+  test-getting-started-script:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Install jq
+        run: sudo apt-get install -y jq
+
+      - name: Checkout code
+        uses: actions/checkout@v3
+
+      - name: run script
+        run: NETBIRD_DOMAIN=use-ip bash -x infrastructure_files/getting-started-with-zitadel.sh
+
+      - name: test Caddy file gen
+        run: test -f Caddyfile
+      - name: test docker-compose file gen
+        run: test -f docker-compose.yml
+      - name: test management.json file gen
+        run: test -f management.json
+      - name: test turnserver.conf file gen
+        run: test -f turnserver.conf
+      - name: test zitadel.env file gen
+        run: test -f zitadel.env
+      - name: test dashboard.env file gen
+        run: test -f dashboard.env

.github/workflows/update-docs.yml

@@ -0,0 +1,22 @@
+name: update docs
+
+on:
+  push:
+    tags:
+      - 'v*'
+    paths:
+      - 'management/server/http/api/openapi.yml'
+
+jobs:
+  trigger_docs_api_update:
+    runs-on: ubuntu-latest
+    if: startsWith(github.ref, 'refs/tags/')
+    steps:
+      - name: Trigger API pages generation
+        uses: benc-uk/workflow-dispatch@v1
+        with:
+          workflow: generate api pages
+          repo: netbirdio/docs
+          ref: "refs/heads/main"
+          token: ${{ secrets.SIGN_GITHUB_TOKEN }}
+          inputs: '{ "tag": "${{ github.ref }}" }'

.gitignore

@@ -7,8 +7,15 @@ bin/
 conf.json
 http-cmds.sh
 infrastructure_files/management.json
+infrastructure_files/management-*.json
 infrastructure_files/docker-compose.yml
+infrastructure_files/openid-configuration.json
+infrastructure_files/turnserver.conf
+management/management
+client/client
+client/client.exe
 *.syso
 client/.distfiles/
 infrastructure_files/setup.env
+infrastructure_files/setup-*.env
 .vscode

.goreleaser.yaml

@@ -12,11 +12,7 @@ builds:
       - arm
       - amd64
       - arm64
-      - mips
       - 386
-    gomips:
-      - hardfloat
-      - softfloat
     ignore:
       - goos: windows
         goarch: arm64
@@ -30,6 +26,26 @@ builds:
     tags:
       - load_wgnt_from_rsrc
 
+  - id: netbird-static
+    dir: client
+    binary: netbird
+    env: [CGO_ENABLED=0]
+    goos:
+      - linux
+    goarch:
+      - mips
+      - mipsle
+      - mips64
+      - mips64le
+    gomips:
+      - hardfloat
+      - softfloat
+    ldflags:
+      - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
+    mod_timestamp: '{{ .CommitTimestamp }}'
+    tags:
+      - load_wgnt_from_rsrc
+
   - id: netbird-mgmt
     dir: management
     env:
@@ -67,6 +83,7 @@ builds:
 archives:
   - builds:
       - netbird
+      - netbird-static
 
 nfpms:
 
@@ -359,4 +376,14 @@ uploads:
     mode: archive
     target: https://pkgs.wiretrustee.com/yum/{{ .Arch }}{{ if .Arm }}{{ .Arm }}{{ end }}
     username: dev@wiretrustee.com
-    method: PUT
+    method: PUT
+
+checksum:
+  extra_files:
+    - glob: ./infrastructure_files/getting-started-with-zitadel.sh
+    - glob: ./release_files/install.sh
+
+release:
+  extra_files:
+    - glob: ./infrastructure_files/getting-started-with-zitadel.sh
+    - glob: ./release_files/install.sh

.goreleaser_ui.yaml

@@ -11,6 +11,8 @@ builds:
       - amd64
     ldflags:
       - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
+    tags:
+      - legacy_appindicator
     mod_timestamp: '{{ .CommitTimestamp }}'
 
   - id: netbird-ui-windows
@@ -55,9 +57,6 @@ nfpms:
       - src: client/ui/disconnected.png
         dst: /usr/share/pixmaps/netbird.png
     dependencies:
-      - libayatana-appindicator3-1
-      - libgtk-3-dev
-      - libappindicator3-dev
       - netbird
 
   - maintainer: Netbird <dev@netbird.io>
@@ -75,9 +74,6 @@ nfpms:
       - src: client/ui/disconnected.png
         dst: /usr/share/pixmaps/netbird.png
     dependencies:
-      - libayatana-appindicator3-1
-      - libgtk-3-dev
-      - libappindicator3-dev
       - netbird
 
 uploads:

README.md

@@ -59,7 +59,7 @@ NetBird uses [NAT traversal techniques](https://en.wikipedia.org/wiki/Interactiv
 - \[x] Network Activity Monitoring.
 
 **Coming soon:**
--  \[ ] Mobile clients.
+- \[ ] Mobile clients.
 
 ### Secure peer-to-peer VPN with SSO and MFA in minutes
 
@@ -70,9 +70,9 @@ For stable versions, see [releases](https://github.com/netbirdio/netbird/release
 
 ### Start using NetBird
 -  Hosted version: [https://app.netbird.io/](https://app.netbird.io/).
--  See our documentation for [Quickstart Guide](https://netbird.io/docs/getting-started/quickstart).
--  If you are looking to self-host NetBird, check our [Self-Hosting Guide](https://netbird.io/docs/getting-started/self-hosting).
--  Step-by-step [Installation Guide](https://netbird.io/docs/getting-started/installation) for different platforms.
+-  See our documentation for [Quickstart Guide](https://docs.netbird.io/how-to/getting-started).
+-  If you are looking to self-host NetBird, check our [Self-Hosting Guide](https://docs.netbird.io/selfhosted/selfhosted-guide).
+-  Step-by-step [Installation Guide](https://docs.netbird.io/how-to/getting-started#installation) for different platforms.
 -  Web UI [repository](https://github.com/netbirdio/dashboard).
 -  5 min [demo video](https://youtu.be/Tu9tPsUWaY0) on YouTube.
 
@@ -91,7 +91,7 @@ For stable versions, see [releases](https://github.com/netbirdio/netbird/release
   <img src="https://netbird.io/docs/img/architecture/high-level-dia.png" width="700"/>
 </p>
 
-See a complete [architecture overview](https://netbird.io/docs/overview/architecture) for details.
+See a complete [architecture overview](https://docs.netbird.io/about-netbird/how-netbird-works#architecture) for details.
 
 ### Roadmap
 -  [Public Roadmap](https://github.com/netbirdio/netbird/projects/2)

base62/base62.go

@@ -0,0 +1,59 @@
+package base62
+
+import (
+	"fmt"
+	"math"
+	"strings"
+)
+
+const (
+	alphabet = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
+	base     = uint32(len(alphabet))
+)
+
+// Encode encodes a uint32 value to a base62 string.
+func Encode(num uint32) string {
+	if num == 0 {
+		return string(alphabet[0])
+	}
+
+	var encoded strings.Builder
+	remainder := uint32(0)
+
+	for num > 0 {
+		remainder = num % base
+		encoded.WriteByte(alphabet[remainder])
+		num /= base
+	}
+
+	// Reverse the encoded string
+	encodedString := encoded.String()
+	reversed := reverse(encodedString)
+	return reversed
+}
+
+// Decode decodes a base62 string to a uint32 value.
+func Decode(encoded string) (uint32, error) {
+	var decoded uint32
+	strLen := len(encoded)
+
+	for i, char := range encoded {
+		index := strings.IndexRune(alphabet, char)
+		if index < 0 {
+			return 0, fmt.Errorf("invalid character: %c", char)
+		}
+
+		decoded += uint32(index) * uint32(math.Pow(float64(base), float64(strLen-i-1)))
+	}
+
+	return decoded, nil
+}
+
+// Reverse a string.
+func reverse(s string) string {
+	runes := []rune(s)
+	for i, j := 0, len(runes)-1; i < j; i, j = i+1, j-1 {
+		runes[i], runes[j] = runes[j], runes[i]
+	}
+	return string(runes)
+}

base62/base62_test.go

@@ -0,0 +1,31 @@
+package base62
+
+import (
+	"testing"
+)
+
+func TestEncodeDecode(t *testing.T) {
+	tests := []struct {
+		num uint32
+	}{
+		{0},
+		{1},
+		{42},
+		{12345},
+		{99999},
+		{123456789},
+	}
+
+	for _, tt := range tests {
+		encoded := Encode(tt.num)
+		decoded, err := Decode(encoded)
+
+		if err != nil {
+			t.Errorf("Decode error: %v", err)
+		}
+
+		if decoded != tt.num {
+			t.Errorf("Decode(%v) = %v, want %v", encoded, decoded, tt.num)
+		}
+	}
+}

client/android/client.go

@@ -7,7 +7,9 @@ import (
 	log "github.com/sirupsen/logrus"
 
 	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/dns"
 	"github.com/netbirdio/netbird/client/internal/peer"
+	"github.com/netbirdio/netbird/client/internal/routemanager"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
 	"github.com/netbirdio/netbird/client/system"
 	"github.com/netbirdio/netbird/formatter"
@@ -29,6 +31,16 @@ type IFaceDiscover interface {
 	stdnet.ExternalIFaceDiscover
 }
 
+// RouteListener export internal RouteListener for mobile
+type RouteListener interface {
+	routemanager.RouteListener
+}
+
+// DnsReadyListener export internal dns ReadyListener for mobile
+type DnsReadyListener interface {
+	dns.ReadyListener
+}
+
 func init() {
 	formatter.SetLogcatFormatter(log.StandardLogger())
 }
@@ -42,13 +54,12 @@ type Client struct {
 	ctxCancel     context.CancelFunc
 	ctxCancelLock *sync.Mutex
 	deviceName    string
+	routeListener routemanager.RouteListener
+	onHostDnsFn   func([]string)
 }
 
 // NewClient instantiate a new Client
-func NewClient(cfgFile, deviceName string, tunAdapter TunAdapter, iFaceDiscover IFaceDiscover) *Client {
-	lvl, _ := log.ParseLevel("trace")
-	log.SetLevel(lvl)
-
+func NewClient(cfgFile, deviceName string, tunAdapter TunAdapter, iFaceDiscover IFaceDiscover, routeListener RouteListener) *Client {
 	return &Client{
 		cfgFile:       cfgFile,
 		deviceName:    deviceName,
@@ -56,11 +67,12 @@ func NewClient(cfgFile, deviceName string, tunAdapter TunAdapter, iFaceDiscover
 		iFaceDiscover: iFaceDiscover,
 		recorder:      peer.NewRecorder(""),
 		ctxCancelLock: &sync.Mutex{},
+		routeListener: routeListener,
 	}
 }
 
 // Run start the internal client. It is a blocker function
-func (c *Client) Run(urlOpener URLOpener) error {
+func (c *Client) Run(urlOpener URLOpener, dns *DNSList, dnsReadyListener DnsReadyListener) error {
 	cfg, err := internal.UpdateOrCreateConfig(internal.ConfigInput{
 		ConfigPath: c.cfgFile,
 	})
@@ -85,7 +97,8 @@ func (c *Client) Run(urlOpener URLOpener) error {
 
 	// todo do not throw error in case of cancelled context
 	ctx = internal.CtxInitState(ctx)
-	return internal.RunClient(ctx, cfg, c.recorder, c.tunAdapter, c.iFaceDiscover)
+	c.onHostDnsFn = func([]string) {}
+	return internal.RunClientMobile(ctx, cfg, c.recorder, c.tunAdapter, c.iFaceDiscover, c.routeListener, dns.items, dnsReadyListener)
 }
 
 // Stop the internal client and free the resources
@@ -99,6 +112,11 @@ func (c *Client) Stop() {
 	c.ctxCancel()
 }
 
+// SetTraceLogLevel configure the logger to trace level
+func (c *Client) SetTraceLogLevel() {
+	log.SetLevel(log.TraceLevel)
+}
+
 // PeersList return with the list of the PeerInfos
 func (c *Client) PeersList() *PeerInfoArray {
 
@@ -110,14 +128,23 @@ func (c *Client) PeersList() *PeerInfoArray {
 			p.IP,
 			p.FQDN,
 			p.ConnStatus.String(),
-			p.Direct,
 		}
 		peerInfos[n] = pi
 	}
-
 	return &PeerInfoArray{items: peerInfos}
 }
 
+// OnUpdatedHostDNS update the DNS servers addresses for root zones
+func (c *Client) OnUpdatedHostDNS(list *DNSList) error {
+	dnsServer, err := dns.GetServerDns()
+	if err != nil {
+		return err
+	}
+
+	dnsServer.OnUpdatedHostDNSServer(list.items)
+	return nil
+}
+
 // SetConnectionListener set the network connection listener
 func (c *Client) SetConnectionListener(listener ConnectionListener) {
 	c.recorder.SetConnectionListener(listener)

client/android/dns_list.go

@@ -0,0 +1,26 @@
+package android
+
+import "fmt"
+
+// DNSList is a wrapper of []string
+type DNSList struct {
+	items []string
+}
+
+// Add new DNS address to the collection
+func (array *DNSList) Add(s string) {
+	array.items = append(array.items, s)
+}
+
+// Get return an element of the collection
+func (array *DNSList) Get(i int) (string, error) {
+	if i >= len(array.items) || i < 0 {
+		return "", fmt.Errorf("out of range")
+	}
+	return array.items[i], nil
+}
+
+// Size return with the size of the collection
+func (array *DNSList) Size() int {
+	return len(array.items)
+}

client/android/dns_list_test.go

@@ -0,0 +1,24 @@
+package android
+
+import "testing"
+
+func TestDNSList_Get(t *testing.T) {
+	l := DNSList{
+		items: make([]string, 1),
+	}
+
+	_, err := l.Get(0)
+	if err != nil {
+		t.Errorf("invalid error: %s", err)
+	}
+
+	_, err = l.Get(-1)
+	if err == nil {
+		t.Errorf("expected error but got nil")
+	}
+
+	_, err = l.Get(1)
+	if err == nil {
+		t.Errorf("expected error but got nil")
+	}
+}

client/android/gomobile.go

@@ -0,0 +1,5 @@
+package android
+
+import _ "golang.org/x/mobile/bind"
+
+// to keep our CI/CD that checks go.mod and go.sum files happy, we need to import the package above

client/android/login.go

@@ -6,15 +6,14 @@ import (
 	"time"
 
 	"github.com/cenkalti/backoff/v4"
-
 	log "github.com/sirupsen/logrus"
 	"google.golang.org/grpc/codes"
 	gstatus "google.golang.org/grpc/status"
 
 	"github.com/netbirdio/netbird/client/cmd"
-	"github.com/netbirdio/netbird/client/system"
-
 	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/auth"
+	"github.com/netbirdio/netbird/client/system"
 )
 
 // SSOListener is async listener for mobile framework
@@ -87,9 +86,15 @@ func (a *Auth) saveConfigIfSSOSupported() (bool, error) {
 	err := a.withBackOff(a.ctx, func() (err error) {
 		_, err = internal.GetDeviceAuthorizationFlowInfo(a.ctx, a.config.PrivateKey, a.config.ManagementURL)
 		if s, ok := gstatus.FromError(err); ok && s.Code() == codes.NotFound {
-			supportsSSO = false
-			err = nil
+			_, err = internal.GetPKCEAuthorizationFlowInfo(a.ctx, a.config.PrivateKey, a.config.ManagementURL)
+			if s, ok := gstatus.FromError(err); ok && s.Code() == codes.NotFound {
+				supportsSSO = false
+				err = nil
+			}
+
+			return err
 		}
+
 		return err
 	})
 
@@ -183,27 +188,15 @@ func (a *Auth) login(urlOpener URLOpener) error {
 	return nil
 }
 
-func (a *Auth) foregroundGetTokenInfo(urlOpener URLOpener) (*internal.TokenInfo, error) {
-	providerConfig, err := internal.GetDeviceAuthorizationFlowInfo(a.ctx, a.config.PrivateKey, a.config.ManagementURL)
+func (a *Auth) foregroundGetTokenInfo(urlOpener URLOpener) (*auth.TokenInfo, error) {
+	oAuthFlow, err := auth.NewOAuthFlow(a.ctx, a.config)
 	if err != nil {
-		s, ok := gstatus.FromError(err)
-		if ok && s.Code() == codes.NotFound {
-			return nil, fmt.Errorf("no SSO provider returned from management. " +
-				"If you are using hosting Netbird see documentation at " +
-				"https://github.com/netbirdio/netbird/tree/main/management for details")
-		} else if ok && s.Code() == codes.Unimplemented {
-			return nil, fmt.Errorf("the management server, %s, does not support SSO providers, "+
-				"please update your servver or use Setup Keys to login", a.config.ManagementURL)
-		} else {
-			return nil, fmt.Errorf("getting device authorization flow info failed with error: %v", err)
-		}
+		return nil, err
 	}
 
-	hostedClient := internal.NewHostedDeviceFlow(providerConfig.ProviderConfig)
-
-	flowInfo, err := hostedClient.RequestDeviceCode(context.TODO())
+	flowInfo, err := oAuthFlow.RequestAuthInfo(context.TODO())
 	if err != nil {
-		return nil, fmt.Errorf("getting a request device code failed: %v", err)
+		return nil, fmt.Errorf("getting a request OAuth flow info failed: %v", err)
 	}
 
 	go urlOpener.Open(flowInfo.VerificationURIComplete)
@@ -211,7 +204,7 @@ func (a *Auth) foregroundGetTokenInfo(urlOpener URLOpener) (*internal.TokenInfo,
 	waitTimeout := time.Duration(flowInfo.ExpiresIn)
 	waitCTX, cancel := context.WithTimeout(a.ctx, waitTimeout*time.Second)
 	defer cancel()
-	tokenInfo, err := hostedClient.WaitToken(waitCTX, flowInfo)
+	tokenInfo, err := oAuthFlow.WaitToken(waitCTX, flowInfo)
 	if err != nil {
 		return nil, fmt.Errorf("waiting for browser login failed: %v", err)
 	}

client/android/peer_notifier.go

@@ -5,7 +5,6 @@ type PeerInfo struct {
 	IP         string
 	FQDN       string
 	ConnStatus string // Todo replace to enum
-	Direct     bool
 }
 
 // PeerInfoCollection made for Java layer to get non default types as collection

client/cmd/login.go

@@ -3,6 +3,8 @@ package cmd
 import (
 	"context"
 	"fmt"
+	"github.com/netbirdio/netbird/client/internal/auth"
+	"strings"
 	"time"
 
 	"github.com/skratchdot/open-golang/open"
@@ -15,6 +17,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/internal"
 	"github.com/netbirdio/netbird/client/proto"
+	"github.com/netbirdio/netbird/client/system"
 )
 
 var loginCmd = &cobra.Command{
@@ -32,6 +35,11 @@ var loginCmd = &cobra.Command{
 
 		ctx := internal.CtxInitState(context.Background())
 
+		if hostName != "" {
+			// nolint
+			ctx = context.WithValue(ctx, system.DeviceNameCtxKey, hostName)
+		}
+
 		// workaround to run without service
 		if logFile == "console" {
 			err = handleRebrand(cmd)
@@ -39,12 +47,16 @@ var loginCmd = &cobra.Command{
 				return err
 			}
 
-			config, err := internal.UpdateOrCreateConfig(internal.ConfigInput{
+			ic := internal.ConfigInput{
 				ManagementURL: managementURL,
 				AdminURL:      adminURL,
 				ConfigPath:    configPath,
-				PreSharedKey:  &preSharedKey,
-			})
+			}
+			if preSharedKey != "" {
+				ic.PreSharedKey = &preSharedKey
+			}
+
+			config, err := internal.UpdateOrCreateConfig(ic)
 			if err != nil {
 				return fmt.Errorf("get config file: %v", err)
 			}
@@ -100,7 +112,7 @@ var loginCmd = &cobra.Command{
 		}
 
 		if loginResp.NeedsSSOLogin {
-			openURL(cmd, loginResp.VerificationURIComplete)
+			openURL(cmd, loginResp.VerificationURIComplete, loginResp.UserCode)
 
 			_, err = client.WaitSSOLogin(ctx, &proto.WaitSSOLoginRequest{UserCode: loginResp.UserCode})
 			if err != nil {
@@ -152,40 +164,24 @@ func foregroundLogin(ctx context.Context, cmd *cobra.Command, config *internal.C
 	return nil
 }
 
-func foregroundGetTokenInfo(ctx context.Context, cmd *cobra.Command, config *internal.Config) (*internal.TokenInfo, error) {
-	providerConfig, err := internal.GetDeviceAuthorizationFlowInfo(ctx, config.PrivateKey, config.ManagementURL)
+func foregroundGetTokenInfo(ctx context.Context, cmd *cobra.Command, config *internal.Config) (*auth.TokenInfo, error) {
+	oAuthFlow, err := auth.NewOAuthFlow(ctx, config)
 	if err != nil {
-		s, ok := gstatus.FromError(err)
-		if ok && s.Code() == codes.NotFound {
-			return nil, fmt.Errorf("no SSO provider returned from management. " +
-				"If you are using hosting Netbird see documentation at " +
-				"https://github.com/netbirdio/netbird/tree/main/management for details")
-		} else if ok && s.Code() == codes.Unimplemented {
-			mgmtURL := managementURL
-			if mgmtURL == "" {
-				mgmtURL = internal.DefaultManagementURL
-			}
-			return nil, fmt.Errorf("the management server, %s, does not support SSO providers, "+
-				"please update your servver or use Setup Keys to login", mgmtURL)
-		} else {
-			return nil, fmt.Errorf("getting device authorization flow info failed with error: %v", err)
-		}
+		return nil, err
 	}
 
-	hostedClient := internal.NewHostedDeviceFlow(providerConfig.ProviderConfig)
-
-	flowInfo, err := hostedClient.RequestDeviceCode(context.TODO())
+	flowInfo, err := oAuthFlow.RequestAuthInfo(context.TODO())
 	if err != nil {
-		return nil, fmt.Errorf("getting a request device code failed: %v", err)
+		return nil, fmt.Errorf("getting a request OAuth flow info failed: %v", err)
 	}
 
-	openURL(cmd, flowInfo.VerificationURIComplete)
+	openURL(cmd, flowInfo.VerificationURIComplete, flowInfo.UserCode)
 
 	waitTimeout := time.Duration(flowInfo.ExpiresIn)
 	waitCTX, c := context.WithTimeout(context.TODO(), waitTimeout*time.Second)
 	defer c()
 
-	tokenInfo, err := hostedClient.WaitToken(waitCTX, flowInfo)
+	tokenInfo, err := oAuthFlow.WaitToken(waitCTX, flowInfo)
 	if err != nil {
 		return nil, fmt.Errorf("waiting for browser login failed: %v", err)
 	}
@@ -193,11 +189,18 @@ func foregroundGetTokenInfo(ctx context.Context, cmd *cobra.Command, config *int
 	return &tokenInfo, nil
 }
 
-func openURL(cmd *cobra.Command, verificationURIComplete string) {
+func openURL(cmd *cobra.Command, verificationURIComplete, userCode string) {
+	var codeMsg string
+	if userCode != "" {
+		if !strings.Contains(verificationURIComplete, userCode) {
+			codeMsg = fmt.Sprintf("and enter the code %s to authenticate.", userCode)
+		}
+	}
+
 	err := open.Run(verificationURIComplete)
 	cmd.Printf("Please do the SSO login in your browser. \n" +
 		"If your browser didn't open automatically, use this URL to log in:\n\n" +
-		" " + verificationURIComplete + " \n\n")
+		" " + verificationURIComplete + " " + codeMsg + " \n\n")
 	if err != nil {
 		cmd.Printf("Alternatively, you may want to use a setup key, see:\n\n https://www.netbird.io/docs/overview/setup-keys\n")
 	}

client/cmd/root.go

@@ -45,6 +45,7 @@ var (
 	managementURL           string
 	adminURL                string
 	setupKey                string
+	hostName                string
 	preSharedKey            string
 	natExternalIPs          []string
 	customDNSAddress        string
@@ -94,6 +95,7 @@ func init() {
 	rootCmd.PersistentFlags().StringVar(&logFile, "log-file", defaultLogFile, "sets Netbird log path. If console is specified the the log will be output to stdout")
 	rootCmd.PersistentFlags().StringVarP(&setupKey, "setup-key", "k", "", "Setup key obtained from the Management Service Dashboard (used to register peer)")
 	rootCmd.PersistentFlags().StringVar(&preSharedKey, "preshared-key", "", "Sets Wireguard PreSharedKey property. If set, then only peers that have the same key can communicate.")
+	rootCmd.PersistentFlags().StringVarP(&hostName, "hostname", "n", "", "Sets a custom hostname for the device")
 	rootCmd.AddCommand(serviceCmd)
 	rootCmd.AddCommand(upCmd)
 	rootCmd.AddCommand(downCmd)

client/cmd/ssh.go

@@ -73,7 +73,8 @@ var sshCmd = &cobra.Command{
 		go func() {
 			// blocking
 			if err := runSSH(sshctx, host, []byte(config.SSHKey), cmd); err != nil {
-				log.Print(err)
+				log.Debug(err)
+				os.Exit(1)
 			}
 			cancel()
 		}()
@@ -92,12 +93,10 @@ func runSSH(ctx context.Context, addr string, pemKey []byte, cmd *cobra.Command)
 	c, err := nbssh.DialWithKey(fmt.Sprintf("%s:%d", addr, port), user, pemKey)
 	if err != nil {
 		cmd.Printf("Error: %v\n", err)
-		cmd.Printf("Couldn't connect. " +
-			"You might be disconnected from the NetBird network, or the NetBird agent isn't running.\n" +
-			"Run the status command: \n\n" +
-			" netbird status\n\n" +
-			"It might also be that the SSH server is disabled on the agent you are trying to connect to.\n")
-		return nil
+		cmd.Printf("Couldn't connect. Please check the connection status or if the ssh server is enabled on the other peer" +
+			"You can verify the connection by running:\n\n" +
+			" netbird status\n\n")
+		return err
 	}
 	go func() {
 		<-ctx.Done()

client/cmd/status.go

@@ -249,7 +249,7 @@ func mapPeers(peers []*proto.PeerState) peersStateOutput {
 			IP:               pbPeerState.GetIP(),
 			PubKey:           pbPeerState.GetPubKey(),
 			Status:           pbPeerState.GetConnStatus(),
-			LastStatusUpdate: timeLocal.UTC(),
+			LastStatusUpdate: timeLocal,
 			ConnType:         connType,
 			Direct:           pbPeerState.GetDirect(),
 			IceCandidateType: iceCandidateType{

client/cmd/status_test.go

@@ -11,6 +11,15 @@ import (
 	"github.com/netbirdio/netbird/version"
 )
 
+func init() {
+	loc, err := time.LoadLocation("UTC")
+	if err != nil {
+		panic(err)
+	}
+
+	time.Local = loc
+}
+
 var resp = &proto.StatusResponse{
 	Status: "Connected",
 	FullStatus: &proto.FullStatus{

client/cmd/testutil.go

@@ -2,21 +2,23 @@ package cmd
 
 import (
 	"context"
-	"github.com/netbirdio/netbird/management/server/activity"
 	"net"
 	"path/filepath"
 	"testing"
 	"time"
 
+	"github.com/netbirdio/netbird/management/server/activity"
+
 	"github.com/netbirdio/netbird/util"
 
+	"google.golang.org/grpc"
+
 	clientProto "github.com/netbirdio/netbird/client/proto"
 	client "github.com/netbirdio/netbird/client/server"
 	mgmtProto "github.com/netbirdio/netbird/management/proto"
 	mgmt "github.com/netbirdio/netbird/management/server"
 	sigProto "github.com/netbirdio/netbird/signal/proto"
 	sig "github.com/netbirdio/netbird/signal/server"
-	"google.golang.org/grpc"
 )
 
 func startTestingServices(t *testing.T) string {
@@ -63,7 +65,7 @@ func startManagement(t *testing.T, config *mgmt.Config) (*grpc.Server, net.Liste
 		t.Fatal(err)
 	}
 	s := grpc.NewServer()
-	store, err := mgmt.NewFileStore(config.Datadir)
+	store, err := mgmt.NewFileStore(config.Datadir, nil)
 	if err != nil {
 		t.Fatal(err)
 	}

client/cmd/up.go

@@ -15,6 +15,7 @@ import (
 	"github.com/netbirdio/netbird/client/internal"
 	"github.com/netbirdio/netbird/client/internal/peer"
 	"github.com/netbirdio/netbird/client/proto"
+	"github.com/netbirdio/netbird/client/system"
 	"github.com/netbirdio/netbird/util"
 )
 
@@ -55,6 +56,11 @@ func upFunc(cmd *cobra.Command, args []string) error {
 
 	ctx := internal.CtxInitState(cmd.Context())
 
+	if hostName != "" {
+		// nolint
+		ctx = context.WithValue(ctx, system.DeviceNameCtxKey, hostName)
+	}
+
 	if foregroundMode {
 		return runInForegroundMode(ctx, cmd)
 	}
@@ -72,14 +78,18 @@ func runInForegroundMode(ctx context.Context, cmd *cobra.Command) error {
 		return err
 	}
 
-	config, err := internal.UpdateOrCreateConfig(internal.ConfigInput{
+	ic := internal.ConfigInput{
 		ManagementURL:    managementURL,
 		AdminURL:         adminURL,
 		ConfigPath:       configPath,
-		PreSharedKey:     &preSharedKey,
 		NATExternalIPs:   natExternalIPs,
 		CustomDNSAddress: customDNSAddressConverted,
-	})
+	}
+	if preSharedKey != "" {
+		ic.PreSharedKey = &preSharedKey
+	}
+
+	config, err := internal.UpdateOrCreateConfig(ic)
 	if err != nil {
 		return fmt.Errorf("get config file: %v", err)
 	}
@@ -94,7 +104,7 @@ func runInForegroundMode(ctx context.Context, cmd *cobra.Command) error {
 	var cancel context.CancelFunc
 	ctx, cancel = context.WithCancel(ctx)
 	SetupCloseHandler(ctx, cancel)
-	return internal.RunClient(ctx, config, peer.NewRecorder(config.ManagementURL.String()), nil, nil)
+	return internal.RunClient(ctx, config, peer.NewRecorder(config.ManagementURL.String()))
 }
 
 func runInDaemonMode(ctx context.Context, cmd *cobra.Command) error {
@@ -166,7 +176,7 @@ func runInDaemonMode(ctx context.Context, cmd *cobra.Command) error {
 
 	if loginResp.NeedsSSOLogin {
 
-		openURL(cmd, loginResp.VerificationURIComplete)
+		openURL(cmd, loginResp.VerificationURIComplete, loginResp.UserCode)
 
 		_, err = client.WaitSSOLogin(ctx, &proto.WaitSSOLoginRequest{UserCode: loginResp.UserCode})
 		if err != nil {

client/firewall/firewall.go

@@ -13,22 +13,24 @@ type Rule interface {
 	GetRuleID() string
 }
 
-// Direction is the direction of the traffic
-type Direction int
+// RuleDirection is the traffic direction which a rule is applied
+type RuleDirection int
 
 const (
-	// DirectionSrc is the direction of the traffic from the source
-	DirectionSrc Direction = iota
-	// DirectionDst is the direction of the traffic from the destination
-	DirectionDst
+	// RuleDirectionIN applies to filters that handlers incoming traffic
+	RuleDirectionIN RuleDirection = iota
+	// RuleDirectionOUT applies to filters that handlers outgoing traffic
+	RuleDirectionOUT
 )
 
 // Action is the action to be taken on a rule
 type Action int
 
 const (
+	// ActionUnknown is a unknown action
+	ActionUnknown Action = iota
 	// ActionAccept is the action to accept a packet
-	ActionAccept Action = iota
+	ActionAccept
 	// ActionDrop is the action to drop a packet
 	ActionDrop
 )
@@ -39,11 +41,17 @@ const (
 // Netbird client for ACL and routing functionality
 type Manager interface {
 	// AddFiltering rule to the firewall
+	//
+	// If comment argument is empty firewall manager should set
+	// rule ID as comment for the rule
 	AddFiltering(
 		ip net.IP,
-		port *Port,
-		direction Direction,
+		proto Protocol,
+		sPort *Port,
+		dPort *Port,
+		direction RuleDirection,
 		action Action,
+		ipsetName string,
 		comment string,
 	) (Rule, error)
 
@@ -53,5 +61,8 @@ type Manager interface {
 	// Reset firewall to the default state
 	Reset() error
 
+	// Flush the changes to firewall controller
+	Flush() error
+
 	// TODO: migrate routemanager firewal actions to this interface
 }

client/firewall/iptables/manager_linux.go

@@ -8,83 +8,205 @@ import (
 
 	"github.com/coreos/go-iptables/iptables"
 	"github.com/google/uuid"
+	"github.com/nadoo/ipset"
+	log "github.com/sirupsen/logrus"
 
 	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
 )
 
 const (
-	// ChainFilterName is the name of the chain that is used for filtering by the Netbird client
-	ChainFilterName = "NETBIRD-ACL"
+	// ChainInputFilterName is the name of the chain that is used for filtering incoming packets
+	ChainInputFilterName = "NETBIRD-ACL-INPUT"
+
+	// ChainOutputFilterName is the name of the chain that is used for filtering outgoing packets
+	ChainOutputFilterName = "NETBIRD-ACL-OUTPUT"
 )
 
+// dropAllDefaultRule in the Netbird chain
+var dropAllDefaultRule = []string{"-j", "DROP"}
+
 // Manager of iptables firewall
 type Manager struct {
 	mutex sync.Mutex
 
 	ipv4Client *iptables.IPTables
 	ipv6Client *iptables.IPTables
+
+	inputDefaultRuleSpecs  []string
+	outputDefaultRuleSpecs []string
+	wgIface                iFaceMapper
+
+	rulesets map[string]ruleset
+}
+
+// iFaceMapper defines subset methods of interface required for manager
+type iFaceMapper interface {
+	Name() string
+	Address() iface.WGAddress
+}
+
+type ruleset struct {
+	rule *Rule
+	ips  map[string]string
 }
 
 // Create iptables firewall manager
-func Create() (*Manager, error) {
-	m := &Manager{}
+func Create(wgIface iFaceMapper) (*Manager, error) {
+	m := &Manager{
+		wgIface: wgIface,
+		inputDefaultRuleSpecs: []string{
+			"-i", wgIface.Name(), "-j", ChainInputFilterName, "-s", wgIface.Address().String()},
+		outputDefaultRuleSpecs: []string{
+			"-o", wgIface.Name(), "-j", ChainOutputFilterName, "-d", wgIface.Address().String()},
+		rulesets: make(map[string]ruleset),
+	}
+
+	if err := ipset.Init(); err != nil {
+		return nil, fmt.Errorf("init ipset: %w", err)
+	}
 
 	// init clients for booth ipv4 and ipv6
 	ipv4Client, err := iptables.NewWithProtocol(iptables.ProtocolIPv4)
 	if err != nil {
 		return nil, fmt.Errorf("iptables is not installed in the system or not supported")
 	}
-	m.ipv4Client = ipv4Client
+	if isIptablesClientAvailable(ipv4Client) {
+		m.ipv4Client = ipv4Client
+	}
 
 	ipv6Client, err := iptables.NewWithProtocol(iptables.ProtocolIPv6)
 	if err != nil {
-		return nil, fmt.Errorf("ip6tables is not installed in the system or not supported")
+		log.Errorf("ip6tables is not installed in the system or not supported: %v", err)
+	} else {
+		if isIptablesClientAvailable(ipv6Client) {
+			m.ipv6Client = ipv6Client
+		}
 	}
-	m.ipv6Client = ipv6Client
 
 	if err := m.Reset(); err != nil {
-		return nil, fmt.Errorf("failed to reset firewall: %s", err)
+		return nil, fmt.Errorf("failed to reset firewall: %v", err)
 	}
-
 	return m, nil
 }
 
+func isIptablesClientAvailable(client *iptables.IPTables) bool {
+	_, err := client.ListChains("filter")
+	return err == nil
+}
+
 // AddFiltering rule to the firewall
+//
+// If comment is empty rule ID is used as comment
 func (m *Manager) AddFiltering(
 	ip net.IP,
-	port *fw.Port,
-	direction fw.Direction,
+	protocol fw.Protocol,
+	sPort *fw.Port,
+	dPort *fw.Port,
+	direction fw.RuleDirection,
 	action fw.Action,
+	ipsetName string,
 	comment string,
 ) (fw.Rule, error) {
 	m.mutex.Lock()
 	defer m.mutex.Unlock()
-	client := m.client(ip)
-	ok, err := client.ChainExists("filter", ChainFilterName)
+
+	client, err := m.client(ip)
 	if err != nil {
-		return nil, fmt.Errorf("failed to check if chain exists: %s", err)
-	}
-	if !ok {
-		if err := client.NewChain("filter", ChainFilterName); err != nil {
-			return nil, fmt.Errorf("failed to create chain: %s", err)
-		}
-	}
-	if port == nil || port.Values == nil || (port.IsRange && len(port.Values) != 2) {
-		return nil, fmt.Errorf("invalid port definition")
-	}
-	pv := strconv.Itoa(port.Values[0])
-	if port.IsRange {
-		pv += ":" + strconv.Itoa(port.Values[1])
-	}
-	specs := m.filterRuleSpecs("filter", ChainFilterName, ip, pv, direction, action, comment)
-	if err := client.AppendUnique("filter", ChainFilterName, specs...); err != nil {
 		return nil, err
 	}
-	rule := &Rule{
-		id:    uuid.New().String(),
-		specs: specs,
-		v6:    ip.To4() == nil,
+
+	var dPortVal, sPortVal string
+	if dPort != nil && dPort.Values != nil {
+		// TODO: we support only one port per rule in current implementation of ACLs
+		dPortVal = strconv.Itoa(dPort.Values[0])
 	}
+	if sPort != nil && sPort.Values != nil {
+		sPortVal = strconv.Itoa(sPort.Values[0])
+	}
+	ipsetName = m.transformIPsetName(ipsetName, sPortVal, dPortVal)
+
+	ruleID := uuid.New().String()
+	if comment == "" {
+		comment = ruleID
+	}
+
+	if ipsetName != "" {
+		rs, rsExists := m.rulesets[ipsetName]
+		if !rsExists {
+			if err := ipset.Flush(ipsetName); err != nil {
+				log.Errorf("flush ipset %q before use it: %v", ipsetName, err)
+			}
+			if err := ipset.Create(ipsetName); err != nil {
+				return nil, fmt.Errorf("failed to create ipset: %w", err)
+			}
+		}
+
+		if err := ipset.Add(ipsetName, ip.String()); err != nil {
+			return nil, fmt.Errorf("failed to add IP to ipset: %w", err)
+		}
+
+		if rsExists {
+			// if ruleset already exists it means we already have the firewall rule
+			// so we need to update IPs in the ruleset and return new fw.Rule object for ACL manager.
+			rs.ips[ip.String()] = ruleID
+			return &Rule{
+				ruleID:    ruleID,
+				ipsetName: ipsetName,
+				ip:        ip.String(),
+				dst:       direction == fw.RuleDirectionOUT,
+				v6:        ip.To4() == nil,
+			}, nil
+		}
+		// this is new ipset so we need to create firewall rule for it
+	}
+
+	specs := m.filterRuleSpecs("filter", ip, string(protocol), sPortVal, dPortVal,
+		direction, action, comment, ipsetName)
+
+	if direction == fw.RuleDirectionOUT {
+		ok, err := client.Exists("filter", ChainOutputFilterName, specs...)
+		if err != nil {
+			return nil, fmt.Errorf("check is output rule already exists: %w", err)
+		}
+		if ok {
+			return nil, fmt.Errorf("input rule already exists")
+		}
+
+		if err := client.Insert("filter", ChainOutputFilterName, 1, specs...); err != nil {
+			return nil, err
+		}
+	} else {
+		ok, err := client.Exists("filter", ChainInputFilterName, specs...)
+		if err != nil {
+			return nil, fmt.Errorf("check is input rule already exists: %w", err)
+		}
+		if ok {
+			return nil, fmt.Errorf("input rule already exists")
+		}
+
+		if err := client.Insert("filter", ChainInputFilterName, 1, specs...); err != nil {
+			return nil, err
+		}
+	}
+
+	rule := &Rule{
+		ruleID:    ruleID,
+		specs:     specs,
+		ipsetName: ipsetName,
+		ip:        ip.String(),
+		dst:       direction == fw.RuleDirectionOUT,
+		v6:        ip.To4() == nil,
+	}
+	if ipsetName != "" {
+		// ipset name is defined and it means that this rule was created
+		// for it, need to assosiate it with ruleset
+		m.rulesets[ipsetName] = ruleset{
+			rule: rule,
+			ips:  map[string]string{rule.ip: ruleID},
+		}
+	}
+
 	return rule, nil
 }
 
@@ -92,64 +214,224 @@ func (m *Manager) AddFiltering(
 func (m *Manager) DeleteRule(rule fw.Rule) error {
 	m.mutex.Lock()
 	defer m.mutex.Unlock()
+
 	r, ok := rule.(*Rule)
 	if !ok {
 		return fmt.Errorf("invalid rule type")
 	}
+
 	client := m.ipv4Client
 	if r.v6 {
+		if m.ipv6Client == nil {
+			return fmt.Errorf("ipv6 is not supported")
+		}
 		client = m.ipv6Client
 	}
-	return client.Delete("filter", ChainFilterName, r.specs...)
+
+	if rs, ok := m.rulesets[r.ipsetName]; ok {
+		// delete IP from ruleset IPs list and ipset
+		if _, ok := rs.ips[r.ip]; ok {
+			if err := ipset.Del(r.ipsetName, r.ip); err != nil {
+				return fmt.Errorf("failed to delete ip from ipset: %w", err)
+			}
+			delete(rs.ips, r.ip)
+		}
+
+		// if after delete, set still contains other IPs,
+		// no need to delete firewall rule and we should exit here
+		if len(rs.ips) != 0 {
+			return nil
+		}
+
+		// we delete last IP from the set, that means we need to delete
+		// set itself and assosiated firewall rule too
+		delete(m.rulesets, r.ipsetName)
+
+		if err := ipset.Destroy(r.ipsetName); err != nil {
+			log.Errorf("delete empty ipset: %v", err)
+		}
+		r = rs.rule
+	}
+
+	if r.dst {
+		return client.Delete("filter", ChainOutputFilterName, r.specs...)
+	}
+	return client.Delete("filter", ChainInputFilterName, r.specs...)
 }
 
 // Reset firewall to the default state
 func (m *Manager) Reset() error {
 	m.mutex.Lock()
 	defer m.mutex.Unlock()
-	if err := m.reset(m.ipv4Client, "filter", ChainFilterName); err != nil {
-		return fmt.Errorf("clean ipv4 firewall ACL chain: %w", err)
+
+	if err := m.reset(m.ipv4Client, "filter"); err != nil {
+		return fmt.Errorf("clean ipv4 firewall ACL input chain: %w", err)
 	}
-	if err := m.reset(m.ipv6Client, "filter", ChainFilterName); err != nil {
-		return fmt.Errorf("clean ipv6 firewall ACL chain: %w", err)
+	if m.ipv6Client != nil {
+		if err := m.reset(m.ipv6Client, "filter"); err != nil {
+			return fmt.Errorf("clean ipv6 firewall ACL input chain: %w", err)
+		}
 	}
+
 	return nil
 }
 
+// Flush doesn't need to be implemented for this manager
+func (m *Manager) Flush() error { return nil }
+
 // reset firewall chain, clear it and drop it
-func (m *Manager) reset(client *iptables.IPTables, table, chain string) error {
-	ok, err := client.ChainExists(table, chain)
+func (m *Manager) reset(client *iptables.IPTables, table string) error {
+	ok, err := client.ChainExists(table, ChainInputFilterName)
 	if err != nil {
-		return fmt.Errorf("failed to check if chain exists: %w", err)
+		return fmt.Errorf("failed to check if input chain exists: %w", err)
 	}
-	if !ok {
+	if ok {
+		if ok, err := client.Exists("filter", "INPUT", m.inputDefaultRuleSpecs...); err != nil {
+			return err
+		} else if ok {
+			if err := client.Delete("filter", "INPUT", m.inputDefaultRuleSpecs...); err != nil {
+				log.WithError(err).Errorf("failed to delete default input rule: %v", err)
+			}
+		}
+	}
+
+	ok, err = client.ChainExists(table, ChainOutputFilterName)
+	if err != nil {
+		return fmt.Errorf("failed to check if output chain exists: %w", err)
+	}
+	if ok {
+		if ok, err := client.Exists("filter", "OUTPUT", m.outputDefaultRuleSpecs...); err != nil {
+			return err
+		} else if ok {
+			if err := client.Delete("filter", "OUTPUT", m.outputDefaultRuleSpecs...); err != nil {
+				log.WithError(err).Errorf("failed to delete default output rule: %v", err)
+			}
+		}
+	}
+
+	if err := client.ClearAndDeleteChain(table, ChainInputFilterName); err != nil {
+		log.Errorf("failed to clear and delete input chain: %v", err)
 		return nil
 	}
-	if err := client.ClearChain(table, ChainFilterName); err != nil {
-		return fmt.Errorf("failed to clear chain: %w", err)
+
+	if err := client.ClearAndDeleteChain(table, ChainOutputFilterName); err != nil {
+		log.Errorf("failed to clear and delete input chain: %v", err)
+		return nil
 	}
-	return client.DeleteChain(table, ChainFilterName)
+
+	for ipsetName := range m.rulesets {
+		if err := ipset.Flush(ipsetName); err != nil {
+			log.Errorf("flush ipset %q during reset: %v", ipsetName, err)
+		}
+		if err := ipset.Destroy(ipsetName); err != nil {
+			log.Errorf("delete ipset %q during reset: %v", ipsetName, err)
+		}
+		delete(m.rulesets, ipsetName)
+	}
+
+	return nil
 }
 
 // filterRuleSpecs returns the specs of a filtering rule
 func (m *Manager) filterRuleSpecs(
-	table string, chain string, ip net.IP, port string,
-	direction fw.Direction, action fw.Action, comment string,
+	table string, ip net.IP, protocol string, sPort, dPort string,
+	direction fw.RuleDirection, action fw.Action, comment string,
+	ipsetName string,
 ) (specs []string) {
-	if direction == fw.DirectionSrc {
-		specs = append(specs, "-s", ip.String())
+	matchByIP := true
+	// don't use IP matching if IP is ip 0.0.0.0
+	if s := ip.String(); s == "0.0.0.0" || s == "::" {
+		matchByIP = false
+	}
+	switch direction {
+	case fw.RuleDirectionIN:
+		if matchByIP {
+			if ipsetName != "" {
+				specs = append(specs, "-m", "set", "--set", ipsetName, "src")
+			} else {
+				specs = append(specs, "-s", ip.String())
+			}
+		}
+	case fw.RuleDirectionOUT:
+		if matchByIP {
+			if ipsetName != "" {
+				specs = append(specs, "-m", "set", "--set", ipsetName, "dst")
+			} else {
+				specs = append(specs, "-d", ip.String())
+			}
+		}
+	}
+	if protocol != "all" {
+		specs = append(specs, "-p", protocol)
+	}
+	if sPort != "" {
+		specs = append(specs, "--sport", sPort)
+	}
+	if dPort != "" {
+		specs = append(specs, "--dport", dPort)
 	}
-	specs = append(specs, "-p", "tcp", "--dport", port)
 	specs = append(specs, "-j", m.actionToStr(action))
 	return append(specs, "-m", "comment", "--comment", comment)
 }
 
-// client returns corresponding iptables client for the given ip
-func (m *Manager) client(ip net.IP) *iptables.IPTables {
+// rawClient returns corresponding iptables client for the given ip
+func (m *Manager) rawClient(ip net.IP) (*iptables.IPTables, error) {
 	if ip.To4() != nil {
-		return m.ipv4Client
+		return m.ipv4Client, nil
 	}
-	return m.ipv6Client
+	if m.ipv6Client == nil {
+		return nil, fmt.Errorf("ipv6 is not supported")
+	}
+	return m.ipv6Client, nil
+}
+
+// client returns client with initialized chain and default rules
+func (m *Manager) client(ip net.IP) (*iptables.IPTables, error) {
+	client, err := m.rawClient(ip)
+	if err != nil {
+		return nil, err
+	}
+
+	ok, err := client.ChainExists("filter", ChainInputFilterName)
+	if err != nil {
+		return nil, fmt.Errorf("failed to check if chain exists: %w", err)
+	}
+
+	if !ok {
+		if err := client.NewChain("filter", ChainInputFilterName); err != nil {
+			return nil, fmt.Errorf("failed to create input chain: %w", err)
+		}
+
+		if err := client.AppendUnique("filter", ChainInputFilterName, dropAllDefaultRule...); err != nil {
+			return nil, fmt.Errorf("failed to create default drop all in netbird input chain: %w", err)
+		}
+
+		if err := client.AppendUnique("filter", "INPUT", m.inputDefaultRuleSpecs...); err != nil {
+			return nil, fmt.Errorf("failed to create input chain jump rule: %w", err)
+		}
+
+	}
+
+	ok, err = client.ChainExists("filter", ChainOutputFilterName)
+	if err != nil {
+		return nil, fmt.Errorf("failed to check if chain exists: %w", err)
+	}
+
+	if !ok {
+		if err := client.NewChain("filter", ChainOutputFilterName); err != nil {
+			return nil, fmt.Errorf("failed to create output chain: %w", err)
+		}
+
+		if err := client.AppendUnique("filter", ChainOutputFilterName, dropAllDefaultRule...); err != nil {
+			return nil, fmt.Errorf("failed to create default drop all in netbird output chain: %w", err)
+		}
+
+		if err := client.AppendUnique("filter", "OUTPUT", m.outputDefaultRuleSpecs...); err != nil {
+			return nil, fmt.Errorf("failed to create output chain jump rule: %w", err)
+		}
+	}
+
+	return client, nil
 }
 
 func (m *Manager) actionToStr(action fw.Action) string {
@@ -158,3 +440,16 @@ func (m *Manager) actionToStr(action fw.Action) string {
 	}
 	return "DROP"
 }
+
+func (m *Manager) transformIPsetName(ipsetName string, sPort, dPort string) string {
+	if ipsetName == "" {
+		return ""
+	} else if sPort != "" && dPort != "" {
+		return ipsetName + "-sport-dport"
+	} else if sPort != "" {
+		return ipsetName + "-sport"
+	} else if dPort != "" {
+		return ipsetName + "-dport"
+	}
+	return ipsetName
+}

client/firewall/iptables/manager_linux_test.go

@@ -1,105 +1,261 @@
 package iptables
 
 import (
+	"fmt"
 	"net"
 	"testing"
+	"time"
 
 	"github.com/coreos/go-iptables/iptables"
+	"github.com/stretchr/testify/require"
+
 	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
 )
 
-func TestNewManager(t *testing.T) {
+// iFaceMapper defines subset methods of interface required for manager
+type iFaceMock struct {
+	NameFunc    func() string
+	AddressFunc func() iface.WGAddress
+}
+
+func (i *iFaceMock) Name() string {
+	if i.NameFunc != nil {
+		return i.NameFunc()
+	}
+	panic("NameFunc is not set")
+}
+
+func (i *iFaceMock) Address() iface.WGAddress {
+	if i.AddressFunc != nil {
+		return i.AddressFunc()
+	}
+	panic("AddressFunc is not set")
+}
+
+func TestIptablesManager(t *testing.T) {
 	ipv4Client, err := iptables.NewWithProtocol(iptables.ProtocolIPv4)
-	if err != nil {
-		t.Fatal(err)
+	require.NoError(t, err)
+
+	mock := &iFaceMock{
+		NameFunc: func() string {
+			return "lo"
+		},
+		AddressFunc: func() iface.WGAddress {
+			return iface.WGAddress{
+				IP: net.ParseIP("10.20.0.1"),
+				Network: &net.IPNet{
+					IP:   net.ParseIP("10.20.0.0"),
+					Mask: net.IPv4Mask(255, 255, 255, 0),
+				},
+			}
+		},
 	}
 
-	manager, err := Create()
-	if err != nil {
-		t.Fatal(err)
-	}
+	// just check on the local interface
+	manager, err := Create(mock)
+	require.NoError(t, err)
+
+	time.Sleep(time.Second)
+
+	defer func() {
+		err := manager.Reset()
+		require.NoError(t, err, "clear the manager state")
+
+		time.Sleep(time.Second)
+	}()
 
 	var rule1 fw.Rule
 	t.Run("add first rule", func(t *testing.T) {
 		ip := net.ParseIP("10.20.0.2")
-		port := &fw.Port{Proto: fw.PortProtocolTCP, Values: []int{8080}}
-		rule1, err = manager.AddFiltering(ip, port, fw.DirectionDst, fw.ActionAccept, "accept HTTP traffic")
-		if err != nil {
-			t.Errorf("failed to add rule: %v", err)
-		}
+		port := &fw.Port{Values: []int{8080}}
+		rule1, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionOUT, fw.ActionAccept, "", "accept HTTP traffic")
+		require.NoError(t, err, "failed to add rule")
 
-		checkRuleSpecs(t, ipv4Client, true, rule1.(*Rule).specs...)
+		checkRuleSpecs(t, ipv4Client, ChainOutputFilterName, true, rule1.(*Rule).specs...)
 	})
 
 	var rule2 fw.Rule
 	t.Run("add second rule", func(t *testing.T) {
 		ip := net.ParseIP("10.20.0.3")
 		port := &fw.Port{
-			Proto:  fw.PortProtocolTCP,
 			Values: []int{8043: 8046},
 		}
 		rule2, err = manager.AddFiltering(
-			ip, port, fw.DirectionDst, fw.ActionAccept, "accept HTTPS traffic from ports range")
-		if err != nil {
-			t.Errorf("failed to add rule: %v", err)
-		}
+			ip, "tcp", port, nil, fw.RuleDirectionIN, fw.ActionAccept, "", "accept HTTPS traffic from ports range")
+		require.NoError(t, err, "failed to add rule")
 
-		checkRuleSpecs(t, ipv4Client, true, rule2.(*Rule).specs...)
+		checkRuleSpecs(t, ipv4Client, ChainInputFilterName, true, rule2.(*Rule).specs...)
 	})
 
 	t.Run("delete first rule", func(t *testing.T) {
-		if err := manager.DeleteRule(rule1); err != nil {
-			t.Errorf("failed to delete rule: %v", err)
-		}
+		err := manager.DeleteRule(rule1)
+		require.NoError(t, err, "failed to delete rule")
 
-		checkRuleSpecs(t, ipv4Client, false, rule1.(*Rule).specs...)
+		checkRuleSpecs(t, ipv4Client, ChainOutputFilterName, false, rule1.(*Rule).specs...)
 	})
 
 	t.Run("delete second rule", func(t *testing.T) {
-		if err := manager.DeleteRule(rule2); err != nil {
-			t.Errorf("failed to delete rule: %v", err)
-		}
+		err := manager.DeleteRule(rule2)
+		require.NoError(t, err, "failed to delete rule")
 
-		checkRuleSpecs(t, ipv4Client, false, rule2.(*Rule).specs...)
+		require.Empty(t, manager.rulesets, "rulesets index after removed second rule must be empty")
 	})
 
 	t.Run("reset check", func(t *testing.T) {
 		// add second rule
 		ip := net.ParseIP("10.20.0.3")
-		port := &fw.Port{Proto: fw.PortProtocolUDP, Values: []int{5353}}
-		_, err = manager.AddFiltering(ip, port, fw.DirectionDst, fw.ActionAccept, "accept Fake DNS traffic")
-		if err != nil {
-			t.Errorf("failed to add rule: %v", err)
-		}
+		port := &fw.Port{Values: []int{5353}}
+		_, err = manager.AddFiltering(ip, "udp", nil, port, fw.RuleDirectionOUT, fw.ActionAccept, "", "accept Fake DNS traffic")
+		require.NoError(t, err, "failed to add rule")
 
-		if err := manager.Reset(); err != nil {
-			t.Errorf("failed to reset: %v", err)
-		}
+		err = manager.Reset()
+		require.NoError(t, err, "failed to reset")
 
-		ok, err := ipv4Client.ChainExists("filter", ChainFilterName)
-		if err != nil {
-			t.Errorf("failed to drop chain: %v", err)
-		}
+		ok, err := ipv4Client.ChainExists("filter", ChainInputFilterName)
+		require.NoError(t, err, "failed check chain exists")
 
 		if ok {
-			t.Errorf("chain '%v' still exists after Reset", ChainFilterName)
+			require.NoErrorf(t, err, "chain '%v' still exists after Reset", ChainInputFilterName)
 		}
 	})
 }
 
-func checkRuleSpecs(t *testing.T, ipv4Client *iptables.IPTables, mustExists bool, rulespec ...string) {
-	exists, err := ipv4Client.Exists("filter", ChainFilterName, rulespec...)
-	if err != nil {
-		t.Errorf("failed to check rule: %v", err)
-		return
+func TestIptablesManagerIPSet(t *testing.T) {
+	ipv4Client, err := iptables.NewWithProtocol(iptables.ProtocolIPv4)
+	require.NoError(t, err)
+
+	mock := &iFaceMock{
+		NameFunc: func() string {
+			return "lo"
+		},
+		AddressFunc: func() iface.WGAddress {
+			return iface.WGAddress{
+				IP: net.ParseIP("10.20.0.1"),
+				Network: &net.IPNet{
+					IP:   net.ParseIP("10.20.0.0"),
+					Mask: net.IPv4Mask(255, 255, 255, 0),
+				},
+			}
+		},
 	}
 
-	if !exists && mustExists {
-		t.Errorf("rule '%v' does not exist", rulespec)
-		return
+	// just check on the local interface
+	manager, err := Create(mock)
+	require.NoError(t, err)
+
+	time.Sleep(time.Second)
+
+	defer func() {
+		err := manager.Reset()
+		require.NoError(t, err, "clear the manager state")
+
+		time.Sleep(time.Second)
+	}()
+
+	var rule1 fw.Rule
+	t.Run("add first rule with set", func(t *testing.T) {
+		ip := net.ParseIP("10.20.0.2")
+		port := &fw.Port{Values: []int{8080}}
+		rule1, err = manager.AddFiltering(
+			ip, "tcp", nil, port, fw.RuleDirectionOUT,
+			fw.ActionAccept, "default", "accept HTTP traffic",
+		)
+		require.NoError(t, err, "failed to add rule")
+
+		checkRuleSpecs(t, ipv4Client, ChainOutputFilterName, true, rule1.(*Rule).specs...)
+		require.Equal(t, rule1.(*Rule).ipsetName, "default-dport", "ipset name must be set")
+		require.Equal(t, rule1.(*Rule).ip, "10.20.0.2", "ipset IP must be set")
+	})
+
+	var rule2 fw.Rule
+	t.Run("add second rule", func(t *testing.T) {
+		ip := net.ParseIP("10.20.0.3")
+		port := &fw.Port{
+			Values: []int{443},
+		}
+		rule2, err = manager.AddFiltering(
+			ip, "tcp", port, nil, fw.RuleDirectionIN, fw.ActionAccept,
+			"default", "accept HTTPS traffic from ports range",
+		)
+		require.NoError(t, err, "failed to add rule")
+		require.Equal(t, rule2.(*Rule).ipsetName, "default-sport", "ipset name must be set")
+		require.Equal(t, rule2.(*Rule).ip, "10.20.0.3", "ipset IP must be set")
+	})
+
+	t.Run("delete first rule", func(t *testing.T) {
+		err := manager.DeleteRule(rule1)
+		require.NoError(t, err, "failed to delete rule")
+
+		require.NotContains(t, manager.rulesets, rule1.(*Rule).ruleID, "rule must be removed form the ruleset index")
+	})
+
+	t.Run("delete second rule", func(t *testing.T) {
+		err := manager.DeleteRule(rule2)
+		require.NoError(t, err, "failed to delete rule")
+
+		require.Empty(t, manager.rulesets, "rulesets index after removed second rule must be empty")
+	})
+
+	t.Run("reset check", func(t *testing.T) {
+		err = manager.Reset()
+		require.NoError(t, err, "failed to reset")
+	})
+}
+
+func checkRuleSpecs(t *testing.T, ipv4Client *iptables.IPTables, chainName string, mustExists bool, rulespec ...string) {
+	exists, err := ipv4Client.Exists("filter", chainName, rulespec...)
+	require.NoError(t, err, "failed to check rule")
+	require.Falsef(t, !exists && mustExists, "rule '%v' does not exist", rulespec)
+	require.Falsef(t, exists && !mustExists, "rule '%v' exist", rulespec)
+}
+
+func TestIptablesCreatePerformance(t *testing.T) {
+	mock := &iFaceMock{
+		NameFunc: func() string {
+			return "lo"
+		},
+		AddressFunc: func() iface.WGAddress {
+			return iface.WGAddress{
+				IP: net.ParseIP("10.20.0.1"),
+				Network: &net.IPNet{
+					IP:   net.ParseIP("10.20.0.0"),
+					Mask: net.IPv4Mask(255, 255, 255, 0),
+				},
+			}
+		},
 	}
-	if exists && !mustExists {
-		t.Errorf("rule '%v' exist", rulespec)
-		return
+
+	for _, testMax := range []int{10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000} {
+		t.Run(fmt.Sprintf("Testing %d rules", testMax), func(t *testing.T) {
+			// just check on the local interface
+			manager, err := Create(mock)
+			require.NoError(t, err)
+			time.Sleep(time.Second)
+
+			defer func() {
+				err := manager.Reset()
+				require.NoError(t, err, "clear the manager state")
+
+				time.Sleep(time.Second)
+			}()
+
+			_, err = manager.client(net.ParseIP("10.20.0.100"))
+			require.NoError(t, err)
+
+			ip := net.ParseIP("10.20.0.100")
+			start := time.Now()
+			for i := 0; i < testMax; i++ {
+				port := &fw.Port{Values: []int{1000 + i}}
+				if i%2 == 0 {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionOUT, fw.ActionAccept, "", "accept HTTP traffic")
+				} else {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionIN, fw.ActionAccept, "", "accept HTTP traffic")
+				}
+
+				require.NoError(t, err, "failed to add rule")
+			}
+			t.Logf("execution avg per rule: %s", time.Since(start)/time.Duration(testMax))
+		})
 	}
 }

client/firewall/iptables/rule.go

@@ -2,12 +2,16 @@ package iptables
 
 // Rule to handle management of rules
 type Rule struct {
-	id    string
+	ruleID    string
+	ipsetName string
+
 	specs []string
+	ip    string
+	dst   bool
 	v6    bool
 }
 
 // GetRuleID returns the rule id
 func (r *Rule) GetRuleID() string {
-	return r.id
+	return r.ruleID
 }

client/firewall/nftables/manager_linux.go

@@ -0,0 +1,758 @@
+package nftables
+
+import (
+	"bytes"
+	"encoding/binary"
+	"fmt"
+	"net"
+	"net/netip"
+	"strconv"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/google/nftables"
+	"github.com/google/nftables/expr"
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/sys/unix"
+
+	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
+)
+
+const (
+	// FilterTableName is the name of the table that is used for filtering by the Netbird client
+	FilterTableName = "netbird-acl"
+
+	// FilterInputChainName is the name of the chain that is used for filtering incoming packets
+	FilterInputChainName = "netbird-acl-input-filter"
+
+	// FilterOutputChainName is the name of the chain that is used for filtering outgoing packets
+	FilterOutputChainName = "netbird-acl-output-filter"
+)
+
+var anyIP = []byte{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
+
+// Manager of iptables firewall
+type Manager struct {
+	mutex sync.Mutex
+
+	rConn     *nftables.Conn
+	sConn     *nftables.Conn
+	tableIPv4 *nftables.Table
+	tableIPv6 *nftables.Table
+
+	filterInputChainIPv4  *nftables.Chain
+	filterOutputChainIPv4 *nftables.Chain
+
+	filterInputChainIPv6  *nftables.Chain
+	filterOutputChainIPv6 *nftables.Chain
+
+	rulesetManager *rulesetManager
+	setRemovedIPs  map[string]struct{}
+	setRemoved     map[string]*nftables.Set
+
+	wgIface iFaceMapper
+}
+
+// iFaceMapper defines subset methods of interface required for manager
+type iFaceMapper interface {
+	Name() string
+	Address() iface.WGAddress
+}
+
+// Create nftables firewall manager
+func Create(wgIface iFaceMapper) (*Manager, error) {
+	// sConn is used for creating sets and adding/removing elements from them
+	// it's differ then rConn (which does create new conn for each flush operation)
+	// and is permanent. Using same connection for booth type of operations
+	// overloads netlink with high amount of rules ( > 10000)
+	sConn, err := nftables.New(nftables.AsLasting())
+	if err != nil {
+		return nil, err
+	}
+
+	m := &Manager{
+		rConn: &nftables.Conn{},
+		sConn: sConn,
+
+		rulesetManager: newRuleManager(),
+		setRemovedIPs:  map[string]struct{}{},
+		setRemoved:     map[string]*nftables.Set{},
+
+		wgIface: wgIface,
+	}
+
+	if err := m.Reset(); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
+
+// AddFiltering rule to the firewall
+//
+// If comment argument is empty firewall manager should set
+// rule ID as comment for the rule
+func (m *Manager) AddFiltering(
+	ip net.IP,
+	proto fw.Protocol,
+	sPort *fw.Port,
+	dPort *fw.Port,
+	direction fw.RuleDirection,
+	action fw.Action,
+	ipsetName string,
+	comment string,
+) (fw.Rule, error) {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	var (
+		err   error
+		ipset *nftables.Set
+		table *nftables.Table
+		chain *nftables.Chain
+	)
+
+	if direction == fw.RuleDirectionOUT {
+		table, chain, err = m.chain(
+			ip,
+			FilterOutputChainName,
+			nftables.ChainHookOutput,
+			nftables.ChainPriorityFilter,
+			nftables.ChainTypeFilter)
+	} else {
+		table, chain, err = m.chain(
+			ip,
+			FilterInputChainName,
+			nftables.ChainHookInput,
+			nftables.ChainPriorityFilter,
+			nftables.ChainTypeFilter)
+	}
+	if err != nil {
+		return nil, err
+	}
+
+	rawIP := ip.To4()
+	if rawIP == nil {
+		rawIP = ip.To16()
+	}
+
+	rulesetID := m.getRulesetID(ip, proto, sPort, dPort, direction, action, ipsetName)
+
+	if ipsetName != "" {
+		// if we already have set with given name, just add ip to the set
+		// and return rule with new ID in other case let's create rule
+		// with fresh created set and set element
+
+		var isSetNew bool
+		ipset, err = m.rConn.GetSetByName(table, ipsetName)
+		if err != nil {
+			if ipset, err = m.createSet(table, rawIP, ipsetName); err != nil {
+				return nil, fmt.Errorf("get set name: %v", err)
+			}
+			isSetNew = true
+		}
+
+		if err := m.sConn.SetAddElements(ipset, []nftables.SetElement{{Key: rawIP}}); err != nil {
+			return nil, fmt.Errorf("add set element for the first time: %v", err)
+		}
+		if err := m.sConn.Flush(); err != nil {
+			return nil, fmt.Errorf("flush add elements: %v", err)
+		}
+
+		if !isSetNew {
+			// if we already have nftables rules with set for given direction
+			// just add new rule to the ruleset and return new fw.Rule object
+
+			if ruleset, ok := m.rulesetManager.getRuleset(rulesetID); ok {
+				return m.rulesetManager.addRule(ruleset, rawIP)
+			}
+			// if ipset exists but it is not linked to rule for given direction
+			// create new rule for direction and bind ipset to it later
+		}
+	}
+
+	ifaceKey := expr.MetaKeyIIFNAME
+	if direction == fw.RuleDirectionOUT {
+		ifaceKey = expr.MetaKeyOIFNAME
+	}
+	expressions := []expr.Any{
+		&expr.Meta{Key: ifaceKey, Register: 1},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     ifname(m.wgIface.Name()),
+		},
+	}
+
+	if proto != "all" {
+		expressions = append(expressions, &expr.Payload{
+			DestRegister: 1,
+			Base:         expr.PayloadBaseNetworkHeader,
+			Offset:       uint32(9),
+			Len:          uint32(1),
+		})
+
+		var protoData []byte
+		switch proto {
+		case fw.ProtocolTCP:
+			protoData = []byte{unix.IPPROTO_TCP}
+		case fw.ProtocolUDP:
+			protoData = []byte{unix.IPPROTO_UDP}
+		case fw.ProtocolICMP:
+			protoData = []byte{unix.IPPROTO_ICMP}
+		default:
+			return nil, fmt.Errorf("unsupported protocol: %s", proto)
+		}
+		expressions = append(expressions, &expr.Cmp{
+			Register: 1,
+			Op:       expr.CmpOpEq,
+			Data:     protoData,
+		})
+	}
+
+	// check if rawIP contains zeroed IPv4 0.0.0.0 or same IPv6 value
+	// in that case not add IP match expression into the rule definition
+	if !bytes.HasPrefix(anyIP, rawIP) {
+		// source address position
+		addrLen := uint32(len(rawIP))
+		addrOffset := uint32(12)
+		if addrLen == 16 {
+			addrOffset = 8
+		}
+
+		// change to destination address position if need
+		if direction == fw.RuleDirectionOUT {
+			addrOffset += addrLen
+		}
+
+		expressions = append(expressions,
+			&expr.Payload{
+				DestRegister: 1,
+				Base:         expr.PayloadBaseNetworkHeader,
+				Offset:       addrOffset,
+				Len:          addrLen,
+			},
+		)
+		// add individual IP for match if no ipset defined
+		if ipset == nil {
+			expressions = append(expressions,
+				&expr.Cmp{
+					Op:       expr.CmpOpEq,
+					Register: 1,
+					Data:     rawIP,
+				},
+			)
+		} else {
+			expressions = append(expressions,
+				&expr.Lookup{
+					SourceRegister: 1,
+					SetName:        ipsetName,
+					SetID:          ipset.ID,
+				},
+			)
+		}
+	}
+
+	if sPort != nil && len(sPort.Values) != 0 {
+		expressions = append(expressions,
+			&expr.Payload{
+				DestRegister: 1,
+				Base:         expr.PayloadBaseTransportHeader,
+				Offset:       0,
+				Len:          2,
+			},
+			&expr.Cmp{
+				Op:       expr.CmpOpEq,
+				Register: 1,
+				Data:     encodePort(*sPort),
+			},
+		)
+	}
+
+	if dPort != nil && len(dPort.Values) != 0 {
+		expressions = append(expressions,
+			&expr.Payload{
+				DestRegister: 1,
+				Base:         expr.PayloadBaseTransportHeader,
+				Offset:       2,
+				Len:          2,
+			},
+			&expr.Cmp{
+				Op:       expr.CmpOpEq,
+				Register: 1,
+				Data:     encodePort(*dPort),
+			},
+		)
+	}
+
+	if action == fw.ActionAccept {
+		expressions = append(expressions, &expr.Verdict{Kind: expr.VerdictAccept})
+	} else {
+		expressions = append(expressions, &expr.Verdict{Kind: expr.VerdictDrop})
+	}
+
+	userData := []byte(strings.Join([]string{rulesetID, comment}, " "))
+
+	rule := m.rConn.InsertRule(&nftables.Rule{
+		Table:    table,
+		Chain:    chain,
+		Position: 0,
+		Exprs:    expressions,
+		UserData: userData,
+	})
+	if err := m.rConn.Flush(); err != nil {
+		return nil, fmt.Errorf("flush insert rule: %v", err)
+	}
+
+	ruleset := m.rulesetManager.createRuleset(rulesetID, rule, ipset)
+	return m.rulesetManager.addRule(ruleset, rawIP)
+}
+
+// getRulesetID returns ruleset ID based on given parameters
+func (m *Manager) getRulesetID(
+	ip net.IP,
+	proto fw.Protocol,
+	sPort *fw.Port,
+	dPort *fw.Port,
+	direction fw.RuleDirection,
+	action fw.Action,
+	ipsetName string,
+) string {
+	rulesetID := ":" + strconv.Itoa(int(direction)) + ":"
+	if sPort != nil {
+		rulesetID += sPort.String()
+	}
+	rulesetID += ":"
+	if dPort != nil {
+		rulesetID += dPort.String()
+	}
+	rulesetID += ":"
+	rulesetID += strconv.Itoa(int(action))
+	if ipsetName == "" {
+		return "ip:" + ip.String() + rulesetID
+	}
+	return "set:" + ipsetName + rulesetID
+}
+
+// createSet in given table by name
+func (m *Manager) createSet(
+	table *nftables.Table,
+	rawIP []byte,
+	name string,
+) (*nftables.Set, error) {
+	keyType := nftables.TypeIPAddr
+	if len(rawIP) == 16 {
+		keyType = nftables.TypeIP6Addr
+	}
+	// else we create new ipset and continue creating rule
+	ipset := &nftables.Set{
+		Name:    name,
+		Table:   table,
+		Dynamic: true,
+		KeyType: keyType,
+	}
+
+	if err := m.rConn.AddSet(ipset, nil); err != nil {
+		return nil, fmt.Errorf("create set: %v", err)
+	}
+
+	if err := m.rConn.Flush(); err != nil {
+		return nil, fmt.Errorf("flush created set: %v", err)
+	}
+
+	return ipset, nil
+}
+
+// chain returns the chain for the given IP address with specific settings
+func (m *Manager) chain(
+	ip net.IP,
+	name string,
+	hook nftables.ChainHook,
+	priority nftables.ChainPriority,
+	cType nftables.ChainType,
+) (*nftables.Table, *nftables.Chain, error) {
+	var err error
+
+	getChain := func(c *nftables.Chain, tf nftables.TableFamily) (*nftables.Chain, error) {
+		if c != nil {
+			return c, nil
+		}
+		return m.createChainIfNotExists(tf, name, hook, priority, cType)
+	}
+
+	if ip.To4() != nil {
+		if name == FilterInputChainName {
+			m.filterInputChainIPv4, err = getChain(m.filterInputChainIPv4, nftables.TableFamilyIPv4)
+			return m.tableIPv4, m.filterInputChainIPv4, err
+		}
+		m.filterOutputChainIPv4, err = getChain(m.filterOutputChainIPv4, nftables.TableFamilyIPv4)
+		return m.tableIPv4, m.filterOutputChainIPv4, err
+	}
+	if name == FilterInputChainName {
+		m.filterInputChainIPv6, err = getChain(m.filterInputChainIPv6, nftables.TableFamilyIPv6)
+		return m.tableIPv4, m.filterInputChainIPv6, err
+	}
+	m.filterOutputChainIPv6, err = getChain(m.filterOutputChainIPv6, nftables.TableFamilyIPv6)
+	return m.tableIPv4, m.filterOutputChainIPv6, err
+}
+
+// table returns the table for the given family of the IP address
+func (m *Manager) table(family nftables.TableFamily) (*nftables.Table, error) {
+	if family == nftables.TableFamilyIPv4 {
+		if m.tableIPv4 != nil {
+			return m.tableIPv4, nil
+		}
+
+		table, err := m.createTableIfNotExists(nftables.TableFamilyIPv4)
+		if err != nil {
+			return nil, err
+		}
+		m.tableIPv4 = table
+		return m.tableIPv4, nil
+	}
+
+	if m.tableIPv6 != nil {
+		return m.tableIPv6, nil
+	}
+
+	table, err := m.createTableIfNotExists(nftables.TableFamilyIPv6)
+	if err != nil {
+		return nil, err
+	}
+	m.tableIPv6 = table
+	return m.tableIPv6, nil
+}
+
+func (m *Manager) createTableIfNotExists(family nftables.TableFamily) (*nftables.Table, error) {
+	tables, err := m.rConn.ListTablesOfFamily(family)
+	if err != nil {
+		return nil, fmt.Errorf("list of tables: %w", err)
+	}
+
+	for _, t := range tables {
+		if t.Name == FilterTableName {
+			return t, nil
+		}
+	}
+
+	table := m.rConn.AddTable(&nftables.Table{Name: FilterTableName, Family: nftables.TableFamilyIPv4})
+	if err := m.rConn.Flush(); err != nil {
+		return nil, err
+	}
+	return table, nil
+}
+
+func (m *Manager) createChainIfNotExists(
+	family nftables.TableFamily,
+	name string,
+	hooknum nftables.ChainHook,
+	priority nftables.ChainPriority,
+	chainType nftables.ChainType,
+) (*nftables.Chain, error) {
+	table, err := m.table(family)
+	if err != nil {
+		return nil, err
+	}
+
+	chains, err := m.rConn.ListChainsOfTableFamily(family)
+	if err != nil {
+		return nil, fmt.Errorf("list of chains: %w", err)
+	}
+
+	for _, c := range chains {
+		if c.Name == name && c.Table.Name == table.Name {
+			return c, nil
+		}
+	}
+
+	polAccept := nftables.ChainPolicyAccept
+	chain := &nftables.Chain{
+		Name:     name,
+		Table:    table,
+		Hooknum:  hooknum,
+		Priority: priority,
+		Type:     chainType,
+		Policy:   &polAccept,
+	}
+
+	chain = m.rConn.AddChain(chain)
+
+	ifaceKey := expr.MetaKeyIIFNAME
+	shiftDSTAddr := 0
+	if name == FilterOutputChainName {
+		ifaceKey = expr.MetaKeyOIFNAME
+		shiftDSTAddr = 1
+	}
+
+	expressions := []expr.Any{
+		&expr.Meta{Key: ifaceKey, Register: 1},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     ifname(m.wgIface.Name()),
+		},
+	}
+
+	mask, _ := netip.AddrFromSlice(m.wgIface.Address().Network.Mask)
+	if m.wgIface.Address().IP.To4() == nil {
+		ip, _ := netip.AddrFromSlice(m.wgIface.Address().Network.IP.To16())
+		expressions = append(expressions,
+			&expr.Payload{
+				DestRegister: 2,
+				Base:         expr.PayloadBaseNetworkHeader,
+				Offset:       uint32(8 + (16 * shiftDSTAddr)),
+				Len:          16,
+			},
+			&expr.Bitwise{
+				SourceRegister: 2,
+				DestRegister:   2,
+				Len:            16,
+				Xor:            []byte{0x0, 0x0, 0x0, 0x0},
+				Mask:           mask.Unmap().AsSlice(),
+			},
+			&expr.Cmp{
+				Op:       expr.CmpOpNeq,
+				Register: 2,
+				Data:     ip.Unmap().AsSlice(),
+			},
+			&expr.Verdict{Kind: expr.VerdictAccept},
+		)
+	} else {
+		ip, _ := netip.AddrFromSlice(m.wgIface.Address().Network.IP.To4())
+		expressions = append(expressions,
+			&expr.Payload{
+				DestRegister: 2,
+				Base:         expr.PayloadBaseNetworkHeader,
+				Offset:       uint32(12 + (4 * shiftDSTAddr)),
+				Len:          4,
+			},
+			&expr.Bitwise{
+				SourceRegister: 2,
+				DestRegister:   2,
+				Len:            4,
+				Xor:            []byte{0x0, 0x0, 0x0, 0x0},
+				Mask:           m.wgIface.Address().Network.Mask,
+			},
+			&expr.Cmp{
+				Op:       expr.CmpOpNeq,
+				Register: 2,
+				Data:     ip.Unmap().AsSlice(),
+			},
+			&expr.Verdict{Kind: expr.VerdictAccept},
+		)
+	}
+
+	_ = m.rConn.AddRule(&nftables.Rule{
+		Table: table,
+		Chain: chain,
+		Exprs: expressions,
+	})
+
+	expressions = []expr.Any{
+		&expr.Meta{Key: ifaceKey, Register: 1},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     ifname(m.wgIface.Name()),
+		},
+		&expr.Verdict{Kind: expr.VerdictDrop},
+	}
+	_ = m.rConn.AddRule(&nftables.Rule{
+		Table: table,
+		Chain: chain,
+		Exprs: expressions,
+	})
+
+	if err := m.rConn.Flush(); err != nil {
+		return nil, err
+	}
+
+	return chain, nil
+}
+
+// DeleteRule from the firewall by rule definition
+func (m *Manager) DeleteRule(rule fw.Rule) error {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	nativeRule, ok := rule.(*Rule)
+	if !ok {
+		return fmt.Errorf("invalid rule type")
+	}
+
+	if nativeRule.nftRule == nil {
+		return nil
+	}
+
+	if nativeRule.nftSet != nil {
+		// call twice of delete set element raises error
+		// so we need to check if element is already removed
+		key := fmt.Sprintf("%s:%v", nativeRule.nftSet.Name, nativeRule.ip)
+		if _, ok := m.setRemovedIPs[key]; !ok {
+			err := m.sConn.SetDeleteElements(nativeRule.nftSet, []nftables.SetElement{{Key: nativeRule.ip}})
+			if err != nil {
+				log.Errorf("delete elements for set %q: %v", nativeRule.nftSet.Name, err)
+			}
+			if err := m.sConn.Flush(); err != nil {
+				return err
+			}
+			m.setRemovedIPs[key] = struct{}{}
+		}
+	}
+
+	if m.rulesetManager.deleteRule(nativeRule) {
+		// deleteRule indicates that we still have IP in the ruleset
+		// it means we should not remove the nftables rule but need to update set
+		// so we prepare IP to be removed from set on the next flush call
+		return nil
+	}
+
+	// ruleset doesn't contain IP anymore (or contains only one), remove nft rule
+	if err := m.rConn.DelRule(nativeRule.nftRule); err != nil {
+		log.Errorf("failed to delete rule: %v", err)
+	}
+	if err := m.rConn.Flush(); err != nil {
+		return err
+	}
+	nativeRule.nftRule = nil
+
+	if nativeRule.nftSet != nil {
+		if _, ok := m.setRemoved[nativeRule.nftSet.Name]; !ok {
+			m.setRemoved[nativeRule.nftSet.Name] = nativeRule.nftSet
+		}
+		nativeRule.nftSet = nil
+	}
+
+	return nil
+}
+
+// Reset firewall to the default state
+func (m *Manager) Reset() error {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	chains, err := m.rConn.ListChains()
+	if err != nil {
+		return fmt.Errorf("list of chains: %w", err)
+	}
+	for _, c := range chains {
+		if c.Name == FilterInputChainName || c.Name == FilterOutputChainName {
+			m.rConn.DelChain(c)
+		}
+	}
+
+	tables, err := m.rConn.ListTables()
+	if err != nil {
+		return fmt.Errorf("list of tables: %w", err)
+	}
+	for _, t := range tables {
+		if t.Name == FilterTableName {
+			m.rConn.DelTable(t)
+		}
+	}
+
+	return m.rConn.Flush()
+}
+
+// Flush rule/chain/set operations from the buffer
+//
+// Method also get all rules after flush and refreshes handle values in the rulesets
+func (m *Manager) Flush() error {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	if err := m.flushWithBackoff(); err != nil {
+		return err
+	}
+
+	// set must be removed after flush rule changes
+	// otherwise we will get error
+	for _, s := range m.setRemoved {
+		m.rConn.FlushSet(s)
+		m.rConn.DelSet(s)
+	}
+
+	if len(m.setRemoved) > 0 {
+		if err := m.flushWithBackoff(); err != nil {
+			return err
+		}
+	}
+
+	m.setRemovedIPs = map[string]struct{}{}
+	m.setRemoved = map[string]*nftables.Set{}
+
+	if err := m.refreshRuleHandles(m.tableIPv4, m.filterInputChainIPv4); err != nil {
+		log.Errorf("failed to refresh rule handles ipv4 input chain: %v", err)
+	}
+
+	if err := m.refreshRuleHandles(m.tableIPv4, m.filterOutputChainIPv4); err != nil {
+		log.Errorf("failed to refresh rule handles IPv4 output chain: %v", err)
+	}
+
+	if err := m.refreshRuleHandles(m.tableIPv6, m.filterInputChainIPv6); err != nil {
+		log.Errorf("failed to refresh rule handles IPv6 input chain: %v", err)
+	}
+
+	if err := m.refreshRuleHandles(m.tableIPv6, m.filterOutputChainIPv6); err != nil {
+		log.Errorf("failed to refresh rule handles IPv6 output chain: %v", err)
+	}
+
+	return nil
+}
+
+func (m *Manager) flushWithBackoff() (err error) {
+	backoff := 4
+	backoffTime := 1000 * time.Millisecond
+	for i := 0; ; i++ {
+		err = m.rConn.Flush()
+		if err != nil {
+			if !strings.Contains(err.Error(), "busy") {
+				return
+			}
+			log.Error("failed to flush nftables, retrying...")
+			if i == backoff-1 {
+				return err
+			}
+			time.Sleep(backoffTime)
+			backoffTime = backoffTime * 2
+			continue
+		}
+		break
+	}
+	return
+}
+
+func (m *Manager) refreshRuleHandles(table *nftables.Table, chain *nftables.Chain) error {
+	if table == nil || chain == nil {
+		return nil
+	}
+
+	list, err := m.rConn.GetRules(table, chain)
+	if err != nil {
+		return err
+	}
+
+	for _, rule := range list {
+		if len(rule.UserData) != 0 {
+			if err := m.rulesetManager.setNftRuleHandle(rule); err != nil {
+				log.Errorf("failed to set rule handle: %v", err)
+			}
+		}
+	}
+
+	return nil
+}
+
+func encodePort(port fw.Port) []byte {
+	bs := make([]byte, 2)
+	binary.BigEndian.PutUint16(bs, uint16(port.Values[0]))
+	return bs
+}
+
+func ifname(n string) []byte {
+	b := make([]byte, 16)
+	copy(b, []byte(n+"\x00"))
+	return b
+}

client/firewall/nftables/manager_linux_test.go

@@ -0,0 +1,207 @@
+package nftables
+
+import (
+	"fmt"
+	"net"
+	"net/netip"
+	"testing"
+	"time"
+
+	"github.com/google/nftables"
+	"github.com/google/nftables/expr"
+	"github.com/stretchr/testify/require"
+	"golang.org/x/sys/unix"
+
+	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
+)
+
+// iFaceMapper defines subset methods of interface required for manager
+type iFaceMock struct {
+	NameFunc    func() string
+	AddressFunc func() iface.WGAddress
+}
+
+func (i *iFaceMock) Name() string {
+	if i.NameFunc != nil {
+		return i.NameFunc()
+	}
+	panic("NameFunc is not set")
+}
+
+func (i *iFaceMock) Address() iface.WGAddress {
+	if i.AddressFunc != nil {
+		return i.AddressFunc()
+	}
+	panic("AddressFunc is not set")
+}
+
+func TestNftablesManager(t *testing.T) {
+	mock := &iFaceMock{
+		NameFunc: func() string {
+			return "lo"
+		},
+		AddressFunc: func() iface.WGAddress {
+			return iface.WGAddress{
+				IP: net.ParseIP("100.96.0.1"),
+				Network: &net.IPNet{
+					IP:   net.ParseIP("100.96.0.0"),
+					Mask: net.IPv4Mask(255, 255, 255, 0),
+				},
+			}
+		},
+	}
+
+	// just check on the local interface
+	manager, err := Create(mock)
+	require.NoError(t, err)
+	time.Sleep(time.Second * 3)
+
+	defer func() {
+		err = manager.Reset()
+		require.NoError(t, err, "failed to reset")
+		time.Sleep(time.Second)
+	}()
+
+	ip := net.ParseIP("100.96.0.1")
+
+	testClient := &nftables.Conn{}
+
+	rule, err := manager.AddFiltering(
+		ip,
+		fw.ProtocolTCP,
+		nil,
+		&fw.Port{Values: []int{53}},
+		fw.RuleDirectionIN,
+		fw.ActionDrop,
+		"",
+		"",
+	)
+	require.NoError(t, err, "failed to add rule")
+
+	err = manager.Flush()
+	require.NoError(t, err, "failed to flush")
+
+	rules, err := testClient.GetRules(manager.tableIPv4, manager.filterInputChainIPv4)
+	require.NoError(t, err, "failed to get rules")
+
+	// test expectations:
+	// 1) regular rule
+	// 2) "accept extra routed traffic rule" for the interface
+	// 3) "drop all rule" for the interface
+	require.Len(t, rules, 3, "expected 3 rules")
+
+	ipToAdd, _ := netip.AddrFromSlice(ip)
+	add := ipToAdd.Unmap()
+	expectedExprs := []expr.Any{
+		&expr.Meta{Key: expr.MetaKeyIIFNAME, Register: 1},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     ifname("lo"),
+		},
+		&expr.Payload{
+			DestRegister: 1,
+			Base:         expr.PayloadBaseNetworkHeader,
+			Offset:       uint32(9),
+			Len:          uint32(1),
+		},
+		&expr.Cmp{
+			Register: 1,
+			Op:       expr.CmpOpEq,
+			Data:     []byte{unix.IPPROTO_TCP},
+		},
+		&expr.Payload{
+			DestRegister: 1,
+			Base:         expr.PayloadBaseNetworkHeader,
+			Offset:       12,
+			Len:          4,
+		},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     add.AsSlice(),
+		},
+		&expr.Payload{
+			DestRegister: 1,
+			Base:         expr.PayloadBaseTransportHeader,
+			Offset:       2,
+			Len:          2,
+		},
+		&expr.Cmp{
+			Op:       expr.CmpOpEq,
+			Register: 1,
+			Data:     []byte{0, 53},
+		},
+		&expr.Verdict{Kind: expr.VerdictDrop},
+	}
+	require.ElementsMatch(t, rules[0].Exprs, expectedExprs, "expected the same expressions")
+
+	err = manager.DeleteRule(rule)
+	require.NoError(t, err, "failed to delete rule")
+
+	err = manager.Flush()
+	require.NoError(t, err, "failed to flush")
+
+	rules, err = testClient.GetRules(manager.tableIPv4, manager.filterInputChainIPv4)
+	require.NoError(t, err, "failed to get rules")
+	// test expectations:
+	// 1) "accept extra routed traffic rule" for the interface
+	// 2) "drop all rule" for the interface
+	require.Len(t, rules, 2, "expected 2 rules after deleteion")
+
+	err = manager.Reset()
+	require.NoError(t, err, "failed to reset")
+}
+
+func TestNFtablesCreatePerformance(t *testing.T) {
+	mock := &iFaceMock{
+		NameFunc: func() string {
+			return "lo"
+		},
+		AddressFunc: func() iface.WGAddress {
+			return iface.WGAddress{
+				IP: net.ParseIP("100.96.0.1"),
+				Network: &net.IPNet{
+					IP:   net.ParseIP("100.96.0.0"),
+					Mask: net.IPv4Mask(255, 255, 255, 0),
+				},
+			}
+		},
+	}
+
+	for _, testMax := range []int{10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000} {
+		t.Run(fmt.Sprintf("Testing %d rules", testMax), func(t *testing.T) {
+			// just check on the local interface
+			manager, err := Create(mock)
+			require.NoError(t, err)
+			time.Sleep(time.Second * 3)
+
+			defer func() {
+				if err := manager.Reset(); err != nil {
+					t.Errorf("clear the manager state: %v", err)
+				}
+				time.Sleep(time.Second)
+			}()
+
+			ip := net.ParseIP("10.20.0.100")
+			start := time.Now()
+			for i := 0; i < testMax; i++ {
+				port := &fw.Port{Values: []int{1000 + i}}
+				if i%2 == 0 {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionOUT, fw.ActionAccept, "", "accept HTTP traffic")
+				} else {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionIN, fw.ActionAccept, "", "accept HTTP traffic")
+				}
+				require.NoError(t, err, "failed to add rule")
+
+				if i%100 == 0 {
+					err = manager.Flush()
+					require.NoError(t, err, "failed to flush")
+				}
+			}
+
+			t.Logf("execution avg per rule: %s", time.Since(start)/time.Duration(testMax))
+		})
+	}
+}

client/firewall/nftables/rule_linux.go

@@ -0,0 +1,19 @@
+package nftables
+
+import (
+	"github.com/google/nftables"
+)
+
+// Rule to handle management of rules
+type Rule struct {
+	nftRule *nftables.Rule
+	nftSet  *nftables.Set
+
+	ruleID string
+	ip     []byte
+}
+
+// GetRuleID returns the rule id
+func (r *Rule) GetRuleID() string {
+	return r.ruleID
+}

client/firewall/nftables/ruleset_linux.go

@@ -0,0 +1,115 @@
+package nftables
+
+import (
+	"bytes"
+	"fmt"
+
+	"github.com/google/nftables"
+	"github.com/rs/xid"
+)
+
+// nftRuleset links native firewall rule and ipset to ACL generated rules
+type nftRuleset struct {
+	nftRule     *nftables.Rule
+	nftSet      *nftables.Set
+	issuedRules map[string]*Rule
+	rulesetID   string
+}
+
+type rulesetManager struct {
+	rulesets map[string]*nftRuleset
+
+	nftSetName2rulesetID   map[string]string
+	issuedRuleID2rulesetID map[string]string
+}
+
+func newRuleManager() *rulesetManager {
+	return &rulesetManager{
+		rulesets: map[string]*nftRuleset{},
+
+		nftSetName2rulesetID:   map[string]string{},
+		issuedRuleID2rulesetID: map[string]string{},
+	}
+}
+
+func (r *rulesetManager) getRuleset(rulesetID string) (*nftRuleset, bool) {
+	ruleset, ok := r.rulesets[rulesetID]
+	return ruleset, ok
+}
+
+func (r *rulesetManager) createRuleset(
+	rulesetID string,
+	nftRule *nftables.Rule,
+	nftSet *nftables.Set,
+) *nftRuleset {
+	ruleset := nftRuleset{
+		rulesetID:   rulesetID,
+		nftRule:     nftRule,
+		nftSet:      nftSet,
+		issuedRules: map[string]*Rule{},
+	}
+	r.rulesets[ruleset.rulesetID] = &ruleset
+	if nftSet != nil {
+		r.nftSetName2rulesetID[nftSet.Name] = ruleset.rulesetID
+	}
+	return &ruleset
+}
+
+func (r *rulesetManager) addRule(
+	ruleset *nftRuleset,
+	ip []byte,
+) (*Rule, error) {
+	if _, ok := r.rulesets[ruleset.rulesetID]; !ok {
+		return nil, fmt.Errorf("ruleset not found")
+	}
+
+	rule := Rule{
+		nftRule: ruleset.nftRule,
+		nftSet:  ruleset.nftSet,
+		ruleID:  xid.New().String(),
+		ip:      ip,
+	}
+
+	ruleset.issuedRules[rule.ruleID] = &rule
+	r.issuedRuleID2rulesetID[rule.ruleID] = ruleset.rulesetID
+
+	return &rule, nil
+}
+
+// deleteRule from ruleset and returns true if contains other rules
+func (r *rulesetManager) deleteRule(rule *Rule) bool {
+	rulesetID, ok := r.issuedRuleID2rulesetID[rule.ruleID]
+	if !ok {
+		return false
+	}
+
+	ruleset := r.rulesets[rulesetID]
+	if ruleset.nftRule == nil {
+		return false
+	}
+	delete(r.issuedRuleID2rulesetID, rule.ruleID)
+	delete(ruleset.issuedRules, rule.ruleID)
+
+	if len(ruleset.issuedRules) == 0 {
+		delete(r.rulesets, ruleset.rulesetID)
+		if rule.nftSet != nil {
+			delete(r.nftSetName2rulesetID, rule.nftSet.Name)
+		}
+		return false
+	}
+	return true
+}
+
+// setNftRuleHandle finds rule by userdata which contains rulesetID and updates it's handle number
+//
+// This is important to do, because after we add rule to the nftables we can't update it until
+// we set correct handle value to it.
+func (r *rulesetManager) setNftRuleHandle(nftRule *nftables.Rule) error {
+	split := bytes.Split(nftRule.UserData, []byte(" "))
+	ruleset, ok := r.rulesets[string(split[0])]
+	if !ok {
+		return fmt.Errorf("ruleset not found")
+	}
+	*ruleset.nftRule = *nftRule
+	return nil
+}

client/firewall/nftables/ruleset_linux_test.go

@@ -0,0 +1,122 @@
+package nftables
+
+import (
+	"testing"
+
+	"github.com/google/nftables"
+	"github.com/stretchr/testify/require"
+)
+
+func TestRulesetManager_createRuleset(t *testing.T) {
+	// Create a ruleset manager.
+	rulesetManager := newRuleManager()
+
+	// Create a ruleset.
+	rulesetID := "ruleset-1"
+	nftRule := nftables.Rule{
+		UserData: []byte(rulesetID),
+	}
+	ruleset := rulesetManager.createRuleset(rulesetID, &nftRule, nil)
+	require.NotNil(t, ruleset, "createRuleset() failed")
+	require.Equal(t, ruleset.rulesetID, rulesetID, "rulesetID is incorrect")
+	require.Equal(t, ruleset.nftRule, &nftRule, "nftRule is incorrect")
+}
+
+func TestRulesetManager_addRule(t *testing.T) {
+	// Create a ruleset manager.
+	rulesetManager := newRuleManager()
+
+	// Create a ruleset.
+	rulesetID := "ruleset-1"
+	nftRule := nftables.Rule{}
+	ruleset := rulesetManager.createRuleset(rulesetID, &nftRule, nil)
+
+	// Add a rule to the ruleset.
+	ip := []byte("192.168.1.1")
+	rule, err := rulesetManager.addRule(ruleset, ip)
+	require.NoError(t, err, "addRule() failed")
+	require.NotNil(t, rule, "rule should not be nil")
+	require.NotEqual(t, rule.ruleID, "ruleID is empty")
+	require.EqualValues(t, rule.ip, ip, "ip is incorrect")
+	require.Contains(t, ruleset.issuedRules, rule.ruleID, "ruleID already exists in ruleset")
+	require.Contains(t, rulesetManager.issuedRuleID2rulesetID, rule.ruleID, "ruleID already exists in ruleset manager")
+
+	ruleset2 := &nftRuleset{
+		rulesetID: "ruleset-2",
+	}
+	_, err = rulesetManager.addRule(ruleset2, ip)
+	require.Error(t, err, "addRule() should have failed")
+}
+
+func TestRulesetManager_deleteRule(t *testing.T) {
+	// Create a ruleset manager.
+	rulesetManager := newRuleManager()
+
+	// Create a ruleset.
+	rulesetID := "ruleset-1"
+	nftRule := nftables.Rule{}
+	ruleset := rulesetManager.createRuleset(rulesetID, &nftRule, nil)
+
+	// Add a rule to the ruleset.
+	ip := []byte("192.168.1.1")
+	rule, err := rulesetManager.addRule(ruleset, ip)
+	require.NoError(t, err, "addRule() failed")
+	require.NotNil(t, rule, "rule should not be nil")
+
+	ip2 := []byte("192.168.1.1")
+	rule2, err := rulesetManager.addRule(ruleset, ip2)
+	require.NoError(t, err, "addRule() failed")
+	require.NotNil(t, rule2, "rule should not be nil")
+
+	hasNext := rulesetManager.deleteRule(rule)
+	require.True(t, hasNext, "deleteRule() should have returned true")
+
+	// Check that the rule is no longer in the manager.
+	require.NotContains(t, rulesetManager.issuedRuleID2rulesetID, rule.ruleID, "rule should have been deleted")
+
+	hasNext = rulesetManager.deleteRule(rule2)
+	require.False(t, hasNext, "deleteRule() should have returned false")
+}
+
+func TestRulesetManager_setNftRuleHandle(t *testing.T) {
+	// Create a ruleset manager.
+	rulesetManager := newRuleManager()
+	// Create a ruleset.
+	rulesetID := "ruleset-1"
+	nftRule := nftables.Rule{}
+	ruleset := rulesetManager.createRuleset(rulesetID, &nftRule, nil)
+	// Add a rule to the ruleset.
+	ip := []byte("192.168.0.1")
+
+	rule, err := rulesetManager.addRule(ruleset, ip)
+	require.NoError(t, err, "addRule() failed")
+	require.NotNil(t, rule, "rule should not be nil")
+
+	nftRuleCopy := nftRule
+	nftRuleCopy.Handle = 2
+	nftRuleCopy.UserData = []byte(rulesetID)
+	err = rulesetManager.setNftRuleHandle(&nftRuleCopy)
+	require.NoError(t, err, "setNftRuleHandle() failed")
+	// check correct work with references
+	require.Equal(t, nftRule.Handle, uint64(2), "nftRule.Handle is incorrect")
+}
+
+func TestRulesetManager_getRuleset(t *testing.T) {
+	// Create a ruleset manager.
+	rulesetManager := newRuleManager()
+	// Create a ruleset.
+	rulesetID := "ruleset-1"
+	nftRule := nftables.Rule{}
+	nftSet := nftables.Set{
+		ID: 2,
+	}
+	ruleset := rulesetManager.createRuleset(rulesetID, &nftRule, &nftSet)
+	require.NotNil(t, ruleset, "createRuleset() failed")
+
+	find, ok := rulesetManager.getRuleset(rulesetID)
+	require.True(t, ok, "getRuleset() failed")
+	require.Equal(t, ruleset, find, "getRulesetBySetID() failed")
+
+	_, ok = rulesetManager.getRuleset("does-not-exist")
+	require.False(t, ok, "getRuleset() failed")
+}

client/firewall/port.go

@@ -1,14 +1,27 @@
 package firewall
 
-// PortProtocol is the protocol of the port
-type PortProtocol string
+import (
+	"strconv"
+)
+
+// Protocol is the protocol of the port
+type Protocol string
 
 const (
-	// PortProtocolTCP is the TCP protocol
-	PortProtocolTCP PortProtocol = "tcp"
+	// ProtocolTCP is the TCP protocol
+	ProtocolTCP Protocol = "tcp"
 
-	// PortProtocolUDP is the UDP protocol
-	PortProtocolUDP PortProtocol = "udp"
+	// ProtocolUDP is the UDP protocol
+	ProtocolUDP Protocol = "udp"
+
+	// ProtocolICMP is the ICMP protocol
+	ProtocolICMP Protocol = "icmp"
+
+	// ProtocolALL cover all supported protocols
+	ProtocolALL Protocol = "all"
+
+	// ProtocolUnknown unknown protocol
+	ProtocolUnknown Protocol = "unknown"
 )
 
 // Port of the address for firewall rule
@@ -18,7 +31,16 @@ type Port struct {
 
 	// Values contains one value for single port, multiple values for the list of ports, or two values for the range of ports
 	Values []int
-
-	// Proto is the protocol of the port
-	Proto PortProtocol
+}
+
+// String interface implementation
+func (p *Port) String() string {
+	var ports string
+	for _, port := range p.Values {
+		if ports != "" {
+			ports += ","
+		}
+		ports += strconv.Itoa(port)
+	}
+	return ports
 }

client/firewall/uspfilter/rule.go

@@ -0,0 +1,30 @@
+package uspfilter
+
+import (
+	"net"
+
+	"github.com/google/gopacket"
+
+	fw "github.com/netbirdio/netbird/client/firewall"
+)
+
+// Rule to handle management of rules
+type Rule struct {
+	id         string
+	ip         net.IP
+	ipLayer    gopacket.LayerType
+	matchByIP  bool
+	protoLayer gopacket.LayerType
+	direction  fw.RuleDirection
+	sPort      uint16
+	dPort      uint16
+	drop       bool
+	comment    string
+
+	udpHook func([]byte) bool
+}
+
+// GetRuleID returns the rule id
+func (r *Rule) GetRuleID() string {
+	return r.id
+}

client/firewall/uspfilter/uspfilter.go

@@ -0,0 +1,377 @@
+package uspfilter
+
+import (
+	"fmt"
+	"net"
+	"sync"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/google/uuid"
+	log "github.com/sirupsen/logrus"
+
+	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
+)
+
+const layerTypeAll = 0
+
+// IFaceMapper defines subset methods of interface required for manager
+type IFaceMapper interface {
+	SetFilter(iface.PacketFilter) error
+}
+
+// RuleSet is a set of rules grouped by a string key
+type RuleSet map[string]Rule
+
+// Manager userspace firewall manager
+type Manager struct {
+	outgoingRules map[string]RuleSet
+	incomingRules map[string]RuleSet
+	wgNetwork     *net.IPNet
+	decoders      sync.Pool
+
+	mutex sync.RWMutex
+}
+
+// decoder for packages
+type decoder struct {
+	eth     layers.Ethernet
+	ip4     layers.IPv4
+	ip6     layers.IPv6
+	tcp     layers.TCP
+	udp     layers.UDP
+	icmp4   layers.ICMPv4
+	icmp6   layers.ICMPv6
+	decoded []gopacket.LayerType
+	parser  *gopacket.DecodingLayerParser
+}
+
+// Create userspace firewall manager constructor
+func Create(iface IFaceMapper) (*Manager, error) {
+	m := &Manager{
+		decoders: sync.Pool{
+			New: func() any {
+				d := &decoder{
+					decoded: []gopacket.LayerType{},
+				}
+				d.parser = gopacket.NewDecodingLayerParser(
+					layers.LayerTypeIPv4,
+					&d.eth, &d.ip4, &d.ip6, &d.icmp4, &d.icmp6, &d.tcp, &d.udp,
+				)
+				d.parser.IgnoreUnsupported = true
+				return d
+			},
+		},
+		outgoingRules: make(map[string]RuleSet),
+		incomingRules: make(map[string]RuleSet),
+	}
+
+	if err := iface.SetFilter(m); err != nil {
+		return nil, err
+	}
+	return m, nil
+}
+
+// AddFiltering rule to the firewall
+//
+// If comment argument is empty firewall manager should set
+// rule ID as comment for the rule
+func (m *Manager) AddFiltering(
+	ip net.IP,
+	proto fw.Protocol,
+	sPort *fw.Port,
+	dPort *fw.Port,
+	direction fw.RuleDirection,
+	action fw.Action,
+	ipsetName string,
+	comment string,
+) (fw.Rule, error) {
+	r := Rule{
+		id:        uuid.New().String(),
+		ip:        ip,
+		ipLayer:   layers.LayerTypeIPv6,
+		matchByIP: true,
+		direction: direction,
+		drop:      action == fw.ActionDrop,
+		comment:   comment,
+	}
+	if ipNormalized := ip.To4(); ipNormalized != nil {
+		r.ipLayer = layers.LayerTypeIPv4
+		r.ip = ipNormalized
+	}
+
+	if s := r.ip.String(); s == "0.0.0.0" || s == "::" {
+		r.matchByIP = false
+	}
+
+	if sPort != nil && len(sPort.Values) == 1 {
+		r.sPort = uint16(sPort.Values[0])
+	}
+
+	if dPort != nil && len(dPort.Values) == 1 {
+		r.dPort = uint16(dPort.Values[0])
+	}
+
+	switch proto {
+	case fw.ProtocolTCP:
+		r.protoLayer = layers.LayerTypeTCP
+	case fw.ProtocolUDP:
+		r.protoLayer = layers.LayerTypeUDP
+	case fw.ProtocolICMP:
+		r.protoLayer = layers.LayerTypeICMPv4
+		if r.ipLayer == layers.LayerTypeIPv6 {
+			r.protoLayer = layers.LayerTypeICMPv6
+		}
+	case fw.ProtocolALL:
+		r.protoLayer = layerTypeAll
+	}
+
+	m.mutex.Lock()
+	if direction == fw.RuleDirectionIN {
+		if _, ok := m.incomingRules[r.ip.String()]; !ok {
+			m.incomingRules[r.ip.String()] = make(RuleSet)
+		}
+		m.incomingRules[r.ip.String()][r.id] = r
+	} else {
+		if _, ok := m.outgoingRules[r.ip.String()]; !ok {
+			m.outgoingRules[r.ip.String()] = make(RuleSet)
+		}
+		m.outgoingRules[r.ip.String()][r.id] = r
+	}
+	m.mutex.Unlock()
+
+	return &r, nil
+}
+
+// DeleteRule from the firewall by rule definition
+func (m *Manager) DeleteRule(rule fw.Rule) error {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	r, ok := rule.(*Rule)
+	if !ok {
+		return fmt.Errorf("delete rule: invalid rule type: %T", rule)
+	}
+
+	if r.direction == fw.RuleDirectionIN {
+		_, ok := m.incomingRules[r.ip.String()][r.id]
+		if !ok {
+			return fmt.Errorf("delete rule: no rule with such id: %v", r.id)
+		}
+		delete(m.incomingRules[r.ip.String()], r.id)
+	} else {
+		_, ok := m.outgoingRules[r.ip.String()][r.id]
+		if !ok {
+			return fmt.Errorf("delete rule: no rule with such id: %v", r.id)
+		}
+		delete(m.outgoingRules[r.ip.String()], r.id)
+	}
+
+	return nil
+}
+
+// Reset firewall to the default state
+func (m *Manager) Reset() error {
+	m.mutex.Lock()
+	defer m.mutex.Unlock()
+
+	m.outgoingRules = make(map[string]RuleSet)
+	m.incomingRules = make(map[string]RuleSet)
+
+	return nil
+}
+
+// Flush doesn't need to be implemented for this manager
+func (m *Manager) Flush() error { return nil }
+
+// DropOutgoing filter outgoing packets
+func (m *Manager) DropOutgoing(packetData []byte) bool {
+	return m.dropFilter(packetData, m.outgoingRules, false)
+}
+
+// DropIncoming filter incoming packets
+func (m *Manager) DropIncoming(packetData []byte) bool {
+	return m.dropFilter(packetData, m.incomingRules, true)
+}
+
+// dropFilter imlements same logic for booth direction of the traffic
+func (m *Manager) dropFilter(packetData []byte, rules map[string]RuleSet, isIncomingPacket bool) bool {
+	m.mutex.RLock()
+	defer m.mutex.RUnlock()
+
+	d := m.decoders.Get().(*decoder)
+	defer m.decoders.Put(d)
+
+	if err := d.parser.DecodeLayers(packetData, &d.decoded); err != nil {
+		log.Tracef("couldn't decode layer, err: %s", err)
+		return true
+	}
+
+	if len(d.decoded) < 2 {
+		log.Tracef("not enough levels in network packet")
+		return true
+	}
+
+	ipLayer := d.decoded[0]
+
+	switch ipLayer {
+	case layers.LayerTypeIPv4:
+		if !m.wgNetwork.Contains(d.ip4.SrcIP) || !m.wgNetwork.Contains(d.ip4.DstIP) {
+			return false
+		}
+	case layers.LayerTypeIPv6:
+		if !m.wgNetwork.Contains(d.ip6.SrcIP) || !m.wgNetwork.Contains(d.ip6.DstIP) {
+			return false
+		}
+	default:
+		log.Errorf("unknown layer: %v", d.decoded[0])
+		return true
+	}
+
+	var ip net.IP
+	switch ipLayer {
+	case layers.LayerTypeIPv4:
+		if isIncomingPacket {
+			ip = d.ip4.SrcIP
+		} else {
+			ip = d.ip4.DstIP
+		}
+	case layers.LayerTypeIPv6:
+		if isIncomingPacket {
+			ip = d.ip6.SrcIP
+		} else {
+			ip = d.ip6.DstIP
+		}
+	}
+
+	filter, ok := validateRule(ip, packetData, rules[ip.String()], d)
+	if ok {
+		return filter
+	}
+	filter, ok = validateRule(ip, packetData, rules["0.0.0.0"], d)
+	if ok {
+		return filter
+	}
+	filter, ok = validateRule(ip, packetData, rules["::"], d)
+	if ok {
+		return filter
+	}
+
+	// default policy is DROP ALL
+	return true
+}
+
+func validateRule(ip net.IP, packetData []byte, rules map[string]Rule, d *decoder) (bool, bool) {
+	payloadLayer := d.decoded[1]
+	for _, rule := range rules {
+		if rule.matchByIP && !ip.Equal(rule.ip) {
+			continue
+		}
+
+		if rule.protoLayer == layerTypeAll {
+			return rule.drop, true
+		}
+
+		if payloadLayer != rule.protoLayer {
+			continue
+		}
+
+		switch payloadLayer {
+		case layers.LayerTypeTCP:
+			if rule.sPort == 0 && rule.dPort == 0 {
+				return rule.drop, true
+			}
+			if rule.sPort != 0 && rule.sPort == uint16(d.tcp.SrcPort) {
+				return rule.drop, true
+			}
+			if rule.dPort != 0 && rule.dPort == uint16(d.tcp.DstPort) {
+				return rule.drop, true
+			}
+		case layers.LayerTypeUDP:
+			// if rule has UDP hook (and if we are here we match this rule)
+			// we ignore rule.drop and call this hook
+			if rule.udpHook != nil {
+				return rule.udpHook(packetData), true
+			}
+
+			if rule.sPort == 0 && rule.dPort == 0 {
+				return rule.drop, true
+			}
+			if rule.sPort != 0 && rule.sPort == uint16(d.udp.SrcPort) {
+				return rule.drop, true
+			}
+			if rule.dPort != 0 && rule.dPort == uint16(d.udp.DstPort) {
+				return rule.drop, true
+			}
+			return rule.drop, true
+		case layers.LayerTypeICMPv4, layers.LayerTypeICMPv6:
+			return rule.drop, true
+		}
+	}
+	return false, false
+}
+
+// SetNetwork of the wireguard interface to which filtering applied
+func (m *Manager) SetNetwork(network *net.IPNet) {
+	m.wgNetwork = network
+}
+
+// AddUDPPacketHook calls hook when UDP packet from given direction matched
+//
+// Hook function returns flag which indicates should be the matched package dropped or not
+func (m *Manager) AddUDPPacketHook(
+	in bool, ip net.IP, dPort uint16, hook func([]byte) bool,
+) string {
+	r := Rule{
+		id:         uuid.New().String(),
+		ip:         ip,
+		protoLayer: layers.LayerTypeUDP,
+		dPort:      dPort,
+		ipLayer:    layers.LayerTypeIPv6,
+		direction:  fw.RuleDirectionOUT,
+		comment:    fmt.Sprintf("UDP Hook direction: %v, ip:%v, dport:%d", in, ip, dPort),
+		udpHook:    hook,
+	}
+
+	if ip.To4() != nil {
+		r.ipLayer = layers.LayerTypeIPv4
+	}
+
+	m.mutex.Lock()
+	if in {
+		r.direction = fw.RuleDirectionIN
+		if _, ok := m.incomingRules[r.ip.String()]; !ok {
+			m.incomingRules[r.ip.String()] = make(map[string]Rule)
+		}
+		m.incomingRules[r.ip.String()][r.id] = r
+	} else {
+		if _, ok := m.outgoingRules[r.ip.String()]; !ok {
+			m.outgoingRules[r.ip.String()] = make(map[string]Rule)
+		}
+		m.outgoingRules[r.ip.String()][r.id] = r
+	}
+
+	m.mutex.Unlock()
+
+	return r.id
+}
+
+// RemovePacketHook removes packet hook by given ID
+func (m *Manager) RemovePacketHook(hookID string) error {
+	for _, arr := range m.incomingRules {
+		for _, r := range arr {
+			if r.id == hookID {
+				return m.DeleteRule(&r)
+			}
+		}
+	}
+	for _, arr := range m.outgoingRules {
+		for _, r := range arr {
+			if r.id == hookID {
+				return m.DeleteRule(&r)
+			}
+		}
+	}
+	return fmt.Errorf("hook with given id not found")
+}

client/firewall/uspfilter/uspfilter_test.go

@@ -0,0 +1,403 @@
+package uspfilter
+
+import (
+	"fmt"
+	"net"
+	"testing"
+	"time"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/stretchr/testify/require"
+
+	fw "github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/iface"
+)
+
+type IFaceMock struct {
+	SetFilterFunc func(iface.PacketFilter) error
+}
+
+func (i *IFaceMock) SetFilter(iface iface.PacketFilter) error {
+	if i.SetFilterFunc == nil {
+		return fmt.Errorf("not implemented")
+	}
+	return i.SetFilterFunc(iface)
+}
+
+func TestManagerCreate(t *testing.T) {
+	ifaceMock := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error { return nil },
+	}
+
+	m, err := Create(ifaceMock)
+	if err != nil {
+		t.Errorf("failed to create Manager: %v", err)
+		return
+	}
+
+	if m == nil {
+		t.Error("Manager is nil")
+	}
+}
+
+func TestManagerAddFiltering(t *testing.T) {
+	isSetFilterCalled := false
+	ifaceMock := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error {
+			isSetFilterCalled = true
+			return nil
+		},
+	}
+
+	m, err := Create(ifaceMock)
+	if err != nil {
+		t.Errorf("failed to create Manager: %v", err)
+		return
+	}
+
+	ip := net.ParseIP("192.168.1.1")
+	proto := fw.ProtocolTCP
+	port := &fw.Port{Values: []int{80}}
+	direction := fw.RuleDirectionOUT
+	action := fw.ActionDrop
+	comment := "Test rule"
+
+	rule, err := m.AddFiltering(ip, proto, nil, port, direction, action, "", comment)
+	if err != nil {
+		t.Errorf("failed to add filtering: %v", err)
+		return
+	}
+
+	if rule == nil {
+		t.Error("Rule is nil")
+		return
+	}
+
+	if !isSetFilterCalled {
+		t.Error("SetFilter was not called")
+		return
+	}
+}
+
+func TestManagerDeleteRule(t *testing.T) {
+	ifaceMock := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error { return nil },
+	}
+
+	m, err := Create(ifaceMock)
+	if err != nil {
+		t.Errorf("failed to create Manager: %v", err)
+		return
+	}
+
+	ip := net.ParseIP("192.168.1.1")
+	proto := fw.ProtocolTCP
+	port := &fw.Port{Values: []int{80}}
+	direction := fw.RuleDirectionOUT
+	action := fw.ActionDrop
+	comment := "Test rule"
+
+	rule, err := m.AddFiltering(ip, proto, nil, port, direction, action, "", comment)
+	if err != nil {
+		t.Errorf("failed to add filtering: %v", err)
+		return
+	}
+
+	ip = net.ParseIP("192.168.1.1")
+	proto = fw.ProtocolTCP
+	port = &fw.Port{Values: []int{80}}
+	direction = fw.RuleDirectionIN
+	action = fw.ActionDrop
+	comment = "Test rule 2"
+
+	rule2, err := m.AddFiltering(ip, proto, nil, port, direction, action, "", comment)
+	if err != nil {
+		t.Errorf("failed to add filtering: %v", err)
+		return
+	}
+
+	err = m.DeleteRule(rule)
+	if err != nil {
+		t.Errorf("failed to delete rule: %v", err)
+		return
+	}
+
+	if _, ok := m.incomingRules[ip.String()][rule2.GetRuleID()]; !ok {
+		t.Errorf("rule2 is not in the incomingRules")
+	}
+
+	err = m.DeleteRule(rule2)
+	if err != nil {
+		t.Errorf("failed to delete rule: %v", err)
+		return
+	}
+
+	if _, ok := m.incomingRules[ip.String()][rule2.GetRuleID()]; ok {
+		t.Errorf("rule2 is not in the incomingRules")
+	}
+}
+
+func TestAddUDPPacketHook(t *testing.T) {
+	tests := []struct {
+		name       string
+		in         bool
+		expDir     fw.RuleDirection
+		ip         net.IP
+		dPort      uint16
+		hook       func([]byte) bool
+		expectedID string
+	}{
+		{
+			name:   "Test Outgoing UDP Packet Hook",
+			in:     false,
+			expDir: fw.RuleDirectionOUT,
+			ip:     net.IPv4(10, 168, 0, 1),
+			dPort:  8000,
+			hook:   func([]byte) bool { return true },
+		},
+		{
+			name:   "Test Incoming UDP Packet Hook",
+			in:     true,
+			expDir: fw.RuleDirectionIN,
+			ip:     net.IPv6loopback,
+			dPort:  9000,
+			hook:   func([]byte) bool { return false },
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			manager := &Manager{
+				incomingRules: map[string]RuleSet{},
+				outgoingRules: map[string]RuleSet{},
+			}
+
+			manager.AddUDPPacketHook(tt.in, tt.ip, tt.dPort, tt.hook)
+
+			var addedRule Rule
+			if tt.in {
+				if len(manager.incomingRules[tt.ip.String()]) != 1 {
+					t.Errorf("expected 1 incoming rule, got %d", len(manager.incomingRules))
+					return
+				}
+				for _, rule := range manager.incomingRules[tt.ip.String()] {
+					addedRule = rule
+				}
+			} else {
+				if len(manager.outgoingRules) != 1 {
+					t.Errorf("expected 1 outgoing rule, got %d", len(manager.outgoingRules))
+					return
+				}
+				for _, rule := range manager.outgoingRules[tt.ip.String()] {
+					addedRule = rule
+				}
+			}
+
+			if !tt.ip.Equal(addedRule.ip) {
+				t.Errorf("expected ip %s, got %s", tt.ip, addedRule.ip)
+				return
+			}
+			if tt.dPort != addedRule.dPort {
+				t.Errorf("expected dPort %d, got %d", tt.dPort, addedRule.dPort)
+				return
+			}
+			if layers.LayerTypeUDP != addedRule.protoLayer {
+				t.Errorf("expected protoLayer %s, got %s", layers.LayerTypeUDP, addedRule.protoLayer)
+				return
+			}
+			if tt.expDir != addedRule.direction {
+				t.Errorf("expected direction %d, got %d", tt.expDir, addedRule.direction)
+				return
+			}
+			if addedRule.udpHook == nil {
+				t.Errorf("expected udpHook to be set")
+				return
+			}
+		})
+	}
+}
+
+func TestManagerReset(t *testing.T) {
+	ifaceMock := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error { return nil },
+	}
+
+	m, err := Create(ifaceMock)
+	if err != nil {
+		t.Errorf("failed to create Manager: %v", err)
+		return
+	}
+
+	ip := net.ParseIP("192.168.1.1")
+	proto := fw.ProtocolTCP
+	port := &fw.Port{Values: []int{80}}
+	direction := fw.RuleDirectionOUT
+	action := fw.ActionDrop
+	comment := "Test rule"
+
+	_, err = m.AddFiltering(ip, proto, nil, port, direction, action, "", comment)
+	if err != nil {
+		t.Errorf("failed to add filtering: %v", err)
+		return
+	}
+
+	err = m.Reset()
+	if err != nil {
+		t.Errorf("failed to reset Manager: %v", err)
+		return
+	}
+
+	if len(m.outgoingRules) != 0 || len(m.incomingRules) != 0 {
+		t.Errorf("rules is not empty")
+	}
+}
+
+func TestNotMatchByIP(t *testing.T) {
+	ifaceMock := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error { return nil },
+	}
+
+	m, err := Create(ifaceMock)
+	if err != nil {
+		t.Errorf("failed to create Manager: %v", err)
+		return
+	}
+	m.wgNetwork = &net.IPNet{
+		IP:   net.ParseIP("100.10.0.0"),
+		Mask: net.CIDRMask(16, 32),
+	}
+
+	ip := net.ParseIP("0.0.0.0")
+	proto := fw.ProtocolUDP
+	direction := fw.RuleDirectionOUT
+	action := fw.ActionAccept
+	comment := "Test rule"
+
+	_, err = m.AddFiltering(ip, proto, nil, nil, direction, action, "", comment)
+	if err != nil {
+		t.Errorf("failed to add filtering: %v", err)
+		return
+	}
+
+	ipv4 := &layers.IPv4{
+		TTL:      64,
+		Version:  4,
+		SrcIP:    net.ParseIP("100.10.0.1"),
+		DstIP:    net.ParseIP("100.10.0.100"),
+		Protocol: layers.IPProtocolUDP,
+	}
+	udp := &layers.UDP{
+		SrcPort: 51334,
+		DstPort: 53,
+	}
+
+	if err := udp.SetNetworkLayerForChecksum(ipv4); err != nil {
+		t.Errorf("failed to set network layer for checksum: %v", err)
+		return
+	}
+	payload := gopacket.Payload([]byte("test"))
+
+	buf := gopacket.NewSerializeBuffer()
+	opts := gopacket.SerializeOptions{
+		ComputeChecksums: true,
+		FixLengths:       true,
+	}
+	if err = gopacket.SerializeLayers(buf, opts, ipv4, udp, payload); err != nil {
+		t.Errorf("failed to serialize packet: %v", err)
+		return
+	}
+
+	if m.dropFilter(buf.Bytes(), m.outgoingRules, false) {
+		t.Errorf("expected packet to be accepted")
+		return
+	}
+
+	if err = m.Reset(); err != nil {
+		t.Errorf("failed to reset Manager: %v", err)
+		return
+	}
+}
+
+// TestRemovePacketHook tests the functionality of the RemovePacketHook method
+func TestRemovePacketHook(t *testing.T) {
+	// creating mock iface
+	iface := &IFaceMock{
+		SetFilterFunc: func(iface.PacketFilter) error { return nil },
+	}
+
+	// creating manager instance
+	manager, err := Create(iface)
+	if err != nil {
+		t.Fatalf("Failed to create Manager: %s", err)
+	}
+
+	// Add a UDP packet hook
+	hookFunc := func(data []byte) bool { return true }
+	hookID := manager.AddUDPPacketHook(false, net.IPv4(192, 168, 0, 1), 8080, hookFunc)
+
+	// Assert the hook is added by finding it in the manager's outgoing rules
+	found := false
+	for _, arr := range manager.outgoingRules {
+		for _, rule := range arr {
+			if rule.id == hookID {
+				found = true
+				break
+			}
+		}
+	}
+
+	if !found {
+		t.Fatalf("The hook was not added properly.")
+	}
+
+	// Now remove the packet hook
+	err = manager.RemovePacketHook(hookID)
+	if err != nil {
+		t.Fatalf("Failed to remove hook: %s", err)
+	}
+
+	// Assert the hook is removed by checking it in the manager's outgoing rules
+	for _, arr := range manager.outgoingRules {
+		for _, rule := range arr {
+			if rule.id == hookID {
+				t.Fatalf("The hook was not removed properly.")
+			}
+		}
+	}
+}
+
+func TestUSPFilterCreatePerformance(t *testing.T) {
+	for _, testMax := range []int{10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000} {
+		t.Run(fmt.Sprintf("Testing %d rules", testMax), func(t *testing.T) {
+			// just check on the local interface
+			ifaceMock := &IFaceMock{
+				SetFilterFunc: func(iface.PacketFilter) error { return nil },
+			}
+			manager, err := Create(ifaceMock)
+			require.NoError(t, err)
+			time.Sleep(time.Second)
+
+			defer func() {
+				if err := manager.Reset(); err != nil {
+					t.Errorf("clear the manager state: %v", err)
+				}
+				time.Sleep(time.Second)
+			}()
+
+			ip := net.ParseIP("10.20.0.100")
+			start := time.Now()
+			for i := 0; i < testMax; i++ {
+				port := &fw.Port{Values: []int{1000 + i}}
+				if i%2 == 0 {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionOUT, fw.ActionAccept, "", "accept HTTP traffic")
+				} else {
+					_, err = manager.AddFiltering(ip, "tcp", nil, port, fw.RuleDirectionIN, fw.ActionAccept, "", "accept HTTP traffic")
+				}
+
+				require.NoError(t, err, "failed to add rule")
+			}
+			t.Logf("execution avg per rule: %s", time.Since(start)/time.Duration(testMax))
+		})
+	}
+}

client/internal/acl/manager.go

@@ -0,0 +1,493 @@
+package acl
+
+import (
+	"crypto/md5"
+	"encoding/hex"
+	"fmt"
+	"net"
+	"strconv"
+	"sync"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/client/ssh"
+	"github.com/netbirdio/netbird/iface"
+	mgmProto "github.com/netbirdio/netbird/management/proto"
+)
+
+// IFaceMapper defines subset methods of interface required for manager
+type IFaceMapper interface {
+	Name() string
+	Address() iface.WGAddress
+	IsUserspaceBind() bool
+	SetFilter(iface.PacketFilter) error
+}
+
+// Manager is a ACL rules manager
+type Manager interface {
+	ApplyFiltering(networkMap *mgmProto.NetworkMap)
+	Stop()
+}
+
+// DefaultManager uses firewall manager to handle
+type DefaultManager struct {
+	manager      firewall.Manager
+	ipsetCounter int
+	rulesPairs   map[string][]firewall.Rule
+	mutex        sync.Mutex
+}
+
+type ipsetInfo struct {
+	name    string
+	ipCount int
+}
+
+func newDefaultManager(fm firewall.Manager) *DefaultManager {
+	return &DefaultManager{
+		manager:    fm,
+		rulesPairs: make(map[string][]firewall.Rule),
+	}
+}
+
+// ApplyFiltering firewall rules to the local firewall manager processed by ACL policy.
+//
+// If allowByDefault is ture it appends allow ALL traffic rules to input and output chains.
+func (d *DefaultManager) ApplyFiltering(networkMap *mgmProto.NetworkMap) {
+	d.mutex.Lock()
+	defer d.mutex.Unlock()
+
+	start := time.Now()
+	defer func() {
+		total := 0
+		for _, pairs := range d.rulesPairs {
+			total += len(pairs)
+		}
+		log.Infof(
+			"ACL rules processed in: %v, total rules count: %d",
+			time.Since(start), total)
+	}()
+
+	if d.manager == nil {
+		log.Debug("firewall manager is not supported, skipping firewall rules")
+		return
+	}
+
+	defer func() {
+		if err := d.manager.Flush(); err != nil {
+			log.Error("failed to flush firewall rules: ", err)
+		}
+	}()
+
+	rules, squashedProtocols := d.squashAcceptRules(networkMap)
+
+	enableSSH := (networkMap.PeerConfig != nil &&
+		networkMap.PeerConfig.SshConfig != nil &&
+		networkMap.PeerConfig.SshConfig.SshEnabled)
+	if _, ok := squashedProtocols[mgmProto.FirewallRule_ALL]; ok {
+		enableSSH = enableSSH && !ok
+	}
+	if _, ok := squashedProtocols[mgmProto.FirewallRule_TCP]; ok {
+		enableSSH = enableSSH && !ok
+	}
+
+	// if TCP protocol rules not squashed and SSH enabled
+	// we add default firewall rule which accepts connection to any peer
+	// in the network by SSH (TCP 22 port).
+	if enableSSH {
+		rules = append(rules, &mgmProto.FirewallRule{
+			PeerIP:    "0.0.0.0",
+			Direction: mgmProto.FirewallRule_IN,
+			Action:    mgmProto.FirewallRule_ACCEPT,
+			Protocol:  mgmProto.FirewallRule_TCP,
+			Port:      strconv.Itoa(ssh.DefaultSSHPort),
+		})
+	}
+
+	// if we got empty rules list but management not set networkMap.FirewallRulesIsEmpty flag
+	// we have old version of management without rules handling, we should allow all traffic
+	if len(networkMap.FirewallRules) == 0 && !networkMap.FirewallRulesIsEmpty {
+		log.Warn("this peer is connected to a NetBird Management service with an older version. Allowing all traffic from connected peers")
+		rules = append(rules,
+			&mgmProto.FirewallRule{
+				PeerIP:    "0.0.0.0",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			&mgmProto.FirewallRule{
+				PeerIP:    "0.0.0.0",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+		)
+	}
+
+	applyFailed := false
+	newRulePairs := make(map[string][]firewall.Rule)
+	ipsetByRuleSelectors := make(map[string]*ipsetInfo)
+
+	// calculate which IP's can be grouped in by which ipset
+	// to do that we use rule selector (which is just rule properties without IP's)
+	for _, r := range rules {
+		selector := d.getRuleGroupingSelector(r)
+		ipset, ok := ipsetByRuleSelectors[selector]
+		if !ok {
+			ipset = &ipsetInfo{}
+		}
+
+		ipset.ipCount++
+		ipsetByRuleSelectors[selector] = ipset
+	}
+
+	for _, r := range rules {
+		// if this rule is member of rule selection with more than DefaultIPsCountForSet
+		// it's IP address can be used in the ipset for firewall manager which supports it
+		ipset := ipsetByRuleSelectors[d.getRuleGroupingSelector(r)]
+		ipsetName := ""
+		if ipset.name == "" {
+			d.ipsetCounter++
+			ipset.name = fmt.Sprintf("nb%07d", d.ipsetCounter)
+		}
+		ipsetName = ipset.name
+		pairID, rulePair, err := d.protoRuleToFirewallRule(r, ipsetName)
+		if err != nil {
+			log.Errorf("failed to apply firewall rule: %+v, %v", r, err)
+			applyFailed = true
+			break
+		}
+		newRulePairs[pairID] = rulePair
+	}
+	if applyFailed {
+		log.Error("failed to apply firewall rules, rollback ACL to previous state")
+		for _, rules := range newRulePairs {
+			for _, rule := range rules {
+				if err := d.manager.DeleteRule(rule); err != nil {
+					log.Errorf("failed to delete new firewall rule (id: %v) during rollback: %v", rule.GetRuleID(), err)
+					continue
+				}
+			}
+		}
+		return
+	}
+
+	for pairID, rules := range d.rulesPairs {
+		if _, ok := newRulePairs[pairID]; !ok {
+			for _, rule := range rules {
+				if err := d.manager.DeleteRule(rule); err != nil {
+					log.Errorf("failed to delete firewall rule: %v", err)
+					continue
+				}
+			}
+			delete(d.rulesPairs, pairID)
+		}
+	}
+	d.rulesPairs = newRulePairs
+}
+
+// Stop ACL controller and clear firewall state
+func (d *DefaultManager) Stop() {
+	d.mutex.Lock()
+	defer d.mutex.Unlock()
+
+	if err := d.manager.Reset(); err != nil {
+		log.WithError(err).Error("reset firewall state")
+	}
+}
+
+func (d *DefaultManager) protoRuleToFirewallRule(
+	r *mgmProto.FirewallRule,
+	ipsetName string,
+) (string, []firewall.Rule, error) {
+	ip := net.ParseIP(r.PeerIP)
+	if ip == nil {
+		return "", nil, fmt.Errorf("invalid IP address, skipping firewall rule")
+	}
+
+	protocol := convertToFirewallProtocol(r.Protocol)
+	if protocol == firewall.ProtocolUnknown {
+		return "", nil, fmt.Errorf("invalid protocol type: %d, skipping firewall rule", r.Protocol)
+	}
+
+	action := convertFirewallAction(r.Action)
+	if action == firewall.ActionUnknown {
+		return "", nil, fmt.Errorf("invalid action type: %d, skipping firewall rule", r.Action)
+	}
+
+	var port *firewall.Port
+	if r.Port != "" {
+		value, err := strconv.Atoi(r.Port)
+		if err != nil {
+			return "", nil, fmt.Errorf("invalid port, skipping firewall rule")
+		}
+		port = &firewall.Port{
+			Values: []int{value},
+		}
+	}
+
+	ruleID := d.getRuleID(ip, protocol, int(r.Direction), port, action, "")
+	if rulesPair, ok := d.rulesPairs[ruleID]; ok {
+		return ruleID, rulesPair, nil
+	}
+
+	var rules []firewall.Rule
+	var err error
+	switch r.Direction {
+	case mgmProto.FirewallRule_IN:
+		rules, err = d.addInRules(ip, protocol, port, action, ipsetName, "")
+	case mgmProto.FirewallRule_OUT:
+		rules, err = d.addOutRules(ip, protocol, port, action, ipsetName, "")
+	default:
+		return "", nil, fmt.Errorf("invalid direction, skipping firewall rule")
+	}
+
+	if err != nil {
+		return "", nil, err
+	}
+
+	d.rulesPairs[ruleID] = rules
+	return ruleID, rules, nil
+}
+
+func (d *DefaultManager) addInRules(
+	ip net.IP,
+	protocol firewall.Protocol,
+	port *firewall.Port,
+	action firewall.Action,
+	ipsetName string,
+	comment string,
+) ([]firewall.Rule, error) {
+	var rules []firewall.Rule
+	rule, err := d.manager.AddFiltering(
+		ip, protocol, nil, port, firewall.RuleDirectionIN, action, ipsetName, comment)
+	if err != nil {
+		return nil, fmt.Errorf("failed to add firewall rule: %v", err)
+	}
+	rules = append(rules, rule)
+
+	if shouldSkipInvertedRule(protocol, port) {
+		return rules, nil
+	}
+
+	rule, err = d.manager.AddFiltering(
+		ip, protocol, port, nil, firewall.RuleDirectionOUT, action, ipsetName, comment)
+	if err != nil {
+		return nil, fmt.Errorf("failed to add firewall rule: %v", err)
+	}
+
+	return append(rules, rule), nil
+}
+
+func (d *DefaultManager) addOutRules(
+	ip net.IP,
+	protocol firewall.Protocol,
+	port *firewall.Port,
+	action firewall.Action,
+	ipsetName string,
+	comment string,
+) ([]firewall.Rule, error) {
+	var rules []firewall.Rule
+	rule, err := d.manager.AddFiltering(
+		ip, protocol, nil, port, firewall.RuleDirectionOUT, action, ipsetName, comment)
+	if err != nil {
+		return nil, fmt.Errorf("failed to add firewall rule: %v", err)
+	}
+	rules = append(rules, rule)
+
+	if shouldSkipInvertedRule(protocol, port) {
+		return rules, nil
+	}
+
+	rule, err = d.manager.AddFiltering(
+		ip, protocol, port, nil, firewall.RuleDirectionIN, action, ipsetName, comment)
+	if err != nil {
+		return nil, fmt.Errorf("failed to add firewall rule: %v", err)
+	}
+
+	return append(rules, rule), nil
+}
+
+// getRuleID() returns unique ID for the rule based on its parameters.
+func (d *DefaultManager) getRuleID(
+	ip net.IP,
+	proto firewall.Protocol,
+	direction int,
+	port *firewall.Port,
+	action firewall.Action,
+	comment string,
+) string {
+	idStr := ip.String() + string(proto) + strconv.Itoa(direction) + strconv.Itoa(int(action)) + comment
+	if port != nil {
+		idStr += port.String()
+	}
+
+	return hex.EncodeToString(md5.New().Sum([]byte(idStr)))
+}
+
+// squashAcceptRules does complex logic to convert many rules which allows connection by traffic type
+// to all peers in the network map to one rule which just accepts that type of the traffic.
+//
+// NOTE: It will not squash two rules for same protocol if one covers all peers in the network,
+// but other has port definitions or has drop policy.
+func (d *DefaultManager) squashAcceptRules(
+	networkMap *mgmProto.NetworkMap,
+) ([]*mgmProto.FirewallRule, map[mgmProto.FirewallRuleProtocol]struct{}) {
+	totalIPs := 0
+	for _, p := range append(networkMap.RemotePeers, networkMap.OfflinePeers...) {
+		for range p.AllowedIps {
+			totalIPs++
+		}
+	}
+
+	type protoMatch map[mgmProto.FirewallRuleProtocol]map[string]int
+
+	in := protoMatch{}
+	out := protoMatch{}
+
+	// trace which type of protocols was squashed
+	squashedRules := []*mgmProto.FirewallRule{}
+	squashedProtocols := map[mgmProto.FirewallRuleProtocol]struct{}{}
+
+	// this function we use to do calculation, can we squash the rules by protocol or not.
+	// We summ amount of Peers IP for given protocol we found in original rules list.
+	// But we zeroed the IP's for protocol if:
+	// 1. Any of the rule has DROP action type.
+	// 2. Any of rule contains Port.
+	//
+	// We zeroed this to notify squash function that this protocol can't be squashed.
+	addRuleToCalculationMap := func(i int, r *mgmProto.FirewallRule, protocols protoMatch) {
+		drop := r.Action == mgmProto.FirewallRule_DROP || r.Port != ""
+		if drop {
+			protocols[r.Protocol] = map[string]int{}
+			return
+		}
+		if _, ok := protocols[r.Protocol]; !ok {
+			protocols[r.Protocol] = map[string]int{}
+		}
+
+		// special case, when we recieve this all network IP address
+		// it means that rules for that protocol was already optimized on the
+		// management side
+		if r.PeerIP == "0.0.0.0" {
+			squashedRules = append(squashedRules, r)
+			squashedProtocols[r.Protocol] = struct{}{}
+			return
+		}
+
+		ipset := protocols[r.Protocol]
+
+		if _, ok := ipset[r.PeerIP]; ok {
+			return
+		}
+		ipset[r.PeerIP] = i
+	}
+
+	for i, r := range networkMap.FirewallRules {
+		// calculate squash for different directions
+		if r.Direction == mgmProto.FirewallRule_IN {
+			addRuleToCalculationMap(i, r, in)
+		} else {
+			addRuleToCalculationMap(i, r, out)
+		}
+	}
+
+	// order of squashing by protocol is important
+	// only for ther first element ALL, it must be done first
+	protocolOrders := []mgmProto.FirewallRuleProtocol{
+		mgmProto.FirewallRule_ALL,
+		mgmProto.FirewallRule_ICMP,
+		mgmProto.FirewallRule_TCP,
+		mgmProto.FirewallRule_UDP,
+	}
+
+	squash := func(matches protoMatch, direction mgmProto.FirewallRuleDirection) {
+		for _, protocol := range protocolOrders {
+			if ipset, ok := matches[protocol]; !ok || len(ipset) != totalIPs || len(ipset) < 2 {
+				// don't squash if :
+				// 1. Rules not cover all peers in the network
+				// 2. Rules cover only one peer in the network.
+				continue
+			}
+
+			// add special rule 0.0.0.0 which allows all IP's in our firewall implementations
+			squashedRules = append(squashedRules, &mgmProto.FirewallRule{
+				PeerIP:    "0.0.0.0",
+				Direction: direction,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  protocol,
+			})
+			squashedProtocols[protocol] = struct{}{}
+
+			if protocol == mgmProto.FirewallRule_ALL {
+				// if we have ALL traffic type squashed rule
+				// it allows all other type of traffic, so we can stop processing
+				break
+			}
+		}
+	}
+
+	squash(in, mgmProto.FirewallRule_IN)
+	squash(out, mgmProto.FirewallRule_OUT)
+
+	// if all protocol was squashed everything is allow and we can ignore all other rules
+	if _, ok := squashedProtocols[mgmProto.FirewallRule_ALL]; ok {
+		return squashedRules, squashedProtocols
+	}
+
+	if len(squashedRules) == 0 {
+		return networkMap.FirewallRules, squashedProtocols
+	}
+
+	var rules []*mgmProto.FirewallRule
+	// filter out rules which was squashed from final list
+	// if we also have other not squashed rules.
+	for i, r := range networkMap.FirewallRules {
+		if _, ok := squashedProtocols[r.Protocol]; ok {
+			if m, ok := in[r.Protocol]; ok && m[r.PeerIP] == i {
+				continue
+			} else if m, ok := out[r.Protocol]; ok && m[r.PeerIP] == i {
+				continue
+			}
+		}
+		rules = append(rules, r)
+	}
+
+	return append(rules, squashedRules...), squashedProtocols
+}
+
+// getRuleGroupingSelector takes all rule properties except IP address to build selector
+func (d *DefaultManager) getRuleGroupingSelector(rule *mgmProto.FirewallRule) string {
+	return fmt.Sprintf("%v:%v:%v:%s", strconv.Itoa(int(rule.Direction)), rule.Action, rule.Protocol, rule.Port)
+}
+
+func convertToFirewallProtocol(protocol mgmProto.FirewallRuleProtocol) firewall.Protocol {
+	switch protocol {
+	case mgmProto.FirewallRule_TCP:
+		return firewall.ProtocolTCP
+	case mgmProto.FirewallRule_UDP:
+		return firewall.ProtocolUDP
+	case mgmProto.FirewallRule_ICMP:
+		return firewall.ProtocolICMP
+	case mgmProto.FirewallRule_ALL:
+		return firewall.ProtocolALL
+	default:
+		return firewall.ProtocolUnknown
+	}
+}
+
+func shouldSkipInvertedRule(protocol firewall.Protocol, port *firewall.Port) bool {
+	return protocol == firewall.ProtocolALL || protocol == firewall.ProtocolICMP || port == nil
+}
+
+func convertFirewallAction(action mgmProto.FirewallRuleAction) firewall.Action {
+	switch action {
+	case mgmProto.FirewallRule_ACCEPT:
+		return firewall.ActionAccept
+	case mgmProto.FirewallRule_DROP:
+		return firewall.ActionDrop
+	default:
+		return firewall.ActionUnknown
+	}
+}

client/internal/acl/manager_create.go

@@ -0,0 +1,23 @@
+//go:build !linux
+
+package acl
+
+import (
+	"fmt"
+	"runtime"
+
+	"github.com/netbirdio/netbird/client/firewall/uspfilter"
+)
+
+// Create creates a firewall manager instance
+func Create(iface IFaceMapper) (manager *DefaultManager, err error) {
+	if iface.IsUserspaceBind() {
+		// use userspace packet filtering firewall
+		fm, err := uspfilter.Create(iface)
+		if err != nil {
+			return nil, err
+		}
+		return newDefaultManager(fm), nil
+	}
+	return nil, fmt.Errorf("not implemented for this OS: %s", runtime.GOOS)
+}

client/internal/acl/manager_create_linux.go

@@ -0,0 +1,33 @@
+package acl
+
+import (
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/client/firewall/iptables"
+	"github.com/netbirdio/netbird/client/firewall/nftables"
+	"github.com/netbirdio/netbird/client/firewall/uspfilter"
+)
+
+// Create creates a firewall manager instance for the Linux
+func Create(iface IFaceMapper) (manager *DefaultManager, err error) {
+	var fm firewall.Manager
+	if iface.IsUserspaceBind() {
+		// use userspace packet filtering firewall
+		if fm, err = uspfilter.Create(iface); err != nil {
+			log.Debugf("failed to create userspace filtering firewall: %s", err)
+			return nil, err
+		}
+	} else {
+		if fm, err = nftables.Create(iface); err != nil {
+			log.Debugf("failed to create nftables manager: %s", err)
+			// fallback to iptables
+			if fm, err = iptables.Create(iface); err != nil {
+				log.Errorf("failed to create iptables manager: %s", err)
+				return nil, err
+			}
+		}
+	}
+
+	return newDefaultManager(fm), nil
+}

client/internal/acl/manager_test.go

@@ -0,0 +1,333 @@
+package acl
+
+import (
+	"testing"
+
+	"github.com/golang/mock/gomock"
+
+	"github.com/netbirdio/netbird/client/internal/acl/mocks"
+	mgmProto "github.com/netbirdio/netbird/management/proto"
+)
+
+func TestDefaultManager(t *testing.T) {
+	networkMap := &mgmProto.NetworkMap{
+		FirewallRules: []*mgmProto.FirewallRule{
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_TCP,
+				Port:      "80",
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_DROP,
+				Protocol:  mgmProto.FirewallRule_UDP,
+				Port:      "53",
+			},
+		},
+	}
+
+	ctrl := gomock.NewController(t)
+	defer ctrl.Finish()
+
+	iface := mocks.NewMockIFaceMapper(ctrl)
+	iface.EXPECT().IsUserspaceBind().Return(true)
+	// iface.EXPECT().Name().Return("lo")
+	iface.EXPECT().SetFilter(gomock.Any())
+
+	// we receive one rule from the management so for testing purposes ignore it
+	acl, err := Create(iface)
+	if err != nil {
+		t.Errorf("create ACL manager: %v", err)
+		return
+	}
+	defer acl.Stop()
+
+	t.Run("apply firewall rules", func(t *testing.T) {
+		acl.ApplyFiltering(networkMap)
+
+		if len(acl.rulesPairs) != 2 {
+			t.Errorf("firewall rules not applied: %v", acl.rulesPairs)
+			return
+		}
+	})
+
+	t.Run("add extra rules", func(t *testing.T) {
+		existedPairs := map[string]struct{}{}
+		for id := range acl.rulesPairs {
+			existedPairs[id] = struct{}{}
+		}
+
+		// remove first rule
+		networkMap.FirewallRules = networkMap.FirewallRules[1:]
+		networkMap.FirewallRules = append(
+			networkMap.FirewallRules,
+			&mgmProto.FirewallRule{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_DROP,
+				Protocol:  mgmProto.FirewallRule_ICMP,
+			},
+		)
+
+		acl.ApplyFiltering(networkMap)
+
+		// we should have one old and one new rule in the existed rules
+		if len(acl.rulesPairs) != 2 {
+			t.Errorf("firewall rules not applied")
+			return
+		}
+
+		// check that old rule was removed
+		previousCount := 0
+		for id := range acl.rulesPairs {
+			if _, ok := existedPairs[id]; ok {
+				previousCount++
+			}
+		}
+		if previousCount != 1 {
+			t.Errorf("old rule was not removed")
+		}
+	})
+
+	t.Run("handle default rules", func(t *testing.T) {
+		networkMap.FirewallRules = networkMap.FirewallRules[:0]
+
+		networkMap.FirewallRulesIsEmpty = true
+		if acl.ApplyFiltering(networkMap); len(acl.rulesPairs) != 0 {
+			t.Errorf("rules should be empty if FirewallRulesIsEmpty is set, got: %v", len(acl.rulesPairs))
+			return
+		}
+
+		networkMap.FirewallRulesIsEmpty = false
+		acl.ApplyFiltering(networkMap)
+		if len(acl.rulesPairs) != 2 {
+			t.Errorf("rules should contain 2 rules if FirewallRulesIsEmpty is not set, got: %v", len(acl.rulesPairs))
+			return
+		}
+	})
+}
+
+func TestDefaultManagerSquashRules(t *testing.T) {
+	networkMap := &mgmProto.NetworkMap{
+		RemotePeers: []*mgmProto.RemotePeerConfig{
+			{AllowedIps: []string{"10.93.0.1"}},
+			{AllowedIps: []string{"10.93.0.2"}},
+			{AllowedIps: []string{"10.93.0.3"}},
+			{AllowedIps: []string{"10.93.0.4"}},
+		},
+		FirewallRules: []*mgmProto.FirewallRule{
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.4",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.4",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+		},
+	}
+
+	manager := &DefaultManager{}
+	rules, _ := manager.squashAcceptRules(networkMap)
+	if len(rules) != 2 {
+		t.Errorf("rules should contain 2, got: %v", rules)
+		return
+	}
+
+	r := rules[0]
+	if r.PeerIP != "0.0.0.0" {
+		t.Errorf("IP should be 0.0.0.0, got: %v", r.PeerIP)
+		return
+	} else if r.Direction != mgmProto.FirewallRule_IN {
+		t.Errorf("direction should be IN, got: %v", r.Direction)
+		return
+	} else if r.Protocol != mgmProto.FirewallRule_ALL {
+		t.Errorf("protocol should be ALL, got: %v", r.Protocol)
+		return
+	} else if r.Action != mgmProto.FirewallRule_ACCEPT {
+		t.Errorf("action should be ACCEPT, got: %v", r.Action)
+		return
+	}
+
+	r = rules[1]
+	if r.PeerIP != "0.0.0.0" {
+		t.Errorf("IP should be 0.0.0.0, got: %v", r.PeerIP)
+		return
+	} else if r.Direction != mgmProto.FirewallRule_OUT {
+		t.Errorf("direction should be OUT, got: %v", r.Direction)
+		return
+	} else if r.Protocol != mgmProto.FirewallRule_ALL {
+		t.Errorf("protocol should be ALL, got: %v", r.Protocol)
+		return
+	} else if r.Action != mgmProto.FirewallRule_ACCEPT {
+		t.Errorf("action should be ACCEPT, got: %v", r.Action)
+		return
+	}
+}
+
+func TestDefaultManagerSquashRulesNoAffect(t *testing.T) {
+	networkMap := &mgmProto.NetworkMap{
+		RemotePeers: []*mgmProto.RemotePeerConfig{
+			{AllowedIps: []string{"10.93.0.1"}},
+			{AllowedIps: []string{"10.93.0.2"}},
+			{AllowedIps: []string{"10.93.0.3"}},
+			{AllowedIps: []string{"10.93.0.4"}},
+		},
+		FirewallRules: []*mgmProto.FirewallRule{
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.4",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_TCP,
+			},
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_ALL,
+			},
+			{
+				PeerIP:    "10.93.0.4",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_UDP,
+			},
+		},
+	}
+
+	manager := &DefaultManager{}
+	if rules, _ := manager.squashAcceptRules(networkMap); len(rules) != len(networkMap.FirewallRules) {
+		t.Errorf("we should got same amount of rules as intput, got %v", len(rules))
+	}
+}
+
+func TestDefaultManagerEnableSSHRules(t *testing.T) {
+	networkMap := &mgmProto.NetworkMap{
+		PeerConfig: &mgmProto.PeerConfig{
+			SshConfig: &mgmProto.SSHConfig{
+				SshEnabled: true,
+			},
+		},
+		RemotePeers: []*mgmProto.RemotePeerConfig{
+			{AllowedIps: []string{"10.93.0.1"}},
+			{AllowedIps: []string{"10.93.0.2"}},
+			{AllowedIps: []string{"10.93.0.3"}},
+		},
+		FirewallRules: []*mgmProto.FirewallRule{
+			{
+				PeerIP:    "10.93.0.1",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_TCP,
+			},
+			{
+				PeerIP:    "10.93.0.2",
+				Direction: mgmProto.FirewallRule_IN,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_TCP,
+			},
+			{
+				PeerIP:    "10.93.0.3",
+				Direction: mgmProto.FirewallRule_OUT,
+				Action:    mgmProto.FirewallRule_ACCEPT,
+				Protocol:  mgmProto.FirewallRule_UDP,
+			},
+		},
+	}
+
+	ctrl := gomock.NewController(t)
+	defer ctrl.Finish()
+
+	iface := mocks.NewMockIFaceMapper(ctrl)
+	iface.EXPECT().IsUserspaceBind().Return(true)
+	// iface.EXPECT().Name().Return("lo")
+	iface.EXPECT().SetFilter(gomock.Any())
+
+	// we receive one rule from the management so for testing purposes ignore it
+	acl, err := Create(iface)
+	if err != nil {
+		t.Errorf("create ACL manager: %v", err)
+		return
+	}
+	defer acl.Stop()
+
+	acl.ApplyFiltering(networkMap)
+
+	if len(acl.rulesPairs) != 4 {
+		t.Errorf("expect 4 rules (last must be SSH), got: %d", len(acl.rulesPairs))
+		return
+	}
+}

client/internal/acl/mocks/README.md

@@ -0,0 +1,7 @@
+## Mocks
+
+To generate (or refresh) mocks from acl package please install [mockgen](https://github.com/golang/mock).
+Run this command from the `./client/internal/acl` folder to update iface mapper interface mock:
+```bash
+mockgen -destination mocks/iface_mapper.go -package mocks . IFaceMapper
+```

client/internal/acl/mocks/iface_mapper.go

@@ -0,0 +1,91 @@
+// Code generated by MockGen. DO NOT EDIT.
+// Source: github.com/netbirdio/netbird/client/internal/acl (interfaces: IFaceMapper)
+
+// Package mocks is a generated GoMock package.
+package mocks
+
+import (
+	reflect "reflect"
+
+	gomock "github.com/golang/mock/gomock"
+	iface "github.com/netbirdio/netbird/iface"
+)
+
+// MockIFaceMapper is a mock of IFaceMapper interface.
+type MockIFaceMapper struct {
+	ctrl     *gomock.Controller
+	recorder *MockIFaceMapperMockRecorder
+}
+
+// MockIFaceMapperMockRecorder is the mock recorder for MockIFaceMapper.
+type MockIFaceMapperMockRecorder struct {
+	mock *MockIFaceMapper
+}
+
+// NewMockIFaceMapper creates a new mock instance.
+func NewMockIFaceMapper(ctrl *gomock.Controller) *MockIFaceMapper {
+	mock := &MockIFaceMapper{ctrl: ctrl}
+	mock.recorder = &MockIFaceMapperMockRecorder{mock}
+	return mock
+}
+
+// EXPECT returns an object that allows the caller to indicate expected use.
+func (m *MockIFaceMapper) EXPECT() *MockIFaceMapperMockRecorder {
+	return m.recorder
+}
+
+// Address mocks base method.
+func (m *MockIFaceMapper) Address() iface.WGAddress {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "Address")
+	ret0, _ := ret[0].(iface.WGAddress)
+	return ret0
+}
+
+// Address indicates an expected call of Address.
+func (mr *MockIFaceMapperMockRecorder) Address() *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Address", reflect.TypeOf((*MockIFaceMapper)(nil).Address))
+}
+
+// IsUserspaceBind mocks base method.
+func (m *MockIFaceMapper) IsUserspaceBind() bool {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "IsUserspaceBind")
+	ret0, _ := ret[0].(bool)
+	return ret0
+}
+
+// IsUserspaceBind indicates an expected call of IsUserspaceBind.
+func (mr *MockIFaceMapperMockRecorder) IsUserspaceBind() *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "IsUserspaceBind", reflect.TypeOf((*MockIFaceMapper)(nil).IsUserspaceBind))
+}
+
+// Name mocks base method.
+func (m *MockIFaceMapper) Name() string {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "Name")
+	ret0, _ := ret[0].(string)
+	return ret0
+}
+
+// Name indicates an expected call of Name.
+func (mr *MockIFaceMapperMockRecorder) Name() *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Name", reflect.TypeOf((*MockIFaceMapper)(nil).Name))
+}
+
+// SetFilter mocks base method.
+func (m *MockIFaceMapper) SetFilter(arg0 iface.PacketFilter) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "SetFilter", arg0)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// SetFilter indicates an expected call of SetFilter.
+func (mr *MockIFaceMapperMockRecorder) SetFilter(arg0 interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SetFilter", reflect.TypeOf((*MockIFaceMapper)(nil).SetFilter), arg0)
+}

client/internal/auth/device_flow.go

@@ -0,0 +1,202 @@
+package auth
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"github.com/netbirdio/netbird/client/internal"
+	"io"
+	"net/http"
+	"net/url"
+	"strings"
+	"time"
+)
+
+// HostedGrantType grant type for device flow on Hosted
+const (
+	HostedGrantType = "urn:ietf:params:oauth:grant-type:device_code"
+)
+
+var _ OAuthFlow = &DeviceAuthorizationFlow{}
+
+// DeviceAuthorizationFlow implements the OAuthFlow interface,
+// for the Device Authorization Flow.
+type DeviceAuthorizationFlow struct {
+	providerConfig internal.DeviceAuthProviderConfig
+
+	HTTPClient HTTPClient
+}
+
+// RequestDeviceCodePayload used for request device code payload for auth0
+type RequestDeviceCodePayload struct {
+	Audience string `json:"audience"`
+	ClientID string `json:"client_id"`
+	Scope    string `json:"scope"`
+}
+
+// TokenRequestPayload used for requesting the auth0 token
+type TokenRequestPayload struct {
+	GrantType    string `json:"grant_type"`
+	DeviceCode   string `json:"device_code,omitempty"`
+	ClientID     string `json:"client_id"`
+	RefreshToken string `json:"refresh_token,omitempty"`
+}
+
+// TokenRequestResponse used for parsing Hosted token's response
+type TokenRequestResponse struct {
+	Error            string `json:"error"`
+	ErrorDescription string `json:"error_description"`
+	TokenInfo
+}
+
+// NewDeviceAuthorizationFlow returns device authorization flow client
+func NewDeviceAuthorizationFlow(config internal.DeviceAuthProviderConfig) (*DeviceAuthorizationFlow, error) {
+	httpTransport := http.DefaultTransport.(*http.Transport).Clone()
+	httpTransport.MaxIdleConns = 5
+
+	httpClient := &http.Client{
+		Timeout:   10 * time.Second,
+		Transport: httpTransport,
+	}
+
+	return &DeviceAuthorizationFlow{
+		providerConfig: config,
+		HTTPClient:     httpClient,
+	}, nil
+}
+
+// GetClientID returns the provider client id
+func (d *DeviceAuthorizationFlow) GetClientID(ctx context.Context) string {
+	return d.providerConfig.ClientID
+}
+
+// RequestAuthInfo requests a device code login flow information from Hosted
+func (d *DeviceAuthorizationFlow) RequestAuthInfo(ctx context.Context) (AuthFlowInfo, error) {
+	form := url.Values{}
+	form.Add("client_id", d.providerConfig.ClientID)
+	form.Add("audience", d.providerConfig.Audience)
+	form.Add("scope", d.providerConfig.Scope)
+	req, err := http.NewRequest("POST", d.providerConfig.DeviceAuthEndpoint,
+		strings.NewReader(form.Encode()))
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("creating request failed with error: %v", err)
+	}
+	req.Header.Add("Content-Type", "application/x-www-form-urlencoded")
+
+	res, err := d.HTTPClient.Do(req)
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("doing request failed with error: %v", err)
+	}
+
+	defer res.Body.Close()
+	body, err := io.ReadAll(res.Body)
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("reading body failed with error: %v", err)
+	}
+
+	if res.StatusCode != 200 {
+		return AuthFlowInfo{}, fmt.Errorf("request device code returned status %d error: %s", res.StatusCode, string(body))
+	}
+
+	deviceCode := AuthFlowInfo{}
+	err = json.Unmarshal(body, &deviceCode)
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("unmarshaling response failed with error: %v", err)
+	}
+
+	// Fallback to the verification_uri if the IdP doesn't support verification_uri_complete
+	if deviceCode.VerificationURIComplete == "" {
+		deviceCode.VerificationURIComplete = deviceCode.VerificationURI
+	}
+
+	return deviceCode, err
+}
+
+func (d *DeviceAuthorizationFlow) requestToken(info AuthFlowInfo) (TokenRequestResponse, error) {
+	form := url.Values{}
+	form.Add("client_id", d.providerConfig.ClientID)
+	form.Add("grant_type", HostedGrantType)
+	form.Add("device_code", info.DeviceCode)
+
+	req, err := http.NewRequest("POST", d.providerConfig.TokenEndpoint, strings.NewReader(form.Encode()))
+	if err != nil {
+		return TokenRequestResponse{}, fmt.Errorf("failed to create request access token: %v", err)
+	}
+	req.Header.Add("Content-Type", "application/x-www-form-urlencoded")
+
+	res, err := d.HTTPClient.Do(req)
+	if err != nil {
+		return TokenRequestResponse{}, fmt.Errorf("failed to request access token with error: %v", err)
+	}
+
+	defer func() {
+		err := res.Body.Close()
+		if err != nil {
+			return
+		}
+	}()
+
+	body, err := io.ReadAll(res.Body)
+	if err != nil {
+		return TokenRequestResponse{}, fmt.Errorf("failed reading access token response body with error: %v", err)
+	}
+
+	if res.StatusCode > 499 {
+		return TokenRequestResponse{}, fmt.Errorf("access token response returned code: %s", string(body))
+	}
+
+	tokenResponse := TokenRequestResponse{}
+	err = json.Unmarshal(body, &tokenResponse)
+	if err != nil {
+		return TokenRequestResponse{}, fmt.Errorf("parsing token response failed with error: %v", err)
+	}
+
+	return tokenResponse, nil
+}
+
+// WaitToken waits user's login and authorize the app. Once the user's authorize
+// it retrieves the access token from Hosted's endpoint and validates it before returning
+func (d *DeviceAuthorizationFlow) WaitToken(ctx context.Context, info AuthFlowInfo) (TokenInfo, error) {
+	interval := time.Duration(info.Interval) * time.Second
+	ticker := time.NewTicker(interval)
+	for {
+		select {
+		case <-ctx.Done():
+			return TokenInfo{}, ctx.Err()
+		case <-ticker.C:
+
+			tokenResponse, err := d.requestToken(info)
+			if err != nil {
+				return TokenInfo{}, fmt.Errorf("parsing token response failed with error: %v", err)
+			}
+
+			if tokenResponse.Error != "" {
+				if tokenResponse.Error == "authorization_pending" {
+					continue
+				} else if tokenResponse.Error == "slow_down" {
+					interval = interval + (3 * time.Second)
+					ticker.Reset(interval)
+					continue
+				}
+
+				return TokenInfo{}, fmt.Errorf(tokenResponse.ErrorDescription)
+			}
+
+			tokenInfo := TokenInfo{
+				AccessToken:  tokenResponse.AccessToken,
+				TokenType:    tokenResponse.TokenType,
+				RefreshToken: tokenResponse.RefreshToken,
+				IDToken:      tokenResponse.IDToken,
+				ExpiresIn:    tokenResponse.ExpiresIn,
+				UseIDToken:   d.providerConfig.UseIDToken,
+			}
+
+			err = isValidAccessToken(tokenInfo.GetTokenToUse(), d.providerConfig.Audience)
+			if err != nil {
+				return TokenInfo{}, fmt.Errorf("validate access token failed with error: %v", err)
+			}
+
+			return tokenInfo, err
+		}
+	}
+}

client/internal/auth/device_flow_test.go

@@ -1,17 +1,17 @@
-package internal
+package auth
 
 import (
 	"context"
 	"fmt"
+	"github.com/golang-jwt/jwt"
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/stretchr/testify/require"
 	"io"
 	"net/http"
 	"net/url"
 	"strings"
 	"testing"
 	"time"
-
-	"github.com/golang-jwt/jwt"
-	"github.com/stretchr/testify/require"
 )
 
 type mockHTTPClient struct {
@@ -53,7 +53,7 @@ func TestHosted_RequestDeviceCode(t *testing.T) {
 		testingErrFunc   require.ErrorAssertionFunc
 		expectedErrorMSG string
 		testingFunc      require.ComparisonAssertionFunc
-		expectedOut      DeviceAuthInfo
+		expectedOut      AuthFlowInfo
 		expectedMSG      string
 		expectPayload    string
 	}
@@ -92,7 +92,7 @@ func TestHosted_RequestDeviceCode(t *testing.T) {
 		testingFunc:      require.EqualValues,
 		expectPayload:    expectPayload,
 	}
-	testCase4Out := DeviceAuthInfo{ExpiresIn: 10}
+	testCase4Out := AuthFlowInfo{ExpiresIn: 10}
 	testCase4 := test{
 		name:           "Got Device Code",
 		inputResBody:   fmt.Sprintf("{\"expires_in\":%d}", testCase4Out.ExpiresIn),
@@ -113,8 +113,8 @@ func TestHosted_RequestDeviceCode(t *testing.T) {
 				err:     testCase.inputReqError,
 			}
 
-			hosted := Hosted{
-				providerConfig: ProviderConfig{
+			deviceFlow := &DeviceAuthorizationFlow{
+				providerConfig: internal.DeviceAuthProviderConfig{
 					Audience:           expectedAudience,
 					ClientID:           expectedClientID,
 					Scope:              expectedScope,
@@ -125,7 +125,7 @@ func TestHosted_RequestDeviceCode(t *testing.T) {
 				HTTPClient: &httpClient,
 			}
 
-			authInfo, err := hosted.RequestDeviceCode(context.TODO())
+			authInfo, err := deviceFlow.RequestAuthInfo(context.TODO())
 			testCase.testingErrFunc(t, err, testCase.expectedErrorMSG)
 
 			require.EqualValues(t, expectPayload, httpClient.reqBody, "payload should match")
@@ -145,7 +145,7 @@ func TestHosted_WaitToken(t *testing.T) {
 		inputMaxReqs      int
 		inputCountResBody string
 		inputTimeout      time.Duration
-		inputInfo         DeviceAuthInfo
+		inputInfo         AuthFlowInfo
 		inputAudience     string
 		testingErrFunc    require.ErrorAssertionFunc
 		expectedErrorMSG  string
@@ -155,7 +155,7 @@ func TestHosted_WaitToken(t *testing.T) {
 		expectPayload     string
 	}
 
-	defaultInfo := DeviceAuthInfo{
+	defaultInfo := AuthFlowInfo{
 		DeviceCode: "test",
 		ExpiresIn:  10,
 		Interval:   1,
@@ -278,8 +278,8 @@ func TestHosted_WaitToken(t *testing.T) {
 				countResBody: testCase.inputCountResBody,
 			}
 
-			hosted := Hosted{
-				providerConfig: ProviderConfig{
+			deviceFlow := DeviceAuthorizationFlow{
+				providerConfig: internal.DeviceAuthProviderConfig{
 					Audience:           testCase.inputAudience,
 					ClientID:           clientID,
 					TokenEndpoint:      "test.hosted.com/token",
@@ -287,11 +287,12 @@ func TestHosted_WaitToken(t *testing.T) {
 					Scope:              "openid",
 					UseIDToken:         false,
 				},
-				HTTPClient: &httpClient}
+				HTTPClient: &httpClient,
+			}
 
 			ctx, cancel := context.WithTimeout(context.TODO(), testCase.inputTimeout)
 			defer cancel()
-			tokenInfo, err := hosted.WaitToken(ctx, testCase.inputInfo)
+			tokenInfo, err := deviceFlow.WaitToken(ctx, testCase.inputInfo)
 			testCase.testingErrFunc(t, err, testCase.expectedErrorMSG)
 
 			require.EqualValues(t, testCase.expectPayload, httpClient.reqBody, "payload should match")

client/internal/auth/oauth.go

@@ -0,0 +1,90 @@
+package auth
+
+import (
+	"context"
+	"fmt"
+	"net/http"
+
+	log "github.com/sirupsen/logrus"
+	"google.golang.org/grpc/codes"
+	gstatus "google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/client/internal"
+)
+
+// OAuthFlow represents an interface for authorization using different OAuth 2.0 flows
+type OAuthFlow interface {
+	RequestAuthInfo(ctx context.Context) (AuthFlowInfo, error)
+	WaitToken(ctx context.Context, info AuthFlowInfo) (TokenInfo, error)
+	GetClientID(ctx context.Context) string
+}
+
+// HTTPClient http client interface for API calls
+type HTTPClient interface {
+	Do(req *http.Request) (*http.Response, error)
+}
+
+// AuthFlowInfo holds information for the OAuth 2.0  authorization flow
+type AuthFlowInfo struct {
+	DeviceCode              string `json:"device_code"`
+	UserCode                string `json:"user_code"`
+	VerificationURI         string `json:"verification_uri"`
+	VerificationURIComplete string `json:"verification_uri_complete"`
+	ExpiresIn               int    `json:"expires_in"`
+	Interval                int    `json:"interval"`
+}
+
+// Claims used when validating the access token
+type Claims struct {
+	Audience interface{} `json:"aud"`
+}
+
+// TokenInfo holds information of issued access token
+type TokenInfo struct {
+	AccessToken  string `json:"access_token"`
+	RefreshToken string `json:"refresh_token"`
+	IDToken      string `json:"id_token"`
+	TokenType    string `json:"token_type"`
+	ExpiresIn    int    `json:"expires_in"`
+	UseIDToken   bool   `json:"-"`
+}
+
+// GetTokenToUse returns either the access or id token based on UseIDToken field
+func (t TokenInfo) GetTokenToUse() string {
+	if t.UseIDToken {
+		return t.IDToken
+	}
+	return t.AccessToken
+}
+
+// NewOAuthFlow initializes and returns the appropriate OAuth flow based on the management configuration.
+func NewOAuthFlow(ctx context.Context, config *internal.Config) (OAuthFlow, error) {
+	log.Debug("getting device authorization flow info")
+
+	// Try to initialize the Device Authorization Flow
+	deviceFlowInfo, err := internal.GetDeviceAuthorizationFlowInfo(ctx, config.PrivateKey, config.ManagementURL)
+	if err == nil {
+		return NewDeviceAuthorizationFlow(deviceFlowInfo.ProviderConfig)
+	}
+
+	log.Debugf("getting device authorization flow info failed with error: %v", err)
+	log.Debugf("falling back to pkce authorization flow info")
+
+	// If Device Authorization Flow failed, try the PKCE Authorization Flow
+	pkceFlowInfo, err := internal.GetPKCEAuthorizationFlowInfo(ctx, config.PrivateKey, config.ManagementURL)
+	if err != nil {
+		s, ok := gstatus.FromError(err)
+		if ok && s.Code() == codes.NotFound {
+			return nil, fmt.Errorf("no SSO provider returned from management. " +
+				"If you are using hosting Netbird see documentation at " +
+				"https://github.com/netbirdio/netbird/tree/main/management for details")
+		} else if ok && s.Code() == codes.Unimplemented {
+			return nil, fmt.Errorf("the management server, %s, does not support SSO providers, "+
+				"please update your server or use Setup Keys to login", config.ManagementURL)
+		} else {
+			return nil, fmt.Errorf("getting pkce authorization flow info failed with error: %v", err)
+		}
+	}
+
+	return NewPKCEAuthorizationFlow(pkceFlowInfo.ProviderConfig)
+}

client/internal/auth/pkce_flow.go

@@ -0,0 +1,245 @@
+package auth
+
+import (
+	"context"
+	"crypto/sha256"
+	"crypto/subtle"
+	"encoding/base64"
+	"fmt"
+	"html/template"
+	"net"
+	"net/http"
+	"net/url"
+	"strings"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/oauth2"
+
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/templates"
+)
+
+var _ OAuthFlow = &PKCEAuthorizationFlow{}
+
+const (
+	queryState                = "state"
+	queryCode                 = "code"
+	queryError                = "error"
+	queryErrorDesc            = "error_description"
+	defaultPKCETimeoutSeconds = 300
+)
+
+// PKCEAuthorizationFlow implements the OAuthFlow interface for
+// the Authorization Code Flow with PKCE.
+type PKCEAuthorizationFlow struct {
+	providerConfig internal.PKCEAuthProviderConfig
+	state          string
+	codeVerifier   string
+	oAuthConfig    *oauth2.Config
+}
+
+// NewPKCEAuthorizationFlow returns new PKCE authorization code flow.
+func NewPKCEAuthorizationFlow(config internal.PKCEAuthProviderConfig) (*PKCEAuthorizationFlow, error) {
+	var availableRedirectURL string
+
+	// find the first available redirect URL
+	for _, redirectURL := range config.RedirectURLs {
+		if !isRedirectURLPortUsed(redirectURL) {
+			availableRedirectURL = redirectURL
+			break
+		}
+	}
+
+	if availableRedirectURL == "" {
+		return nil, fmt.Errorf("no available port found from configured redirect URLs: %q", config.RedirectURLs)
+	}
+
+	cfg := &oauth2.Config{
+		ClientID:     config.ClientID,
+		ClientSecret: config.ClientSecret,
+		Endpoint: oauth2.Endpoint{
+			AuthURL:  config.AuthorizationEndpoint,
+			TokenURL: config.TokenEndpoint,
+		},
+		RedirectURL: availableRedirectURL,
+		Scopes:      strings.Split(config.Scope, " "),
+	}
+
+	return &PKCEAuthorizationFlow{
+		providerConfig: config,
+		oAuthConfig:    cfg,
+	}, nil
+}
+
+// GetClientID returns the provider client id
+func (p *PKCEAuthorizationFlow) GetClientID(_ context.Context) string {
+	return p.providerConfig.ClientID
+}
+
+// RequestAuthInfo requests a authorization code login flow information.
+func (p *PKCEAuthorizationFlow) RequestAuthInfo(_ context.Context) (AuthFlowInfo, error) {
+	state, err := randomBytesInHex(24)
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("could not generate random state: %v", err)
+	}
+	p.state = state
+
+	codeVerifier, err := randomBytesInHex(64)
+	if err != nil {
+		return AuthFlowInfo{}, fmt.Errorf("could not create a code verifier: %v", err)
+	}
+	p.codeVerifier = codeVerifier
+
+	codeChallenge := createCodeChallenge(codeVerifier)
+	authURL := p.oAuthConfig.AuthCodeURL(
+		state,
+		oauth2.SetAuthURLParam("code_challenge_method", "S256"),
+		oauth2.SetAuthURLParam("code_challenge", codeChallenge),
+		oauth2.SetAuthURLParam("audience", p.providerConfig.Audience),
+	)
+
+	return AuthFlowInfo{
+		VerificationURIComplete: authURL,
+		ExpiresIn:               defaultPKCETimeoutSeconds,
+	}, nil
+}
+
+// WaitToken waits for the OAuth token in the PKCE Authorization Flow.
+// It starts an HTTP server to receive the OAuth token callback and waits for the token or an error.
+// Once the token is received, it is converted to TokenInfo and validated before returning.
+func (p *PKCEAuthorizationFlow) WaitToken(ctx context.Context, _ AuthFlowInfo) (TokenInfo, error) {
+	tokenChan := make(chan *oauth2.Token, 1)
+	errChan := make(chan error, 1)
+
+	go p.startServer(tokenChan, errChan)
+
+	select {
+	case <-ctx.Done():
+		return TokenInfo{}, ctx.Err()
+	case token := <-tokenChan:
+		return p.handleOAuthToken(token)
+	case err := <-errChan:
+		return TokenInfo{}, err
+	}
+}
+
+func (p *PKCEAuthorizationFlow) startServer(tokenChan chan<- *oauth2.Token, errChan chan<- error) {
+	parsedURL, err := url.Parse(p.oAuthConfig.RedirectURL)
+	if err != nil {
+		errChan <- fmt.Errorf("failed to parse redirect URL: %v", err)
+		return
+	}
+	port := parsedURL.Port()
+
+	server := http.Server{Addr: fmt.Sprintf(":%s", port)}
+	defer func() {
+		if err := server.Shutdown(context.Background()); err != nil {
+			log.Errorf("error while shutting down pkce flow server: %v", err)
+		}
+	}()
+
+	http.HandleFunc("/", func(w http.ResponseWriter, req *http.Request) {
+		tokenValidatorFunc := func() (*oauth2.Token, error) {
+			query := req.URL.Query()
+
+			if authError := query.Get(queryError); authError != "" {
+				authErrorDesc := query.Get(queryErrorDesc)
+				return nil, fmt.Errorf("%s.%s", authError, authErrorDesc)
+			}
+
+			// Prevent timing attacks on state
+			if state := query.Get(queryState); subtle.ConstantTimeCompare([]byte(p.state), []byte(state)) == 0 {
+				return nil, fmt.Errorf("invalid state")
+			}
+
+			code := query.Get(queryCode)
+			if code == "" {
+				return nil, fmt.Errorf("missing code")
+			}
+
+			return p.oAuthConfig.Exchange(
+				req.Context(),
+				code,
+				oauth2.SetAuthURLParam("code_verifier", p.codeVerifier),
+			)
+		}
+
+		token, err := tokenValidatorFunc()
+		if err != nil {
+			renderPKCEFlowTmpl(w, err)
+			errChan <- fmt.Errorf("PKCE authorization flow failed: %v", err)
+			return
+		}
+
+		renderPKCEFlowTmpl(w, nil)
+		tokenChan <- token
+	})
+
+	if err := server.ListenAndServe(); err != nil {
+		errChan <- err
+	}
+}
+
+func (p *PKCEAuthorizationFlow) handleOAuthToken(token *oauth2.Token) (TokenInfo, error) {
+	tokenInfo := TokenInfo{
+		AccessToken:  token.AccessToken,
+		RefreshToken: token.RefreshToken,
+		TokenType:    token.TokenType,
+		ExpiresIn:    token.Expiry.Second(),
+		UseIDToken:   p.providerConfig.UseIDToken,
+	}
+	if idToken, ok := token.Extra("id_token").(string); ok {
+		tokenInfo.IDToken = idToken
+	}
+
+	if err := isValidAccessToken(tokenInfo.GetTokenToUse(), p.providerConfig.Audience); err != nil {
+		return TokenInfo{}, fmt.Errorf("validate access token failed with error: %v", err)
+	}
+
+	return tokenInfo, nil
+}
+
+func createCodeChallenge(codeVerifier string) string {
+	sha2 := sha256.Sum256([]byte(codeVerifier))
+	return base64.RawURLEncoding.EncodeToString(sha2[:])
+}
+
+// isRedirectURLPortUsed checks if the port used in the redirect URL is in use.
+func isRedirectURLPortUsed(redirectURL string) bool {
+	parsedURL, err := url.Parse(redirectURL)
+	if err != nil {
+		log.Errorf("failed to parse redirect URL: %v", err)
+		return true
+	}
+
+	addr := fmt.Sprintf(":%s", parsedURL.Port())
+	conn, err := net.DialTimeout("tcp", addr, 3*time.Second)
+	if err != nil {
+		return false
+	}
+	defer func() {
+		if err := conn.Close(); err != nil {
+			log.Errorf("error while closing the connection: %v", err)
+		}
+	}()
+
+	return true
+}
+
+func renderPKCEFlowTmpl(w http.ResponseWriter, authError error) {
+	tmpl, err := template.New("pkce-auth-flow").Parse(templates.PKCEAuthMsgTmpl)
+	if err != nil {
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	data := make(map[string]string)
+	if authError != nil {
+		data["Error"] = authError.Error()
+	}
+
+	if err := tmpl.Execute(w, data); err != nil {
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+	}
+}

client/internal/auth/util.go

@@ -0,0 +1,62 @@
+package auth
+
+import (
+	"crypto/rand"
+	"encoding/base64"
+	"encoding/hex"
+	"encoding/json"
+	"fmt"
+	"io"
+	"reflect"
+	"strings"
+)
+
+func randomBytesInHex(count int) (string, error) {
+	buf := make([]byte, count)
+	_, err := io.ReadFull(rand.Reader, buf)
+	if err != nil {
+		return "", fmt.Errorf("could not generate %d random bytes: %v", count, err)
+	}
+
+	return hex.EncodeToString(buf), nil
+}
+
+// isValidAccessToken is a simple validation of the access token
+func isValidAccessToken(token string, audience string) error {
+	if token == "" {
+		return fmt.Errorf("token received is empty")
+	}
+
+	encodedClaims := strings.Split(token, ".")[1]
+	claimsString, err := base64.RawURLEncoding.DecodeString(encodedClaims)
+	if err != nil {
+		return err
+	}
+
+	claims := Claims{}
+	err = json.Unmarshal(claimsString, &claims)
+	if err != nil {
+		return err
+	}
+
+	if claims.Audience == nil {
+		return fmt.Errorf("required token field audience is absent")
+	}
+
+	// Audience claim of JWT can be a string or an array of strings
+	typ := reflect.TypeOf(claims.Audience)
+	switch typ.Kind() {
+	case reflect.String:
+		if claims.Audience == audience {
+			return nil
+		}
+	case reflect.Slice:
+		for _, aud := range claims.Audience.([]interface{}) {
+			if audience == aud {
+				return nil
+			}
+		}
+	}
+
+	return fmt.Errorf("invalid JWT token audience field")
+}

client/internal/config.go

@@ -215,10 +215,12 @@ func update(input ConfigInput) (*Config, error) {
 	}
 
 	if input.PreSharedKey != nil && config.PreSharedKey != *input.PreSharedKey {
-		log.Infof("new pre-shared key provided, updated to %s (old value %s)",
-			*input.PreSharedKey, config.PreSharedKey)
-		config.PreSharedKey = *input.PreSharedKey
-		refresh = true
+		if *input.PreSharedKey != "" {
+			log.Infof("new pre-shared key provides, updated to %s (old value %s)",
+				*input.PreSharedKey, config.PreSharedKey)
+			config.PreSharedKey = *input.PreSharedKey
+			refresh = true
+		}
 	}
 
 	if config.SSHKey == "" {

client/internal/config_test.go

@@ -63,7 +63,22 @@ func TestGetConfig(t *testing.T) {
 	assert.Equal(t, config.ManagementURL.String(), managementURL)
 	assert.Equal(t, config.PreSharedKey, preSharedKey)
 
-	// case 4: existing config, but new managementURL has been provided -> update config
+	// case 4: new empty pre-shared key config -> fetch it
+	newPreSharedKey := ""
+	config, err = UpdateOrCreateConfig(ConfigInput{
+		ManagementURL: managementURL,
+		AdminURL:      adminURL,
+		ConfigPath:    path,
+		PreSharedKey:  &newPreSharedKey,
+	})
+	if err != nil {
+		return
+	}
+
+	assert.Equal(t, config.ManagementURL.String(), managementURL)
+	assert.Equal(t, config.PreSharedKey, preSharedKey)
+
+	// case 5: existing config, but new managementURL has been provided -> update config
 	newManagementURL := "https://test.newManagement.url:33071"
 	config, err = UpdateOrCreateConfig(ConfigInput{
 		ManagementURL: newManagementURL,

client/internal/connect.go

@@ -12,7 +12,9 @@ import (
 	"google.golang.org/grpc/codes"
 	gstatus "google.golang.org/grpc/status"
 
+	"github.com/netbirdio/netbird/client/internal/dns"
 	"github.com/netbirdio/netbird/client/internal/peer"
+	"github.com/netbirdio/netbird/client/internal/routemanager"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
 	"github.com/netbirdio/netbird/client/ssh"
 	"github.com/netbirdio/netbird/client/system"
@@ -23,7 +25,24 @@ import (
 )
 
 // RunClient with main logic.
-func RunClient(ctx context.Context, config *Config, statusRecorder *peer.Status, tunAdapter iface.TunAdapter, iFaceDiscover stdnet.ExternalIFaceDiscover) error {
+func RunClient(ctx context.Context, config *Config, statusRecorder *peer.Status) error {
+	return runClient(ctx, config, statusRecorder, MobileDependency{})
+}
+
+// RunClientMobile with main logic on mobile system
+func RunClientMobile(ctx context.Context, config *Config, statusRecorder *peer.Status, tunAdapter iface.TunAdapter, iFaceDiscover stdnet.ExternalIFaceDiscover, routeListener routemanager.RouteListener, dnsAddresses []string, dnsReadyListener dns.ReadyListener) error {
+	// in case of non Android os these variables will be nil
+	mobileDependency := MobileDependency{
+		TunAdapter:       tunAdapter,
+		IFaceDiscover:    iFaceDiscover,
+		RouteListener:    routeListener,
+		HostDNSAddresses: dnsAddresses,
+		DnsReadyListener: dnsReadyListener,
+	}
+	return runClient(ctx, config, statusRecorder, mobileDependency)
+}
+
+func runClient(ctx context.Context, config *Config, statusRecorder *peer.Status, mobileDependency MobileDependency) error {
 	backOff := &backoff.ExponentialBackOff{
 		InitialInterval:     time.Second,
 		RandomizationFactor: 1,
@@ -144,13 +163,13 @@ func RunClient(ctx context.Context, config *Config, statusRecorder *peer.Status,
 
 		peerConfig := loginResp.GetPeerConfig()
 
-		engineConfig, err := createEngineConfig(myPrivateKey, config, peerConfig, tunAdapter, iFaceDiscover)
+		engineConfig, err := createEngineConfig(myPrivateKey, config, peerConfig)
 		if err != nil {
 			log.Error(err)
 			return wrapErr(err)
 		}
 
-		engine := NewEngine(engineCtx, cancel, signalClient, mgmClient, engineConfig, statusRecorder)
+		engine := NewEngine(engineCtx, cancel, signalClient, mgmClient, engineConfig, mobileDependency, statusRecorder)
 		err = engine.Start()
 		if err != nil {
 			log.Errorf("error while starting Netbird Connection Engine: %s", err)
@@ -194,13 +213,10 @@ func RunClient(ctx context.Context, config *Config, statusRecorder *peer.Status,
 }
 
 // createEngineConfig converts configuration received from Management Service to EngineConfig
-func createEngineConfig(key wgtypes.Key, config *Config, peerConfig *mgmProto.PeerConfig, tunAdapter iface.TunAdapter, iFaceDiscover stdnet.ExternalIFaceDiscover) (*EngineConfig, error) {
-
+func createEngineConfig(key wgtypes.Key, config *Config, peerConfig *mgmProto.PeerConfig) (*EngineConfig, error) {
 	engineConf := &EngineConfig{
 		WgIfaceName:          config.WgIface,
 		WgAddr:               peerConfig.Address,
-		TunAdapter:           tunAdapter,
-		IFaceDiscover:        iFaceDiscover,
 		IFaceBlackList:       config.IFaceBlackList,
 		DisableIPv6Discovery: config.DisableIPv6Discovery,
 		WgPrivateKey:         key,

client/internal/device_auth.go

@@ -16,11 +16,11 @@ import (
 // DeviceAuthorizationFlow represents Device Authorization Flow information
 type DeviceAuthorizationFlow struct {
 	Provider       string
-	ProviderConfig ProviderConfig
+	ProviderConfig DeviceAuthProviderConfig
 }
 
-// ProviderConfig has all attributes needed to initiate a device authorization flow
-type ProviderConfig struct {
+// DeviceAuthProviderConfig has all attributes needed to initiate a device authorization flow
+type DeviceAuthProviderConfig struct {
 	// ClientID An IDP application client id
 	ClientID string
 	// ClientSecret An IDP application client secret
@@ -88,7 +88,7 @@ func GetDeviceAuthorizationFlowInfo(ctx context.Context, privateKey string, mgmU
 	deviceAuthorizationFlow := DeviceAuthorizationFlow{
 		Provider: protoDeviceAuthorizationFlow.Provider.String(),
 
-		ProviderConfig: ProviderConfig{
+		ProviderConfig: DeviceAuthProviderConfig{
 			Audience:           protoDeviceAuthorizationFlow.GetProviderConfig().GetAudience(),
 			ClientID:           protoDeviceAuthorizationFlow.GetProviderConfig().GetClientID(),
 			ClientSecret:       protoDeviceAuthorizationFlow.GetProviderConfig().GetClientSecret(),
@@ -105,7 +105,7 @@ func GetDeviceAuthorizationFlowInfo(ctx context.Context, privateKey string, mgmU
 		deviceAuthorizationFlow.ProviderConfig.Scope = "openid"
 	}
 
-	err = isProviderConfigValid(deviceAuthorizationFlow.ProviderConfig)
+	err = isDeviceAuthProviderConfigValid(deviceAuthorizationFlow.ProviderConfig)
 	if err != nil {
 		return DeviceAuthorizationFlow{}, err
 	}
@@ -113,7 +113,7 @@ func GetDeviceAuthorizationFlowInfo(ctx context.Context, privateKey string, mgmU
 	return deviceAuthorizationFlow, nil
 }
 
-func isProviderConfigValid(config ProviderConfig) error {
+func isDeviceAuthProviderConfigValid(config DeviceAuthProviderConfig) error {
 	errorMSGFormat := "invalid provider configuration received from management: %s value is empty. Contact your NetBird administrator"
 	if config.Audience == "" {
 		return fmt.Errorf(errorMSGFormat, "Audience")

client/internal/dns/dbus_linux.go

@@ -1,3 +1,5 @@
+//go:build !android
+
 package dns
 
 import (

client/internal/dns/file_linux.go

@@ -1,3 +1,5 @@
+//go:build !android
+
 package dns
 
 import (
@@ -13,7 +15,8 @@ const (
 	fileGeneratedResolvConfSearchBeginContent = "search "
 	fileGeneratedResolvConfContentFormat      = fileGeneratedResolvConfContentHeader +
 		"\n# If needed you can restore the original file by copying back %s\n\nnameserver %s\n" +
-		fileGeneratedResolvConfSearchBeginContent + "%s\n"
+		fileGeneratedResolvConfSearchBeginContent + "%s\n\n" +
+		"%s\n"
 )
 
 const (
@@ -32,6 +35,10 @@ func newFileConfigurator() (hostManager, error) {
 	return &fileConfigurator{}, nil
 }
 
+func (f *fileConfigurator) supportCustomPort() bool {
+	return false
+}
+
 func (f *fileConfigurator) applyDNSConfig(config hostDNSConfig) error {
 	backupFileExist := false
 	_, err := os.Stat(fileDefaultResolvConfBackupLocation)
@@ -85,7 +92,12 @@ func (f *fileConfigurator) applyDNSConfig(config hostDNSConfig) error {
 		searchDomains += " " + dConf.domain
 		appendedDomains++
 	}
-	content := fmt.Sprintf(fileGeneratedResolvConfContentFormat, fileDefaultResolvConfBackupLocation, config.serverIP, searchDomains)
+
+	originalContent, err := os.ReadFile(fileDefaultResolvConfBackupLocation)
+	if err != nil {
+		log.Errorf("Could not read existing resolv.conf")
+	}
+	content := fmt.Sprintf(fileGeneratedResolvConfContentFormat, fileDefaultResolvConfBackupLocation, config.serverIP, searchDomains, string(originalContent))
 	err = writeDNSConfig(content, defaultResolvConfPath, f.originalPerms)
 	if err != nil {
 		err = f.restore()

client/internal/dns/host.go

@@ -10,6 +10,7 @@ import (
 type hostManager interface {
 	applyDNSConfig(config hostDNSConfig) error
 	restoreHostDNS() error
+	supportCustomPort() bool
 }
 
 type hostDNSConfig struct {
@@ -26,8 +27,9 @@ type domainConfig struct {
 }
 
 type mockHostConfigurator struct {
-	applyDNSConfigFunc func(config hostDNSConfig) error
-	restoreHostDNSFunc func() error
+	applyDNSConfigFunc    func(config hostDNSConfig) error
+	restoreHostDNSFunc    func() error
+	supportCustomPortFunc func() bool
 }
 
 func (m *mockHostConfigurator) applyDNSConfig(config hostDNSConfig) error {
@@ -44,10 +46,18 @@ func (m *mockHostConfigurator) restoreHostDNS() error {
 	return fmt.Errorf("method restoreHostDNS is not implemented")
 }
 
+func (m *mockHostConfigurator) supportCustomPort() bool {
+	if m.supportCustomPortFunc != nil {
+		return m.supportCustomPortFunc()
+	}
+	return false
+}
+
 func newNoopHostMocker() hostManager {
 	return &mockHostConfigurator{
-		applyDNSConfigFunc: func(config hostDNSConfig) error { return nil },
-		restoreHostDNSFunc: func() error { return nil },
+		applyDNSConfigFunc:    func(config hostDNSConfig) error { return nil },
+		restoreHostDNSFunc:    func() error { return nil },
+		supportCustomPortFunc: func() bool { return true },
 	}
 }
 

client/internal/dns/host_android.go

@@ -0,0 +1,20 @@
+package dns
+
+type androidHostManager struct {
+}
+
+func newHostManager(wgInterface WGIface) (hostManager, error) {
+	return &androidHostManager{}, nil
+}
+
+func (a androidHostManager) applyDNSConfig(config hostDNSConfig) error {
+	return nil
+}
+
+func (a androidHostManager) restoreHostDNS() error {
+	return nil
+}
+
+func (a androidHostManager) supportCustomPort() bool {
+	return false
+}

client/internal/dns/host_darwin.go

@@ -8,7 +8,6 @@ import (
 	"strconv"
 	"strings"
 
-	"github.com/netbirdio/netbird/iface"
 	log "github.com/sirupsen/logrus"
 )
 
@@ -33,12 +32,16 @@ type systemConfigurator struct {
 	createdKeys      map[string]struct{}
 }
 
-func newHostManager(_ *iface.WGIface) (hostManager, error) {
+func newHostManager(_ WGIface) (hostManager, error) {
 	return &systemConfigurator{
 		createdKeys: make(map[string]struct{}),
 	}, nil
 }
 
+func (s *systemConfigurator) supportCustomPort() bool {
+	return true
+}
+
 func (s *systemConfigurator) applyDNSConfig(config hostDNSConfig) error {
 	var err error
 
@@ -179,12 +182,11 @@ func (s *systemConfigurator) addDNSState(state, domains, dnsServer string, port
 }
 
 func (s *systemConfigurator) addDNSSetupForAll(dnsServer string, port int) error {
-	primaryServiceKey := s.getPrimaryService()
+	primaryServiceKey, existingNameserver := s.getPrimaryService()
 	if primaryServiceKey == "" {
 		return fmt.Errorf("couldn't find the primary service key")
 	}
-
-	err := s.addDNSSetup(getKeyWithInput(primaryServiceSetupKeyFormat, primaryServiceKey), dnsServer, port)
+	err := s.addDNSSetup(getKeyWithInput(primaryServiceSetupKeyFormat, primaryServiceKey), dnsServer, port, existingNameserver)
 	if err != nil {
 		return err
 	}
@@ -193,27 +195,32 @@ func (s *systemConfigurator) addDNSSetupForAll(dnsServer string, port int) error
 	return nil
 }
 
-func (s *systemConfigurator) getPrimaryService() string {
+func (s *systemConfigurator) getPrimaryService() (string, string) {
 	line := buildCommandLine("show", globalIPv4State, "")
 	stdinCommands := wrapCommand(line)
 	b, err := runSystemConfigCommand(stdinCommands)
 	if err != nil {
 		log.Error("got error while sending the command: ", err)
-		return ""
+		return "", ""
 	}
 	scanner := bufio.NewScanner(bytes.NewReader(b))
+	primaryService := ""
+	router := ""
 	for scanner.Scan() {
 		text := scanner.Text()
 		if strings.Contains(text, "PrimaryService") {
-			return strings.TrimSpace(strings.Split(text, ":")[1])
+			primaryService = strings.TrimSpace(strings.Split(text, ":")[1])
+		}
+		if strings.Contains(text, "Router") {
+			router = strings.TrimSpace(strings.Split(text, ":")[1])
 		}
 	}
-	return ""
+	return primaryService, router
 }
 
-func (s *systemConfigurator) addDNSSetup(setupKey, dnsServer string, port int) error {
+func (s *systemConfigurator) addDNSSetup(setupKey, dnsServer string, port int, existingDNSServer string) error {
 	lines := buildAddCommandLine(keySupplementalMatchDomainsNoSearch, digitSymbol+strconv.Itoa(0))
-	lines += buildAddCommandLine(keyServerAddresses, arraySymbol+dnsServer)
+	lines += buildAddCommandLine(keyServerAddresses, arraySymbol+dnsServer+" "+existingDNSServer)
 	lines += buildAddCommandLine(keyServerPort, digitSymbol+strconv.Itoa(port))
 	addDomainCommand := buildCreateStateWithOperation(setupKey, lines)
 	stdinCommands := wrapCommand(addDomainCommand)

client/internal/dns/host_linux.go

@@ -1,12 +1,14 @@
+//go:build !android
+
 package dns
 
 import (
 	"bufio"
 	"fmt"
-	"github.com/netbirdio/netbird/iface"
-	log "github.com/sirupsen/logrus"
 	"os"
 	"strings"
+
+	log "github.com/sirupsen/logrus"
 )
 
 const (
@@ -23,7 +25,7 @@ const (
 
 type osManagerType int
 
-func newHostManager(wgInterface *iface.WGIface) (hostManager, error) {
+func newHostManager(wgInterface WGIface) (hostManager, error) {
 	osManager, err := getOSDNSManagerType()
 	if err != nil {
 		return nil, err

client/internal/dns/host_windows.go

@@ -4,7 +4,6 @@ import (
 	"fmt"
 	"strings"
 
-	"github.com/netbirdio/netbird/iface"
 	log "github.com/sirupsen/logrus"
 	"golang.org/x/sys/windows/registry"
 )
@@ -32,7 +31,7 @@ type registryConfigurator struct {
 	existingSearchDomains []string
 }
 
-func newHostManager(wgInterface *iface.WGIface) (hostManager, error) {
+func newHostManager(wgInterface WGIface) (hostManager, error) {
 	guid, err := wgInterface.GetInterfaceGUIDString()
 	if err != nil {
 		return nil, err
@@ -42,6 +41,10 @@ func newHostManager(wgInterface *iface.WGIface) (hostManager, error) {
 	}, nil
 }
 
+func (s *registryConfigurator) supportCustomPort() bool {
+	return false
+}
+
 func (r *registryConfigurator) applyDNSConfig(config hostDNSConfig) error {
 	var err error
 	if config.routeAll {

client/internal/dns/local.go

@@ -2,10 +2,12 @@ package dns
 
 import (
 	"fmt"
-	"github.com/miekg/dns"
-	nbdns "github.com/netbirdio/netbird/dns"
-	log "github.com/sirupsen/logrus"
 	"sync"
+
+	"github.com/miekg/dns"
+	log "github.com/sirupsen/logrus"
+
+	nbdns "github.com/netbirdio/netbird/dns"
 )
 
 type registrationMap map[string]struct{}
@@ -15,9 +17,12 @@ type localResolver struct {
 	records       sync.Map
 }
 
+func (d *localResolver) stop() {
+}
+
 // ServeDNS handles a DNS request
 func (d *localResolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
-	log.Tracef("received question: %#v\n", r.Question[0])
+	log.Tracef("received question: %#v", r.Question[0])
 	replyMessage := &dns.Msg{}
 	replyMessage.SetReply(r)
 	replyMessage.RecursionAvailable = true

client/internal/dns/mockServer.go

@@ -7,16 +7,17 @@ import (
 
 // MockServer is the mock instance of a dns server
 type MockServer struct {
-	StartFunc           func()
+	InitializeFunc      func() error
 	StopFunc            func()
 	UpdateDNSServerFunc func(serial uint64, update nbdns.Config) error
 }
 
-// Start mock implementation of Start from Server interface
-func (m *MockServer) Start() {
-	if m.StartFunc != nil {
-		m.StartFunc()
+// Initialize mock implementation of Initialize from Server interface
+func (m *MockServer) Initialize() error {
+	if m.InitializeFunc != nil {
+		return m.InitializeFunc()
 	}
+	return nil
 }
 
 // Stop mock implementation of Stop from Server interface
@@ -26,6 +27,15 @@ func (m *MockServer) Stop() {
 	}
 }
 
+func (m *MockServer) DnsIP() string {
+	return ""
+}
+
+func (m *MockServer) OnUpdatedHostDNSServer(strings []string) {
+	//TODO implement me
+	panic("implement me")
+}
+
 // UpdateDNSServer mock implementation of UpdateDNSServer from Server interface
 func (m *MockServer) UpdateDNSServer(serial uint64, update nbdns.Config) error {
 	if m.UpdateDNSServerFunc != nil {

client/internal/dns/network_manager_linux.go

@@ -1,3 +1,5 @@
+//go:build !android
+
 package dns
 
 import (
@@ -11,7 +13,6 @@ import (
 	"github.com/godbus/dbus/v5"
 	"github.com/hashicorp/go-version"
 	"github.com/miekg/dns"
-	"github.com/netbirdio/netbird/iface"
 	log "github.com/sirupsen/logrus"
 )
 
@@ -69,7 +70,7 @@ func (s networkManagerConnSettings) cleanDeprecatedSettings() {
 	}
 }
 
-func newNetworkManagerDbusConfigurator(wgInterface *iface.WGIface) (hostManager, error) {
+func newNetworkManagerDbusConfigurator(wgInterface WGIface) (hostManager, error) {
 	obj, closeConn, err := getDbusObject(networkManagerDest, networkManagerDbusObjectNode)
 	if err != nil {
 		return nil, err
@@ -88,6 +89,10 @@ func newNetworkManagerDbusConfigurator(wgInterface *iface.WGIface) (hostManager,
 	}, nil
 }
 
+func (n *networkManagerDbusConfigurator) supportCustomPort() bool {
+	return false
+}
+
 func (n *networkManagerDbusConfigurator) applyDNSConfig(config hostDNSConfig) error {
 	connSettings, configVersion, err := n.getAppliedConnectionSettings()
 	if err != nil {

client/internal/dns/resolvconf_linux.go

@@ -1,11 +1,13 @@
+//go:build !android
+
 package dns
 
 import (
 	"fmt"
+	"os"
 	"os/exec"
 	"strings"
 
-	"github.com/netbirdio/netbird/iface"
 	log "github.com/sirupsen/logrus"
 )
 
@@ -15,12 +17,16 @@ type resolvconf struct {
 	ifaceName string
 }
 
-func newResolvConfConfigurator(wgInterface *iface.WGIface) (hostManager, error) {
+func newResolvConfConfigurator(wgInterface WGIface) (hostManager, error) {
 	return &resolvconf{
 		ifaceName: wgInterface.Name(),
 	}, nil
 }
 
+func (r *resolvconf) supportCustomPort() bool {
+	return false
+}
+
 func (r *resolvconf) applyDNSConfig(config hostDNSConfig) error {
 	var err error
 	if !config.routeAll {
@@ -54,7 +60,11 @@ func (r *resolvconf) applyDNSConfig(config hostDNSConfig) error {
 		appendedDomains++
 	}
 
-	content := fmt.Sprintf(fileGeneratedResolvConfContentFormat, fileDefaultResolvConfBackupLocation, config.serverIP, searchDomains)
+	originalContent, err := os.ReadFile(fileDefaultResolvConfBackupLocation)
+	if err != nil {
+		log.Errorf("Could not read existing resolv.conf")
+	}
+	content := fmt.Sprintf(fileGeneratedResolvConfContentFormat, fileDefaultResolvConfBackupLocation, config.serverIP, searchDomains, string(originalContent))
 
 	err = r.applyConfig(content)
 	if err != nil {

client/internal/dns/response_writer.go

@@ -0,0 +1,103 @@
+package dns
+
+import (
+	"fmt"
+	"net"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/miekg/dns"
+	"golang.zx2c4.com/wireguard/tun"
+)
+
+type responseWriter struct {
+	local  net.Addr
+	remote net.Addr
+	packet gopacket.Packet
+	device tun.Device
+}
+
+// LocalAddr returns the net.Addr of the server
+func (r *responseWriter) LocalAddr() net.Addr {
+	return r.local
+}
+
+// RemoteAddr returns the net.Addr of the client that sent the current request.
+func (r *responseWriter) RemoteAddr() net.Addr {
+	return r.remote
+}
+
+// WriteMsg writes a reply back to the client.
+func (r *responseWriter) WriteMsg(msg *dns.Msg) error {
+	buff, err := msg.Pack()
+	if err != nil {
+		return err
+	}
+	_, err = r.Write(buff)
+	return err
+}
+
+// Write writes a raw buffer back to the client.
+func (r *responseWriter) Write(data []byte) (int, error) {
+	var ip gopacket.SerializableLayer
+
+	// Get the UDP layer
+	udpLayer := r.packet.Layer(layers.LayerTypeUDP)
+	udp := udpLayer.(*layers.UDP)
+
+	// Swap the source and destination addresses for the response
+	udp.SrcPort, udp.DstPort = udp.DstPort, udp.SrcPort
+
+	// Check if it's an IPv4 packet
+	if ipv4Layer := r.packet.Layer(layers.LayerTypeIPv4); ipv4Layer != nil {
+		ipv4 := ipv4Layer.(*layers.IPv4)
+		ipv4.SrcIP, ipv4.DstIP = ipv4.DstIP, ipv4.SrcIP
+		ip = ipv4
+	} else if ipv6Layer := r.packet.Layer(layers.LayerTypeIPv6); ipv6Layer != nil {
+		ipv6 := ipv6Layer.(*layers.IPv6)
+		ipv6.SrcIP, ipv6.DstIP = ipv6.DstIP, ipv6.SrcIP
+		ip = ipv6
+	}
+
+	if err := udp.SetNetworkLayerForChecksum(ip.(gopacket.NetworkLayer)); err != nil {
+		return 0, fmt.Errorf("failed to set network layer for checksum: %v", err)
+	}
+
+	// Serialize the packet
+	buffer := gopacket.NewSerializeBuffer()
+	options := gopacket.SerializeOptions{
+		ComputeChecksums: true,
+		FixLengths:       true,
+	}
+
+	payload := gopacket.Payload(data)
+	err := gopacket.SerializeLayers(buffer, options, ip, udp, payload)
+	if err != nil {
+		return 0, fmt.Errorf("failed to serialize packet: %v", err)
+	}
+
+	send := buffer.Bytes()
+	sendBuffer := make([]byte, 40, len(send)+40)
+	sendBuffer = append(sendBuffer, send...)
+
+	return r.device.Write([][]byte{sendBuffer}, 40)
+}
+
+// Close closes the connection.
+func (r *responseWriter) Close() error {
+	return nil
+}
+
+// TsigStatus returns the status of the Tsig.
+func (r *responseWriter) TsigStatus() error {
+	return nil
+}
+
+// TsigTimersOnly sets the tsig timers only boolean.
+func (r *responseWriter) TsigTimersOnly(bool) {
+}
+
+// Hijack lets the caller take over the connection.
+// After a call to Hijack(), the DNS package will not do anything with the connection.
+func (r *responseWriter) Hijack() {
+}

client/internal/dns/response_writer_test.go

@@ -0,0 +1,93 @@
+package dns
+
+import (
+	"net"
+	"testing"
+
+	"github.com/golang/mock/gomock"
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/miekg/dns"
+
+	"github.com/netbirdio/netbird/iface/mocks"
+)
+
+func TestResponseWriterLocalAddr(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	defer ctrl.Finish()
+
+	device := mocks.NewMockDevice(ctrl)
+	device.EXPECT().Write(gomock.Any(), gomock.Any())
+
+	request := &dns.Msg{
+		Question: []dns.Question{{
+			Name:   "google.com.",
+			Qtype:  dns.TypeA,
+			Qclass: dns.TypeA,
+		}},
+	}
+
+	replyMessage := &dns.Msg{}
+	replyMessage.SetReply(request)
+	replyMessage.RecursionAvailable = true
+	replyMessage.Rcode = dns.RcodeSuccess
+	replyMessage.Answer = []dns.RR{
+		&dns.A{
+			A: net.IPv4(8, 8, 8, 8),
+		},
+	}
+
+	ipv4 := &layers.IPv4{
+		Protocol: layers.IPProtocolUDP,
+		SrcIP:    net.IPv4(127, 0, 0, 1),
+		DstIP:    net.IPv4(127, 0, 0, 2),
+	}
+	udp := &layers.UDP{
+		DstPort: 53,
+		SrcPort: 45223,
+	}
+	if err := udp.SetNetworkLayerForChecksum(ipv4); err != nil {
+		t.Error("failed to set network layer for checksum")
+		return
+	}
+
+	// Serialize the packet
+	buffer := gopacket.NewSerializeBuffer()
+	options := gopacket.SerializeOptions{
+		ComputeChecksums: true,
+		FixLengths:       true,
+	}
+
+	requestData, err := request.Pack()
+	if err != nil {
+		t.Errorf("got an error while packing the request message, error: %v", err)
+		return
+	}
+	payload := gopacket.Payload(requestData)
+
+	if err := gopacket.SerializeLayers(buffer, options, ipv4, udp, payload); err != nil {
+		t.Errorf("failed to serialize packet: %v", err)
+		return
+	}
+
+	rw := &responseWriter{
+		local: &net.UDPAddr{
+			IP:   net.IPv4(127, 0, 0, 1),
+			Port: 55223,
+		},
+		remote: &net.UDPAddr{
+			IP:   net.IPv4(127, 0, 0, 1),
+			Port: 53,
+		},
+		packet: gopacket.NewPacket(
+			buffer.Bytes(),
+			layers.LayerTypeIPv4,
+			gopacket.Default,
+		),
+		device: device,
+	}
+	if err := rw.WriteMsg(replyMessage); err != nil {
+		t.Errorf("got an error while writing the local resolver response, error: %v", err)
+		return
+	}
+}

client/internal/dns/server.go

@@ -1,10 +1,469 @@
 package dns
 
-import nbdns "github.com/netbirdio/netbird/dns"
+import (
+	"context"
+	"fmt"
+	"net/netip"
+	"sync"
+
+	"github.com/miekg/dns"
+	"github.com/mitchellh/hashstructure/v2"
+	log "github.com/sirupsen/logrus"
+
+	nbdns "github.com/netbirdio/netbird/dns"
+)
+
+// ReadyListener is a notification mechanism what indicate the server is ready to handle host dns address changes
+type ReadyListener interface {
+	OnReady()
+}
 
 // Server is a dns server interface
 type Server interface {
-	Start()
+	Initialize() error
 	Stop()
+	DnsIP() string
 	UpdateDNSServer(serial uint64, update nbdns.Config) error
+	OnUpdatedHostDNSServer(strings []string)
+}
+
+type registeredHandlerMap map[string]handlerWithStop
+
+// DefaultServer dns server object
+type DefaultServer struct {
+	ctx                context.Context
+	ctxCancel          context.CancelFunc
+	mux                sync.Mutex
+	service            service
+	dnsMuxMap          registeredHandlerMap
+	localResolver      *localResolver
+	wgInterface        WGIface
+	hostManager        hostManager
+	updateSerial       uint64
+	previousConfigHash uint64
+	currentConfig      hostDNSConfig
+
+	// permanent related properties
+	permanent        bool
+	hostsDnsList     []string
+	hostsDnsListLock sync.Mutex
+}
+
+type handlerWithStop interface {
+	dns.Handler
+	stop()
+}
+
+type muxUpdate struct {
+	domain  string
+	handler handlerWithStop
+}
+
+// NewDefaultServer returns a new dns server
+func NewDefaultServer(ctx context.Context, wgInterface WGIface, customAddress string) (*DefaultServer, error) {
+	var addrPort *netip.AddrPort
+	if customAddress != "" {
+		parsedAddrPort, err := netip.ParseAddrPort(customAddress)
+		if err != nil {
+			return nil, fmt.Errorf("unable to parse the custom dns address, got error: %s", err)
+		}
+		addrPort = &parsedAddrPort
+	}
+
+	var dnsService service
+	if wgInterface.IsUserspaceBind() {
+		dnsService = newServiceViaMemory(wgInterface)
+	} else {
+		dnsService = newServiceViaListener(wgInterface, addrPort)
+	}
+
+	return newDefaultServer(ctx, wgInterface, dnsService), nil
+}
+
+// NewDefaultServerPermanentUpstream returns a new dns server. It optimized for mobile systems
+func NewDefaultServerPermanentUpstream(ctx context.Context, wgInterface WGIface, hostsDnsList []string) *DefaultServer {
+	log.Debugf("host dns address list is: %v", hostsDnsList)
+	ds := newDefaultServer(ctx, wgInterface, newServiceViaMemory(wgInterface))
+	ds.permanent = true
+	ds.hostsDnsList = hostsDnsList
+	ds.addHostRootZone()
+	setServerDns(ds)
+	return ds
+}
+
+func newDefaultServer(ctx context.Context, wgInterface WGIface, dnsService service) *DefaultServer {
+	ctx, stop := context.WithCancel(ctx)
+	defaultServer := &DefaultServer{
+		ctx:       ctx,
+		ctxCancel: stop,
+		service:   dnsService,
+		dnsMuxMap: make(registeredHandlerMap),
+		localResolver: &localResolver{
+			registeredMap: make(registrationMap),
+		},
+		wgInterface: wgInterface,
+	}
+
+	return defaultServer
+}
+
+// Initialize instantiate host manager and the dns service
+func (s *DefaultServer) Initialize() (err error) {
+	s.mux.Lock()
+	defer s.mux.Unlock()
+
+	if s.hostManager != nil {
+		return nil
+	}
+
+	if s.permanent {
+		err = s.service.Listen()
+		if err != nil {
+			return err
+		}
+	}
+
+	s.hostManager, err = newHostManager(s.wgInterface)
+	return
+}
+
+// DnsIP returns the DNS resolver server IP address
+//
+// When kernel space interface used it return real DNS server listener IP address
+// For bind interface, fake DNS resolver address returned (second last IP address from Nebird network)
+func (s *DefaultServer) DnsIP() string {
+	return s.service.RuntimeIP()
+}
+
+// Stop stops the server
+func (s *DefaultServer) Stop() {
+	s.mux.Lock()
+	defer s.mux.Unlock()
+	s.ctxCancel()
+
+	if s.hostManager != nil {
+		err := s.hostManager.restoreHostDNS()
+		if err != nil {
+			log.Error(err)
+		}
+	}
+
+	s.service.Stop()
+}
+
+// OnUpdatedHostDNSServer update the DNS servers addresses for root zones
+// It will be applied if the mgm server do not enforce DNS settings for root zone
+func (s *DefaultServer) OnUpdatedHostDNSServer(hostsDnsList []string) {
+	s.hostsDnsListLock.Lock()
+	defer s.hostsDnsListLock.Unlock()
+
+	s.hostsDnsList = hostsDnsList
+	_, ok := s.dnsMuxMap[nbdns.RootZone]
+	if ok {
+		log.Debugf("on new host DNS config but skip to apply it")
+		return
+	}
+	log.Debugf("update host DNS settings: %+v", hostsDnsList)
+	s.addHostRootZone()
+}
+
+// UpdateDNSServer processes an update received from the management service
+func (s *DefaultServer) UpdateDNSServer(serial uint64, update nbdns.Config) error {
+	select {
+	case <-s.ctx.Done():
+		log.Infof("not updating DNS server as context is closed")
+		return s.ctx.Err()
+	default:
+		if serial < s.updateSerial {
+			return fmt.Errorf("not applying dns update, error: "+
+				"network update is %d behind the last applied update", s.updateSerial-serial)
+		}
+		s.mux.Lock()
+		defer s.mux.Unlock()
+
+		if s.hostManager == nil {
+			return fmt.Errorf("dns service is not initialized yet")
+		}
+
+		hash, err := hashstructure.Hash(update, hashstructure.FormatV2, &hashstructure.HashOptions{
+			ZeroNil:         true,
+			IgnoreZeroValue: true,
+			SlicesAsSets:    true,
+			UseStringer:     true,
+		})
+		if err != nil {
+			log.Errorf("unable to hash the dns configuration update, got error: %s", err)
+		}
+
+		if s.previousConfigHash == hash {
+			log.Debugf("not applying the dns configuration update as there is nothing new")
+			s.updateSerial = serial
+			return nil
+		}
+
+		if err := s.applyConfiguration(update); err != nil {
+			return err
+		}
+
+		s.updateSerial = serial
+		s.previousConfigHash = hash
+
+		return nil
+	}
+}
+
+func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
+	// is the service should be disabled, we stop the listener or fake resolver
+	// and proceed with a regular update to clean up the handlers and records
+	if update.ServiceEnable {
+		_ = s.service.Listen()
+	} else if !s.permanent {
+		s.service.Stop()
+	}
+
+	localMuxUpdates, localRecords, err := s.buildLocalHandlerUpdate(update.CustomZones)
+	if err != nil {
+		return fmt.Errorf("not applying dns update, error: %v", err)
+	}
+	upstreamMuxUpdates, err := s.buildUpstreamHandlerUpdate(update.NameServerGroups)
+	if err != nil {
+		return fmt.Errorf("not applying dns update, error: %v", err)
+	}
+	muxUpdates := append(localMuxUpdates, upstreamMuxUpdates...)
+
+	s.updateMux(muxUpdates)
+	s.updateLocalResolver(localRecords)
+	s.currentConfig = dnsConfigToHostDNSConfig(update, s.service.RuntimeIP(), s.service.RuntimePort())
+
+	hostUpdate := s.currentConfig
+	if s.service.RuntimePort() != defaultPort && !s.hostManager.supportCustomPort() {
+		log.Warnf("the DNS manager of this peer doesn't support custom port. Disabling primary DNS setup. " +
+			"Learn more at: https://netbird.io/docs/how-to-guides/nameservers#local-resolver")
+		hostUpdate.routeAll = false
+	}
+
+	if err = s.hostManager.applyDNSConfig(hostUpdate); err != nil {
+		log.Error(err)
+	}
+
+	return nil
+}
+
+func (s *DefaultServer) buildLocalHandlerUpdate(customZones []nbdns.CustomZone) ([]muxUpdate, map[string]nbdns.SimpleRecord, error) {
+	var muxUpdates []muxUpdate
+	localRecords := make(map[string]nbdns.SimpleRecord, 0)
+
+	for _, customZone := range customZones {
+
+		if len(customZone.Records) == 0 {
+			return nil, nil, fmt.Errorf("received an empty list of records")
+		}
+
+		muxUpdates = append(muxUpdates, muxUpdate{
+			domain:  customZone.Domain,
+			handler: s.localResolver,
+		})
+
+		for _, record := range customZone.Records {
+			var class uint16 = dns.ClassINET
+			if record.Class != nbdns.DefaultClass {
+				return nil, nil, fmt.Errorf("received an invalid class type: %s", record.Class)
+			}
+			key := buildRecordKey(record.Name, class, uint16(record.Type))
+			localRecords[key] = record
+		}
+	}
+	return muxUpdates, localRecords, nil
+}
+
+func (s *DefaultServer) buildUpstreamHandlerUpdate(nameServerGroups []*nbdns.NameServerGroup) ([]muxUpdate, error) {
+
+	var muxUpdates []muxUpdate
+	for _, nsGroup := range nameServerGroups {
+		if len(nsGroup.NameServers) == 0 {
+			log.Warn("received a nameserver group with empty nameserver list")
+			continue
+		}
+
+		handler := newUpstreamResolver(s.ctx)
+		for _, ns := range nsGroup.NameServers {
+			if ns.NSType != nbdns.UDPNameServerType {
+				log.Warnf("skiping nameserver %s with type %s, this peer supports only %s",
+					ns.IP.String(), ns.NSType.String(), nbdns.UDPNameServerType.String())
+				continue
+			}
+			handler.upstreamServers = append(handler.upstreamServers, getNSHostPort(ns))
+		}
+
+		if len(handler.upstreamServers) == 0 {
+			handler.stop()
+			log.Errorf("received a nameserver group with an invalid nameserver list")
+			continue
+		}
+
+		// when upstream fails to resolve domain several times over all it servers
+		// it will calls this hook to exclude self from the configuration and
+		// reapply DNS settings, but it not touch the original configuration and serial number
+		// because it is temporal deactivation until next try
+		//
+		// after some period defined by upstream it trys to reactivate self by calling this hook
+		// everything we need here is just to re-apply current configuration because it already
+		// contains this upstream settings (temporal deactivation not removed it)
+		handler.deactivate, handler.reactivate = s.upstreamCallbacks(nsGroup, handler)
+
+		if nsGroup.Primary {
+			muxUpdates = append(muxUpdates, muxUpdate{
+				domain:  nbdns.RootZone,
+				handler: handler,
+			})
+			continue
+		}
+
+		if len(nsGroup.Domains) == 0 {
+			handler.stop()
+			return nil, fmt.Errorf("received a non primary nameserver group with an empty domain list")
+		}
+
+		for _, domain := range nsGroup.Domains {
+			if domain == "" {
+				handler.stop()
+				return nil, fmt.Errorf("received a nameserver group with an empty domain element")
+			}
+			muxUpdates = append(muxUpdates, muxUpdate{
+				domain:  domain,
+				handler: handler,
+			})
+		}
+	}
+	return muxUpdates, nil
+}
+
+func (s *DefaultServer) updateMux(muxUpdates []muxUpdate) {
+	muxUpdateMap := make(registeredHandlerMap)
+
+	var isContainRootUpdate bool
+
+	for _, update := range muxUpdates {
+		s.service.RegisterMux(update.domain, update.handler)
+		muxUpdateMap[update.domain] = update.handler
+		if existingHandler, ok := s.dnsMuxMap[update.domain]; ok {
+			existingHandler.stop()
+		}
+
+		if update.domain == nbdns.RootZone {
+			isContainRootUpdate = true
+		}
+	}
+
+	for key, existingHandler := range s.dnsMuxMap {
+		_, found := muxUpdateMap[key]
+		if !found {
+			if !isContainRootUpdate && key == nbdns.RootZone {
+				s.hostsDnsListLock.Lock()
+				s.addHostRootZone()
+				s.hostsDnsListLock.Unlock()
+				existingHandler.stop()
+			} else {
+				existingHandler.stop()
+				s.service.DeregisterMux(key)
+			}
+		}
+	}
+
+	s.dnsMuxMap = muxUpdateMap
+}
+
+func (s *DefaultServer) updateLocalResolver(update map[string]nbdns.SimpleRecord) {
+	for key := range s.localResolver.registeredMap {
+		_, found := update[key]
+		if !found {
+			s.localResolver.deleteRecord(key)
+		}
+	}
+
+	updatedMap := make(registrationMap)
+	for key, record := range update {
+		err := s.localResolver.registerRecord(record)
+		if err != nil {
+			log.Warnf("got an error while registering the record (%s), error: %v", record.String(), err)
+		}
+		updatedMap[key] = struct{}{}
+	}
+
+	s.localResolver.registeredMap = updatedMap
+}
+
+func getNSHostPort(ns nbdns.NameServer) string {
+	return fmt.Sprintf("%s:%d", ns.IP.String(), ns.Port)
+}
+
+// upstreamCallbacks returns two functions, the first one is used to deactivate
+// the upstream resolver from the configuration, the second one is used to
+// reactivate it. Not allowed to call reactivate before deactivate.
+func (s *DefaultServer) upstreamCallbacks(
+	nsGroup *nbdns.NameServerGroup,
+	handler dns.Handler,
+) (deactivate func(), reactivate func()) {
+	var removeIndex map[string]int
+	deactivate = func() {
+		s.mux.Lock()
+		defer s.mux.Unlock()
+
+		l := log.WithField("nameservers", nsGroup.NameServers)
+		l.Info("temporary deactivate nameservers group due timeout")
+
+		removeIndex = make(map[string]int)
+		for _, domain := range nsGroup.Domains {
+			removeIndex[domain] = -1
+		}
+		if nsGroup.Primary {
+			removeIndex[nbdns.RootZone] = -1
+			s.currentConfig.routeAll = false
+		}
+
+		for i, item := range s.currentConfig.domains {
+			if _, found := removeIndex[item.domain]; found {
+				s.currentConfig.domains[i].disabled = true
+				s.service.DeregisterMux(item.domain)
+				removeIndex[item.domain] = i
+			}
+		}
+		if err := s.hostManager.applyDNSConfig(s.currentConfig); err != nil {
+			l.WithError(err).Error("fail to apply nameserver deactivation on the host")
+		}
+	}
+	reactivate = func() {
+		s.mux.Lock()
+		defer s.mux.Unlock()
+
+		for domain, i := range removeIndex {
+			if i == -1 || i >= len(s.currentConfig.domains) || s.currentConfig.domains[i].domain != domain {
+				continue
+			}
+			s.currentConfig.domains[i].disabled = false
+			s.service.RegisterMux(domain, handler)
+		}
+
+		l := log.WithField("nameservers", nsGroup.NameServers)
+		l.Debug("reactivate temporary disabled nameserver group")
+
+		if nsGroup.Primary {
+			s.currentConfig.routeAll = true
+		}
+		if err := s.hostManager.applyDNSConfig(s.currentConfig); err != nil {
+			l.WithError(err).Error("reactivate temporary disabled nameserver group, DNS update apply")
+		}
+	}
+	return
+}
+
+func (s *DefaultServer) addHostRootZone() {
+	handler := newUpstreamResolver(s.ctx)
+	handler.upstreamServers = make([]string, len(s.hostsDnsList))
+	for n, ua := range s.hostsDnsList {
+		handler.upstreamServers[n] = fmt.Sprintf("%s:53", ua)
+	}
+	handler.deactivate = func() {}
+	handler.reactivate = func() {}
+	s.service.RegisterMux(nbdns.RootZone, handler)
 }

client/internal/dns/server_android.go

@@ -1,32 +0,0 @@
-package dns
-
-import (
-	"context"
-
-	nbdns "github.com/netbirdio/netbird/dns"
-	"github.com/netbirdio/netbird/iface"
-)
-
-// DefaultServer dummy dns server
-type DefaultServer struct {
-}
-
-// NewDefaultServer On Android the DNS feature is not supported yet
-func NewDefaultServer(ctx context.Context, wgInterface *iface.WGIface, customAddress string) (*DefaultServer, error) {
-	return &DefaultServer{}, nil
-}
-
-// Start dummy implementation
-func (s DefaultServer) Start() {
-
-}
-
-// Stop dummy implementation
-func (s DefaultServer) Stop() {
-
-}
-
-// UpdateDNSServer dummy implementation
-func (s DefaultServer) UpdateDNSServer(serial uint64, update nbdns.Config) error {
-	return nil
-}

client/internal/dns/server_export.go

@@ -0,0 +1,29 @@
+package dns
+
+import (
+	"fmt"
+	"sync"
+)
+
+var (
+	mutex  sync.Mutex
+	server Server
+)
+
+// GetServerDns export the DNS server instance in static way. It used by the Mobile client
+func GetServerDns() (Server, error) {
+	mutex.Lock()
+	if server == nil {
+		mutex.Unlock()
+		return nil, fmt.Errorf("DNS server not instantiated yet")
+	}
+	s := server
+	mutex.Unlock()
+	return s, nil
+}
+
+func setServerDns(newServerServer Server) {
+	mutex.Lock()
+	server = newServerServer
+	defer mutex.Unlock()
+}

client/internal/dns/server_export_test.go

@@ -0,0 +1,24 @@
+package dns
+
+import (
+	"testing"
+)
+
+func TestGetServerDns(t *testing.T) {
+	_, err := GetServerDns()
+	if err == nil {
+		t.Errorf("invalid dns server instance")
+	}
+
+	srv := &MockServer{}
+	setServerDns(srv)
+
+	srvB, err := GetServerDns()
+	if err != nil {
+		t.Errorf("invalid dns server instance: %s", err)
+	}
+
+	if srvB != srv {
+		t.Errorf("missmatch dns instances")
+	}
+}

client/internal/dns/server_nonandroid.go

@@ -1,465 +0,0 @@
-//go:build !android
-
-package dns
-
-import (
-	"context"
-	"fmt"
-	"net"
-	"net/netip"
-	"runtime"
-	"sync"
-	"time"
-
-	"github.com/miekg/dns"
-	"github.com/mitchellh/hashstructure/v2"
-	nbdns "github.com/netbirdio/netbird/dns"
-	"github.com/netbirdio/netbird/iface"
-	log "github.com/sirupsen/logrus"
-)
-
-const (
-	defaultPort = 53
-	customPort  = 5053
-	defaultIP   = "127.0.0.1"
-	customIP    = "127.0.0.153"
-)
-
-// DefaultServer dns server object
-type DefaultServer struct {
-	ctx                context.Context
-	ctxCancel          context.CancelFunc
-	upstreamCtxCancel  context.CancelFunc
-	mux                sync.Mutex
-	server             *dns.Server
-	dnsMux             *dns.ServeMux
-	dnsMuxMap          registrationMap
-	localResolver      *localResolver
-	wgInterface        *iface.WGIface
-	hostManager        hostManager
-	updateSerial       uint64
-	listenerIsRunning  bool
-	runtimePort        int
-	runtimeIP          string
-	previousConfigHash uint64
-	currentConfig      hostDNSConfig
-	customAddress      *netip.AddrPort
-}
-
-type muxUpdate struct {
-	domain  string
-	handler dns.Handler
-}
-
-// NewDefaultServer returns a new dns server
-func NewDefaultServer(ctx context.Context, wgInterface *iface.WGIface, customAddress string) (*DefaultServer, error) {
-	mux := dns.NewServeMux()
-
-	dnsServer := &dns.Server{
-		Net:     "udp",
-		Handler: mux,
-		UDPSize: 65535,
-	}
-
-	ctx, stop := context.WithCancel(ctx)
-
-	var addrPort *netip.AddrPort
-	if customAddress != "" {
-		parsedAddrPort, err := netip.ParseAddrPort(customAddress)
-		if err != nil {
-			stop()
-			return nil, fmt.Errorf("unable to parse the custom dns address, got error: %s", err)
-		}
-		addrPort = &parsedAddrPort
-	}
-
-	defaultServer := &DefaultServer{
-		ctx:       ctx,
-		ctxCancel: stop,
-		server:    dnsServer,
-		dnsMux:    mux,
-		dnsMuxMap: make(registrationMap),
-		localResolver: &localResolver{
-			registeredMap: make(registrationMap),
-		},
-		wgInterface:   wgInterface,
-		runtimePort:   defaultPort,
-		customAddress: addrPort,
-	}
-
-	hostmanager, err := newHostManager(wgInterface)
-	if err != nil {
-		stop()
-		return nil, err
-	}
-	defaultServer.hostManager = hostmanager
-	return defaultServer, err
-}
-
-// Start runs the listener in a go routine
-func (s *DefaultServer) Start() {
-	if s.customAddress != nil {
-		s.runtimeIP = s.customAddress.Addr().String()
-		s.runtimePort = int(s.customAddress.Port())
-	} else {
-		ip, port, err := s.getFirstListenerAvailable()
-		if err != nil {
-			log.Error(err)
-			return
-		}
-		s.runtimeIP = ip
-		s.runtimePort = port
-	}
-
-	s.server.Addr = fmt.Sprintf("%s:%d", s.runtimeIP, s.runtimePort)
-
-	log.Debugf("starting dns on %s", s.server.Addr)
-
-	go func() {
-		s.setListenerStatus(true)
-		defer s.setListenerStatus(false)
-
-		err := s.server.ListenAndServe()
-		if err != nil {
-			log.Errorf("dns server running with %d port returned an error: %v. Will not retry", s.runtimePort, err)
-		}
-	}()
-}
-
-func (s *DefaultServer) getFirstListenerAvailable() (string, int, error) {
-	ips := []string{defaultIP, customIP}
-	if runtime.GOOS != "darwin" && s.wgInterface != nil {
-		ips = append([]string{s.wgInterface.Address().IP.String()}, ips...)
-	}
-	ports := []int{defaultPort, customPort}
-	for _, port := range ports {
-		for _, ip := range ips {
-			addrString := fmt.Sprintf("%s:%d", ip, port)
-			udpAddr := net.UDPAddrFromAddrPort(netip.MustParseAddrPort(addrString))
-			probeListener, err := net.ListenUDP("udp", udpAddr)
-			if err == nil {
-				err = probeListener.Close()
-				if err != nil {
-					log.Errorf("got an error closing the probe listener, error: %s", err)
-				}
-				return ip, port, nil
-			}
-			log.Warnf("binding dns on %s is not available, error: %s", addrString, err)
-		}
-	}
-	return "", 0, fmt.Errorf("unable to find an unused ip and port combination. IPs tested: %v and ports %v", ips, ports)
-}
-
-func (s *DefaultServer) setListenerStatus(running bool) {
-	s.listenerIsRunning = running
-}
-
-// Stop stops the server
-func (s *DefaultServer) Stop() {
-	s.mux.Lock()
-	defer s.mux.Unlock()
-	s.ctxCancel()
-
-	err := s.hostManager.restoreHostDNS()
-	if err != nil {
-		log.Error(err)
-	}
-
-	err = s.stopListener()
-	if err != nil {
-		log.Error(err)
-	}
-}
-
-func (s *DefaultServer) stopListener() error {
-	if !s.listenerIsRunning {
-		return nil
-	}
-
-	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-	defer cancel()
-
-	err := s.server.ShutdownContext(ctx)
-	if err != nil {
-		return fmt.Errorf("stopping dns server listener returned an error: %v", err)
-	}
-	return nil
-}
-
-// UpdateDNSServer processes an update received from the management service
-func (s *DefaultServer) UpdateDNSServer(serial uint64, update nbdns.Config) error {
-	select {
-	case <-s.ctx.Done():
-		log.Infof("not updating DNS server as context is closed")
-		return s.ctx.Err()
-	default:
-		if serial < s.updateSerial {
-			return fmt.Errorf("not applying dns update, error: "+
-				"network update is %d behind the last applied update", s.updateSerial-serial)
-		}
-		s.mux.Lock()
-		defer s.mux.Unlock()
-
-		hash, err := hashstructure.Hash(update, hashstructure.FormatV2, &hashstructure.HashOptions{
-			ZeroNil:         true,
-			IgnoreZeroValue: true,
-			SlicesAsSets:    true,
-			UseStringer:     true,
-		})
-		if err != nil {
-			log.Errorf("unable to hash the dns configuration update, got error: %s", err)
-		}
-
-		if s.previousConfigHash == hash {
-			log.Debugf("not applying the dns configuration update as there is nothing new")
-			s.updateSerial = serial
-			return nil
-		}
-
-		if err := s.applyConfiguration(update); err != nil {
-			return err
-		}
-
-		s.updateSerial = serial
-		s.previousConfigHash = hash
-
-		return nil
-	}
-}
-
-func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
-	// is the service should be disabled, we stop the listener
-	// and proceed with a regular update to clean up the handlers and records
-	if !update.ServiceEnable {
-		err := s.stopListener()
-		if err != nil {
-			log.Error(err)
-		}
-	} else if !s.listenerIsRunning {
-		s.Start()
-	}
-
-	localMuxUpdates, localRecords, err := s.buildLocalHandlerUpdate(update.CustomZones)
-	if err != nil {
-		return fmt.Errorf("not applying dns update, error: %v", err)
-	}
-	upstreamMuxUpdates, err := s.buildUpstreamHandlerUpdate(update.NameServerGroups)
-	if err != nil {
-		return fmt.Errorf("not applying dns update, error: %v", err)
-	}
-
-	muxUpdates := append(localMuxUpdates, upstreamMuxUpdates...)
-
-	s.updateMux(muxUpdates)
-	s.updateLocalResolver(localRecords)
-	s.currentConfig = dnsConfigToHostDNSConfig(update, s.runtimeIP, s.runtimePort)
-
-	if err = s.hostManager.applyDNSConfig(s.currentConfig); err != nil {
-		log.Error(err)
-	}
-
-	return nil
-}
-
-func (s *DefaultServer) buildLocalHandlerUpdate(customZones []nbdns.CustomZone) ([]muxUpdate, map[string]nbdns.SimpleRecord, error) {
-	var muxUpdates []muxUpdate
-	localRecords := make(map[string]nbdns.SimpleRecord, 0)
-
-	for _, customZone := range customZones {
-
-		if len(customZone.Records) == 0 {
-			return nil, nil, fmt.Errorf("received an empty list of records")
-		}
-
-		muxUpdates = append(muxUpdates, muxUpdate{
-			domain:  customZone.Domain,
-			handler: s.localResolver,
-		})
-
-		for _, record := range customZone.Records {
-			var class uint16 = dns.ClassINET
-			if record.Class != nbdns.DefaultClass {
-				return nil, nil, fmt.Errorf("received an invalid class type: %s", record.Class)
-			}
-			key := buildRecordKey(record.Name, class, uint16(record.Type))
-			localRecords[key] = record
-		}
-	}
-	return muxUpdates, localRecords, nil
-}
-
-func (s *DefaultServer) buildUpstreamHandlerUpdate(nameServerGroups []*nbdns.NameServerGroup) ([]muxUpdate, error) {
-	// clean up the previous upstream resolver
-	if s.upstreamCtxCancel != nil {
-		s.upstreamCtxCancel()
-	}
-
-	var muxUpdates []muxUpdate
-	for _, nsGroup := range nameServerGroups {
-		if len(nsGroup.NameServers) == 0 {
-			log.Warn("received a nameserver group with empty nameserver list")
-			continue
-		}
-
-		var ctx context.Context
-		ctx, s.upstreamCtxCancel = context.WithCancel(s.ctx)
-
-		handler := newUpstreamResolver(ctx)
-		for _, ns := range nsGroup.NameServers {
-			if ns.NSType != nbdns.UDPNameServerType {
-				log.Warnf("skiping nameserver %s with type %s, this peer supports only %s",
-					ns.IP.String(), ns.NSType.String(), nbdns.UDPNameServerType.String())
-				continue
-			}
-			handler.upstreamServers = append(handler.upstreamServers, getNSHostPort(ns))
-		}
-
-		if len(handler.upstreamServers) == 0 {
-			log.Errorf("received a nameserver group with an invalid nameserver list")
-			continue
-		}
-
-		// when upstream fails to resolve domain several times over all it servers
-		// it will calls this hook to exclude self from the configuration and
-		// reapply DNS settings, but it not touch the original configuration and serial number
-		// because it is temporal deactivation until next try
-		//
-		// after some period defined by upstream it trys to reactivate self by calling this hook
-		// everything we need here is just to re-apply current configuration because it already
-		// contains this upstream settings (temporal deactivation not removed it)
-		handler.deactivate, handler.reactivate = s.upstreamCallbacks(nsGroup, handler)
-
-		if nsGroup.Primary {
-			muxUpdates = append(muxUpdates, muxUpdate{
-				domain:  nbdns.RootZone,
-				handler: handler,
-			})
-			continue
-		}
-
-		if len(nsGroup.Domains) == 0 {
-			return nil, fmt.Errorf("received a non primary nameserver group with an empty domain list")
-		}
-
-		for _, domain := range nsGroup.Domains {
-			if domain == "" {
-				return nil, fmt.Errorf("received a nameserver group with an empty domain element")
-			}
-			muxUpdates = append(muxUpdates, muxUpdate{
-				domain:  domain,
-				handler: handler,
-			})
-		}
-	}
-	return muxUpdates, nil
-}
-
-func (s *DefaultServer) updateMux(muxUpdates []muxUpdate) {
-	muxUpdateMap := make(registrationMap)
-
-	for _, update := range muxUpdates {
-		s.registerMux(update.domain, update.handler)
-		muxUpdateMap[update.domain] = struct{}{}
-	}
-
-	for key := range s.dnsMuxMap {
-		_, found := muxUpdateMap[key]
-		if !found {
-			s.deregisterMux(key)
-		}
-	}
-
-	s.dnsMuxMap = muxUpdateMap
-}
-
-func (s *DefaultServer) updateLocalResolver(update map[string]nbdns.SimpleRecord) {
-	for key := range s.localResolver.registeredMap {
-		_, found := update[key]
-		if !found {
-			s.localResolver.deleteRecord(key)
-		}
-	}
-
-	updatedMap := make(registrationMap)
-	for key, record := range update {
-		err := s.localResolver.registerRecord(record)
-		if err != nil {
-			log.Warnf("got an error while registering the record (%s), error: %v", record.String(), err)
-		}
-		updatedMap[key] = struct{}{}
-	}
-
-	s.localResolver.registeredMap = updatedMap
-}
-
-func getNSHostPort(ns nbdns.NameServer) string {
-	return fmt.Sprintf("%s:%d", ns.IP.String(), ns.Port)
-}
-
-func (s *DefaultServer) registerMux(pattern string, handler dns.Handler) {
-	s.dnsMux.Handle(pattern, handler)
-}
-
-func (s *DefaultServer) deregisterMux(pattern string) {
-	s.dnsMux.HandleRemove(pattern)
-}
-
-// upstreamCallbacks returns two functions, the first one is used to deactivate
-// the upstream resolver from the configuration, the second one is used to
-// reactivate it. Not allowed to call reactivate before deactivate.
-func (s *DefaultServer) upstreamCallbacks(
-	nsGroup *nbdns.NameServerGroup,
-	handler dns.Handler,
-) (deactivate func(), reactivate func()) {
-	var removeIndex map[string]int
-	deactivate = func() {
-		s.mux.Lock()
-		defer s.mux.Unlock()
-
-		l := log.WithField("nameservers", nsGroup.NameServers)
-		l.Info("temporary deactivate nameservers group due timeout")
-
-		removeIndex = make(map[string]int)
-		for _, domain := range nsGroup.Domains {
-			removeIndex[domain] = -1
-		}
-		if nsGroup.Primary {
-			removeIndex[nbdns.RootZone] = -1
-			s.currentConfig.routeAll = false
-		}
-
-		for i, item := range s.currentConfig.domains {
-			if _, found := removeIndex[item.domain]; found {
-				s.currentConfig.domains[i].disabled = true
-				s.deregisterMux(item.domain)
-				removeIndex[item.domain] = i
-			}
-		}
-		if err := s.hostManager.applyDNSConfig(s.currentConfig); err != nil {
-			l.WithError(err).Error("fail to apply nameserver deactivation on the host")
-		}
-	}
-	reactivate = func() {
-		s.mux.Lock()
-		defer s.mux.Unlock()
-
-		for domain, i := range removeIndex {
-			if i == -1 || i >= len(s.currentConfig.domains) || s.currentConfig.domains[i].domain != domain {
-				continue
-			}
-			s.currentConfig.domains[i].disabled = false
-			s.registerMux(domain, handler)
-		}
-
-		l := log.WithField("nameservers", nsGroup.NameServers)
-		l.Debug("reactivate temporary disabled nameserver group")
-
-		if nsGroup.Primary {
-			s.currentConfig.routeAll = true
-		}
-		if err := s.hostManager.applyDNSConfig(s.currentConfig); err != nil {
-			l.WithError(err).Error("reactivate temporary disabled nameserver group, DNS update apply")
-		}
-	}
-	return
-}

client/internal/dns/server_test.go

@@ -5,18 +5,59 @@ import (
 	"fmt"
 	"net"
 	"net/netip"
+	"os"
 	"strings"
 	"testing"
 	"time"
 
+	"github.com/golang/mock/gomock"
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/client/firewall/uspfilter"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
-
-	"github.com/miekg/dns"
-
 	nbdns "github.com/netbirdio/netbird/dns"
+	"github.com/netbirdio/netbird/formatter"
 	"github.com/netbirdio/netbird/iface"
+	pfmock "github.com/netbirdio/netbird/iface/mocks"
 )
 
+type mocWGIface struct {
+	filter iface.PacketFilter
+}
+
+func (w *mocWGIface) Name() string {
+	panic("implement me")
+}
+
+func (w *mocWGIface) Address() iface.WGAddress {
+	ip, network, _ := net.ParseCIDR("100.66.100.0/24")
+	return iface.WGAddress{
+		IP:      ip,
+		Network: network,
+	}
+}
+
+func (w *mocWGIface) GetFilter() iface.PacketFilter {
+	return w.filter
+}
+
+func (w *mocWGIface) GetDevice() *iface.DeviceWrapper {
+	panic("implement me")
+}
+
+func (w *mocWGIface) GetInterfaceGUIDString() (string, error) {
+	panic("implement me")
+}
+
+func (w *mocWGIface) IsUserspaceBind() bool {
+	return false
+}
+
+func (w *mocWGIface) SetFilter(filter iface.PacketFilter) error {
+	w.filter = filter
+	return nil
+}
+
 var zoneRecords = []nbdns.SimpleRecord{
 	{
 		Name:  "peera.netbird.cloud",
@@ -27,6 +68,11 @@ var zoneRecords = []nbdns.SimpleRecord{
 	},
 }
 
+func init() {
+	log.SetLevel(log.TraceLevel)
+	formatter.SetTextFormatter(log.StandardLogger())
+}
+
 func TestUpdateDNSServer(t *testing.T) {
 	nameServers := []nbdns.NameServer{
 		{
@@ -41,21 +87,23 @@ func TestUpdateDNSServer(t *testing.T) {
 		},
 	}
 
+	dummyHandler := &localResolver{}
+
 	testCases := []struct {
 		name                string
-		initUpstreamMap     registrationMap
+		initUpstreamMap     registeredHandlerMap
 		initLocalMap        registrationMap
 		initSerial          uint64
 		inputSerial         uint64
 		inputUpdate         nbdns.Config
 		shouldFail          bool
-		expectedUpstreamMap registrationMap
+		expectedUpstreamMap registeredHandlerMap
 		expectedLocalMap    registrationMap
 	}{
 		{
 			name:            "Initial Config Should Succeed",
 			initLocalMap:    make(registrationMap),
-			initUpstreamMap: make(registrationMap),
+			initUpstreamMap: make(registeredHandlerMap),
 			initSerial:      0,
 			inputSerial:     1,
 			inputUpdate: nbdns.Config{
@@ -77,13 +125,13 @@ func TestUpdateDNSServer(t *testing.T) {
 					},
 				},
 			},
-			expectedUpstreamMap: registrationMap{"netbird.io": struct{}{}, "netbird.cloud": struct{}{}, nbdns.RootZone: struct{}{}},
+			expectedUpstreamMap: registeredHandlerMap{"netbird.io": dummyHandler, "netbird.cloud": dummyHandler, nbdns.RootZone: dummyHandler},
 			expectedLocalMap:    registrationMap{buildRecordKey(zoneRecords[0].Name, 1, 1): struct{}{}},
 		},
 		{
 			name:            "New Config Should Succeed",
 			initLocalMap:    registrationMap{"netbird.cloud": struct{}{}},
-			initUpstreamMap: registrationMap{buildRecordKey(zoneRecords[0].Name, 1, 1): struct{}{}},
+			initUpstreamMap: registeredHandlerMap{buildRecordKey(zoneRecords[0].Name, 1, 1): dummyHandler},
 			initSerial:      0,
 			inputSerial:     1,
 			inputUpdate: nbdns.Config{
@@ -101,13 +149,13 @@ func TestUpdateDNSServer(t *testing.T) {
 					},
 				},
 			},
-			expectedUpstreamMap: registrationMap{"netbird.io": struct{}{}, "netbird.cloud": struct{}{}},
+			expectedUpstreamMap: registeredHandlerMap{"netbird.io": dummyHandler, "netbird.cloud": dummyHandler},
 			expectedLocalMap:    registrationMap{buildRecordKey(zoneRecords[0].Name, 1, 1): struct{}{}},
 		},
 		{
 			name:            "Smaller Config Serial Should Be Skipped",
 			initLocalMap:    make(registrationMap),
-			initUpstreamMap: make(registrationMap),
+			initUpstreamMap: make(registeredHandlerMap),
 			initSerial:      2,
 			inputSerial:     1,
 			shouldFail:      true,
@@ -115,7 +163,7 @@ func TestUpdateDNSServer(t *testing.T) {
 		{
 			name:            "Empty NS Group Domain Or Not Primary Element Should Fail",
 			initLocalMap:    make(registrationMap),
-			initUpstreamMap: make(registrationMap),
+			initUpstreamMap: make(registeredHandlerMap),
 			initSerial:      0,
 			inputSerial:     1,
 			inputUpdate: nbdns.Config{
@@ -137,7 +185,7 @@ func TestUpdateDNSServer(t *testing.T) {
 		{
 			name:            "Invalid NS Group Nameservers list Should Fail",
 			initLocalMap:    make(registrationMap),
-			initUpstreamMap: make(registrationMap),
+			initUpstreamMap: make(registeredHandlerMap),
 			initSerial:      0,
 			inputSerial:     1,
 			inputUpdate: nbdns.Config{
@@ -159,7 +207,7 @@ func TestUpdateDNSServer(t *testing.T) {
 		{
 			name:            "Invalid Custom Zone Records list Should Fail",
 			initLocalMap:    make(registrationMap),
-			initUpstreamMap: make(registrationMap),
+			initUpstreamMap: make(registeredHandlerMap),
 			initSerial:      0,
 			inputSerial:     1,
 			inputUpdate: nbdns.Config{
@@ -181,21 +229,21 @@ func TestUpdateDNSServer(t *testing.T) {
 		{
 			name:                "Empty Config Should Succeed and Clean Maps",
 			initLocalMap:        registrationMap{"netbird.cloud": struct{}{}},
-			initUpstreamMap:     registrationMap{zoneRecords[0].Name: struct{}{}},
+			initUpstreamMap:     registeredHandlerMap{zoneRecords[0].Name: dummyHandler},
 			initSerial:          0,
 			inputSerial:         1,
 			inputUpdate:         nbdns.Config{ServiceEnable: true},
-			expectedUpstreamMap: make(registrationMap),
+			expectedUpstreamMap: make(registeredHandlerMap),
 			expectedLocalMap:    make(registrationMap),
 		},
 		{
 			name:                "Disabled Service Should clean map",
 			initLocalMap:        registrationMap{"netbird.cloud": struct{}{}},
-			initUpstreamMap:     registrationMap{zoneRecords[0].Name: struct{}{}},
+			initUpstreamMap:     registeredHandlerMap{zoneRecords[0].Name: dummyHandler},
 			initSerial:          0,
 			inputSerial:         1,
 			inputUpdate:         nbdns.Config{ServiceEnable: false},
-			expectedUpstreamMap: make(registrationMap),
+			expectedUpstreamMap: make(registeredHandlerMap),
 			expectedLocalMap:    make(registrationMap),
 		},
 	}
@@ -224,6 +272,10 @@ func TestUpdateDNSServer(t *testing.T) {
 			if err != nil {
 				t.Fatal(err)
 			}
+			err = dnsServer.Initialize()
+			if err != nil {
+				t.Fatal(err)
+			}
 			defer func() {
 				err = dnsServer.hostManager.restoreHostDNS()
 				if err != nil {
@@ -234,8 +286,6 @@ func TestUpdateDNSServer(t *testing.T) {
 			dnsServer.dnsMuxMap = testCase.initUpstreamMap
 			dnsServer.localResolver.registeredMap = testCase.initLocalMap
 			dnsServer.updateSerial = testCase.initSerial
-			// pretend we are running
-			dnsServer.listenerIsRunning = true
 
 			err = dnsServer.UpdateDNSServer(testCase.inputSerial, testCase.inputUpdate)
 			if err != nil {
@@ -270,6 +320,133 @@ func TestUpdateDNSServer(t *testing.T) {
 	}
 }
 
+func TestDNSFakeResolverHandleUpdates(t *testing.T) {
+	ov := os.Getenv("NB_WG_KERNEL_DISABLED")
+	defer os.Setenv("NB_WG_KERNEL_DISABLED", ov)
+
+	_ = os.Setenv("NB_WG_KERNEL_DISABLED", "true")
+	newNet, err := stdnet.NewNet(nil)
+	if err != nil {
+		t.Errorf("create stdnet: %v", err)
+		return
+	}
+
+	wgIface, err := iface.NewWGIFace("utun2301", "100.66.100.1/32", iface.DefaultMTU, nil, newNet)
+	if err != nil {
+		t.Errorf("build interface wireguard: %v", err)
+		return
+	}
+
+	err = wgIface.Create()
+	if err != nil {
+		t.Errorf("crate and init wireguard interface: %v", err)
+		return
+	}
+	defer func() {
+		if err = wgIface.Close(); err != nil {
+			t.Logf("close wireguard interface: %v", err)
+		}
+	}()
+
+	ctrl := gomock.NewController(t)
+	defer ctrl.Finish()
+
+	_, ipNet, err := net.ParseCIDR("100.66.100.1/32")
+	if err != nil {
+		t.Errorf("parse CIDR: %v", err)
+		return
+	}
+
+	packetfilter := pfmock.NewMockPacketFilter(ctrl)
+	packetfilter.EXPECT().DropOutgoing(gomock.Any()).AnyTimes()
+	packetfilter.EXPECT().AddUDPPacketHook(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any())
+	packetfilter.EXPECT().RemovePacketHook(gomock.Any())
+	packetfilter.EXPECT().SetNetwork(ipNet)
+
+	if err := wgIface.SetFilter(packetfilter); err != nil {
+		t.Errorf("set packet filter: %v", err)
+		return
+	}
+
+	dnsServer, err := NewDefaultServer(context.Background(), wgIface, "")
+	if err != nil {
+		t.Errorf("create DNS server: %v", err)
+		return
+	}
+
+	err = dnsServer.Initialize()
+	if err != nil {
+		t.Errorf("run DNS server: %v", err)
+		return
+	}
+	defer func() {
+		if err = dnsServer.hostManager.restoreHostDNS(); err != nil {
+			t.Logf("restore DNS settings on the host: %v", err)
+			return
+		}
+	}()
+
+	dnsServer.dnsMuxMap = registeredHandlerMap{zoneRecords[0].Name: &localResolver{}}
+	dnsServer.localResolver.registeredMap = registrationMap{"netbird.cloud": struct{}{}}
+	dnsServer.updateSerial = 0
+
+	nameServers := []nbdns.NameServer{
+		{
+			IP:     netip.MustParseAddr("8.8.8.8"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+		{
+			IP:     netip.MustParseAddr("8.8.4.4"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+	}
+
+	update := nbdns.Config{
+		ServiceEnable: true,
+		CustomZones: []nbdns.CustomZone{
+			{
+				Domain:  "netbird.cloud",
+				Records: zoneRecords,
+			},
+		},
+		NameServerGroups: []*nbdns.NameServerGroup{
+			{
+				Domains:     []string{"netbird.io"},
+				NameServers: nameServers,
+			},
+			{
+				NameServers: nameServers,
+				Primary:     true,
+			},
+		},
+	}
+
+	// Start the server with regular configuration
+	if err := dnsServer.UpdateDNSServer(1, update); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+
+	update2 := update
+	update2.ServiceEnable = false
+	// Disable the server, stop the listener
+	if err := dnsServer.UpdateDNSServer(2, update2); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+
+	update3 := update2
+	update3.NameServerGroups = update3.NameServerGroups[:1]
+	// But service still get updates and we checking that we handle
+	// internal state in the right way
+	if err := dnsServer.UpdateDNSServer(3, update3); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+}
+
 func TestDNSServerStartStop(t *testing.T) {
 	testCases := []struct {
 		name     string
@@ -286,21 +463,23 @@ func TestDNSServerStartStop(t *testing.T) {
 
 	for _, testCase := range testCases {
 		t.Run(testCase.name, func(t *testing.T) {
-			dnsServer := getDefaultServerWithNoHostManager(t, testCase.addrPort)
-
-			dnsServer.hostManager = newNoopHostMocker()
-			dnsServer.Start()
-			time.Sleep(100 * time.Millisecond)
-			if !dnsServer.listenerIsRunning {
-				t.Fatal("dns server listener is not running")
+			dnsServer, err := NewDefaultServer(context.Background(), &mocWGIface{}, testCase.addrPort)
+			if err != nil {
+				t.Fatalf("%v", err)
 			}
+			dnsServer.hostManager = newNoopHostMocker()
+			err = dnsServer.service.Listen()
+			if err != nil {
+				t.Fatalf("dns server is not running: %s", err)
+			}
+			time.Sleep(100 * time.Millisecond)
 			defer dnsServer.Stop()
-			err := dnsServer.localResolver.registerRecord(zoneRecords[0])
+			err = dnsServer.localResolver.registerRecord(zoneRecords[0])
 			if err != nil {
 				t.Error(err)
 			}
 
-			dnsServer.dnsMux.Handle("netbird.cloud", dnsServer.localResolver)
+			dnsServer.service.RegisterMux("netbird.cloud", dnsServer.localResolver)
 
 			resolver := &net.Resolver{
 				PreferGo: true,
@@ -308,7 +487,7 @@ func TestDNSServerStartStop(t *testing.T) {
 					d := net.Dialer{
 						Timeout: time.Second * 5,
 					}
-					addr := fmt.Sprintf("%s:%d", dnsServer.runtimeIP, dnsServer.runtimePort)
+					addr := fmt.Sprintf("%s:%d", dnsServer.service.RuntimeIP(), dnsServer.service.RuntimePort())
 					conn, err := d.DialContext(ctx, network, addr)
 					if err != nil {
 						t.Log(err)
@@ -343,7 +522,7 @@ func TestDNSServerStartStop(t *testing.T) {
 func TestDNSServerUpstreamDeactivateCallback(t *testing.T) {
 	hostManager := &mockHostConfigurator{}
 	server := DefaultServer{
-		dnsMux: dns.DefaultServeMux,
+		service: newServiceViaMemory(&mocWGIface{}),
 		localResolver: &localResolver{
 			registeredMap: make(registrationMap),
 		},
@@ -406,35 +585,237 @@ func TestDNSServerUpstreamDeactivateCallback(t *testing.T) {
 	}
 }
 
-func getDefaultServerWithNoHostManager(t *testing.T, addrPort string) *DefaultServer {
-	mux := dns.NewServeMux()
-
-	var parsedAddrPort *netip.AddrPort
-	if addrPort != "" {
-		parsed, err := netip.ParseAddrPort(addrPort)
-		if err != nil {
-			t.Fatal(err)
-		}
-		parsedAddrPort = &parsed
+func TestDNSPermanent_updateHostDNS_emptyUpstream(t *testing.T) {
+	wgIFace, err := createWgInterfaceWithBind(t)
+	if err != nil {
+		t.Fatal("failed to initialize wg interface")
 	}
+	defer wgIFace.Close()
 
-	dnsServer := &dns.Server{
-		Net:     "udp",
-		Handler: mux,
-		UDPSize: 65535,
+	var dnsList []string
+	dnsServer := NewDefaultServerPermanentUpstream(context.Background(), wgIFace, dnsList)
+	err = dnsServer.Initialize()
+	if err != nil {
+		t.Errorf("failed to initialize DNS server: %v", err)
+		return
 	}
+	defer dnsServer.Stop()
 
-	ctx, cancel := context.WithCancel(context.TODO())
+	dnsServer.OnUpdatedHostDNSServer([]string{"8.8.8.8"})
 
-	return &DefaultServer{
-		ctx:       ctx,
-		ctxCancel: cancel,
-		server:    dnsServer,
-		dnsMux:    mux,
-		dnsMuxMap: make(registrationMap),
-		localResolver: &localResolver{
-			registeredMap: make(registrationMap),
-		},
-		customAddress: parsedAddrPort,
+	resolver := newDnsResolver(dnsServer.service.RuntimeIP(), dnsServer.service.RuntimePort())
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+}
+
+func TestDNSPermanent_updateUpstream(t *testing.T) {
+	wgIFace, err := createWgInterfaceWithBind(t)
+	if err != nil {
+		t.Fatal("failed to initialize wg interface")
+	}
+	defer wgIFace.Close()
+
+	dnsServer := NewDefaultServerPermanentUpstream(context.Background(), wgIFace, []string{"8.8.8.8"})
+	err = dnsServer.Initialize()
+	if err != nil {
+		t.Errorf("failed to initialize DNS server: %v", err)
+		return
+	}
+	defer dnsServer.Stop()
+
+	// check initial state
+	resolver := newDnsResolver(dnsServer.service.RuntimeIP(), dnsServer.service.RuntimePort())
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+
+	update := nbdns.Config{
+		ServiceEnable: true,
+		CustomZones: []nbdns.CustomZone{
+			{
+				Domain:  "netbird.cloud",
+				Records: zoneRecords,
+			},
+		},
+		NameServerGroups: []*nbdns.NameServerGroup{
+			{
+				NameServers: []nbdns.NameServer{
+					{
+						IP:     netip.MustParseAddr("8.8.4.4"),
+						NSType: nbdns.UDPNameServerType,
+						Port:   53,
+					},
+				},
+				Enabled: true,
+				Primary: true,
+			},
+		},
+	}
+
+	err = dnsServer.UpdateDNSServer(1, update)
+	if err != nil {
+		t.Errorf("failed to update dns server: %s", err)
+	}
+
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+	ips, err := resolver.LookupHost(context.Background(), zoneRecords[0].Name)
+	if err != nil {
+		t.Fatalf("failed resolve zone record: %v", err)
+	}
+	if ips[0] != zoneRecords[0].RData {
+		t.Fatalf("invalid zone record: %v", err)
+	}
+
+	update2 := nbdns.Config{
+		ServiceEnable: true,
+		CustomZones: []nbdns.CustomZone{
+			{
+				Domain:  "netbird.cloud",
+				Records: zoneRecords,
+			},
+		},
+		NameServerGroups: []*nbdns.NameServerGroup{},
+	}
+
+	err = dnsServer.UpdateDNSServer(2, update2)
+	if err != nil {
+		t.Errorf("failed to update dns server: %s", err)
+	}
+
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+
+	ips, err = resolver.LookupHost(context.Background(), zoneRecords[0].Name)
+	if err != nil {
+		t.Fatalf("failed resolve zone record: %v", err)
+	}
+	if ips[0] != zoneRecords[0].RData {
+		t.Fatalf("invalid zone record: %v", err)
+	}
+}
+
+func TestDNSPermanent_matchOnly(t *testing.T) {
+	wgIFace, err := createWgInterfaceWithBind(t)
+	if err != nil {
+		t.Fatal("failed to initialize wg interface")
+	}
+	defer wgIFace.Close()
+
+	dnsServer := NewDefaultServerPermanentUpstream(context.Background(), wgIFace, []string{"8.8.8.8"})
+	err = dnsServer.Initialize()
+	if err != nil {
+		t.Errorf("failed to initialize DNS server: %v", err)
+		return
+	}
+	defer dnsServer.Stop()
+
+	// check initial state
+	resolver := newDnsResolver(dnsServer.service.RuntimeIP(), dnsServer.service.RuntimePort())
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+
+	update := nbdns.Config{
+		ServiceEnable: true,
+		CustomZones: []nbdns.CustomZone{
+			{
+				Domain:  "netbird.cloud",
+				Records: zoneRecords,
+			},
+		},
+		NameServerGroups: []*nbdns.NameServerGroup{
+			{
+				NameServers: []nbdns.NameServer{
+					{
+						IP:     netip.MustParseAddr("8.8.4.4"),
+						NSType: nbdns.UDPNameServerType,
+						Port:   53,
+					},
+				},
+				Domains: []string{"customdomain.com"},
+				Primary: false,
+			},
+		},
+	}
+
+	err = dnsServer.UpdateDNSServer(1, update)
+	if err != nil {
+		t.Errorf("failed to update dns server: %s", err)
+	}
+
+	_, err = resolver.LookupHost(context.Background(), "netbird.io")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+	ips, err := resolver.LookupHost(context.Background(), zoneRecords[0].Name)
+	if err != nil {
+		t.Fatalf("failed resolve zone record: %v", err)
+	}
+	if ips[0] != zoneRecords[0].RData {
+		t.Fatalf("invalid zone record: %v", err)
+	}
+	_, err = resolver.LookupHost(context.Background(), "customdomain.com")
+	if err != nil {
+		t.Errorf("failed to resolve: %s", err)
+	}
+}
+
+func createWgInterfaceWithBind(t *testing.T) (*iface.WGIface, error) {
+	ov := os.Getenv("NB_WG_KERNEL_DISABLED")
+	defer os.Setenv("NB_WG_KERNEL_DISABLED", ov)
+
+	_ = os.Setenv("NB_WG_KERNEL_DISABLED", "true")
+	newNet, err := stdnet.NewNet(nil)
+	if err != nil {
+		t.Fatalf("create stdnet: %v", err)
+		return nil, nil
+	}
+
+	wgIface, err := iface.NewWGIFace("utun2301", "100.66.100.2/24", iface.DefaultMTU, nil, newNet)
+	if err != nil {
+		t.Fatalf("build interface wireguard: %v", err)
+		return nil, err
+	}
+
+	err = wgIface.Create()
+	if err != nil {
+		t.Fatalf("crate and init wireguard interface: %v", err)
+		return nil, err
+	}
+
+	pf, err := uspfilter.Create(wgIface)
+	if err != nil {
+		t.Fatalf("failed to create uspfilter: %v", err)
+		return nil, err
+	}
+
+	err = wgIface.SetFilter(pf)
+	if err != nil {
+		t.Fatalf("set packet filter: %v", err)
+		return nil, err
+	}
+
+	return wgIface, nil
+}
+
+func newDnsResolver(ip string, port int) *net.Resolver {
+	return &net.Resolver{
+		PreferGo: true,
+		Dial: func(ctx context.Context, network, address string) (net.Conn, error) {
+			d := net.Dialer{
+				Timeout: time.Second * 3,
+			}
+			addr := fmt.Sprintf("%s:%d", ip, port)
+			return d.DialContext(ctx, network, addr)
+		},
 	}
 }

client/internal/dns/service.go

@@ -0,0 +1,18 @@
+package dns
+
+import (
+	"github.com/miekg/dns"
+)
+
+const (
+	defaultPort = 53
+)
+
+type service interface {
+	Listen() error
+	Stop()
+	RegisterMux(domain string, handler dns.Handler)
+	DeregisterMux(key string)
+	RuntimePort() int
+	RuntimeIP() string
+}

client/internal/dns/service_listener.go

@@ -0,0 +1,145 @@
+package dns
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"net/netip"
+	"runtime"
+	"sync"
+	"time"
+
+	"github.com/miekg/dns"
+	log "github.com/sirupsen/logrus"
+)
+
+const (
+	customPort = 5053
+	defaultIP  = "127.0.0.1"
+	customIP   = "127.0.0.153"
+)
+
+type serviceViaListener struct {
+	wgInterface       WGIface
+	dnsMux            *dns.ServeMux
+	customAddr        *netip.AddrPort
+	server            *dns.Server
+	runtimeIP         string
+	runtimePort       int
+	listenerIsRunning bool
+	listenerFlagLock  sync.Mutex
+}
+
+func newServiceViaListener(wgIface WGIface, customAddr *netip.AddrPort) *serviceViaListener {
+	mux := dns.NewServeMux()
+
+	s := &serviceViaListener{
+		wgInterface: wgIface,
+		dnsMux:      mux,
+		customAddr:  customAddr,
+		server: &dns.Server{
+			Net:     "udp",
+			Handler: mux,
+			UDPSize: 65535,
+		},
+	}
+	return s
+}
+
+func (s *serviceViaListener) Listen() error {
+	s.listenerFlagLock.Lock()
+	defer s.listenerFlagLock.Unlock()
+
+	if s.listenerIsRunning {
+		return nil
+	}
+
+	var err error
+	s.runtimeIP, s.runtimePort, err = s.evalRuntimeAddress()
+	if err != nil {
+		log.Errorf("failed to eval runtime address: %s", err)
+		return err
+	}
+	s.server.Addr = fmt.Sprintf("%s:%d", s.runtimeIP, s.runtimePort)
+
+	log.Debugf("starting dns on %s", s.server.Addr)
+	go func() {
+		s.setListenerStatus(true)
+		defer s.setListenerStatus(false)
+
+		err := s.server.ListenAndServe()
+		if err != nil {
+			log.Errorf("dns server running with %d port returned an error: %v. Will not retry", s.runtimePort, err)
+		}
+	}()
+	return nil
+}
+
+func (s *serviceViaListener) Stop() {
+	s.listenerFlagLock.Lock()
+	defer s.listenerFlagLock.Unlock()
+
+	if !s.listenerIsRunning {
+		return
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	err := s.server.ShutdownContext(ctx)
+	if err != nil {
+		log.Errorf("stopping dns server listener returned an error: %v", err)
+	}
+}
+
+func (s *serviceViaListener) RegisterMux(pattern string, handler dns.Handler) {
+	s.dnsMux.Handle(pattern, handler)
+}
+
+func (s *serviceViaListener) DeregisterMux(pattern string) {
+	s.dnsMux.HandleRemove(pattern)
+}
+
+func (s *serviceViaListener) RuntimePort() int {
+	return s.runtimePort
+}
+
+func (s *serviceViaListener) RuntimeIP() string {
+	return s.runtimeIP
+}
+
+func (s *serviceViaListener) setListenerStatus(running bool) {
+	s.listenerIsRunning = running
+}
+
+func (s *serviceViaListener) getFirstListenerAvailable() (string, int, error) {
+	ips := []string{defaultIP, customIP}
+	if runtime.GOOS != "darwin" {
+		ips = append([]string{s.wgInterface.Address().IP.String()}, ips...)
+	}
+	ports := []int{defaultPort, customPort}
+	for _, port := range ports {
+		for _, ip := range ips {
+			addrString := fmt.Sprintf("%s:%d", ip, port)
+			udpAddr := net.UDPAddrFromAddrPort(netip.MustParseAddrPort(addrString))
+			probeListener, err := net.ListenUDP("udp", udpAddr)
+			if err == nil {
+				err = probeListener.Close()
+				if err != nil {
+					log.Errorf("got an error closing the probe listener, error: %s", err)
+				}
+				return ip, port, nil
+			}
+			log.Warnf("binding dns on %s is not available, error: %s", addrString, err)
+		}
+	}
+	return "", 0, fmt.Errorf("unable to find an unused ip and port combination. IPs tested: %v and ports %v", ips, ports)
+}
+
+func (s *serviceViaListener) evalRuntimeAddress() (string, int, error) {
+	if s.customAddr != nil {
+		return s.customAddr.Addr().String(), int(s.customAddr.Port()), nil
+	}
+
+	return s.getFirstListenerAvailable()
+}

client/internal/dns/service_memory.go

@@ -0,0 +1,139 @@
+package dns
+
+import (
+	"fmt"
+	"math/big"
+	"net"
+	"sync"
+
+	"github.com/google/gopacket"
+	"github.com/google/gopacket/layers"
+	"github.com/miekg/dns"
+	log "github.com/sirupsen/logrus"
+)
+
+type serviceViaMemory struct {
+	wgInterface       WGIface
+	dnsMux            *dns.ServeMux
+	runtimeIP         string
+	runtimePort       int
+	udpFilterHookID   string
+	listenerIsRunning bool
+	listenerFlagLock  sync.Mutex
+}
+
+func newServiceViaMemory(wgIface WGIface) *serviceViaMemory {
+	s := &serviceViaMemory{
+		wgInterface: wgIface,
+		dnsMux:      dns.NewServeMux(),
+
+		runtimeIP:   getLastIPFromNetwork(wgIface.Address().Network, 1),
+		runtimePort: defaultPort,
+	}
+	return s
+}
+
+func (s *serviceViaMemory) Listen() error {
+	s.listenerFlagLock.Lock()
+	defer s.listenerFlagLock.Unlock()
+
+	if s.listenerIsRunning {
+		return nil
+	}
+
+	var err error
+	s.udpFilterHookID, err = s.filterDNSTraffic()
+	if err != nil {
+		return err
+	}
+	s.listenerIsRunning = true
+
+	log.Debugf("dns service listening on: %s", s.RuntimeIP())
+	return nil
+}
+
+func (s *serviceViaMemory) Stop() {
+	s.listenerFlagLock.Lock()
+	defer s.listenerFlagLock.Unlock()
+
+	if !s.listenerIsRunning {
+		return
+	}
+
+	if err := s.wgInterface.GetFilter().RemovePacketHook(s.udpFilterHookID); err != nil {
+		log.Errorf("unable to remove DNS packet hook: %s", err)
+	}
+
+	s.listenerIsRunning = false
+}
+
+func (s *serviceViaMemory) RegisterMux(pattern string, handler dns.Handler) {
+	s.dnsMux.Handle(pattern, handler)
+}
+
+func (s *serviceViaMemory) DeregisterMux(pattern string) {
+	s.dnsMux.HandleRemove(pattern)
+}
+
+func (s *serviceViaMemory) RuntimePort() int {
+	return s.runtimePort
+}
+
+func (s *serviceViaMemory) RuntimeIP() string {
+	return s.runtimeIP
+}
+
+func (s *serviceViaMemory) filterDNSTraffic() (string, error) {
+	filter := s.wgInterface.GetFilter()
+	if filter == nil {
+		return "", fmt.Errorf("can't set DNS filter, filter not initialized")
+	}
+
+	firstLayerDecoder := layers.LayerTypeIPv4
+	if s.wgInterface.Address().Network.IP.To4() == nil {
+		firstLayerDecoder = layers.LayerTypeIPv6
+	}
+
+	hook := func(packetData []byte) bool {
+		// Decode the packet
+		packet := gopacket.NewPacket(packetData, firstLayerDecoder, gopacket.Default)
+
+		// Get the UDP layer
+		udpLayer := packet.Layer(layers.LayerTypeUDP)
+		udp := udpLayer.(*layers.UDP)
+
+		msg := new(dns.Msg)
+		if err := msg.Unpack(udp.Payload); err != nil {
+			log.Tracef("parse DNS request: %v", err)
+			return true
+		}
+
+		writer := responseWriter{
+			packet: packet,
+			device: s.wgInterface.GetDevice().Device,
+		}
+		go s.dnsMux.ServeDNS(&writer, msg)
+		return true
+	}
+
+	return filter.AddUDPPacketHook(false, net.ParseIP(s.runtimeIP), uint16(s.runtimePort), hook), nil
+}
+
+func getLastIPFromNetwork(network *net.IPNet, fromEnd int) string {
+	// Calculate the last IP in the CIDR range
+	var endIP net.IP
+	for i := 0; i < len(network.IP); i++ {
+		endIP = append(endIP, network.IP[i]|^network.Mask[i])
+	}
+
+	// convert to big.Int
+	endInt := big.NewInt(0)
+	endInt.SetBytes(endIP)
+
+	// subtract fromEnd from the last ip
+	fromEndBig := big.NewInt(int64(fromEnd))
+	resultInt := big.NewInt(0)
+	resultInt.Sub(endInt, fromEndBig)
+
+	return net.IP(resultInt.Bytes()).String()
+}

client/internal/dns/service_memory_test.go

@@ -0,0 +1,31 @@
+package dns
+
+import (
+	"net"
+	"testing"
+)
+
+func TestGetLastIPFromNetwork(t *testing.T) {
+	tests := []struct {
+		addr string
+		ip   string
+	}{
+		{"2001:db8::/32", "2001:db8:ffff:ffff:ffff:ffff:ffff:fffe"},
+		{"192.168.0.0/30", "192.168.0.2"},
+		{"192.168.0.0/16", "192.168.255.254"},
+		{"192.168.0.0/24", "192.168.0.254"},
+	}
+
+	for _, tt := range tests {
+		_, ipnet, err := net.ParseCIDR(tt.addr)
+		if err != nil {
+			t.Errorf("Error parsing CIDR: %v", err)
+			return
+		}
+
+		lastIP := getLastIPFromNetwork(ipnet, 1)
+		if lastIP != tt.ip {
+			t.Errorf("wrong IP address, expected %s: got %s", tt.ip, lastIP)
+		}
+	}
+}

client/internal/dns/systemd_linux.go

@@ -1,3 +1,5 @@
+//go:build !android
+
 package dns
 
 import (
@@ -9,10 +11,10 @@ import (
 
 	"github.com/godbus/dbus/v5"
 	"github.com/miekg/dns"
-	nbdns "github.com/netbirdio/netbird/dns"
-	"github.com/netbirdio/netbird/iface"
 	log "github.com/sirupsen/logrus"
 	"golang.org/x/sys/unix"
+
+	nbdns "github.com/netbirdio/netbird/dns"
 )
 
 const (
@@ -50,7 +52,7 @@ type systemdDbusLinkDomainsInput struct {
 	MatchOnly bool
 }
 
-func newSystemdDbusConfigurator(wgInterface *iface.WGIface) (hostManager, error) {
+func newSystemdDbusConfigurator(wgInterface WGIface) (hostManager, error) {
 	iface, err := net.InterfaceByName(wgInterface.Name())
 	if err != nil {
 		return nil, err
@@ -75,6 +77,10 @@ func newSystemdDbusConfigurator(wgInterface *iface.WGIface) (hostManager, error)
 	}, nil
 }
 
+func (s *systemdDbusConfigurator) supportCustomPort() bool {
+	return true
+}
+
 func (s *systemdDbusConfigurator) applyDNSConfig(config hostDNSConfig) error {
 	parsedIP, err := netip.ParseAddr(config.serverIP)
 	if err != nil {

client/internal/dns/upstream.go

@@ -3,24 +3,31 @@ package dns
 import (
 	"context"
 	"errors"
+	"fmt"
 	"net"
 	"sync"
 	"sync/atomic"
 	"time"
 
+	"github.com/cenkalti/backoff/v4"
 	"github.com/miekg/dns"
 	log "github.com/sirupsen/logrus"
 )
 
 const (
-	failsTillDeact   = int32(3)
-	reactivatePeriod = time.Minute
+	failsTillDeact   = int32(5)
+	reactivatePeriod = 30 * time.Second
 	upstreamTimeout  = 15 * time.Second
 )
 
+type upstreamClient interface {
+	ExchangeContext(ctx context.Context, m *dns.Msg, a string) (r *dns.Msg, rtt time.Duration, err error)
+}
+
 type upstreamResolver struct {
 	ctx              context.Context
-	upstreamClient   *dns.Client
+	cancel           context.CancelFunc
+	upstreamClient   upstreamClient
 	upstreamServers  []string
 	disabled         bool
 	failsCount       atomic.Int32
@@ -33,9 +40,11 @@ type upstreamResolver struct {
 	reactivate func()
 }
 
-func newUpstreamResolver(ctx context.Context) *upstreamResolver {
+func newUpstreamResolver(parentCTX context.Context) *upstreamResolver {
+	ctx, cancel := context.WithCancel(parentCTX)
 	return &upstreamResolver{
 		ctx:              ctx,
+		cancel:           cancel,
 		upstreamClient:   &dns.Client{},
 		upstreamTimeout:  upstreamTimeout,
 		reactivatePeriod: reactivatePeriod,
@@ -43,6 +52,11 @@ func newUpstreamResolver(ctx context.Context) *upstreamResolver {
 	}
 }
 
+func (u *upstreamResolver) stop() {
+	log.Debugf("stoping serving DNS for upstreams %s", u.upstreamServers)
+	u.cancel()
+}
+
 // ServeDNS handles a DNS request
 func (u *upstreamResolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
 	defer u.checkUpstreamFails()
@@ -107,28 +121,57 @@ func (u *upstreamResolver) checkUpstreamFails() {
 		log.Warnf("upstream resolving is disabled for %v", reactivatePeriod)
 		u.deactivate()
 		u.disabled = true
-		go u.waitUntilReactivation()
+		go u.waitUntilResponse()
 	}
 }
 
-// waitUntilReactivation reset fails counter and activates upstream resolving
-func (u *upstreamResolver) waitUntilReactivation() {
-	timer := time.NewTimer(u.reactivatePeriod)
-	defer func() {
-		if !timer.Stop() {
-			<-timer.C
-		}
-	}()
-
-	select {
-	case <-u.ctx.Done():
-		return
-	case <-timer.C:
-		log.Info("upstream resolving is reactivated")
-		u.failsCount.Store(0)
-		u.reactivate()
-		u.disabled = false
+// waitUntilResponse retries, in an exponential interval, querying the upstream servers until it gets a positive response
+func (u *upstreamResolver) waitUntilResponse() {
+	exponentialBackOff := &backoff.ExponentialBackOff{
+		InitialInterval:     500 * time.Millisecond,
+		RandomizationFactor: 0.5,
+		Multiplier:          1.1,
+		MaxInterval:         u.reactivatePeriod,
+		MaxElapsedTime:      0,
+		Stop:                backoff.Stop,
+		Clock:               backoff.SystemClock,
 	}
+
+	r := new(dns.Msg).SetQuestion("netbird.io.", dns.TypeA)
+
+	operation := func() error {
+		select {
+		case <-u.ctx.Done():
+			return backoff.Permanent(fmt.Errorf("exiting upstream retry loop for upstreams %s: parent context has been canceled", u.upstreamServers))
+		default:
+		}
+
+		var err error
+		for _, upstream := range u.upstreamServers {
+			ctx, cancel := context.WithTimeout(u.ctx, u.upstreamTimeout)
+			_, _, err = u.upstreamClient.ExchangeContext(ctx, r, upstream)
+
+			cancel()
+
+			if err == nil {
+				return nil
+			}
+		}
+
+		log.Tracef("checking connectivity with upstreams %s failed with error: %s. Retrying in %s", err, u.upstreamServers, exponentialBackOff.NextBackOff())
+		return fmt.Errorf("got an error from upstream check call")
+	}
+
+	err := backoff.Retry(operation, exponentialBackOff)
+	if err != nil {
+		log.Warn(err)
+		return
+	}
+
+	log.Infof("upstreams %s are responsive again. Adding them back to system", u.upstreamServers)
+	u.failsCount.Store(0)
+	u.reactivate()
+	u.disabled = false
 }
 
 // isTimeout returns true if the given error is a network timeout error.

client/internal/dns/upstream_test.go

@@ -2,10 +2,11 @@ package dns
 
 import (
 	"context"
-	"github.com/miekg/dns"
 	"strings"
 	"testing"
 	"time"
+
+	"github.com/miekg/dns"
 )
 
 func TestUpstreamResolver_ServeDNS(t *testing.T) {
@@ -106,8 +107,29 @@ func TestUpstreamResolver_ServeDNS(t *testing.T) {
 	}
 }
 
+type mockUpstreamResolver struct {
+	r   *dns.Msg
+	rtt time.Duration
+	err error
+}
+
+// ExchangeContext mock implementation of ExchangeContext from upstreamResolver
+func (c mockUpstreamResolver) ExchangeContext(_ context.Context, _ *dns.Msg, _ string) (r *dns.Msg, rtt time.Duration, err error) {
+	return c.r, c.rtt, c.err
+}
+
 func TestUpstreamResolver_DeactivationReactivation(t *testing.T) {
-	resolver := newUpstreamResolver(context.TODO())
+	resolver := &upstreamResolver{
+		ctx: context.TODO(),
+		upstreamClient: &mockUpstreamResolver{
+			err: nil,
+			r:   new(dns.Msg),
+			rtt: time.Millisecond,
+		},
+		upstreamTimeout:  upstreamTimeout,
+		reactivatePeriod: reactivatePeriod,
+		failsTillDeact:   failsTillDeact,
+	}
 	resolver.upstreamServers = []string{"0.0.0.0:-1"}
 	resolver.failsTillDeact = 0
 	resolver.reactivatePeriod = time.Microsecond * 100

client/internal/dns/wgiface.go

@@ -0,0 +1,14 @@
+//go:build !windows
+
+package dns
+
+import "github.com/netbirdio/netbird/iface"
+
+// WGIface defines subset methods of interface required for manager
+type WGIface interface {
+	Name() string
+	Address() iface.WGAddress
+	IsUserspaceBind() bool
+	GetFilter() iface.PacketFilter
+	GetDevice() *iface.DeviceWrapper
+}

client/internal/dns/wgiface_windows.go

@@ -0,0 +1,13 @@
+package dns
+
+import "github.com/netbirdio/netbird/iface"
+
+// WGIface defines subset methods of interface required for manager
+type WGIface interface {
+	Name() string
+	Address() iface.WGAddress
+	IsUserspaceBind() bool
+	GetFilter() iface.PacketFilter
+	GetDevice() *iface.DeviceWrapper
+	GetInterfaceGUIDString() (string, error)
+}

client/internal/engine.go

@@ -3,6 +3,7 @@ package internal
 import (
 	"context"
 	"fmt"
+	"io"
 	"math/rand"
 	"net"
 	"net/netip"
@@ -16,16 +17,19 @@ import (
 	log "github.com/sirupsen/logrus"
 	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
 
+	"github.com/netbirdio/netbird/client/internal/acl"
 	"github.com/netbirdio/netbird/client/internal/dns"
 	"github.com/netbirdio/netbird/client/internal/peer"
 	"github.com/netbirdio/netbird/client/internal/routemanager"
-	"github.com/netbirdio/netbird/client/internal/stdnet"
+	"github.com/netbirdio/netbird/client/internal/wgproxy"
 	nbssh "github.com/netbirdio/netbird/client/ssh"
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/iface"
+	"github.com/netbirdio/netbird/iface/bind"
 	mgm "github.com/netbirdio/netbird/management/client"
 	mgmProto "github.com/netbirdio/netbird/management/proto"
 	"github.com/netbirdio/netbird/route"
+	"github.com/netbirdio/netbird/sharedsock"
 	signal "github.com/netbirdio/netbird/signal/client"
 	sProto "github.com/netbirdio/netbird/signal/proto"
 	"github.com/netbirdio/netbird/util"
@@ -46,10 +50,6 @@ var ErrResetConnection = fmt.Errorf("reset connection")
 type EngineConfig struct {
 	WgPort      int
 	WgIfaceName string
-	// TunAdapter is option. It is necessary for mobile version.
-	TunAdapter iface.TunAdapter
-
-	IFaceDiscover stdnet.ExternalIFaceDiscover
 
 	// WgAddr is a Wireguard local address (Netbird Network IP)
 	WgAddr string
@@ -89,7 +89,9 @@ type Engine struct {
 	// syncMsgMux is used to guarantee sequential Management Service message processing
 	syncMsgMux *sync.Mutex
 
-	config *EngineConfig
+	config    *EngineConfig
+	mobileDep MobileDependency
+
 	// STUNs is a list of STUN servers used by ICE
 	STUNs []*ice.URL
 	// TURNs is a list of STUN servers used by ICE
@@ -99,12 +101,11 @@ type Engine struct {
 
 	ctx context.Context
 
-	wgInterface *iface.WGIface
+	wgInterface    *iface.WGIface
+	wgProxyFactory *wgproxy.Factory
 
-	udpMux          ice.UDPMux
-	udpMuxSrflx     ice.UniversalUDPMux
-	udpMuxConn      *net.UDPConn
-	udpMuxConnSrflx *net.UDPConn
+	udpMux     *bind.UniversalUDPMuxDefault
+	udpMuxConn io.Closer
 
 	// networkSerial is the latest CurrentSerial (state ID) of the network sent by the Management service
 	networkSerial uint64
@@ -115,6 +116,7 @@ type Engine struct {
 	statusRecorder *peer.Status
 
 	routeManager routemanager.Manager
+	acl          acl.Manager
 
 	dnsServer dns.Server
 }
@@ -129,8 +131,9 @@ type Peer struct {
 func NewEngine(
 	ctx context.Context, cancel context.CancelFunc,
 	signalClient signal.Client, mgmClient mgm.Client,
-	config *EngineConfig, statusRecorder *peer.Status,
+	config *EngineConfig, mobileDep MobileDependency, statusRecorder *peer.Status,
 ) *Engine {
+
 	return &Engine{
 		ctx:            ctx,
 		cancel:         cancel,
@@ -139,11 +142,13 @@ func NewEngine(
 		peerConns:      make(map[string]*peer.Conn),
 		syncMsgMux:     &sync.Mutex{},
 		config:         config,
+		mobileDep:      mobileDep,
 		STUNs:          []*ice.URL{},
 		TURNs:          []*ice.URL{},
 		networkSerial:  0,
 		sshServerFunc:  nbssh.DefaultSSHServer,
 		statusRecorder: statusRecorder,
+		wgProxyFactory: wgproxy.NewFactory(config.WgPort),
 	}
 }
 
@@ -178,15 +183,48 @@ func (e *Engine) Start() error {
 	var err error
 	transportNet, err := e.newStdNet()
 	if err != nil {
-		log.Warnf("failed to create pion's stdnet: %s", err)
+		log.Errorf("failed to create pion's stdnet: %s", err)
 	}
-	e.wgInterface, err = iface.NewWGIFace(wgIFaceName, wgAddr, iface.DefaultMTU, e.config.TunAdapter, transportNet)
+
+	e.wgInterface, err = iface.NewWGIFace(wgIFaceName, wgAddr, iface.DefaultMTU, e.mobileDep.TunAdapter, transportNet)
 	if err != nil {
 		log.Errorf("failed creating wireguard interface instance %s: [%s]", wgIFaceName, err.Error())
 		return err
 	}
 
-	err = e.wgInterface.Create()
+	var routes []*route.Route
+
+	if runtime.GOOS == "android" {
+		routes, err = e.readInitialSettings()
+		if err != nil {
+			return err
+		}
+		if e.dnsServer == nil {
+			e.dnsServer = dns.NewDefaultServerPermanentUpstream(e.ctx, e.wgInterface, e.mobileDep.HostDNSAddresses)
+			go e.mobileDep.DnsReadyListener.OnReady()
+		}
+	} else {
+		// todo fix custom address
+		if e.dnsServer == nil {
+			e.dnsServer, err = dns.NewDefaultServer(e.ctx, e.wgInterface, e.config.CustomDNSAddress)
+			if err != nil {
+				e.close()
+				return err
+			}
+		}
+	}
+
+	e.routeManager = routemanager.NewManager(e.ctx, e.config.WgPrivateKey.PublicKey().String(), e.wgInterface, e.statusRecorder, routes)
+	e.routeManager.SetRouteChangeListener(e.mobileDep.RouteListener)
+
+	if runtime.GOOS != "android" {
+		err = e.wgInterface.Create()
+	} else {
+		err = e.wgInterface.CreateOnMobile(iface.MobileIFaceArguments{
+			Routes: e.routeManager.InitialRouteRange(),
+			Dns:    e.dnsServer.DnsIP(),
+		})
+	}
 	if err != nil {
 		log.Errorf("failed creating tunnel interface %s: [%s]", wgIFaceName, err.Error())
 		e.close()
@@ -207,45 +245,29 @@ func (e *Engine) Start() error {
 			e.close()
 			return err
 		}
-		e.udpMux = udpMux.UDPMuxDefault
-		e.udpMuxSrflx = udpMux
+		e.udpMux = udpMux
 		log.Infof("using userspace bind mode %s", udpMux.LocalAddr().String())
 	} else {
-		networkName := "udp"
-		if e.config.DisableIPv6Discovery {
-			networkName = "udp4"
-		}
-		e.udpMuxConn, err = net.ListenUDP(networkName, &net.UDPAddr{Port: e.config.UDPMuxPort})
+		rawSock, err := sharedsock.Listen(e.config.WgPort, sharedsock.NewIncomingSTUNFilter())
 		if err != nil {
-			log.Errorf("failed listening on UDP port %d: [%s]", e.config.UDPMuxPort, err.Error())
-			e.close()
 			return err
 		}
-		udpMuxParams := ice.UDPMuxParams{
-			UDPConn: e.udpMuxConn,
-			Net:     transportNet,
-		}
-		e.udpMux = ice.NewUDPMuxDefault(udpMuxParams)
-
-		e.udpMuxConnSrflx, err = net.ListenUDP(networkName, &net.UDPAddr{Port: e.config.UDPMuxSrflxPort})
-		if err != nil {
-			log.Errorf("failed listening on UDP port %d: [%s]", e.config.UDPMuxSrflxPort, err.Error())
-			e.close()
-			return err
-		}
-		e.udpMuxSrflx = ice.NewUniversalUDPMuxDefault(ice.UniversalUDPMuxParams{UDPConn: e.udpMuxConnSrflx, Net: transportNet})
+		mux := bind.NewUniversalUDPMuxDefault(bind.UniversalUDPMuxParams{UDPConn: rawSock, Net: transportNet})
+		go mux.ReadFromConn(e.ctx)
+		e.udpMuxConn = rawSock
+		e.udpMux = mux
 	}
 
-	e.routeManager = routemanager.NewManager(e.ctx, e.config.WgPrivateKey.PublicKey().String(), e.wgInterface, e.statusRecorder)
+	if acl, err := acl.Create(e.wgInterface); err != nil {
+		log.Errorf("failed to create ACL manager, policy will not work: %s", err.Error())
+	} else {
+		e.acl = acl
+	}
 
-	if e.dnsServer == nil {
-		// todo fix custom address
-		dnsServer, err := dns.NewDefaultServer(e.ctx, e.wgInterface, e.config.CustomDNSAddress)
-		if err != nil {
-			e.close()
-			return err
-		}
-		e.dnsServer = dnsServer
+	err = e.dnsServer.Initialize()
+	if err != nil {
+		e.close()
+		return err
 	}
 
 	e.receiveSignalEvents()
@@ -263,7 +285,7 @@ func (e *Engine) modifyPeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
 	for _, p := range peersUpdate {
 		peerPubKey := p.GetWgPubKey()
 		if peerConn, ok := e.peerConns[peerPubKey]; ok {
-			if peerConn.GetConf().ProxyConfig.AllowedIps != strings.Join(p.AllowedIps, ",") {
+			if peerConn.WgConfig().AllowedIps != strings.Join(p.AllowedIps, ",") {
 				modified = append(modified, p)
 				continue
 			}
@@ -395,9 +417,6 @@ func SignalOfferAnswer(offerAnswer peer.OfferAnswer, myKey wgtypes.Key, remoteKe
 		return err
 	}
 
-	// indicates message support in gRPC
-	msg.Body.FeaturesSupported = []uint32{signal.DirectCheck}
-
 	err = s.Send(msg)
 	if err != nil {
 		return err
@@ -591,6 +610,7 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 	// cleanup request, most likely our peer has been deleted
 	if networkMap.GetRemotePeersIsEmpty() {
 		err := e.removeAllPeers()
+		e.statusRecorder.FinishPeerListModifications()
 		if err != nil {
 			return err
 		}
@@ -610,6 +630,8 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 			return err
 		}
 
+		e.statusRecorder.FinishPeerListModifications()
+
 		// update SSHServer by adding remote peer SSH keys
 		if !isNil(e.sshServer) {
 			for _, config := range networkMap.GetRemotePeers() {
@@ -641,6 +663,9 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 		log.Errorf("failed to update dns server, err: %v", err)
 	}
 
+	if e.acl != nil {
+		e.acl.ApplyFiltering(networkMap)
+	}
 	e.networkSerial = serial
 	return nil
 }
@@ -742,17 +767,13 @@ func (e *Engine) addNewPeer(peerConfig *mgmProto.RemotePeerConfig) error {
 		}
 		e.peerConns[peerKey] = conn
 
-		err = e.statusRecorder.AddPeer(peerKey)
+		err = e.statusRecorder.AddPeer(peerKey, peerConfig.Fqdn)
 		if err != nil {
 			log.Warnf("error adding peer %s to status recorder, got error: %v", peerKey, err)
 		}
 
 		go e.connWorker(conn, peerKey)
 	}
-	err := e.statusRecorder.UpdatePeerFQDN(peerKey, peerConfig.Fqdn)
-	if err != nil {
-		log.Warnf("error updating peer's %s fqdn in the status recorder, got error: %v", peerKey, err)
-	}
 	return nil
 }
 
@@ -777,9 +798,7 @@ func (e *Engine) connWorker(conn *peer.Conn, peerKey string) {
 
 		// we might have received new STUN and TURN servers meanwhile, so update them
 		e.syncMsgMux.Lock()
-		conf := conn.GetConf()
-		conf.StunTurn = append(e.STUNs, e.TURNs...)
-		conn.UpdateConf(conf)
+		conn.UpdateStunTurn(append(e.STUNs, e.TURNs...))
 		e.syncMsgMux.Unlock()
 
 		err := conn.Open()
@@ -795,19 +814,27 @@ func (e *Engine) connWorker(conn *peer.Conn, peerKey string) {
 	}
 }
 
-func (e Engine) peerExists(peerKey string) bool {
+func (e *Engine) peerExists(peerKey string) bool {
 	e.syncMsgMux.Lock()
 	defer e.syncMsgMux.Unlock()
 	_, ok := e.peerConns[peerKey]
 	return ok
 }
 
-func (e Engine) createPeerConn(pubKey string, allowedIPs string) (*peer.Conn, error) {
+func (e *Engine) createPeerConn(pubKey string, allowedIPs string) (*peer.Conn, error) {
 	log.Debugf("creating peer connection %s", pubKey)
 	var stunTurn []*ice.URL
 	stunTurn = append(stunTurn, e.STUNs...)
 	stunTurn = append(stunTurn, e.TURNs...)
 
+	wgConfig := peer.WgConfig{
+		RemoteKey:    pubKey,
+		WgListenPort: e.config.WgPort,
+		WgInterface:  e.wgInterface,
+		AllowedIps:   allowedIPs,
+		PreSharedKey: e.config.PreSharedKey,
+	}
+
 	// randomize connection timeout
 	timeout := time.Duration(rand.Intn(PeerConnectionTimeoutMax-PeerConnectionTimeoutMin)+PeerConnectionTimeoutMin) * time.Millisecond
 	config := peer.ConnConfig{
@@ -817,14 +844,15 @@ func (e Engine) createPeerConn(pubKey string, allowedIPs string) (*peer.Conn, er
 		InterfaceBlackList:   e.config.IFaceBlackList,
 		DisableIPv6Discovery: e.config.DisableIPv6Discovery,
 		Timeout:              timeout,
-		UDPMux:               e.udpMux,
-		UDPMuxSrflx:          e.udpMuxSrflx,
+		UDPMux:               e.udpMux.UDPMuxDefault,
+		UDPMuxSrflx:          e.udpMux,
+		WgConfig:             wgConfig,
 		LocalWgPort:          e.config.WgPort,
 		NATExternalIPs:       e.parseNATExternalIPMappings(),
-		AllowedIPs:           allowedIPs,
+		UserspaceBind:        e.wgInterface.IsUserspaceBind(),
 	}
 
-	peerConn, err := peer.NewConn(config, e.wgInterface, e.statusRecorder, e.config.TunAdapter, e.config.IFaceDiscover)
+	peerConn, err := peer.NewConn(config, e.statusRecorder, e.wgProxyFactory, e.mobileDep.TunAdapter, e.mobileDep.IFaceDiscover)
 	if err != nil {
 		return nil, err
 	}
@@ -910,18 +938,6 @@ func (e *Engine) receiveSignalEvents() {
 				}
 				conn.OnRemoteCandidate(candidate)
 			case sProto.Body_MODE:
-				protoMode := msg.GetBody().GetMode()
-				if protoMode == nil {
-					return fmt.Errorf("received an empty mode message")
-				}
-
-				err := conn.OnModeMessage(peer.ModeMessage{
-					Direct: protoMode.GetDirect(),
-				})
-				if err != nil {
-					log.Errorf("failed processing a mode message -> %s", err)
-					return err
-				}
 			}
 
 			return nil
@@ -993,6 +1009,10 @@ func (e *Engine) parseNATExternalIPMappings() []string {
 }
 
 func (e *Engine) close() {
+	if err := e.wgProxyFactory.Free(); err != nil {
+		log.Errorf("failed closing ebpf proxy: %s", err)
+	}
+
 	log.Debugf("removing Netbird interface %s", e.config.WgIfaceName)
 	if e.wgInterface != nil {
 		if err := e.wgInterface.Close(); err != nil {
@@ -1012,12 +1032,6 @@ func (e *Engine) close() {
 		}
 	}
 
-	if e.udpMuxConnSrflx != nil {
-		if err := e.udpMuxConnSrflx.Close(); err != nil {
-			log.Debugf("close server reflexive udp mux connection: %v", err)
-		}
-	}
-
 	if !isNil(e.sshServer) {
 		err := e.sshServer.Stop()
 		if err != nil {
@@ -1033,6 +1047,18 @@ func (e *Engine) close() {
 		e.dnsServer.Stop()
 	}
 
+	if e.acl != nil {
+		e.acl.Stop()
+	}
+}
+
+func (e *Engine) readInitialSettings() ([]*route.Route, error) {
+	netMap, err := e.mgmClient.GetNetworkMap()
+	if err != nil {
+		return nil, err
+	}
+	routes := toRoutes(netMap.GetRoutes())
+	return routes, nil
 }
 
 func findIPFromInterfaceName(ifaceName string) (net.IP, error) {

client/internal/engine_stdnet_android.go

@@ -3,5 +3,5 @@ package internal
 import "github.com/netbirdio/netbird/client/internal/stdnet"
 
 func (e *Engine) newStdNet() (*stdnet.Net, error) {
-	return stdnet.NewNetWithDiscover(e.config.IFaceDiscover, e.config.IFaceBlackList)
+	return stdnet.NewNetWithDiscover(e.mobileDep.IFaceDiscover, e.config.IFaceBlackList)
 }

client/internal/engine_test.go

@@ -3,8 +3,6 @@ package internal
 import (
 	"context"
 	"fmt"
-	"github.com/netbirdio/netbird/iface/bind"
-	"github.com/pion/transport/v2/stdnet"
 	"net"
 	"net/netip"
 	"os"
@@ -15,6 +13,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/pion/transport/v2/stdnet"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -29,6 +28,7 @@ import (
 	"github.com/netbirdio/netbird/client/system"
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/iface"
+	"github.com/netbirdio/netbird/iface/bind"
 	mgmt "github.com/netbirdio/netbird/management/client"
 	mgmtProto "github.com/netbirdio/netbird/management/proto"
 	"github.com/netbirdio/netbird/management/server"
@@ -74,7 +74,7 @@ func TestEngine_SSH(t *testing.T) {
 		WgAddr:       "100.64.0.1/24",
 		WgPrivateKey: key,
 		WgPort:       33100,
-	}, peer.NewRecorder("https://mgm"))
+	}, MobileDependency{}, peer.NewRecorder("https://mgm"))
 
 	engine.dnsServer = &dns.MockServer{
 		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
@@ -208,7 +208,7 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
 		WgAddr:       "100.64.0.1/24",
 		WgPrivateKey: key,
 		WgPort:       33100,
-	}, peer.NewRecorder("https://mgm"))
+	}, MobileDependency{}, peer.NewRecorder("https://mgm"))
 	newNet, err := stdnet.NewNet()
 	if err != nil {
 		t.Fatal(err)
@@ -217,7 +217,7 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	engine.routeManager = routemanager.NewManager(ctx, key.PublicKey().String(), engine.wgInterface, engine.statusRecorder)
+	engine.routeManager = routemanager.NewManager(ctx, key.PublicKey().String(), engine.wgInterface, engine.statusRecorder, nil)
 	engine.dnsServer = &dns.MockServer{
 		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
 	}
@@ -367,9 +367,9 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
 					t.Errorf("expecting Engine.peerConns to contain peer %s", p)
 				}
 				expectedAllowedIPs := strings.Join(p.AllowedIps, ",")
-				if conn.GetConf().ProxyConfig.AllowedIps != expectedAllowedIPs {
+				if conn.WgConfig().AllowedIps != expectedAllowedIPs {
 					t.Errorf("expecting peer %s to have AllowedIPs= %s, got %s", p.GetWgPubKey(),
-						expectedAllowedIPs, conn.GetConf().ProxyConfig.AllowedIps)
+						expectedAllowedIPs, conn.WgConfig().AllowedIps)
 				}
 			}
 		})
@@ -404,7 +404,7 @@ func TestEngine_Sync(t *testing.T) {
 		WgAddr:       "100.64.0.1/24",
 		WgPrivateKey: key,
 		WgPort:       33100,
-	}, peer.NewRecorder("https://mgm"))
+	}, MobileDependency{}, peer.NewRecorder("https://mgm"))
 
 	engine.dnsServer = &dns.MockServer{
 		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
@@ -562,7 +562,7 @@ func TestEngine_UpdateNetworkMapWithRoutes(t *testing.T) {
 				WgAddr:       wgAddr,
 				WgPrivateKey: key,
 				WgPort:       33100,
-			}, peer.NewRecorder("https://mgm"))
+			}, MobileDependency{}, peer.NewRecorder("https://mgm"))
 			newNet, err := stdnet.NewNet()
 			if err != nil {
 				t.Fatal(err)
@@ -731,7 +731,7 @@ func TestEngine_UpdateNetworkMapWithDNSUpdate(t *testing.T) {
 				WgAddr:       wgAddr,
 				WgPrivateKey: key,
 				WgPort:       33100,
-			}, peer.NewRecorder("https://mgm"))
+			}, MobileDependency{}, peer.NewRecorder("https://mgm"))
 			newNet, err := stdnet.NewNet()
 			if err != nil {
 				t.Fatal(err)
@@ -1000,7 +1000,7 @@ func createEngine(ctx context.Context, cancel context.CancelFunc, setupKey strin
 		WgPort:       wgPort,
 	}
 
-	return NewEngine(ctx, cancel, signalClient, mgmtClient, conf, peer.NewRecorder("https://mgm")), nil
+	return NewEngine(ctx, cancel, signalClient, mgmtClient, conf, MobileDependency{}, peer.NewRecorder("https://mgm")), nil
 }
 
 func startSignal() (*grpc.Server, string, error) {
@@ -1039,7 +1039,7 @@ func startManagement(dataDir string) (*grpc.Server, string, error) {
 		return nil, "", err
 	}
 	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
-	store, err := server.NewFileStore(config.Datadir)
+	store, err := server.NewFileStore(config.Datadir, nil)
 	if err != nil {
 		log.Fatalf("failed creating a store: %s: %v", config.Datadir, err)
 	}

client/internal/mobile_dependency.go

@@ -0,0 +1,17 @@
+package internal
+
+import (
+	"github.com/netbirdio/netbird/client/internal/dns"
+	"github.com/netbirdio/netbird/client/internal/routemanager"
+	"github.com/netbirdio/netbird/client/internal/stdnet"
+	"github.com/netbirdio/netbird/iface"
+)
+
+// MobileDependency collect all dependencies for mobile platform
+type MobileDependency struct {
+	TunAdapter       iface.TunAdapter
+	IFaceDiscover    stdnet.ExternalIFaceDiscover
+	RouteListener    routemanager.RouteListener
+	HostDNSAddresses []string
+	DnsReadyListener dns.ReadyListener
+}

client/internal/oauth.go

@@ -1,281 +0,0 @@
-package internal
-
-import (
-	"context"
-	"encoding/base64"
-	"encoding/json"
-	"fmt"
-	"io"
-	"net/http"
-	"net/url"
-	"reflect"
-	"strings"
-	"time"
-)
-
-// OAuthClient is a OAuth client interface for various idp providers
-type OAuthClient interface {
-	RequestDeviceCode(ctx context.Context) (DeviceAuthInfo, error)
-	WaitToken(ctx context.Context, info DeviceAuthInfo) (TokenInfo, error)
-	GetClientID(ctx context.Context) string
-}
-
-// HTTPClient http client interface for API calls
-type HTTPClient interface {
-	Do(req *http.Request) (*http.Response, error)
-}
-
-// DeviceAuthInfo holds information for the OAuth device login flow
-type DeviceAuthInfo struct {
-	DeviceCode              string `json:"device_code"`
-	UserCode                string `json:"user_code"`
-	VerificationURI         string `json:"verification_uri"`
-	VerificationURIComplete string `json:"verification_uri_complete"`
-	ExpiresIn               int    `json:"expires_in"`
-	Interval                int    `json:"interval"`
-}
-
-// HostedGrantType grant type for device flow on Hosted
-const (
-	HostedGrantType    = "urn:ietf:params:oauth:grant-type:device_code"
-	HostedRefreshGrant = "refresh_token"
-)
-
-// Hosted client
-type Hosted struct {
-	providerConfig ProviderConfig
-
-	HTTPClient HTTPClient
-}
-
-// RequestDeviceCodePayload used for request device code payload for auth0
-type RequestDeviceCodePayload struct {
-	Audience string `json:"audience"`
-	ClientID string `json:"client_id"`
-	Scope    string `json:"scope"`
-}
-
-// TokenRequestPayload used for requesting the auth0 token
-type TokenRequestPayload struct {
-	GrantType    string `json:"grant_type"`
-	DeviceCode   string `json:"device_code,omitempty"`
-	ClientID     string `json:"client_id"`
-	RefreshToken string `json:"refresh_token,omitempty"`
-}
-
-// TokenRequestResponse used for parsing Hosted token's response
-type TokenRequestResponse struct {
-	Error            string `json:"error"`
-	ErrorDescription string `json:"error_description"`
-	TokenInfo
-}
-
-// Claims used when validating the access token
-type Claims struct {
-	Audience interface{} `json:"aud"`
-}
-
-// TokenInfo holds information of issued access token
-type TokenInfo struct {
-	AccessToken  string `json:"access_token"`
-	RefreshToken string `json:"refresh_token"`
-	IDToken      string `json:"id_token"`
-	TokenType    string `json:"token_type"`
-	ExpiresIn    int    `json:"expires_in"`
-	UseIDToken   bool   `json:"-"`
-}
-
-// GetTokenToUse returns either the access or id token based on UseIDToken field
-func (t TokenInfo) GetTokenToUse() string {
-	if t.UseIDToken {
-		return t.IDToken
-	}
-	return t.AccessToken
-}
-
-// NewHostedDeviceFlow returns an Hosted OAuth client
-func NewHostedDeviceFlow(config ProviderConfig) *Hosted {
-	httpTransport := http.DefaultTransport.(*http.Transport).Clone()
-	httpTransport.MaxIdleConns = 5
-
-	httpClient := &http.Client{
-		Timeout:   10 * time.Second,
-		Transport: httpTransport,
-	}
-
-	return &Hosted{
-		providerConfig: config,
-		HTTPClient:     httpClient,
-	}
-}
-
-// GetClientID returns the provider client id
-func (h *Hosted) GetClientID(ctx context.Context) string {
-	return h.providerConfig.ClientID
-}
-
-// RequestDeviceCode requests a device code login flow information from Hosted
-func (h *Hosted) RequestDeviceCode(ctx context.Context) (DeviceAuthInfo, error) {
-	form := url.Values{}
-	form.Add("client_id", h.providerConfig.ClientID)
-	form.Add("audience", h.providerConfig.Audience)
-	form.Add("scope", h.providerConfig.Scope)
-	req, err := http.NewRequest("POST", h.providerConfig.DeviceAuthEndpoint,
-		strings.NewReader(form.Encode()))
-	if err != nil {
-		return DeviceAuthInfo{}, fmt.Errorf("creating request failed with error: %v", err)
-	}
-	req.Header.Add("Content-Type", "application/x-www-form-urlencoded")
-
-	res, err := h.HTTPClient.Do(req)
-	if err != nil {
-		return DeviceAuthInfo{}, fmt.Errorf("doing request failed with error: %v", err)
-	}
-
-	defer res.Body.Close()
-	body, err := io.ReadAll(res.Body)
-	if err != nil {
-		return DeviceAuthInfo{}, fmt.Errorf("reading body failed with error: %v", err)
-	}
-
-	if res.StatusCode != 200 {
-		return DeviceAuthInfo{}, fmt.Errorf("request device code returned status %d error: %s", res.StatusCode, string(body))
-	}
-
-	deviceCode := DeviceAuthInfo{}
-	err = json.Unmarshal(body, &deviceCode)
-	if err != nil {
-		return DeviceAuthInfo{}, fmt.Errorf("unmarshaling response failed with error: %v", err)
-	}
-
-	return deviceCode, err
-}
-
-func (h *Hosted) requestToken(info DeviceAuthInfo) (TokenRequestResponse, error) {
-	form := url.Values{}
-	form.Add("client_id", h.providerConfig.ClientID)
-	form.Add("grant_type", HostedGrantType)
-	form.Add("device_code", info.DeviceCode)
-	req, err := http.NewRequest("POST", h.providerConfig.TokenEndpoint, strings.NewReader(form.Encode()))
-	if err != nil {
-		return TokenRequestResponse{}, fmt.Errorf("failed to create request access token: %v", err)
-	}
-
-	req.Header.Add("Content-Type", "application/x-www-form-urlencoded")
-
-	res, err := h.HTTPClient.Do(req)
-	if err != nil {
-		return TokenRequestResponse{}, fmt.Errorf("failed to request access token with error: %v", err)
-	}
-
-	defer func() {
-		err := res.Body.Close()
-		if err != nil {
-			return
-		}
-	}()
-
-	body, err := io.ReadAll(res.Body)
-	if err != nil {
-		return TokenRequestResponse{}, fmt.Errorf("failed reading access token response body with error: %v", err)
-	}
-
-	if res.StatusCode > 499 {
-		return TokenRequestResponse{}, fmt.Errorf("access token response returned code: %s", string(body))
-	}
-
-	tokenResponse := TokenRequestResponse{}
-	err = json.Unmarshal(body, &tokenResponse)
-	if err != nil {
-		return TokenRequestResponse{}, fmt.Errorf("parsing token response failed with error: %v", err)
-	}
-
-	return tokenResponse, nil
-}
-
-// WaitToken waits user's login and authorize the app. Once the user's authorize
-// it retrieves the access token from Hosted's endpoint and validates it before returning
-func (h *Hosted) WaitToken(ctx context.Context, info DeviceAuthInfo) (TokenInfo, error) {
-	interval := time.Duration(info.Interval) * time.Second
-	ticker := time.NewTicker(interval)
-	for {
-		select {
-		case <-ctx.Done():
-			return TokenInfo{}, ctx.Err()
-		case <-ticker.C:
-
-			tokenResponse, err := h.requestToken(info)
-			if err != nil {
-				return TokenInfo{}, fmt.Errorf("parsing token response failed with error: %v", err)
-			}
-
-			if tokenResponse.Error != "" {
-				if tokenResponse.Error == "authorization_pending" {
-					continue
-				} else if tokenResponse.Error == "slow_down" {
-					interval = interval + (3 * time.Second)
-					ticker.Reset(interval)
-					continue
-				}
-
-				return TokenInfo{}, fmt.Errorf(tokenResponse.ErrorDescription)
-			}
-
-			tokenInfo := TokenInfo{
-				AccessToken:  tokenResponse.AccessToken,
-				TokenType:    tokenResponse.TokenType,
-				RefreshToken: tokenResponse.RefreshToken,
-				IDToken:      tokenResponse.IDToken,
-				ExpiresIn:    tokenResponse.ExpiresIn,
-				UseIDToken:   h.providerConfig.UseIDToken,
-			}
-
-			err = isValidAccessToken(tokenInfo.GetTokenToUse(), h.providerConfig.Audience)
-			if err != nil {
-				return TokenInfo{}, fmt.Errorf("validate access token failed with error: %v", err)
-			}
-
-			return tokenInfo, err
-		}
-	}
-}
-
-// isValidAccessToken is a simple validation of the access token
-func isValidAccessToken(token string, audience string) error {
-	if token == "" {
-		return fmt.Errorf("token received is empty")
-	}
-
-	encodedClaims := strings.Split(token, ".")[1]
-	claimsString, err := base64.RawURLEncoding.DecodeString(encodedClaims)
-	if err != nil {
-		return err
-	}
-
-	claims := Claims{}
-	err = json.Unmarshal(claimsString, &claims)
-	if err != nil {
-		return err
-	}
-
-	if claims.Audience == nil {
-		return fmt.Errorf("required token field audience is absent")
-	}
-
-	// Audience claim of JWT can be a string or an array of strings
-	typ := reflect.TypeOf(claims.Audience)
-	switch typ.Kind() {
-	case reflect.String:
-		if claims.Audience == audience {
-			return nil
-		}
-	case reflect.Slice:
-		for _, aud := range claims.Audience.([]interface{}) {
-			if audience == aud {
-				return nil
-			}
-		}
-	}
-
-	return fmt.Errorf("invalid JWT token audience field")
-}

client/internal/peer/conn.go

@@ -10,15 +10,32 @@ import (
 
 	"github.com/pion/ice/v2"
 	log "github.com/sirupsen/logrus"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
 
-	"github.com/netbirdio/netbird/client/internal/proxy"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
+	"github.com/netbirdio/netbird/client/internal/wgproxy"
 	"github.com/netbirdio/netbird/iface"
+	"github.com/netbirdio/netbird/iface/bind"
 	signal "github.com/netbirdio/netbird/signal/client"
 	sProto "github.com/netbirdio/netbird/signal/proto"
 	"github.com/netbirdio/netbird/version"
 )
 
+const (
+	iceKeepAliveDefault           = 4 * time.Second
+	iceDisconnectedTimeoutDefault = 6 * time.Second
+
+	defaultWgKeepAlive = 25 * time.Second
+)
+
+type WgConfig struct {
+	WgListenPort int
+	RemoteKey    string
+	WgInterface  *iface.WGIface
+	AllowedIps   string
+	PreSharedKey *wgtypes.Key
+}
+
 // ConnConfig is a peer Connection configuration
 type ConnConfig struct {
 
@@ -37,6 +54,8 @@ type ConnConfig struct {
 
 	Timeout time.Duration
 
+	WgConfig WgConfig
+
 	UDPMux      ice.UDPMux
 	UDPMuxSrflx ice.UniversalUDPMux
 
@@ -44,7 +63,8 @@ type ConnConfig struct {
 
 	NATExternalIPs []string
 
-	AllowedIPs string
+	// UsesBind indicates whether the WireGuard interface is userspace and uses bind.ICEBind
+	UserspaceBind bool
 }
 
 // OfferAnswer represents a session establishment offer or answer
@@ -89,11 +109,12 @@ type Conn struct {
 
 	statusRecorder *Status
 
-	proxy        proxy.Proxy
+	wgProxyFactory *wgproxy.Factory
+	wgProxy        wgproxy.Proxy
+
 	remoteModeCh chan ModeMessage
 	meta         meta
 
-	wgIface       *iface.WGIface
 	adapter       iface.TunAdapter
 	iFaceDiscover stdnet.ExternalIFaceDiscover
 }
@@ -114,14 +135,19 @@ func (conn *Conn) GetConf() ConnConfig {
 	return conn.config
 }
 
-// UpdateConf updates the connection config
-func (conn *Conn) UpdateConf(conf ConnConfig) {
-	conn.config = conf
+// WgConfig returns the WireGuard config
+func (conn *Conn) WgConfig() WgConfig {
+	return conn.config.WgConfig
+}
+
+// UpdateStunTurn update the turn and stun addresses
+func (conn *Conn) UpdateStunTurn(turnStun []*ice.URL) {
+	conn.config.StunTurn = turnStun
 }
 
 // NewConn creates a new not opened Conn to the remote peer.
 // To establish a connection run Conn.Open
-func NewConn(config ConnConfig, wgIface *iface.WGIface, statusRecorder *Status, adapter iface.TunAdapter, iFaceDiscover stdnet.ExternalIFaceDiscover) (*Conn, error) {
+func NewConn(config ConnConfig, statusRecorder *Status, wgProxyFactory *wgproxy.Factory, adapter iface.TunAdapter, iFaceDiscover stdnet.ExternalIFaceDiscover) (*Conn, error) {
 	return &Conn{
 		config:         config,
 		mu:             sync.Mutex{},
@@ -131,9 +157,9 @@ func NewConn(config ConnConfig, wgIface *iface.WGIface, statusRecorder *Status,
 		remoteAnswerCh: make(chan OfferAnswer),
 		statusRecorder: statusRecorder,
 		remoteModeCh:   make(chan ModeMessage, 1),
+		wgProxyFactory: wgProxyFactory,
 		adapter:        adapter,
 		iFaceDiscover:  iFaceDiscover,
-		wgIface:        wgIface,
 	}, nil
 }
 
@@ -146,19 +172,25 @@ func (conn *Conn) reCreateAgent() error {
 	var err error
 	transportNet, err := conn.newStdNet()
 	if err != nil {
-		log.Warnf("failed to create pion's stdnet: %s", err)
+		log.Errorf("failed to create pion's stdnet: %s", err)
 	}
+
+	iceKeepAlive := iceKeepAlive()
+	iceDisconnectedTimeout := iceDisconnectedTimeout()
+
 	agentConfig := &ice.AgentConfig{
-		MulticastDNSMode: ice.MulticastDNSModeDisabled,
-		NetworkTypes:     []ice.NetworkType{ice.NetworkTypeUDP4, ice.NetworkTypeUDP6},
-		Urls:             conn.config.StunTurn,
-		CandidateTypes:   []ice.CandidateType{ice.CandidateTypeHost, ice.CandidateTypeServerReflexive, ice.CandidateTypeRelay},
-		FailedTimeout:    &failedTimeout,
-		InterfaceFilter:  stdnet.InterfaceFilter(conn.config.InterfaceBlackList),
-		UDPMux:           conn.config.UDPMux,
-		UDPMuxSrflx:      conn.config.UDPMuxSrflx,
-		NAT1To1IPs:       conn.config.NATExternalIPs,
-		Net:              transportNet,
+		MulticastDNSMode:    ice.MulticastDNSModeDisabled,
+		NetworkTypes:        []ice.NetworkType{ice.NetworkTypeUDP4, ice.NetworkTypeUDP6},
+		Urls:                conn.config.StunTurn,
+		CandidateTypes:      conn.candidateTypes(),
+		FailedTimeout:       &failedTimeout,
+		InterfaceFilter:     stdnet.InterfaceFilter(conn.config.InterfaceBlackList),
+		UDPMux:              conn.config.UDPMux,
+		UDPMuxSrflx:         conn.config.UDPMuxSrflx,
+		NAT1To1IPs:          conn.config.NATExternalIPs,
+		Net:                 transportNet,
+		DisconnectedTimeout: &iceDisconnectedTimeout,
+		KeepaliveInterval:   &iceKeepAlive,
 	}
 
 	if conn.config.DisableIPv6Discovery {
@@ -189,18 +221,25 @@ func (conn *Conn) reCreateAgent() error {
 	return nil
 }
 
+func (conn *Conn) candidateTypes() []ice.CandidateType {
+	if hasICEForceRelayConn() {
+		return []ice.CandidateType{ice.CandidateTypeRelay}
+	}
+	return []ice.CandidateType{ice.CandidateTypeHost, ice.CandidateTypeServerReflexive, ice.CandidateTypeRelay}
+}
+
 // Open opens connection to the remote peer starting ICE candidate gathering process.
 // Blocks until connection has been closed or connection timeout.
 // ConnStatus will be set accordingly
 func (conn *Conn) Open() error {
 	log.Debugf("trying to connect to peer %s", conn.config.Key)
 
-	peerState := State{PubKey: conn.config.Key}
-
-	peerState.IP = strings.Split(conn.config.AllowedIPs, "/")[0]
-	peerState.ConnStatusUpdate = time.Now()
-	peerState.ConnStatus = conn.status
-
+	peerState := State{
+		PubKey:           conn.config.Key,
+		IP:               strings.Split(conn.config.WgConfig.AllowedIps, "/")[0],
+		ConnStatusUpdate: time.Now(),
+		ConnStatus:       conn.status,
+	}
 	err := conn.statusRecorder.UpdatePeerState(peerState)
 	if err != nil {
 		log.Warnf("erro while updating the state of peer %s,err: %v", conn.config.Key, err)
@@ -255,10 +294,11 @@ func (conn *Conn) Open() error {
 	defer conn.notifyDisconnected()
 	conn.mu.Unlock()
 
-	peerState = State{PubKey: conn.config.Key}
-
-	peerState.ConnStatus = conn.status
-	peerState.ConnStatusUpdate = time.Now()
+	peerState = State{
+		PubKey:           conn.config.Key,
+		ConnStatus:       conn.status,
+		ConnStatusUpdate: time.Now(),
+	}
 	err = conn.statusRecorder.UpdatePeerState(peerState)
 	if err != nil {
 		log.Warnf("erro while updating the state of peer %s,err: %v", conn.config.Key, err)
@@ -289,19 +329,12 @@ func (conn *Conn) Open() error {
 		remoteWgPort = remoteOfferAnswer.WgListenPort
 	}
 	// the ice connection has been established successfully so we are ready to start the proxy
-	err = conn.startProxy(remoteConn, remoteWgPort)
+	remoteAddr, err := conn.configureConnection(remoteConn, remoteWgPort)
 	if err != nil {
 		return err
 	}
 
-	if conn.proxy.Type() == proxy.TypeDirectNoProxy {
-		host, _, _ := net.SplitHostPort(remoteConn.LocalAddr().String())
-		rhost, _, _ := net.SplitHostPort(remoteConn.RemoteAddr().String())
-		// direct Wireguard connection
-		log.Infof("directly connected to peer %s [laddr <-> raddr] [%s:%d <-> %s:%d]", conn.config.Key, host, conn.config.LocalWgPort, rhost, remoteWgPort)
-	} else {
-		log.Infof("connected to peer %s [laddr <-> raddr] [%s <-> %s]", conn.config.Key, remoteConn.LocalAddr().String(), remoteConn.RemoteAddr().String())
-	}
+	log.Infof("connected to peer %s, endpoint address: %s", conn.config.Key, remoteAddr.String())
 
 	// wait until connection disconnected or has been closed externally (upper layer, e.g. engine)
 	select {
@@ -314,177 +347,83 @@ func (conn *Conn) Open() error {
 	}
 }
 
-// useProxy determines whether a direct connection (without a go proxy) is possible
-//
-// There are 3 cases:
-//
-// * When neither candidate is from hard nat and one of the peers has a public IP
-//
-// * both peers are in the same private network
-//
-// * Local peer uses userspace interface with bind.ICEBind and is not relayed
-//
-// Please note, that this check happens when peers were already able to ping each other using ICE layer.
-func shouldUseProxy(pair *ice.CandidatePair, userspaceBind bool) bool {
-
-	if !isRelayCandidate(pair.Local) && userspaceBind {
-		return false
-	}
-
-	if !isHardNATCandidate(pair.Local) && isHostCandidateWithPublicIP(pair.Remote) {
-		return false
-	}
-
-	if !isHardNATCandidate(pair.Remote) && isHostCandidateWithPublicIP(pair.Local) {
-		return false
-	}
-
-	if isHostCandidateWithPrivateIP(pair.Local) && isHostCandidateWithPrivateIP(pair.Remote) && isSameNetworkPrefix(pair) {
-		return false
-	}
-
-	return true
-}
-
-func isSameNetworkPrefix(pair *ice.CandidatePair) bool {
-
-	localIPStr, _, err := net.SplitHostPort(pair.Local.Address())
-	if err != nil {
-		return false
-	}
-	remoteIPStr, _, err := net.SplitHostPort(pair.Remote.Address())
-	if err != nil {
-		return false
-	}
-	localIP := net.ParseIP(localIPStr)
-	remoteIP := net.ParseIP(remoteIPStr)
-	if localIP == nil || remoteIP == nil {
-		return false
-	}
-	// only consider /16 networks
-	mask := net.IPMask{255, 255, 0, 0}
-	return localIP.Mask(mask).Equal(remoteIP.Mask(mask))
-}
-
 func isRelayCandidate(candidate ice.Candidate) bool {
 	return candidate.Type() == ice.CandidateTypeRelay
 }
 
-func isHardNATCandidate(candidate ice.Candidate) bool {
-	return candidate.Type() == ice.CandidateTypeRelay || candidate.Type() == ice.CandidateTypePeerReflexive
-}
-
-func isHostCandidateWithPublicIP(candidate ice.Candidate) bool {
-	return candidate.Type() == ice.CandidateTypeHost && isPublicIP(candidate.Address())
-}
-
-func isHostCandidateWithPrivateIP(candidate ice.Candidate) bool {
-	return candidate.Type() == ice.CandidateTypeHost && !isPublicIP(candidate.Address())
-}
-
-func isPublicIP(address string) bool {
-	ip := net.ParseIP(address)
-	if ip.IsLoopback() || ip.IsLinkLocalUnicast() || ip.IsLinkLocalMulticast() || ip.IsPrivate() {
-		return false
-	}
-	return true
-}
-
-// startProxy starts proxying traffic from/to local Wireguard and sets connection status to StatusConnected
-func (conn *Conn) startProxy(remoteConn net.Conn, remoteWgPort int) error {
+// configureConnection starts proxying traffic from/to local Wireguard and sets connection status to StatusConnected
+func (conn *Conn) configureConnection(remoteConn net.Conn, remoteWgPort int) (net.Addr, error) {
 	conn.mu.Lock()
 	defer conn.mu.Unlock()
 
-	var pair *ice.CandidatePair
 	pair, err := conn.agent.GetSelectedCandidatePair()
 	if err != nil {
-		return err
+		return nil, err
 	}
 
-	peerState := State{PubKey: conn.config.Key}
-	p := conn.getProxyWithMessageExchange(pair, remoteWgPort)
-	conn.proxy = p
-	err = p.Start(remoteConn)
+	var endpoint net.Addr
+	if isRelayCandidate(pair.Local) {
+		log.Debugf("setup relay connection")
+		conn.wgProxy = conn.wgProxyFactory.GetProxy()
+		endpoint, err = conn.wgProxy.AddTurnConn(remoteConn)
+		if err != nil {
+			return nil, err
+		}
+	} else {
+		// To support old version's with direct mode we attempt to punch an additional role with the remote wireguard port
+		go conn.punchRemoteWGPort(pair, remoteWgPort)
+		endpoint = remoteConn.RemoteAddr()
+	}
+
+	endpointUdpAddr, _ := net.ResolveUDPAddr(endpoint.Network(), endpoint.String())
+
+	err = conn.config.WgConfig.WgInterface.UpdatePeer(conn.config.WgConfig.RemoteKey, conn.config.WgConfig.AllowedIps, defaultWgKeepAlive, endpointUdpAddr, conn.config.WgConfig.PreSharedKey)
 	if err != nil {
-		return err
+		if conn.wgProxy != nil {
+			_ = conn.wgProxy.CloseConn()
+		}
+		return nil, err
 	}
 
 	conn.status = StatusConnected
 
-	peerState.ConnStatus = conn.status
-	peerState.ConnStatusUpdate = time.Now()
-	peerState.LocalIceCandidateType = pair.Local.Type().String()
-	peerState.RemoteIceCandidateType = pair.Remote.Type().String()
+	peerState := State{
+		PubKey:                 conn.config.Key,
+		ConnStatus:             conn.status,
+		ConnStatusUpdate:       time.Now(),
+		LocalIceCandidateType:  pair.Local.Type().String(),
+		RemoteIceCandidateType: pair.Remote.Type().String(),
+		Direct:                 !isRelayCandidate(pair.Local),
+	}
 	if pair.Local.Type() == ice.CandidateTypeRelay || pair.Remote.Type() == ice.CandidateTypeRelay {
 		peerState.Relayed = true
 	}
-	peerState.Direct = p.Type() == proxy.TypeDirectNoProxy
 
 	err = conn.statusRecorder.UpdatePeerState(peerState)
 	if err != nil {
 		log.Warnf("unable to save peer's state, got error: %v", err)
 	}
 
-	return nil
+	return endpoint, nil
 }
 
-func (conn *Conn) getProxyWithMessageExchange(pair *ice.CandidatePair, remoteWgPort int) proxy.Proxy {
-	useProxy := shouldUseProxy(pair, conn.wgIface.IsUserspaceBind())
-	localDirectMode := !useProxy
-	remoteDirectMode := localDirectMode
-
-	if conn.meta.protoSupport.DirectCheck {
-		go conn.sendLocalDirectMode(localDirectMode)
-		// will block until message received or timeout
-		remoteDirectMode = conn.receiveRemoteDirectMode()
-	}
-
-	if conn.wgIface.IsUserspaceBind() && localDirectMode {
-		return proxy.NewNoProxy(conn.config.ProxyConfig)
-	}
-
-	if localDirectMode && remoteDirectMode {
-		//wgInterface *iface.WGIface, remoteKey string, allowedIps string, preSharedKey *wgtypes.Key, remoteWgPort int)
-		return proxy.NewDirectNoProxy(conn.wgIface, conn.config.Key, conn.config.AllowedIPs, remoteWgPort)
-	}
-
-	log.Debugf("falling back to local proxy mode with peer %s", conn.config.Key)
-	return proxy.NewWireGuardProxy(conn.config.ProxyConfig)
-}
-
-func (conn *Conn) sendLocalDirectMode(localMode bool) {
-	// todo what happens when we couldn't deliver this message?
-	// we could retry, etc but there is no guarantee
-
-	err := conn.sendSignalMessage(&sProto.Message{
-		Key:       conn.config.LocalKey,
-		RemoteKey: conn.config.Key,
-		Body: &sProto.Body{
-			Type: sProto.Body_MODE,
-			Mode: &sProto.Mode{
-				Direct: &localMode,
-			},
-			NetBirdVersion: version.NetbirdVersion(),
-		},
-	})
+func (conn *Conn) punchRemoteWGPort(pair *ice.CandidatePair, remoteWgPort int) {
+	// wait local endpoint configuration
+	time.Sleep(time.Second)
+	addr, err := net.ResolveUDPAddr("udp", fmt.Sprintf("%s:%d", pair.Remote.Address(), remoteWgPort))
 	if err != nil {
-		log.Errorf("failed to send local proxy mode to remote peer %s, error: %s", conn.config.Key, err)
+		log.Warnf("got an error while resolving the udp address, err: %s", err)
+		return
 	}
-}
 
-func (conn *Conn) receiveRemoteDirectMode() bool {
-	timeout := time.Second
-	timer := time.NewTimer(timeout)
-	defer timer.Stop()
-
-	select {
-	case receivedMSG := <-conn.remoteModeCh:
-		return receivedMSG.Direct
-	case <-timer.C:
-		// we didn't receive a message from remote so we assume that it supports the direct mode to keep the old behaviour
-		log.Debugf("timeout after %s while waiting for remote direct mode message from remote peer %s",
-			timeout, conn.config.Key)
-		return true
+	mux, ok := conn.config.UDPMuxSrflx.(*bind.UniversalUDPMuxDefault)
+	if !ok {
+		log.Warn("invalid udp mux conversion")
+		return
+	}
+	_, err = mux.GetSharedConn().WriteTo([]byte{0x6e, 0x62}, addr)
+	if err != nil {
+		log.Warnf("got an error while sending the punch packet, err: %s", err)
 	}
 }
 
@@ -494,22 +433,22 @@ func (conn *Conn) cleanup() error {
 	conn.mu.Lock()
 	defer conn.mu.Unlock()
 
+	var err1, err2, err3 error
 	if conn.agent != nil {
-		err := conn.agent.Close()
-		if err != nil {
-			return err
+		err1 = conn.agent.Close()
+		if err1 == nil {
+			conn.agent = nil
 		}
-		conn.agent = nil
 	}
 
-	if conn.proxy != nil {
-		err := conn.proxy.Close()
-		if err != nil {
-			return err
-		}
-		conn.proxy = nil
+	if conn.wgProxy != nil {
+		err2 = conn.wgProxy.CloseConn()
+		conn.wgProxy = nil
 	}
 
+	// todo: is it problem if we try to remove a peer what is never existed?
+	err3 = conn.config.WgConfig.WgInterface.RemovePeer(conn.config.WgConfig.RemoteKey)
+
 	if conn.notifyDisconnected != nil {
 		conn.notifyDisconnected()
 		conn.notifyDisconnected = nil
@@ -517,10 +456,11 @@ func (conn *Conn) cleanup() error {
 
 	conn.status = StatusDisconnected
 
-	peerState := State{PubKey: conn.config.Key}
-	peerState.ConnStatus = conn.status
-	peerState.ConnStatusUpdate = time.Now()
-
+	peerState := State{
+		PubKey:           conn.config.Key,
+		ConnStatus:       conn.status,
+		ConnStatusUpdate: time.Now(),
+	}
 	err := conn.statusRecorder.UpdatePeerState(peerState)
 	if err != nil {
 		// pretty common error because by that time Engine can already remove the peer and status won't be available.
@@ -529,8 +469,13 @@ func (conn *Conn) cleanup() error {
 	}
 
 	log.Debugf("cleaned up connection to peer %s", conn.config.Key)
-
-	return nil
+	if err1 != nil {
+		return err1
+	}
+	if err2 != nil {
+		return err2
+	}
+	return err3
 }
 
 // SetSignalOffer sets a handler function to be triggered by Conn when a new connection offer has to be signalled to the remote peer
@@ -708,16 +653,6 @@ func (conn *Conn) GetKey() string {
 	return conn.config.Key
 }
 
-// OnModeMessage unmarshall the payload message and send it to the mode message channel
-func (conn *Conn) OnModeMessage(message ModeMessage) error {
-	select {
-	case conn.remoteModeCh <- message:
-		return nil
-	default:
-		return fmt.Errorf("unable to process mode message: channel busy")
-	}
-}
-
 // RegisterProtoSupportMeta register supported proto message in the connection metadata
 func (conn *Conn) RegisterProtoSupportMeta(support []uint32) {
 	protoSupport := signal.ParseFeaturesSupported(support)

client/internal/peer/conn_test.go

@@ -1,18 +1,16 @@
 package peer
 
 import (
-	"github.com/netbirdio/netbird/client/internal/stdnet"
 	"sync"
 	"testing"
 	"time"
 
 	"github.com/magiconair/properties/assert"
 	"github.com/pion/ice/v2"
-	"golang.org/x/sync/errgroup"
 
-	"github.com/netbirdio/netbird/client/internal/proxy"
+	"github.com/netbirdio/netbird/client/internal/stdnet"
+	"github.com/netbirdio/netbird/client/internal/wgproxy"
 	"github.com/netbirdio/netbird/iface"
-	sproto "github.com/netbirdio/netbird/signal/proto"
 )
 
 var connConf = ConnConfig{
@@ -21,7 +19,6 @@ var connConf = ConnConfig{
 	StunTurn:           []*ice.URL{},
 	InterfaceBlackList: nil,
 	Timeout:            time.Second,
-	ProxyConfig:        proxy.Config{},
 	LocalWgPort:        51820,
 }
 
@@ -38,7 +35,11 @@ func TestNewConn_interfaceFilter(t *testing.T) {
 }
 
 func TestConn_GetKey(t *testing.T) {
-	conn, err := NewConn(connConf, nil, nil, nil)
+	wgProxyFactory := wgproxy.NewFactory(connConf.LocalWgPort)
+	defer func() {
+		_ = wgProxyFactory.Free()
+	}()
+	conn, err := NewConn(connConf, nil, wgProxyFactory, nil, nil)
 	if err != nil {
 		return
 	}
@@ -49,8 +50,11 @@ func TestConn_GetKey(t *testing.T) {
 }
 
 func TestConn_OnRemoteOffer(t *testing.T) {
-
-	conn, err := NewConn(connConf, NewRecorder("https://mgm"), nil, nil)
+	wgProxyFactory := wgproxy.NewFactory(connConf.LocalWgPort)
+	defer func() {
+		_ = wgProxyFactory.Free()
+	}()
+	conn, err := NewConn(connConf, NewRecorder("https://mgm"), wgProxyFactory, nil, nil)
 	if err != nil {
 		return
 	}
@@ -83,8 +87,11 @@ func TestConn_OnRemoteOffer(t *testing.T) {
 }
 
 func TestConn_OnRemoteAnswer(t *testing.T) {
-
-	conn, err := NewConn(connConf, NewRecorder("https://mgm"), nil, nil)
+	wgProxyFactory := wgproxy.NewFactory(connConf.LocalWgPort)
+	defer func() {
+		_ = wgProxyFactory.Free()
+	}()
+	conn, err := NewConn(connConf, NewRecorder("https://mgm"), wgProxyFactory, nil, nil)
 	if err != nil {
 		return
 	}
@@ -116,8 +123,11 @@ func TestConn_OnRemoteAnswer(t *testing.T) {
 	wg.Wait()
 }
 func TestConn_Status(t *testing.T) {
-
-	conn, err := NewConn(connConf, NewRecorder("https://mgm"), nil, nil)
+	wgProxyFactory := wgproxy.NewFactory(connConf.LocalWgPort)
+	defer func() {
+		_ = wgProxyFactory.Free()
+	}()
+	conn, err := NewConn(connConf, NewRecorder("https://mgm"), wgProxyFactory, nil, nil)
 	if err != nil {
 		return
 	}
@@ -143,8 +153,11 @@ func TestConn_Status(t *testing.T) {
 }
 
 func TestConn_Close(t *testing.T) {
-
-	conn, err := NewConn(connConf, NewRecorder("https://mgm"), nil, nil)
+	wgProxyFactory := wgproxy.NewFactory(connConf.LocalWgPort)
+	defer func() {
+		_ = wgProxyFactory.Free()
+	}()
+	conn, err := NewConn(connConf, NewRecorder("https://mgm"), wgProxyFactory, nil, nil)
 	if err != nil {
 		return
 	}
@@ -169,274 +182,3 @@ func TestConn_Close(t *testing.T) {
 
 	wg.Wait()
 }
-
-type mockICECandidate struct {
-	ice.Candidate
-	AddressFunc func() string
-	TypeFunc    func() ice.CandidateType
-}
-
-// Address mocks and overwrite ice.Candidate Address method
-func (m *mockICECandidate) Address() string {
-	if m.AddressFunc != nil {
-		return m.AddressFunc()
-	}
-	return ""
-}
-
-// Type mocks and overwrite ice.Candidate Type method
-func (m *mockICECandidate) Type() ice.CandidateType {
-	if m.TypeFunc != nil {
-		return m.TypeFunc()
-	}
-	return ice.CandidateTypeUnspecified
-}
-
-func TestConn_ShouldUseProxy(t *testing.T) {
-	publicHostCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "8.8.8.8"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeHost
-		},
-	}
-	privateHostCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "10.0.0.1:44576"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeHost
-		},
-	}
-
-	srflxCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "1.1.1.1"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeServerReflexive
-		},
-	}
-
-	prflxCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "1.1.1.1"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypePeerReflexive
-		},
-	}
-
-	relayCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "1.1.1.1"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeRelay
-		},
-	}
-
-	testCases := []struct {
-		name        string
-		candatePair *ice.CandidatePair
-		expected    bool
-	}{
-		{
-			name: "Use Proxy When Local Candidate Is Relay",
-			candatePair: &ice.CandidatePair{
-				Local:  relayCandidate,
-				Remote: privateHostCandidate,
-			},
-			expected: true,
-		},
-		{
-			name: "Use Proxy When Remote Candidate Is Relay",
-			candatePair: &ice.CandidatePair{
-				Local:  privateHostCandidate,
-				Remote: relayCandidate,
-			},
-			expected: true,
-		},
-		{
-			name: "Use Proxy When Local Candidate Is Peer Reflexive",
-			candatePair: &ice.CandidatePair{
-				Local:  prflxCandidate,
-				Remote: privateHostCandidate,
-			},
-			expected: true,
-		},
-		{
-			name: "Use Proxy When Remote Candidate Is Peer Reflexive",
-			candatePair: &ice.CandidatePair{
-				Local:  privateHostCandidate,
-				Remote: prflxCandidate,
-			},
-			expected: true,
-		},
-		{
-			name: "Don't Use Proxy When Local Candidate Is Public And Remote Is Private",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: privateHostCandidate,
-			},
-			expected: false,
-		},
-		{
-			name: "Don't Use Proxy When Remote Candidate Is Public And Local Is Private",
-			candatePair: &ice.CandidatePair{
-				Local:  privateHostCandidate,
-				Remote: publicHostCandidate,
-			},
-			expected: false,
-		},
-		{
-			name: "Don't Use Proxy When Local Candidate is Public And Remote Is Server Reflexive",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: srflxCandidate,
-			},
-			expected: false,
-		},
-		{
-			name: "Don't Use Proxy When Remote Candidate is Public And Local Is Server Reflexive",
-			candatePair: &ice.CandidatePair{
-				Local:  srflxCandidate,
-				Remote: publicHostCandidate,
-			},
-			expected: false,
-		},
-		{
-			name: "Don't Use Proxy When Both Candidates Are Public",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: publicHostCandidate,
-			},
-			expected: false,
-		},
-		{
-			name: "Don't Use Proxy When Both Candidates Are Private",
-			candatePair: &ice.CandidatePair{
-				Local:  privateHostCandidate,
-				Remote: privateHostCandidate,
-			},
-			expected: false,
-		},
-	}
-
-	for _, testCase := range testCases {
-		t.Run(testCase.name, func(t *testing.T) {
-			result := shouldUseProxy(testCase.candatePair, false)
-			if result != testCase.expected {
-				t.Errorf("got a different result. Expected %t Got %t", testCase.expected, result)
-			}
-		})
-	}
-}
-
-func TestGetProxyWithMessageExchange(t *testing.T) {
-	publicHostCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "8.8.8.8"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeHost
-		},
-	}
-	relayCandidate := &mockICECandidate{
-		AddressFunc: func() string {
-			return "1.1.1.1"
-		},
-		TypeFunc: func() ice.CandidateType {
-			return ice.CandidateTypeRelay
-		},
-	}
-
-	testCases := []struct {
-		name                   string
-		candatePair            *ice.CandidatePair
-		inputDirectModeSupport bool
-		inputRemoteModeMessage bool
-		expected               proxy.Type
-	}{
-		{
-			name: "Should Result In Using Wireguard Proxy When Local Eval Is Use Proxy",
-			candatePair: &ice.CandidatePair{
-				Local:  relayCandidate,
-				Remote: publicHostCandidate,
-			},
-			inputDirectModeSupport: true,
-			inputRemoteModeMessage: true,
-			expected:               proxy.TypeWireGuard,
-		},
-		{
-			name: "Should Result In Using Wireguard Proxy When Remote Eval Is Use Proxy",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: publicHostCandidate,
-			},
-			inputDirectModeSupport: true,
-			inputRemoteModeMessage: false,
-			expected:               proxy.TypeWireGuard,
-		},
-		{
-			name: "Should Result In Using Wireguard Proxy When Remote Direct Mode Support Is False And Local Eval Is Use Proxy",
-			candatePair: &ice.CandidatePair{
-				Local:  relayCandidate,
-				Remote: publicHostCandidate,
-			},
-			inputDirectModeSupport: false,
-			inputRemoteModeMessage: false,
-			expected:               proxy.TypeWireGuard,
-		},
-		{
-			name: "Should Result In Using Direct When Remote Direct Mode Support Is False And Local Eval Is No Use Proxy",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: publicHostCandidate,
-			},
-			inputDirectModeSupport: false,
-			inputRemoteModeMessage: false,
-			expected:               proxy.TypeDirectNoProxy,
-		},
-		{
-			name: "Should Result In Using Direct When Local And Remote Eval Is No Proxy",
-			candatePair: &ice.CandidatePair{
-				Local:  publicHostCandidate,
-				Remote: publicHostCandidate,
-			},
-			inputDirectModeSupport: true,
-			inputRemoteModeMessage: true,
-			expected:               proxy.TypeDirectNoProxy,
-		},
-	}
-	for _, testCase := range testCases {
-		t.Run(testCase.name, func(t *testing.T) {
-			g := errgroup.Group{}
-			conn, err := NewConn(connConf, nil, nil, nil)
-			if err != nil {
-				t.Fatal(err)
-			}
-			conn.meta.protoSupport.DirectCheck = testCase.inputDirectModeSupport
-			conn.SetSendSignalMessage(func(message *sproto.Message) error {
-				return nil
-			})
-
-			g.Go(func() error {
-				return conn.OnModeMessage(ModeMessage{
-					Direct: testCase.inputRemoteModeMessage,
-				})
-			})
-
-			resultProxy := conn.getProxyWithMessageExchange(testCase.candatePair, 1000)
-
-			err = g.Wait()
-			if err != nil {
-				t.Error(err)
-			}
-			if resultProxy.Type() != testCase.expected {
-				t.Errorf("result didn't match expected value: Expected: %s, Got: %s", testCase.expected, resultProxy.Type())
-			}
-		})
-	}
-}

client/internal/peer/env_config.go

@@ -0,0 +1,53 @@
+package peer
+
+import (
+	"os"
+	"strconv"
+	"strings"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+)
+
+const (
+	envICEKeepAliveIntervalSec   = "NB_ICE_KEEP_ALIVE_INTERVAL_SEC"
+	envICEDisconnectedTimeoutSec = "NB_ICE_DISCONNECTED_TIMEOUT_SEC"
+	envICEForceRelayConn         = "NB_ICE_FORCE_RELAY_CONN"
+)
+
+func iceKeepAlive() time.Duration {
+	keepAliveEnv := os.Getenv(envICEKeepAliveIntervalSec)
+	if keepAliveEnv == "" {
+		return iceKeepAliveDefault
+	}
+
+	log.Debugf("setting ICE keep alive interval to %s seconds", keepAliveEnv)
+	keepAliveEnvSec, err := strconv.Atoi(keepAliveEnv)
+	if err != nil {
+		log.Warnf("invalid value %s set for %s, using default %v", keepAliveEnv, envICEKeepAliveIntervalSec, iceKeepAliveDefault)
+		return iceKeepAliveDefault
+	}
+
+	return time.Duration(keepAliveEnvSec) * time.Second
+}
+
+func iceDisconnectedTimeout() time.Duration {
+	disconnectedTimeoutEnv := os.Getenv(envICEDisconnectedTimeoutSec)
+	if disconnectedTimeoutEnv == "" {
+		return iceDisconnectedTimeoutDefault
+	}
+
+	log.Debugf("setting ICE disconnected timeout to %s seconds", disconnectedTimeoutEnv)
+	disconnectedTimeoutSec, err := strconv.Atoi(disconnectedTimeoutEnv)
+	if err != nil {
+		log.Warnf("invalid value %s set for %s, using default %v", disconnectedTimeoutEnv, envICEDisconnectedTimeoutSec, iceDisconnectedTimeoutDefault)
+		return iceDisconnectedTimeoutDefault
+	}
+
+	return time.Duration(disconnectedTimeoutSec) * time.Second
+}
+
+func hasICEForceRelayConn() bool {
+	disconnectedTimeoutEnv := os.Getenv(envICEForceRelayConn)
+	return strings.ToLower(disconnectedTimeoutEnv) == "true"
+}

client/internal/peer/status.go

@@ -59,6 +59,11 @@ type Status struct {
 	mgmAddress      string
 	signalAddress   string
 	notifier        *notifier
+
+	// To reduce the number of notification invocation this bool will be true when need to call the notification
+	// Some Peer actions mostly used by in a batch when the network map has been synchronized. In these type of events
+	// set to true this variable and at the end of the processing we will reset it by the FinishPeerListModifications()
+	peerListChangedForNotification bool
 }
 
 // NewRecorder returns a new Status instance
@@ -78,10 +83,13 @@ func (d *Status) ReplaceOfflinePeers(replacement []State) {
 	defer d.mux.Unlock()
 	d.offlinePeers = make([]State, len(replacement))
 	copy(d.offlinePeers, replacement)
+
+	// todo we should set to true in case if the list changed only
+	d.peerListChangedForNotification = true
 }
 
 // AddPeer adds peer to Daemon status map
-func (d *Status) AddPeer(peerPubKey string) error {
+func (d *Status) AddPeer(peerPubKey string, fqdn string) error {
 	d.mux.Lock()
 	defer d.mux.Unlock()
 
@@ -89,7 +97,12 @@ func (d *Status) AddPeer(peerPubKey string) error {
 	if ok {
 		return errors.New("peer already exist")
 	}
-	d.peers[peerPubKey] = State{PubKey: peerPubKey, ConnStatus: StatusDisconnected}
+	d.peers[peerPubKey] = State{
+		PubKey:     peerPubKey,
+		ConnStatus: StatusDisconnected,
+		FQDN:       fqdn,
+	}
+	d.peerListChangedForNotification = true
 	return nil
 }
 
@@ -111,13 +124,13 @@ func (d *Status) RemovePeer(peerPubKey string) error {
 	defer d.mux.Unlock()
 
 	_, ok := d.peers[peerPubKey]
-	if ok {
-		delete(d.peers, peerPubKey)
-		return nil
+	if !ok {
+		return errors.New("no peer with to remove")
 	}
 
-	d.notifyPeerListChanged()
-	return errors.New("no peer with to remove")
+	delete(d.peers, peerPubKey)
+	d.peerListChangedForNotification = true
+	return nil
 }
 
 // UpdatePeerState updates peer status
@@ -134,6 +147,8 @@ func (d *Status) UpdatePeerState(receivedState State) error {
 		peerState.IP = receivedState.IP
 	}
 
+	skipNotification := shouldSkipNotify(receivedState, peerState)
+
 	if receivedState.ConnStatus != peerState.ConnStatus {
 		peerState.ConnStatus = receivedState.ConnStatus
 		peerState.ConnStatusUpdate = receivedState.ConnStatusUpdate
@@ -145,6 +160,10 @@ func (d *Status) UpdatePeerState(receivedState State) error {
 
 	d.peers[receivedState.PubKey] = peerState
 
+	if skipNotification {
+		return nil
+	}
+
 	ch, found := d.changeNotify[receivedState.PubKey]
 	if found && ch != nil {
 		close(ch)
@@ -155,6 +174,19 @@ func (d *Status) UpdatePeerState(receivedState State) error {
 	return nil
 }
 
+func shouldSkipNotify(new, curr State) bool {
+	switch {
+	case new.ConnStatus == StatusConnecting:
+		return true
+	case new.ConnStatus == StatusDisconnected && curr.ConnStatus == StatusConnecting:
+		return true
+	case new.ConnStatus == StatusDisconnected && curr.ConnStatus == StatusDisconnected:
+		return curr.IP != ""
+	default:
+		return false
+	}
+}
+
 // UpdatePeerFQDN update peer's state fqdn only
 func (d *Status) UpdatePeerFQDN(peerPubKey, fqdn string) error {
 	d.mux.Lock()
@@ -168,10 +200,23 @@ func (d *Status) UpdatePeerFQDN(peerPubKey, fqdn string) error {
 	peerState.FQDN = fqdn
 	d.peers[peerPubKey] = peerState
 
-	d.notifyPeerListChanged()
 	return nil
 }
 
+// FinishPeerListModifications this event invoke the notification
+func (d *Status) FinishPeerListModifications() {
+	d.mux.Lock()
+
+	if !d.peerListChangedForNotification {
+		d.mux.Unlock()
+		return
+	}
+	d.peerListChangedForNotification = false
+	d.mux.Unlock()
+
+	d.notifyPeerListChanged()
+}
+
 // GetPeerStateChangeNotifier returns a change notifier channel for a peer
 func (d *Status) GetPeerStateChangeNotifier(peer string) <-chan struct{} {
 	d.mux.Lock()
@@ -308,7 +353,7 @@ func (d *Status) onConnectionChanged() {
 }
 
 func (d *Status) notifyPeerListChanged() {
-	d.notifier.peerListChanged(len(d.peers))
+	d.notifier.peerListChanged(len(d.peers) + len(d.offlinePeers))
 }
 
 func (d *Status) notifyAddressChanged() {

client/internal/peer/status_test.go

@@ -9,13 +9,13 @@ import (
 func TestAddPeer(t *testing.T) {
 	key := "abc"
 	status := NewRecorder("https://mgm")
-	err := status.AddPeer(key)
+	err := status.AddPeer(key, "abc.netbird")
 	assert.NoError(t, err, "shouldn't return error")
 
 	_, exists := status.peers[key]
 	assert.True(t, exists, "value was found")
 
-	err = status.AddPeer(key)
+	err = status.AddPeer(key, "abc.netbird")
 
 	assert.Error(t, err, "should return error on duplicate")
 }
@@ -23,7 +23,7 @@ func TestAddPeer(t *testing.T) {
 func TestGetPeer(t *testing.T) {
 	key := "abc"
 	status := NewRecorder("https://mgm")
-	err := status.AddPeer(key)
+	err := status.AddPeer(key, "abc.netbird")
 	assert.NoError(t, err, "shouldn't return error")
 
 	peerStatus, err := status.GetPeer(key)

client/internal/pkce_auth.go

@@ -0,0 +1,128 @@
+package internal
+
+import (
+	"context"
+	"fmt"
+	"net/url"
+
+	log "github.com/sirupsen/logrus"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/status"
+
+	mgm "github.com/netbirdio/netbird/management/client"
+)
+
+// PKCEAuthorizationFlow represents PKCE Authorization Flow information
+type PKCEAuthorizationFlow struct {
+	ProviderConfig PKCEAuthProviderConfig
+}
+
+// PKCEAuthProviderConfig has all attributes needed to initiate pkce authorization flow
+type PKCEAuthProviderConfig struct {
+	// ClientID An IDP application client id
+	ClientID string
+	// ClientSecret An IDP application client secret
+	ClientSecret string
+	// Audience An Audience for to authorization validation
+	Audience string
+	// TokenEndpoint is the endpoint of an IDP manager where clients can obtain access token
+	TokenEndpoint string
+	// AuthorizationEndpoint is the endpoint of an IDP manager where clients can obtain authorization code
+	AuthorizationEndpoint string
+	// Scopes provides the scopes to be included in the token request
+	Scope string
+	// RedirectURL handles authorization code from IDP manager
+	RedirectURLs []string
+	// UseIDToken indicates if the id token should be used for authentication
+	UseIDToken bool
+}
+
+// GetPKCEAuthorizationFlowInfo initialize a PKCEAuthorizationFlow instance and return with it
+func GetPKCEAuthorizationFlowInfo(ctx context.Context, privateKey string, mgmURL *url.URL) (PKCEAuthorizationFlow, error) {
+	// validate our peer's Wireguard PRIVATE key
+	myPrivateKey, err := wgtypes.ParseKey(privateKey)
+	if err != nil {
+		log.Errorf("failed parsing Wireguard key %s: [%s]", privateKey, err.Error())
+		return PKCEAuthorizationFlow{}, err
+	}
+
+	var mgmTLSEnabled bool
+	if mgmURL.Scheme == "https" {
+		mgmTLSEnabled = true
+	}
+
+	log.Debugf("connecting to Management Service %s", mgmURL.String())
+	mgmClient, err := mgm.NewClient(ctx, mgmURL.Host, myPrivateKey, mgmTLSEnabled)
+	if err != nil {
+		log.Errorf("failed connecting to Management Service %s %v", mgmURL.String(), err)
+		return PKCEAuthorizationFlow{}, err
+	}
+	log.Debugf("connected to the Management service %s", mgmURL.String())
+
+	defer func() {
+		err = mgmClient.Close()
+		if err != nil {
+			log.Warnf("failed to close the Management service client %v", err)
+		}
+	}()
+
+	serverKey, err := mgmClient.GetServerPublicKey()
+	if err != nil {
+		log.Errorf("failed while getting Management Service public key: %v", err)
+		return PKCEAuthorizationFlow{}, err
+	}
+
+	protoPKCEAuthorizationFlow, err := mgmClient.GetPKCEAuthorizationFlow(*serverKey)
+	if err != nil {
+		if s, ok := status.FromError(err); ok && s.Code() == codes.NotFound {
+			log.Warnf("server couldn't find pkce flow, contact admin: %v", err)
+			return PKCEAuthorizationFlow{}, err
+		}
+		log.Errorf("failed to retrieve pkce flow: %v", err)
+		return PKCEAuthorizationFlow{}, err
+	}
+
+	authFlow := PKCEAuthorizationFlow{
+		ProviderConfig: PKCEAuthProviderConfig{
+			Audience:              protoPKCEAuthorizationFlow.GetProviderConfig().GetAudience(),
+			ClientID:              protoPKCEAuthorizationFlow.GetProviderConfig().GetClientID(),
+			ClientSecret:          protoPKCEAuthorizationFlow.GetProviderConfig().GetClientSecret(),
+			TokenEndpoint:         protoPKCEAuthorizationFlow.GetProviderConfig().GetTokenEndpoint(),
+			AuthorizationEndpoint: protoPKCEAuthorizationFlow.GetProviderConfig().GetAuthorizationEndpoint(),
+			Scope:                 protoPKCEAuthorizationFlow.GetProviderConfig().GetScope(),
+			RedirectURLs:          protoPKCEAuthorizationFlow.GetProviderConfig().GetRedirectURLs(),
+			UseIDToken:            protoPKCEAuthorizationFlow.GetProviderConfig().GetUseIDToken(),
+		},
+	}
+
+	err = isPKCEProviderConfigValid(authFlow.ProviderConfig)
+	if err != nil {
+		return PKCEAuthorizationFlow{}, err
+	}
+
+	return authFlow, nil
+}
+
+func isPKCEProviderConfigValid(config PKCEAuthProviderConfig) error {
+	errorMSGFormat := "invalid provider configuration received from management: %s value is empty. Contact your NetBird administrator"
+	if config.Audience == "" {
+		return fmt.Errorf(errorMSGFormat, "Audience")
+	}
+	if config.ClientID == "" {
+		return fmt.Errorf(errorMSGFormat, "Client ID")
+	}
+	if config.TokenEndpoint == "" {
+		return fmt.Errorf(errorMSGFormat, "Token Endpoint")
+	}
+	if config.AuthorizationEndpoint == "" {
+		return fmt.Errorf(errorMSGFormat, "Authorization Auth Endpoint")
+	}
+	if config.Scope == "" {
+		return fmt.Errorf(errorMSGFormat, "PKCE Auth Scopes")
+	}
+	if config.RedirectURLs == nil {
+		return fmt.Errorf(errorMSGFormat, "PKCE Redirect URLs")
+	}
+	return nil
+}

client/internal/proxy/direct.go

@@ -1,67 +0,0 @@
-package proxy
-
-import (
-	"net"
-
-	log "github.com/sirupsen/logrus"
-
-	"github.com/netbirdio/netbird/iface"
-)
-
-// DirectNoProxy is used when there is no need for a proxy between ICE and WireGuard.
-// This is possible in either of these cases:
-// - peers are in the same local network
-// - one of the peers has a public static IP (host)
-// DirectNoProxy will just update remote peer with a remote host and fixed WireGuard port (r.g. 51820).
-// In order DirectNoProxy to work, WireGuard port has to be fixed for the time being.
-type DirectNoProxy struct {
-	wgInterface *iface.WGIface
-
-	remoteKey  string
-	allowedIps string
-
-	// RemoteWgListenPort is a WireGuard port of a remote peer.
-	// It is used instead of the hardcoded 51820 port.
-	remoteWgListenPort int
-}
-
-// NewDirectNoProxy creates a new DirectNoProxy with a provided config and remote peer's WireGuard listen port
-func NewDirectNoProxy(wgInterface *iface.WGIface, remoteKey string, allowedIps string, remoteWgPort int) *DirectNoProxy {
-	return &DirectNoProxy{
-		wgInterface:        wgInterface,
-		remoteKey:          remoteKey,
-		allowedIps:         allowedIps,
-		remoteWgListenPort: remoteWgPort}
-}
-
-// Close removes peer from the WireGuard interface
-func (p *DirectNoProxy) Close() error {
-	err := p.wgInterface.RemovePeer(p.remoteKey)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-// Start just updates WireGuard peer with the remote IP and default WireGuard port
-func (p *DirectNoProxy) Start(remoteConn net.Conn) error {
-
-	log.Debugf("using DirectNoProxy while connecting to peer %s", p.remoteKey)
-	addr, err := net.ResolveUDPAddr("udp", remoteConn.RemoteAddr().String())
-	if err != nil {
-		return err
-	}
-	addr.Port = p.remoteWgListenPort
-	err = p.wgInterface.UpdatePeer(p.remoteKey, p.allowedIps, addr)
-
-	if err != nil {
-		return err
-	}
-
-	return nil
-}
-
-// Type returns the type of this proxy
-func (p *DirectNoProxy) Type() Type {
-	return TypeDirectNoProxy
-}

client/internal/proxy/dummy.go

@@ -1,72 +0,0 @@
-package proxy
-
-import (
-	"context"
-	log "github.com/sirupsen/logrus"
-	"net"
-	"time"
-)
-
-// DummyProxy just sends pings to the RemoteKey peer and reads responses
-type DummyProxy struct {
-	conn   net.Conn
-	remote string
-	ctx    context.Context
-	cancel context.CancelFunc
-}
-
-func NewDummyProxy(remote string) *DummyProxy {
-	p := &DummyProxy{remote: remote}
-	p.ctx, p.cancel = context.WithCancel(context.Background())
-	return p
-}
-
-func (p *DummyProxy) Close() error {
-	p.cancel()
-	return nil
-}
-
-func (p *DummyProxy) Start(remoteConn net.Conn) error {
-	p.conn = remoteConn
-	go func() {
-		buf := make([]byte, 1500)
-		for {
-			select {
-			case <-p.ctx.Done():
-				return
-			default:
-				_, err := p.conn.Read(buf)
-				if err != nil {
-					log.Errorf("error while reading RemoteKey %s proxy %v", p.remote, err)
-					return
-				}
-				//log.Debugf("received %s from %s", string(buf[:n]), p.remote)
-			}
-
-		}
-	}()
-
-	go func() {
-		for {
-			select {
-			case <-p.ctx.Done():
-				return
-			default:
-				_, err := p.conn.Write([]byte("hello"))
-				//log.Debugf("sent ping to %s", p.remote)
-				if err != nil {
-					log.Errorf("error while writing to RemoteKey %s proxy %v", p.remote, err)
-					return
-				}
-				time.Sleep(5 * time.Second)
-			}
-		}
-
-	}()
-
-	return nil
-}
-
-func (p *DummyProxy) Type() Type {
-	return TypeDummy
-}

client/internal/proxy/noproxy.go

@@ -1,42 +0,0 @@
-package proxy
-
-import (
-	log "github.com/sirupsen/logrus"
-	"net"
-)
-
-// NoProxy is used just to configure WireGuard without any local proxy in between.
-// Used when the WireGuard interface is userspace and uses bind.ICEBind
-type NoProxy struct {
-	config Config
-}
-
-// NewNoProxy creates a new NoProxy with a provided config
-func NewNoProxy(config Config) *NoProxy {
-	return &NoProxy{config: config}
-}
-
-// Close removes peer from the WireGuard interface
-func (p *NoProxy) Close() error {
-	err := p.config.WgInterface.RemovePeer(p.config.RemoteKey)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-// Start just updates WireGuard peer with the remote address
-func (p *NoProxy) Start(remoteConn net.Conn) error {
-
-	log.Debugf("using NoProxy to connect to peer %s at %s", p.config.RemoteKey, remoteConn.RemoteAddr().String())
-	addr, err := net.ResolveUDPAddr("udp", remoteConn.RemoteAddr().String())
-	if err != nil {
-		return err
-	}
-	return p.config.WgInterface.UpdatePeer(p.config.RemoteKey, p.config.AllowedIps, DefaultWgKeepAlive,
-		addr, p.config.PreSharedKey)
-}
-
-func (p *NoProxy) Type() Type {
-	return TypeNoProxy
-}

client/internal/proxy/proxy.go

@@ -1,22 +0,0 @@
-package proxy
-
-import (
-	"io"
-	"net"
-)
-
-type Type string
-
-const (
-	TypeDirectNoProxy Type = "DirectNoProxy"
-	TypeWireGuard     Type = "WireGuard"
-	TypeDummy         Type = "Dummy"
-	TypeNoProxy       Type = "NoProxy"
-)
-
-type Proxy interface {
-	io.Closer
-	// Start creates a local remoteConn and starts proxying data from/to remoteConn
-	Start(remoteConn net.Conn) error
-	Type() Type
-}

client/internal/proxy/wireguard.go

@@ -1,128 +0,0 @@
-package proxy
-
-import (
-	"context"
-	log "github.com/sirupsen/logrus"
-	"net"
-)
-
-// WireGuardProxy proxies
-type WireGuardProxy struct {
-	ctx    context.Context
-	cancel context.CancelFunc
-
-	config Config
-
-	remoteConn net.Conn
-	localConn  net.Conn
-}
-
-func NewWireGuardProxy(config Config) *WireGuardProxy {
-	p := &WireGuardProxy{config: config}
-	p.ctx, p.cancel = context.WithCancel(context.Background())
-	return p
-}
-
-func (p *WireGuardProxy) updateEndpoint() error {
-	udpAddr, err := net.ResolveUDPAddr(p.localConn.LocalAddr().Network(), p.localConn.LocalAddr().String())
-	if err != nil {
-		return err
-	}
-	// add local proxy connection as a Wireguard peer
-	err = p.config.WgInterface.UpdatePeer(p.config.RemoteKey, p.config.AllowedIps, DefaultWgKeepAlive,
-		udpAddr, p.config.PreSharedKey)
-	if err != nil {
-		return err
-	}
-
-	return nil
-}
-
-func (p *WireGuardProxy) Start(remoteConn net.Conn) error {
-	p.remoteConn = remoteConn
-
-	var err error
-	p.localConn, err = net.Dial("udp", p.config.WgListenAddr)
-	if err != nil {
-		log.Errorf("failed dialing to local Wireguard port %s", err)
-		return err
-	}
-
-	err = p.updateEndpoint()
-	if err != nil {
-		log.Errorf("error while updating Wireguard peer endpoint [%s] %v", p.config.RemoteKey, err)
-		return err
-	}
-
-	go p.proxyToRemote()
-	go p.proxyToLocal()
-
-	return nil
-}
-
-func (p *WireGuardProxy) Close() error {
-	p.cancel()
-	if c := p.localConn; c != nil {
-		err := p.localConn.Close()
-		if err != nil {
-			return err
-		}
-	}
-	err := p.config.WgInterface.RemovePeer(p.config.RemoteKey)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-// proxyToRemote proxies everything from Wireguard to the RemoteKey peer
-// blocks
-func (p *WireGuardProxy) proxyToRemote() {
-
-	buf := make([]byte, 1500)
-	for {
-		select {
-		case <-p.ctx.Done():
-			log.Debugf("stopped proxying to remote peer %s due to closed connection", p.config.RemoteKey)
-			return
-		default:
-			n, err := p.localConn.Read(buf)
-			if err != nil {
-				continue
-			}
-
-			_, err = p.remoteConn.Write(buf[:n])
-			if err != nil {
-				continue
-			}
-		}
-	}
-}
-
-// proxyToLocal proxies everything from the RemoteKey peer to local Wireguard
-// blocks
-func (p *WireGuardProxy) proxyToLocal() {
-
-	buf := make([]byte, 1500)
-	for {
-		select {
-		case <-p.ctx.Done():
-			log.Debugf("stopped proxying from remote peer %s due to closed connection", p.config.RemoteKey)
-			return
-		default:
-			n, err := p.remoteConn.Read(buf)
-			if err != nil {
-				continue
-			}
-
-			_, err = p.localConn.Write(buf[:n])
-			if err != nil {
-				continue
-			}
-		}
-	}
-}
-
-func (p *WireGuardProxy) Type() Type {
-	return TypeWireGuard
-}

client/internal/routemanager/client.go

@@ -71,7 +71,7 @@ func (c *clientNetwork) getRouterPeerStatuses() map[string]routerPeerStatus {
 }
 
 func (c *clientNetwork) getBestRouteFromStatuses(routePeerStatuses map[string]routerPeerStatus) string {
-	var chosen string
+	chosen := ""
 	chosenScore := 0
 
 	currID := ""
@@ -85,17 +85,26 @@ func (c *clientNetwork) getBestRouteFromStatuses(routePeerStatuses map[string]ro
 		if !found || !peerStatus.connected {
 			continue
 		}
+
 		if r.Metric < route.MaxMetric {
 			metricDiff := route.MaxMetric - r.Metric
 			tempScore = metricDiff * 10
 		}
+
 		if !peerStatus.relayed {
 			tempScore++
 		}
-		if !peerStatus.direct {
+
+		if peerStatus.direct {
 			tempScore++
 		}
-		if tempScore > chosenScore || (tempScore == chosenScore && currID == r.ID) {
+
+		if tempScore > chosenScore || (tempScore == chosenScore && r.ID == currID) {
+			chosen = r.ID
+			chosenScore = tempScore
+		}
+
+		if chosen == "" && currID == "" {
 			chosen = r.ID
 			chosenScore = tempScore
 		}
@@ -106,7 +115,9 @@ func (c *clientNetwork) getBestRouteFromStatuses(routePeerStatuses map[string]ro
 		for _, r := range c.routes {
 			peers = append(peers, r.Peer)
 		}
-		log.Warnf("no route was chosen for network %s because no peers from list %s were connected", c.network, peers)
+
+		log.Warnf("the network %s has not been assigned a routing peer as no peers from the list %s are currently connected", c.network, peers)
+
 	} else if chosen != currID {
 		log.Infof("new chosen route is %s with peer %s with score %d", chosen, c.routes[chosen].Peer, chosenScore)
 	}

client/internal/routemanager/client_test.go

@@ -0,0 +1,199 @@
+package routemanager
+
+import (
+	"net/netip"
+	"testing"
+
+	"github.com/netbirdio/netbird/route"
+)
+
+func TestGetBestrouteFromStatuses(t *testing.T) {
+
+	testCases := []struct {
+		name            string
+		statuses        map[string]routerPeerStatus
+		expectedRouteID string
+		currentRoute    *route.Route
+		existingRoutes  map[string]*route.Route
+	}{
+		{
+			name: "one route",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   false,
+					direct:    true,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+		{
+			name: "one connected routes with relayed and direct",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   true,
+					direct:    true,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+		{
+			name: "one connected routes with relayed and no direct",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   true,
+					direct:    false,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+		{
+			name: "no connected peers",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: false,
+					relayed:   false,
+					direct:    false,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "",
+		},
+		{
+			name: "multiple connected peers with different metrics",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   false,
+					direct:    true,
+				},
+				"route2": {
+					connected: true,
+					relayed:   false,
+					direct:    true,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: 9000,
+					Peer:   "peer1",
+				},
+				"route2": {
+					ID:     "route2",
+					Metric: route.MaxMetric,
+					Peer:   "peer2",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+		{
+			name: "multiple connected peers with one relayed",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   false,
+					direct:    true,
+				},
+				"route2": {
+					connected: true,
+					relayed:   true,
+					direct:    true,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+				"route2": {
+					ID:     "route2",
+					Metric: route.MaxMetric,
+					Peer:   "peer2",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+		{
+			name: "multiple connected peers with one direct",
+			statuses: map[string]routerPeerStatus{
+				"route1": {
+					connected: true,
+					relayed:   false,
+					direct:    true,
+				},
+				"route2": {
+					connected: true,
+					relayed:   false,
+					direct:    false,
+				},
+			},
+			existingRoutes: map[string]*route.Route{
+				"route1": {
+					ID:     "route1",
+					Metric: route.MaxMetric,
+					Peer:   "peer1",
+				},
+				"route2": {
+					ID:     "route2",
+					Metric: route.MaxMetric,
+					Peer:   "peer2",
+				},
+			},
+			currentRoute:    nil,
+			expectedRouteID: "route1",
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			// create new clientNetwork
+			client := &clientNetwork{
+				network:     netip.MustParsePrefix("192.168.0.0/24"),
+				routes:      tc.existingRoutes,
+				chosenRoute: tc.currentRoute,
+			}
+
+			chosenRoute := client.getBestRouteFromStatuses(tc.statuses)
+			if chosenRoute != tc.expectedRouteID {
+				t.Errorf("expected routeID %s, got %s", tc.expectedRouteID, chosenRoute)
+			}
+		})
+	}
+}