Merge branch 'main' into feat-post_quantum_ml_kem

# Conflicts:
#	client/internal/peer/conn.go
#	client/internal/peer/handshaker.go

.coderabbit.yaml

@@ -14,5 +14,15 @@ reviews:
     - "!**/*.ts"
     - "!**/*.js"
     - "!**/*.svg"
+  pre_merge_checks:
+    custom_checks:
+      - name: "No attribution trailers"
+        mode: error
+        instructions: >-
+          Fail when the PR description or any commit message carries an
+          attribution trailer or footer: Co-Authored-By, Claude-Session,
+          Generated-By, or a "Generated with"/"Generated by" tool line.
+          Contributors own their contributions (AGENTS.md); ask for the
+          lines to be removed.
 chat:
   auto_reply: true

.githooks/commit-msg

@@ -0,0 +1,26 @@
+#!/bin/bash
+# Refuses commit messages that carry attribution trailers. Contributors own
+# their contributions (AGENTS.md, "No Co-Authored-By or tool-attribution
+# trailers"); a trailer spreads that ownership onto a tool or a bystander.
+
+msg_file="$1"
+
+# Trailer keys in any casing, with any bullet or emoji in front.
+trailers='^[^[:alnum:]]*(co-authored-by|claude-session|generated-by):'
+# "Generated with/by" footers, including "Generated with <emoji> by".
+footer='^[^[:alnum:]]*generated (with|by)( [^[:alnum:]]*by)? '
+# A footer names a product, so a capitalized word must follow the phrase
+# itself. Prose such as "generated by the protobuf compiler" stays legal.
+tool='[Gg][Ee][Nn][Ee][Rr][Aa][Tt][Ee][Dd] ([Ww][Ii][Tt][Hh]|[Bb][Yy])( [^[:alnum:]]*[Bb][Yy])? [^[:alnum:]]*[A-Z]'
+
+offending=$( {
+    grep -Ein "$trailers" "$msg_file"
+    grep -Ein "$footer" "$msg_file" | grep -E "$tool"
+} | sort -un )
+
+if [ -n "$offending" ]; then
+    echo "commit-msg: attribution trailers are not accepted in this repository:" >&2
+    printf '%s\n' "$offending" | sed 's/^/  /' >&2
+    echo "Remove them and commit again (see AGENTS.md)." >&2
+    exit 1
+fi

.github/dependabot.yml

@@ -46,3 +46,25 @@ updates:
       wireguard:
         patterns:
           - "golang.zx2c4.com/wireguard*"
+
+  # Base images of the source-build Dockerfiles, pinned by digest (Chainguard
+  # publishes only :latest for free). Dockerfile.release files feed goreleaser
+  # and keep the published images as they are, so their bases are left alone.
+  - package-ecosystem: "docker"
+    directories:
+      - "/upload-server"
+    schedule:
+      interval: "weekly"
+    open-pull-requests-limit: 3
+    groups:
+      base-images:
+        patterns:
+          - "*"
+    ignore:
+      - dependency-name: "gcr.io/distroless/base"
+      # Go minor and major versions move with the rest of the repository;
+      # patch releases and new digests of the pinned tag still come through.
+      - dependency-name: "golang"
+        update-types:
+          - "version-update:semver-minor"
+          - "version-update:semver-major"

.github/scripts/test-homebrew-cask.sh

@@ -0,0 +1,338 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+fail() {
+    echo "::error::$*" >&2
+    exit 1
+}
+
+if [[ ${RUNNER_ENVIRONMENT:-} != github-hosted || ${RUNNER_OS:-} != macOS || $(uname -s) != Darwin ]]; then
+    fail "This test installs a system daemon and must run on a disposable GitHub macOS runner."
+fi
+if [[ $EUID == 0 ]]; then
+    fail "Run this script as the Homebrew user, not root."
+fi
+
+readonly test_dir="${RUNNER_TEMP:?}/homebrew-cask"
+readonly results_dir="$test_dir/results"
+readonly app='/Applications/Netbird UI.app'
+readonly plist='/Library/LaunchDaemons/netbird.plist'
+readonly cask='netbirdio/tap/netbird-ui'
+readonly formula='netbirdio/tap/netbird'
+readonly published_cask="$test_dir/published-netbird-ui.rb"
+readonly legacy_cask="$test_dir/legacy-netbird-ui.rb"
+readonly rendered_cask="$test_dir/rendered-netbird-ui.rb"
+readonly fixture_dir="$test_dir/fixture"
+readonly serve_dir="$test_dir/serve"
+readonly fixture_zip="$serve_dir/netbird-ui.zip"
+readonly fixture_port=18080
+readonly fixture_url="http://127.0.0.1:$fixture_port/netbird-ui.zip"
+readonly marker="$test_dir/installer.marker"
+
+mkdir -p "$results_dir" "$fixture_dir/netbird_ui_darwin" "$serve_dir" "$test_dir/downloads"
+exec > >(tee "$results_dir/test.log") 2>&1
+
+sudo -n true
+if command -v netbird || [[ -e "$app" || -e "$plist" ]] || pgrep -x netbird-ui; then
+    fail "The runner already has NetBird installed or running."
+fi
+if sudo launchctl print system/netbird > "$results_dir/initial-service.log" 2>&1; then
+    fail "The runner already has a NetBird service loaded."
+fi
+
+install_attempted=false
+server_pid=''
+daemon_pid=''
+version=''
+
+stop_ui() {
+    local status=0
+    sudo pkill -x netbird-ui || status=$?
+    # pkill returns 1 when the UI is already closed.
+    [[ $status == 0 || $status == 1 ]]
+}
+
+cleanup() {
+    local status=$?
+    trap - EXIT
+    set +e
+
+    if [[ $install_attempted == true ]]; then
+        stop_ui || status=1
+        if [[ -S /var/run/netbird.sock ]]; then
+            sudo netbird down || status=1
+        fi
+        if brew list --cask "$cask" >/dev/null 2>&1 || [[ -e "$app" ]]; then
+            brew uninstall --cask --force "$cask" || status=1
+        fi
+        # A failed cask install can leave a daemon even after Homebrew rolls back the app.
+        if sudo launchctl print system/netbird > "$results_dir/cleanup-service.log" 2>&1; then
+            sudo netbird service stop || status=1
+        fi
+        if [[ -e "$plist" ]]; then
+            sudo netbird service uninstall || status=1
+        fi
+    fi
+    if [[ -f /var/log/netbird/client.log ]]; then
+        sudo cat /var/log/netbird/client.log > "$results_dir/client.log" || status=1
+    fi
+    if command -v netbird >/dev/null; then
+        brew uninstall --formula "$formula" || status=1
+    fi
+    if [[ -n $server_pid ]]; then
+        kill "$server_pid" 2>/dev/null || true
+    fi
+    exit "$status"
+}
+trap cleanup EXIT
+trap 'exit 130' INT
+trap 'exit 143' TERM
+
+run_logged() {
+    local name=$1
+    shift
+    "$@" 2>&1 | tee "$results_dir/$name.log"
+}
+
+cask_field() {
+    local stanza=$1 file=$2
+    sed -nE "s/^[[:space:]]*$stanza \"([^\"]+)\".*/\\1/p" "$file"
+}
+
+release_fields() {
+    local file=$1
+    grep -E '^[[:space:]]*(version|url|sha256|app) ' "$file"
+}
+
+use_cask() {
+    local file=$1
+    cp "$file" "$tap_dir/Casks/netbird-ui.rb"
+}
+
+# The released installer opens the UI as root, which never returns on a headless
+# runner. The cask only needs two script paths and a version argument, so the test
+# ships a stub bundle that records what it received and starts the daemon.
+build_fixture() {
+    local bundle="$fixture_dir/netbird_ui_darwin"
+    printf '#!/bin/sh\nexit 0\n' > "$bundle/netbird-ui"
+    chmod 755 "$bundle/netbird-ui"
+    # After a bootout launchd keeps tearing the previous daemon down for a couple of
+    # seconds, and loading the same label again fails until that finishes.
+    cat > "$bundle/installer.sh" <<EOF
+#!/bin/sh
+set -eu
+export PATH=\$PATH:/usr/local/bin:/opt/homebrew/bin
+printf 'version=%s\\nuid=%s\\n' "\$1" "\$(id -u)" > '$marker'
+netbird service install
+attempt=0
+until netbird service start; do
+    attempt=\$((attempt + 1))
+    [ "\$attempt" -lt 15 ] || exit 1
+    sleep 1
+done
+EOF
+    printf '#!/bin/sh\nexit 0\n' > "$bundle/uninstaller.sh"
+    # Shipped without the executable bit so the 0755 seen after install can only come from the cask.
+    chmod 644 "$bundle/installer.sh" "$bundle/uninstaller.sh"
+    rm -f "$fixture_zip"
+    (cd "$fixture_dir" && zip -qr "$fixture_zip" netbird_ui_darwin)
+}
+
+start_fixture_server() {
+    python3 -m http.server "$fixture_port" --bind 127.0.0.1 --directory "$serve_dir" \
+        > "$results_dir/fixture-server.log" 2>&1 &
+    server_pid=$!
+    local attempt
+    for attempt in {1..20}; do
+        if curl --silent --fail --output /dev/null "$fixture_url"; then
+            return
+        fi
+        sleep 0.5
+    done
+    fail "The fixture HTTP server did not come up on port $fixture_port."
+}
+
+assert_published_layout() {
+    local url archive script
+    while read -r url; do
+        archive="$test_dir/downloads/${url##*/}"
+        curl --fail --location --silent --retry 3 --output "$archive" "$url"
+        for script in installer.sh uninstaller.sh; do
+            unzip -l "$archive" | grep -q " netbird_ui_darwin/$script\$" ||
+                fail "The published archive ${url##*/} has no netbird_ui_darwin/$script."
+        done
+    done < <(cask_field url "$published_cask")
+}
+
+assert_no_deprecations() {
+    if grep -Ei '(postflight|uninstall_preflight).*deprecated|deprecated.*(postflight|uninstall_preflight)' "$@"; then
+        fail "Homebrew reported a deprecated cask lifecycle hook."
+    fi
+}
+
+wait_for_daemon() {
+    local attempt
+    for attempt in {1..30}; do
+        if sudo launchctl print system/netbird > "$results_dir/service.log" 2>&1 &&
+            grep -Eq '^[[:space:]]*state = running$' "$results_dir/service.log"; then
+            return
+        fi
+        sleep 1
+    done
+    cat "$results_dir/service.log"
+    fail "The installed daemon did not reach the running state."
+}
+
+wait_for_exit() {
+    local pid=$1 attempt
+    for attempt in {1..30}; do
+        if ! sudo kill -0 "$pid" 2>/dev/null; then
+            return
+        fi
+        sleep 1
+    done
+    fail "Daemon process $pid is still running after removal."
+}
+
+assert_service_absent() {
+    if sudo launchctl print system/netbird > "$results_dir/removed-service.log" 2>&1; then
+        fail "The NetBird service is still loaded after removal."
+    fi
+}
+
+assert_installed() {
+    local script
+    [[ -f $marker ]] || fail "The cask did not run installer.sh."
+    grep -qx "version=$version" "$marker" || fail "installer.sh did not receive the cask version: $(cat "$marker")"
+    grep -qx 'uid=0' "$marker" || fail "installer.sh did not run as root: $(cat "$marker")"
+    [[ -d "$app" && -x "$app/netbird-ui" ]] || fail "The UI was not installed."
+    for script in installer.sh uninstaller.sh; do
+        [[ $(stat -f '%Lp' "$app/$script") == 755 ]] || fail "Incorrect permissions on $script."
+    done
+    [[ -f "$plist" ]] || fail "The installer did not create the daemon plist."
+    wait_for_daemon
+    daemon_pid=$(awk '/^[[:space:]]*pid = / { print $3; exit }' "$results_dir/service.log")
+    [[ $daemon_pid =~ ^[0-9]+$ ]] || fail "The running daemon has no PID."
+    sudo kill -0 "$daemon_pid"
+}
+
+assert_uninstalled() {
+    local log=$1
+    assert_no_deprecations "$log"
+    [[ ! -e "$app" ]] || fail "The UI app remains after uninstall."
+    [[ ! -e "$plist" ]] || fail "The daemon plist remains after uninstall."
+    assert_service_absent
+    wait_for_exit "$daemon_pid"
+    [[ $(netbird version) == "$version" ]] || fail "Cask uninstall removed the CLI dependency."
+}
+
+installed_caskfiles() {
+    local extension=$1
+    find "$(brew --caskroom)/netbird-ui/.metadata" -name "netbird-ui.$extension" 2>/dev/null
+}
+
+assert_legacy_metadata() {
+    installed_caskfiles rb | grep -q . || fail "The legacy cask did not leave a Ruby caskfile behind."
+}
+
+assert_steps_metadata() {
+    if installed_caskfiles rb | grep -q .; then
+        fail "Homebrew still keeps the legacy Ruby caskfile after reinstall."
+    fi
+    installed_caskfiles json | grep -q . || fail "Homebrew did not save the reinstalled cask as JSON."
+}
+
+brew --version
+sw_vers
+brew tap netbirdio/tap "${GITHUB_WORKSPACE:?}/.homebrew-cask-tap"
+tap_dir=$(brew --repository netbirdio/tap)
+readonly tap_dir
+
+[[ -f "$tap_dir/Casks/netbird-ui.rb" ]] || fail "The tap has no Casks/netbird-ui.rb."
+cp "$tap_dir/Casks/netbird-ui.rb" "$published_cask"
+cp "$published_cask" "$results_dir/published-netbird-ui.rb"
+
+version=$(brew info --json=v2 --formula "$formula" | jq -r '.formulae[0].versions.stable')
+readonly version
+[[ -n $version && $version != null ]] || fail "Could not read the formula version from the tap."
+
+assert_published_layout
+
+build_fixture
+fixture_sha=$(shasum -a 256 "$fixture_zip" | cut -d' ' -f1)
+readonly fixture_sha
+start_fixture_server
+
+export PROJECT=netbird-ui VERSION="$version"
+export AMD="$fixture_zip" ARM="$fixture_zip" AMD_URL="$fixture_url" ARM_URL="$fixture_url"
+gomplate -f "$GITHUB_WORKSPACE/client/ui/netbird-ui.rb.tmpl" -o "$rendered_cask"
+cp "$rendered_cask" "$results_dir/rendered-netbird-ui.rb"
+
+sed -E "s|^([[:space:]]*version) \"[^\"]+\"|\\1 \"$version\"|; s|^([[:space:]]*url) \"[^\"]+\"|\\1 \"$fixture_url\"|; s|^([[:space:]]*sha256) \"[^\"]+\"|\\1 \"$fixture_sha\"|" \
+    "$published_cask" > "$legacy_cask"
+cp "$legacy_cask" "$results_dir/legacy-netbird-ui.rb"
+if ! diff <(release_fields "$legacy_cask") <(release_fields "$rendered_cask"); then
+    fail "The rendered cask changes release data, not only lifecycle stanzas."
+fi
+
+use_cask "$rendered_cask"
+brew info --json=v2 --cask "$cask" > "$results_dir/cask.json" 2> "$results_dir/load.log"
+cat "$results_dir/load.log"
+assert_no_deprecations "$results_dir/load.log"
+run_logged style brew style --cask --only-cops=Cask/InstallSteps "$cask"
+
+run_logged install-cli brew install --formula "$formula"
+[[ $(netbird version) == "$version" ]] || fail "The installed CLI does not report the formula version."
+
+for scenario in running stopped missing; do
+    echo "::group::Uninstall with $scenario service"
+    install_attempted=true
+    sudo rm -f "$marker"
+    run_logged "install-$scenario" brew install --cask "$cask"
+    assert_no_deprecations "$results_dir/install-$scenario.log"
+    assert_installed
+    stop_ui
+
+    case "$scenario" in
+        running) ;;
+        stopped)
+            run_logged stop-daemon sudo netbird service stop
+            wait_for_exit "$daemon_pid"
+            [[ -f "$plist" ]] || fail "Stopping the daemon unexpectedly removed its plist."
+            ;;
+        missing)
+            run_logged stop-missing-daemon sudo netbird service stop
+            run_logged remove-daemon sudo netbird service uninstall
+            wait_for_exit "$daemon_pid"
+            [[ ! -e "$plist" ]] || fail "The missing-service scenario still has a plist."
+            assert_service_absent
+            ;;
+        *) fail "Unknown uninstall scenario: $scenario" ;;
+    esac
+
+    run_logged "uninstall-$scenario" brew uninstall --cask "$cask"
+    assert_uninstalled "$results_dir/uninstall-$scenario.log"
+    echo "::endgroup::"
+done
+
+# Every existing user first meets the new cask through an upgrade of the published
+# one, whose legacy flight blocks Homebrew replays from the saved Ruby caskfile.
+echo "::group::Reinstall over the published legacy cask"
+install_attempted=true
+use_cask "$legacy_cask"
+sudo rm -f "$marker"
+run_logged install-legacy brew install --cask "$cask"
+assert_installed
+assert_legacy_metadata
+stop_ui
+
+use_cask "$rendered_cask"
+sudo rm -f "$marker"
+run_logged reinstall-legacy brew reinstall --cask "$cask"
+assert_installed
+assert_steps_metadata
+stop_ui
+
+run_logged uninstall-legacy brew uninstall --cask "$cask"
+assert_uninstalled "$results_dir/uninstall-legacy.log"
+echo "::endgroup::"

.github/workflows/frontend-ui.yml

@@ -38,12 +38,12 @@ jobs:
           persist-credentials: false
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v7
         with:
           node-version: "22"
 
       - name: Set up pnpm
-        uses: pnpm/action-setup@a3252b78c470c02df07e9d59298aecedc3ccdd6d # v3.0.0
+        uses: pnpm/action-setup@ea17c68df8912ef543352723c149a84f56e3d413 # v6.1.0
         with:
           version: 11
 
@@ -79,7 +79,7 @@ jobs:
         run: echo "path=$(pnpm store path --silent)" >> "$GITHUB_OUTPUT"
 
       - name: Cache pnpm store
-        uses: actions/cache@v4
+        uses: actions/cache@v6
         with:
           path: ${{ steps.pnpm-store.outputs.path }}
           key: ${{ runner.os }}-pnpm-${{ hashFiles('client/ui/frontend/pnpm-lock.yaml') }}

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

@@ -46,15 +46,17 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Test
-        # Exclude client/ui: its main.go uses //go:embed all:frontend/dist,
-        # which fails to compile until the frontend has been built. The Wails UI
-        # has no Go-side unit tests, and its release pipeline runs `pnpm build`
-        # before goreleaser.
+        # Exclude the client/ui package itself: its main.go uses //go:embed
+        # all:frontend/dist, which fails to compile until the frontend has been
+        # built, and its release pipeline runs `pnpm build` before goreleaser.
+        # The pattern is anchored so the subpackages (services, preferences,
+        # i18n, authsession) still run: they hold Go-side unit tests and need no
+        # frontend bundle.
         # `go list -e` lets the listing succeed even though the embed fails to
         # resolve; the grep then drops the broken package by path. Without -e,
         # go list aborts with empty stdout and `go test` falls back to the repo
         # root, which has no Go files.
-        run: NETBIRD_STORE_ENGINE=${{ matrix.store }} CI=true go test -coverprofile=coverage.txt -tags 'devcert privileged' -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' -timeout 5m -p 1 $(go list -e ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e /client/ui -e /client/testutil/privileged)
+        run: NETBIRD_STORE_ENGINE=${{ matrix.store }} CI=true go test -coverprofile=coverage.txt -tags 'devcert privileged' -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' -timeout 5m -p 1 $(go list -e ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e '/client/ui$' -e /client/testutil/privileged)
 
       - name: Upload coverage reports to Codecov
         uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f #v7.0.0

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

@@ -160,9 +160,10 @@ jobs:
 
       - name: Test
         # Exclude client/ui: its main.go uses //go:embed all:frontend/dist,
-        # which fails to compile until the frontend has been built. The Wails UI
-        # has no Go-side unit tests, and its release pipeline runs `pnpm build`
-        # before goreleaser.
+        # which fails to compile until the frontend has been built, and its
+        # release pipeline runs `pnpm build` before goreleaser. The subpackages
+        # go with it because this runner's gtk4 is older than the wails runtime
+        # needs; the Client UI / Unit job below covers them instead.
         # `go list -e` lets the listing succeed even though the embed fails to
         # resolve; the grep then drops the broken package by path. Without -e,
         # go list aborts with empty stdout and `go test` falls back to the repo
@@ -177,6 +178,35 @@ jobs:
           slug: netbirdio/netbird
           flags: unit,client
 
+  test_client_ui:
+    name: "Client UI / Unit"
+    # Pinned to 24.04 rather than the 22.04 the other client jobs use: the wails
+    # runtime's linux cgo layer needs GtkFileDialog, which arrived in gtk4 4.10,
+    # and jammy ships 4.6. Not ubuntu-latest, so a runner image rollover cannot
+    # move this out from under us.
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
+        with:
+          persist-credentials: false
+
+      - name: Install Go
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
+        with:
+          go-version-file: "go.mod"
+          cache: false
+
+      - name: Install dependencies
+        run: sudo apt update && sudo apt install -y -q libgtk-4-dev libwebkitgtk-6.0-dev libsoup-3.0-dev
+
+      - name: Test
+        # client/ui itself stays out: its main.go embeds all:frontend/dist,
+        # which only exists after `pnpm build`. The subpackages carry the
+        # Go-side unit tests, including the window manager re-entrancy
+        # regression test, and need no frontend bundle.
+        run: CGO_ENABLED=1 go test -timeout 5m ./client/ui/authsession/... ./client/ui/i18n/... ./client/ui/preferences/... ./client/ui/services/...
+
   test_client_on_docker:
     name: "Client (Docker) / Unit"
     needs: [build-cache]
@@ -211,6 +241,9 @@ jobs:
             ${{ runner.os }}-gotest-cache-
 
       - name: Run tests in container
+        # Unlike the native job above, this one drops all of client/ui including
+        # the subpackages: the alpine container has no gtk4/webkitgtk, so the
+        # Wails application package they import would fail to link.
         env:
           HOST_GOCACHE: ${{ steps.go-env.outputs.cache_dir }}
           HOST_GOMODCACHE: ${{ steps.go-env.outputs.modcache_dir }}
@@ -481,14 +514,32 @@ jobs:
         if: matrix.store == 'mysql'
         run: docker pull mlsmaycon/warmed-mysql:8
 
+      # The -json stream goes through tools/gotestsummary so the log shows one
+      # line per test, the output of failed tests, the head of a timeout panic
+      # with the still-running tests, and the slowest tests per package.
       - name: Test
+        shell: bash
         run: |
+          set -o pipefail
           CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
           NETBIRD_STORE_ENGINE=${{ matrix.store }} \
             CI=true \
-          go test -tags=devcert -coverprofile=coverage.txt \
+          go test -json -tags=devcert -coverprofile=coverage.txt \
             -exec "sudo --preserve-env=CI,NETBIRD_STORE_ENGINE" \
-            -timeout 20m ./management/... ./shared/management/...
+            -timeout 20m ./management/... ./shared/management/... \
+            | tee management-test-events.jsonl \
+            | go run ./tools/gotestsummary
+
+      # The summary trims long outputs; the raw stream keeps every line for
+      # the failures that need it. A green run has no use for it.
+      - name: Upload raw test events
+        if: failure()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a #v7.0.1
+        with:
+          name: management-unit-test-events-${{ matrix.store }}
+          path: management-test-events.jsonl
+          if-no-files-found: ignore
+          retention-days: 14
 
       - name: Upload coverage reports to Codecov
         if: matrix.arch == 'amd64'
@@ -738,12 +789,27 @@ jobs:
       - name: check git status
         run: git --no-pager diff --exit-code
 
+      # Same summary as the unit job: a timeout here names the tests still
+      # running instead of ending in a goroutine dump.
       - name: Test
+        shell: bash
         run: |
+          set -o pipefail
           CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
           NETBIRD_STORE_ENGINE=${{ matrix.store }} \
           CI=true \
-          mage integrationtest:all -gotestflags="-coverprofile=coverage.txt"
+          mage integrationtest:all -gotestflags="-json -coverprofile=coverage.txt" \
+            | tee management-integration-test-events.jsonl \
+            | go run ./tools/gotestsummary
+
+      - name: Upload raw test events
+        if: failure()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a #v7.0.1
+        with:
+          name: management-integration-test-events-${{ matrix.store }}
+          path: management-integration-test-events.jsonl
+          if-no-files-found: ignore
+          retention-days: 14
 
       - name: Upload coverage reports to Codecov
         if: matrix.arch == 'amd64'

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

@@ -66,15 +66,17 @@ jobs:
       - run: PsExec64 -s -w ${{ github.workspace }} C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe env -w GOCACHE=${{ env.modcache }}
       - run: PsExec64 -s -w ${{ github.workspace }} C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe mod tidy
       - name: Generate test script
-        # Exclude client/ui: its main.go uses //go:embed all:frontend/dist,
-        # which fails to compile until the frontend has been built. The Wails UI
-        # has no Go-side unit tests, and its release pipeline runs `pnpm build`
-        # before goreleaser.
+        # Exclude the client/ui package itself: its main.go uses //go:embed
+        # all:frontend/dist, which fails to compile until the frontend has been
+        # built, and its release pipeline runs `pnpm build` before goreleaser.
+        # The pattern is anchored so the subpackages (services, preferences,
+        # i18n, authsession) still run: they hold Go-side unit tests and need no
+        # frontend bundle.
         # `go list -e` lets the listing succeed even though the embed fails to
         # resolve; the Where-Object pipeline then drops the broken package by
         # path. Without -e, go list aborts with empty stdout.
         run: |
-          $packages = go list -e ./... | Where-Object { $_ -notmatch '/management' } | Where-Object { $_ -notmatch '/relay' } | Where-Object { $_ -notmatch '/signal' } | Where-Object { $_ -notmatch '/proxy' } | Where-Object { $_ -notmatch '/combined' } | Where-Object { $_ -notmatch '/client/ui' }
+          $packages = go list -e ./... | Where-Object { $_ -notmatch '/management' } | Where-Object { $_ -notmatch '/relay' } | Where-Object { $_ -notmatch '/signal' } | Where-Object { $_ -notmatch '/proxy' } | Where-Object { $_ -notmatch '/combined' } | Where-Object { $_ -notmatch '/client/ui$' }
           $goExe = "C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe"
           $cmd = "$goExe test -tags `"devcert privileged`" -timeout 10m -p 1 $($packages -join ' ') > test-out.txt 2>&1"
           Set-Content -Path "${{ github.workspace }}\run-tests.cmd" -Value $cmd

.github/workflows/golangci-lint.yml

@@ -80,3 +80,49 @@ jobs:
           skip-save-cache: true
           cache-invalidation-interval: 0
           args: --timeout=20m
+
+  # Separate job rather than extra rows in the matrix above: those rows pick a
+  # GOOS by picking a runner OS, while android/ios are cross-compiled from
+  # ubuntu — an `include` entry with os: ubuntu-latest would merge into the
+  # Linux row instead of adding one. The package path is restricted because a
+  # whole-repo run under GOOS=android pulls *_linux.go files into packages that
+  # have no android counterpart.
+  golangci-mobile:
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - goos: android
+            goarch: arm64
+            packages: ./client/android/...
+            display_name: Android
+          - goos: ios
+            goarch: arm64
+            packages: ./client/ios/...
+            display_name: iOS
+    name: ${{ matrix.display_name }}
+    runs-on: ubuntu-latest
+    timeout-minutes: 25
+    env:
+      CGO_ENABLED: 0
+      GOOS: ${{ matrix.goos }}
+      GOARCH: ${{ matrix.goarch }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
+        with:
+          persist-credentials: false
+      - name: Install Go
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
+        with:
+          go-version-file: "go.mod"
+          cache: false
+      - name: golangci-lint
+        uses: golangci/golangci-lint-action@82606bf257cbaff209d206a39f5134f0cfbfd2ee #v9.2.1
+        with:
+          version: latest
+          install-mode: binary
+          skip-cache: true
+          skip-save-cache: true
+          cache-invalidation-interval: 0
+          args: --timeout=20m ${{ matrix.packages }}

.github/workflows/mobile-build-validation.yml

@@ -0,0 +1,64 @@
+name: Mobile
+
+on:
+  push:
+    branches:
+      - main
+      - "release-*"
+  pull_request:
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.actor_id }}
+  cancel-in-progress: true
+
+jobs:
+  android_build:
+    name: "Android / Build"
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        goarch: [arm64, arm, amd64, "386"]
+    env:
+      CGO_ENABLED: 0
+      GOOS: android
+      GOARCH: ${{ matrix.goarch }}
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
+        with:
+          persist-credentials: false
+      - name: Install Go
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
+        with:
+          go-version-file: "go.mod"
+      - name: Build Android bridge
+        run: go build ./client/android/...
+      - name: Vet Android bridge
+        if: matrix.goarch == 'arm64'
+        run: go vet ./client/android/...
+
+  ios_build:
+    name: "iOS / Build"
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        goarch: [arm64, amd64]
+    env:
+      CGO_ENABLED: 0
+      GOOS: ios
+      GOARCH: ${{ matrix.goarch }}
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
+        with:
+          persist-credentials: false
+      - name: Install Go
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
+        with:
+          go-version-file: "go.mod"
+      # No `go vet` counterpart: every ios target requires external (cgo)
+      # linking, which needs an Xcode toolchain the runner does not have.
+      - name: Build iOS SDK
+        run: go build ./client/ios/...

.github/workflows/release.yml

@@ -186,6 +186,22 @@ jobs:
         run: bash shared/management/http/api/generate.sh
       - name: check git status
         run: git --no-pager diff --exit-code
+      - name: Generate RPM changelog from git tags
+        # nfpm embeds changelog.yml into the RPM; Red Hat software certification
+        # requires a changelog. Generated, not committed (see .gitignore).
+        # chglog is a go.mod tool directive, so go.sum pins it and its deps.
+        run: bash release_files/rpm-changelog.sh
+      - name: Fill the RPM ISA provide version
+        # nfpm cannot emit rpmbuild's ISA provide and GoReleaser cannot template it.
+        run: bash release_files/rpm-provides.sh
+      - name: Set up Node.js
+        uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
+        with:
+          node-version: '22'
+      - name: Install proxy web dependencies for license collection
+        # proxy/collect-licenses.sh reads the UI's license terms from node_modules.
+        working-directory: proxy/web
+        run: npm ci --ignore-scripts
       - name: Set up QEMU
         uses: docker/setup-qemu-action@06116385d9baf250c9f4dcb4858b16962ea869c3 #v4.1.0
       - name: Set up Docker Buildx
@@ -225,14 +241,18 @@ jobs:
         uses: goreleaser/goreleaser-action@5daf1e915a5f0af01ddbcd89a43b8061ff4f1a89 # v7.2.2
         with:
           version: ${{ env.GORELEASER_VER }}
-          args: release --clean ${{ env.flags }}
+          args: release --config .goreleaser.generated.yaml --clean ${{ env.flags }}
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           HOMEBREW_TAP_GITHUB_TOKEN: ${{ secrets.HOMEBREW_TAP_GITHUB_TOKEN }}
           UPLOAD_DEBIAN_SECRET: ${{ secrets.PKG_UPLOAD_SECRET }}
           UPLOAD_YUM_SECRET: ${{ secrets.PKG_UPLOAD_SECRET }}
           GPG_RPM_KEY_FILE: ${{ env.GPG_RPM_KEY_FILE }}
-          NFPM_NETBIRD_RPM_PASSPHRASE: ${{ secrets.GPG_RPM_PASSPHRASE }}
+          # One per nfpm id: GoReleaser looks the passphrase up as NFPM_<ID>_PASSPHRASE.
+          NFPM_NETBIRD_RPM_AMD64_PASSPHRASE: ${{ secrets.GPG_RPM_PASSPHRASE }}
+          NFPM_NETBIRD_RPM_ARM64_PASSPHRASE: ${{ secrets.GPG_RPM_PASSPHRASE }}
+          NFPM_NETBIRD_RPM_ARM_PASSPHRASE: ${{ secrets.GPG_RPM_PASSPHRASE }}
+          NFPM_NETBIRD_RPM_386_PASSPHRASE: ${{ secrets.GPG_RPM_PASSPHRASE }}
           SKIP_PUBLISH: ${{ env.SKIP_PUBLISH }}
           SKIP_DOCKER_PUSH: ${{ env.SKIP_DOCKER_PUSH }}
       - name: Verify RPM signatures
@@ -289,10 +309,12 @@ jobs:
           tag_and_push() {
             local src="$1" img_name tag dst variant=""
             img_name="${src%%:*}"
-            # Client variants share a repository, so keep their tag suffixes.
+            # Variants share a repository with their default image, so keep
+            # their tag suffixes. Order matters: the first matching pattern wins.
             case "$src" in
               *-rootless-ubi-amd64) variant="-rootless-ubi" ;;
               *-rootless-amd64) variant="-rootless" ;;
+              *-ubi-amd64) variant="-ubi" ;;
             esac
             for tag in $(resolve_tags); do
               dst="${img_name}:${tag}${variant}"
@@ -358,6 +380,132 @@ jobs:
           path: dist/netbird_darwin**
           retention-days: 7
 
+  # Certify and publish the rootless UBI client image in the Red Hat Ecosystem
+  # Catalog. Stable tags only: goreleaser pushes <version>-rootless-ubi to
+  # ghcr.io in the release job above, and preflight submits every architecture
+  # of that manifest list to Pyxis. Auto-publish on the component makes the new
+  # version public once certification passes.
+  redhat_certification:
+    name: "Red Hat / Certify rootless UBI image"
+    needs: release
+    if: |
+      github.repository == 'netbirdio/netbird' &&
+      startsWith(github.ref, 'refs/tags/v') &&
+      !contains(github.ref_name, '-')
+    runs-on: ubuntu-24.04
+    permissions:
+      contents: read
+    env:
+      PREFLIGHT_VERSION: "1.21.0"
+      # sha256 of preflight-linux-amd64 from the 1.21.0 GitHub release.
+      # Red Hat publishes no checksum file, so the value is pinned here.
+      PREFLIGHT_SHA256: "5e653135503c72f8702bbe31d7643197d12937c68086879133dd6b9650a9a449"
+      IMAGE_REPOSITORY: "ghcr.io/netbirdio/netbird"
+      # Component "NetBird Client Container Image (rootless)" in Partner Connect.
+      # Override with the REDHAT_CERT_COMPONENT_ID repository variable if it changes.
+      DEFAULT_COMPONENT_ID: "6aa3ca4b4676aefdf07aaa97"
+    steps:
+      - name: Resolve image reference
+        id: image
+        env:
+          INPUT_VERSION: ${{ github.ref_name }}
+        run: |
+          set -euo pipefail
+          version="${INPUT_VERSION#v}"
+          if [[ ! "$version" =~ ^[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
+            echo "::error::Only stable x.y.z versions are certified, got '${INPUT_VERSION}'"
+            exit 1
+          fi
+          echo "version=${version}" >> "$GITHUB_OUTPUT"
+          echo "ref=${IMAGE_REPOSITORY}:${version}-rootless-ubi" >> "$GITHUB_OUTPUT"
+
+      - name: Verify the multi-arch image is on ghcr.io
+        env:
+          IMAGE_REF: ${{ steps.image.outputs.ref }}
+        run: |
+          set -euo pipefail
+          docker buildx imagetools inspect "$IMAGE_REF" --raw > manifest.json
+          for arch in amd64 arm64; do
+            if ! jq -e --arg a "$arch" '.manifests[] | select(.platform.architecture == $a)' manifest.json > /dev/null; then
+              echo "::error::${IMAGE_REF} has no ${arch} manifest"
+              exit 1
+            fi
+          done
+          echo "Manifest list for ${IMAGE_REF}:"
+          jq -r '.manifests[] | "\(.platform.os)/\(.platform.architecture) \(.digest)"' manifest.json
+
+      - name: Install preflight
+        run: |
+          set -euo pipefail
+          curl -fsSL --proto '=https' --proto-redir '=https' -o preflight \
+            "https://github.com/redhat-openshift-ecosystem/openshift-preflight/releases/download/${PREFLIGHT_VERSION}/preflight-linux-amd64"
+          echo "${PREFLIGHT_SHA256}  preflight" | sha256sum -c -
+          chmod +x preflight
+          ./preflight --version
+
+      - name: Run preflight checks and submit to Red Hat
+        env:
+          IMAGE_REF: ${{ steps.image.outputs.ref }}
+          PFLT_PYXIS_API_TOKEN: ${{ secrets.PYXIS_API_TOKEN }}
+          PFLT_CERTIFICATION_COMPONENT_ID: ${{ vars.REDHAT_CERT_COMPONENT_ID || env.DEFAULT_COMPONENT_ID }}
+          PFLT_ARTIFACTS: artifacts
+          PFLT_LOGFILE: artifacts/preflight.log
+          PFLT_LOGLEVEL: info
+          PFLT_JUNIT: "true"
+        run: |
+          set -euo pipefail
+          # No --platform: preflight walks the manifest list and submits every
+          # architecture in one run, grouped under one manifest-list digest.
+          ./preflight check container "$IMAGE_REF" --submit
+
+      - name: Fail if any check did not pass
+        run: |
+          set -euo pipefail
+          shopt -s nullglob
+          results=(artifacts/results.json artifacts/*/results.json)
+          if [[ ${#results[@]} -eq 0 ]]; then
+            echo "::error::preflight produced no results.json"
+            exit 1
+          fi
+          status=0
+          for f in "${results[@]}"; do
+            arch="$(basename "$(dirname "$f")")"
+            passed="$(jq -r '.passed' "$f")"
+            failed="$(jq -r '[.results.failed[]?.name] | join(", ")' "$f")"
+            echo "${arch}: passed=${passed} ${failed:+failed checks: ${failed}}"
+            [[ "$passed" == "true" ]] || status=1
+          done
+          exit $status
+
+      - name: Upload preflight artifacts
+        if: always()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: redhat-preflight-${{ steps.image.outputs.version }}
+          path: artifacts/
+          retention-days: 30
+
+      - name: Wait for Pyxis to mark both architectures certified
+        env:
+          VERSION: ${{ steps.image.outputs.version }}
+          PFLT_PYXIS_API_TOKEN: ${{ secrets.PYXIS_API_TOKEN }}
+          COMPONENT_ID: ${{ vars.REDHAT_CERT_COMPONENT_ID || env.DEFAULT_COMPONENT_ID }}
+        run: |
+          set -euo pipefail
+          tag="${VERSION}-rootless-ubi"
+          url="https://catalog.redhat.com/api/containers/v1/projects/certification/id/${COMPONENT_ID}/images?page_size=100"
+          for attempt in $(seq 1 20); do
+            certified="$(curl -fsS --proto '=https' --proto-redir '=https' -H "X-API-KEY: ${PFLT_PYXIS_API_TOKEN}" "$url" \
+              | jq -r --arg t "$tag" '[.data[] | select(.repositories[]?.tags[]?.name == $t) | select(.certified == true) | .architecture] | unique | join(",")')"
+            echo "attempt ${attempt}: certified architectures for ${tag}: ${certified:-none}"
+            if [[ "$certified" == "amd64,arm64" ]]; then
+              echo "Both architectures certified. Auto-publish is enabled on the component, so the catalog updates on its own."
+              exit 0
+            fi
+            sleep 30
+          done
+          echo "::warning::Pyxis has not marked both architectures certified after 10 minutes. Check https://connect.redhat.com/component/view/${COMPONENT_ID}/images"
+
   release_ui:
     runs-on: ubuntu-latest
     outputs:
@@ -412,12 +560,12 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v7
         with:
           node-version: '22'
 
       - name: Set up pnpm
-        uses: pnpm/action-setup@a3252b78c470c02df07e9d59298aecedc3ccdd6d # v3.0.0
+        uses: pnpm/action-setup@ea17c68df8912ef543352723c149a84f56e3d413 # v6.1.0
         with:
           version: 11
 
@@ -549,12 +697,12 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Set up Node.js
-        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
+        uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7.0.0
         with:
           node-version: '22'
 
       - name: Set up pnpm
-        uses: pnpm/action-setup@a3252b78c470c02df07e9d59298aecedc3ccdd6d # v3.0.0
+        uses: pnpm/action-setup@ea17c68df8912ef543352723c149a84f56e3d413 # v6.1.0
         with:
           version: 11
 
@@ -646,11 +794,11 @@ jobs:
       - name: check git status
         run: git --no-pager diff --exit-code
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v7
         with:
           node-version: '22'
       - name: Set up pnpm
-        uses: pnpm/action-setup@a3252b78c470c02df07e9d59298aecedc3ccdd6d # v3.0.0
+        uses: pnpm/action-setup@ea17c68df8912ef543352723c149a84f56e3d413 # v6.1.0
         with:
           version: 11
       - name: Install wails3 CLI
@@ -769,7 +917,7 @@ jobs:
         run: 7z x -o"${{ github.workspace }}/NSIS_Plugins" "${{ github.workspace }}/ShellExecAsUser_amd64-Unicode.7z"
 
       - name: Set up Go for wails3 CLI
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v6
         with:
           go-version-file: "go.mod"
           cache: false

.github/workflows/test-homebrew-cask.yml

@@ -0,0 +1,46 @@
+name: Test Homebrew cask
+
+on:
+  pull_request:
+    paths:
+      - "client/ui/netbird-ui.rb.tmpl"
+      - ".github/scripts/test-homebrew-cask.sh"
+      - ".github/workflows/test-homebrew-cask.yml"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.actor_id }}
+  cancel-in-progress: true
+
+jobs:
+  install-uninstall:
+    runs-on: macos-latest
+    timeout-minutes: 20
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7.0.0
+        with:
+          persist-credentials: false
+
+      - name: Clone the Homebrew tap
+        run: git clone https://github.com/netbirdio/homebrew-tap.git .homebrew-cask-tap
+
+      - name: Update Homebrew and install gomplate
+        # The runner image disables auto-update; the cask steps DSL needs Homebrew 6.0.20 or newer.
+        run: |
+          brew update
+          brew install gomplate
+
+      - name: Install and uninstall the cask
+        run: .github/scripts/test-homebrew-cask.sh
+
+      - name: Upload logs
+        if: always()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a #v7.0.1
+        with:
+          name: homebrew-cask-results
+          path: ${{ runner.temp }}/homebrew-cask/results
+          if-no-files-found: ignore

.github/workflows/ui-translations.yml

@@ -32,7 +32,7 @@ jobs:
           persist-credentials: false
 
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v7
         with:
           node-version: "22"
 

.gitignore

@@ -35,3 +35,10 @@ vendor/
 /netbird
 client/netbird-electron/
 management/server/types/testdata/
+
+# generated by chglog in the release workflow, embedded into the RPM
+changelog.yml
+
+# generated by rpm-provides.sh, the config GoReleaser actually runs
+.goreleaser.generated.yaml
+.chglog.yml

.goreleaser.yaml

@@ -40,6 +40,32 @@ builds:
     tags:
       - load_wgnt_from_rsrc
 
+  # Single-arch builds: nfpm provides is not templated, so the RPM splits per arch.
+  - &netbird_rpm_build
+    id: netbird-rpm-amd64
+    dir: client
+    binary: netbird
+    env: [CGO_ENABLED=0]
+    goos: [linux]
+    goarch: [amd64]
+    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
+
+  - <<: *netbird_rpm_build
+    id: netbird-rpm-arm64
+    goarch: [arm64]
+
+  - <<: *netbird_rpm_build
+    id: netbird-rpm-arm
+    goarch: [arm]
+
+  - <<: *netbird_rpm_build
+    id: netbird-rpm-386
+    goarch: [386]
+
   - id: netbird-static
     dir: client
     binary: netbird
@@ -223,23 +249,72 @@ nfpms:
       postinstall: "release_files/post_install.sh"
       preremove: "release_files/pre_remove.sh"
 
-  - maintainer: Netbird <dev@netbird.io>
+  - &netbird_rpm
+    maintainer: Netbird <dev@netbird.io>
     description: Netbird client.
     homepage: https://netbird.io/
     license: BSD-3-Clause
     vendor: NetBird
-    id: netbird_rpm
+    id: netbird_rpm_amd64
     bindir: /usr/bin
-    builds:
-      - netbird
+    ids:
+      - netbird-rpm-amd64
     formats:
       - rpm
+    # Red Hat certification (RPM Version Handling) requires rpmbuild's ISA
+    # provide, which nfpm does not emit. The version is filled in by the release job.
+    provides:
+      - "netbird(x86-64) = @RPM_EVR@"
+    # The client verifies TLS to management and signal against the system trust
+    # store. Red Hat software certification (RPM Dependency Tracking) also
+    # rejects packages that declare no dependencies at all.
+    dependencies:
+      - ca-certificates
+    # Generated in CI by chglog from git tags; Red Hat certification requires an
+    # RPM changelog (RPM Version Handling subtest).
+    changelog: changelog.yml
+    # License, documentation and a config file so the RPM Provenance subtest sees
+    # %license, %doc and %config entries instead of a bare binary.
+    contents:
+      - src: LICENSE
+        dst: /usr/share/licenses/netbird/LICENSE
+        type: license
+      - src: README.md
+        dst: /usr/share/doc/netbird/README.md
+        type: doc
+      - src: release_files/netbird.sysconfig
+        dst: /etc/sysconfig/netbird
+        type: config|noreplace
     scripts:
       postinstall: "release_files/post_install.sh"
       preremove: "release_files/pre_remove.sh"
     rpm:
+      summary: NetBird client
+      group: Applications/Internet
+      packager: NetBird <dev@netbird.io>
       signature:
         key_file: '{{ if index .Env "GPG_RPM_KEY_FILE" }}{{ .Env.GPG_RPM_KEY_FILE }}{{ end }}'
+
+  - <<: *netbird_rpm
+    id: netbird_rpm_arm64
+    ids:
+      - netbird-rpm-arm64
+    provides:
+      - "netbird(aarch-64) = @RPM_EVR@"
+
+  - <<: *netbird_rpm
+    id: netbird_rpm_arm
+    ids:
+      - netbird-rpm-arm
+    provides:
+      - "netbird(armv6hl-32) = @RPM_EVR@"
+
+  - <<: *netbird_rpm
+    id: netbird_rpm_386
+    ids:
+      - netbird-rpm-386
+    provides:
+      - "netbird(x86-32) = @RPM_EVR@"
 dockers_v2:
    - id: netbird
      disable: "{{ .Env.SKIP_DOCKER_PUSH }}"
@@ -402,7 +477,7 @@ dockers_v2:
      tags:
        - "{{ .Version }}"
        - "{{ if eq .Env.SKIP_PUBLISH \"false\" }}latest{{ end }}"
-     dockerfile: upload-server/Dockerfile
+     dockerfile: upload-server/Dockerfile.release
      platforms:
        - linux/amd64
        - linux/arm64
@@ -458,6 +533,41 @@ dockers_v2:
        "org.opencontainers.image.revision": "{{.FullCommit}}"
        "org.opencontainers.image.source": "{{.GitURL}}"
        "maintainer": "dev@netbird.io"
+   - id: proxy-ubi
+     disable: "{{ .Env.SKIP_DOCKER_PUSH }}"
+     ids:
+       - netbird-proxy
+     images:
+       - netbirdio/reverse-proxy
+       - ghcr.io/netbirdio/reverse-proxy
+     tags:
+       - "{{ .Version }}-ubi"
+       - "{{ if eq .Env.SKIP_PUBLISH \"false\" }}ubi-latest{{ end }}"
+     dockerfile: proxy/Dockerfile.ubi
+     platforms:
+       - linux/amd64
+       - linux/arm64
+     build_args:
+       VERSION: "{{ .Version }}"
+       RELEASE: "{{ .Timestamp }}"
+     hooks:
+       pre:
+         - cmd: 'sh proxy/collect-licenses.sh "{{ .ContextDir }}/licenses" amd64 arm64'
+           env:
+             - GOOS=linux
+             - CGO_ENABLED=0
+     labels:
+       "org.opencontainers.image.created": "{{.Date}}"
+       "org.opencontainers.image.version": "{{.Version}}"
+       "org.opencontainers.image.revision": "{{.FullCommit}}"
+       "org.opencontainers.image.source": "{{.GitURL}}"
+     annotations:
+       "org.opencontainers.image.created": "{{.Date}}"
+       "org.opencontainers.image.title": "{{.ProjectName}}"
+       "org.opencontainers.image.version": "{{.Version}}"
+       "org.opencontainers.image.revision": "{{.FullCommit}}"
+       "org.opencontainers.image.source": "{{.GitURL}}"
+       "maintainer": "dev@netbird.io"
 
 brews:
   - ids:
@@ -490,7 +600,10 @@ uploads:
   - name: yum
     skip: "{{ .Env.SKIP_PUBLISH }}"
     ids:
-      - netbird_rpm
+      - netbird_rpm_amd64
+      - netbird_rpm_arm64
+      - netbird_rpm_arm
+      - netbird_rpm_386
     mode: archive
     target: https://pkgs.wiretrustee.com/yum/{{ .Arch }}{{ if .Arm }}{{ .Arm }}{{ end }}
     username: dev@wiretrustee.com

AGENTS.md

@@ -77,7 +77,7 @@ make lint            # golangci-lint on files changed vs origin/main (also the p
 make lint-all        # full-repository lint, matches CI
 make test-unit       # host-safe unit tests, -tags devcert, no sudo
 make test-privileged # privileged-tagged suite in a Docker container with NET_ADMIN
-make setup-hooks     # wire make lint into .githooks/pre-push
+make setup-hooks     # wire .githooks: pre-push runs make lint, commit-msg refuses attribution trailers
 
 # Narrow runs
 go test ./client/internal/dns/...

CLAUDE.md

@@ -1 +1,4 @@
-See [AGENTS.md](AGENTS.md) for the agent guidelines in this repository.
+The agent guidelines live in [AGENTS.md](AGENTS.md). It is imported here so
+every session loads it in full rather than following a pointer.
+
+@AGENTS.md

Makefile

@@ -23,8 +23,8 @@ lint-install: $(GOLANGCI_LINT)
 # Setup git hooks for all developers
 setup-hooks:
 	@git config core.hooksPath .githooks
-	@chmod +x .githooks/pre-push
-	@echo "✅ Git hooks configured! Pre-push will now run 'make lint'"
+	@chmod +x .githooks/pre-push .githooks/commit-msg
+	@echo "✅ Git hooks configured! Pre-push runs 'make lint'; commit-msg refuses attribution trailers"
 
 # Host-safe unit tests: excludes the privileged-tagged tests (root / system-mutating).
 # Runs as a normal user with no sudo and leaves host networking untouched.

README.md

@@ -115,6 +115,56 @@ export NETBIRD_DOMAIN=netbird.example.com; curl -fsSL https://github.com/netbird
 
 See a complete [architecture overview](https://docs.netbird.io/about-netbird/how-netbird-works#architecture) for details.
 
+### Reporting bugs and requesting features
+
+NetBird uses a discussion-first workflow. Bug reports and feature requests start in
+[Discussions](https://github.com/netbirdio/netbird/discussions), not as issues.
+
+| What you want to do | Where to go |
+| --- | --- |
+| Report a bug, regression, or unexpected behavior | [Issue Triage](https://github.com/netbirdio/netbird/discussions/new?category=issue-triage) |
+| Request a feature or share an idea | [Ideas & Feature Requests](https://github.com/netbirdio/netbird/discussions/new?category=ideas-feature-requests) |
+| Ask about setup, configuration, or self-hosting | [Q&A / Support](https://github.com/netbirdio/netbird/discussions/new?category=q-a-support) |
+| Report a security vulnerability | [Security policy](https://github.com/netbirdio/netbird/security/policy), never a public thread |
+
+Our team and maintainers triage discussions, ask follow-up questions, check for duplicates,
+and reproduce bugs. Validated reports are promoted to issues. This keeps the issue tracker a clear
+answer to one question: what is the team working on.
+
+Please search existing discussions and issues first, including closed ones. If something similar
+already exists, upvote it and add your details there instead of opening a duplicate.
+
+For bug reports, include your NetBird version, operating system, deployment type (Cloud,
+self-hosted, Kubernetes, or Docker), reproduction steps, expected and actual behavior, and a debug
+bundle where relevant:
+
+```shell
+netbird version
+netbird status -d -A
+netbird debug for 1m -A -S -U
+```
+
+`-U` uploads the bundle and prints a file key you can paste instead of attaching the archive.
+`-A` anonymizes the output, which matters on a public thread. It masks most identifying details
+but is not full redaction, so read the bundle before posting it. Two levels are available:
+
+| Level | How to select | What it masks |
+| --- | --- | --- |
+| `default` | `-A` / `--anonymize` | Public IP addresses, IPv6 ULA addresses, MAC addresses, and domains other than `netbird.io`, `netbird.cloud`, `netbird.selfhosted`, and `netbird.stage`. IPv4 private, CGNAT, and link-local ranges are kept |
+| `strict` | `--anonymize-level strict` (implies `-A`) | The above, plus IPv4 private, CGNAT, and link-local ranges, peer names in front of `netbird.cloud`, `netbird.selfhosted`, and `netbird.stage`, and WireGuard public keys. Labels under `netbird.io` are kept, since it only hosts infrastructure |
+
+See [collecting a debug bundle](https://docs.netbird.io/help/troubleshooting-client#debug-bundle)
+and the [CLI reference](https://docs.netbird.io/get-started/cli#debug-for) for details.
+
+See [How to use Discussions, Issues, and Pull Requests](https://github.com/netbirdio/netbird/discussions/6075)
+for the full workflow, or [SUPPORT.md](SUPPORT.md) for a shorter version.
+
+### Contributing
+
+Contributions are welcome. Read [CONTRIBUTING.md](CONTRIBUTING.md) first. NetBird works ticket
+first, anything that changes behavior needs an issue the team has agreed on before you open a pull
+request.
+
 ### Community projects
 - [NetBird installer script](https://github.com/physk/netbird-installer)
 - [netbird-tui](https://github.com/n0pashkov/netbird-tui) - terminal UI for managing NetBird peers, routes, and settings

SUPPORT.md

@@ -0,0 +1,121 @@
+# Getting help with NetBird
+
+Where to go depends on what you need. If you are not sure, start with
+[Q&A / Support](https://github.com/netbirdio/netbird/discussions/new?category=q-a-support)
+and we will move it.
+
+## Before you post
+
+1. Search existing [discussions](https://github.com/netbirdio/netbird/discussions) and
+   [issues](https://github.com/netbirdio/netbird/issues), including closed ones.
+2. Check the [documentation](https://docs.netbird.io) and the troubleshooting guides for
+   [clients](https://docs.netbird.io/help/troubleshooting-client) and
+   [self-hosted deployments](https://docs.netbird.io/selfhosted/troubleshooting).
+3. Remove or anonymize sensitive information from logs, screenshots, and configuration.
+
+If a discussion already covers your problem, upvote it and add your details there rather than
+opening a duplicate. Extra reproduction detail, affected versions, and deployment notes are
+useful even on an existing thread.
+
+## Community support
+
+Free, for everyone. Covers the NetBird client, open source self-hosted deployments, and general
+questions.
+
+| What you want to do | Where to go |
+| --- | --- |
+| Report a bug, regression, or unexpected behavior | [Issue Triage](https://github.com/netbirdio/netbird/discussions/new?category=issue-triage) |
+| Request a feature or share an idea | [Ideas & Feature Requests](https://github.com/netbirdio/netbird/discussions/new?category=ideas-feature-requests) |
+| Ask about setup, configuration, or self-hosting | [Q&A / Support](https://github.com/netbirdio/netbird/discussions/new?category=q-a-support) |
+| Chat with the community | [Slack](https://docs.netbird.io/slack-url) |
+
+## Paid support
+
+For NetBird Cloud customers and commercial-license self-hosted deployments, covering the
+dashboard, control plane, billing, and subscriptions, see
+[reporting bugs and issues](https://docs.netbird.io/help/report-bug-issues).
+
+## Security
+
+Do not report security vulnerabilities in public issues or discussions, and do not post secrets,
+private keys, internal hostnames, or sensitive logs. Use the
+[security policy](https://github.com/netbirdio/netbird/security/policy).
+
+## What makes a report we can act on
+
+For a bug, the most useful reports include:
+
+- NetBird version, and component versions where applicable
+- Operating system or environment
+- Deployment type: NetBird Cloud, self-hosted, Kubernetes, Docker, or local development
+- Current behavior and expected behavior
+- The smallest set of steps that reproduces the problem
+- Logs, status output, screenshots, or a debug bundle when relevant
+- Whether this worked before, and the last known working version
+
+For client reports, these commands usually give us what we need:
+
+```shell
+netbird version
+netbird status -d -A
+netbird debug for 1m -A -S -U
+```
+
+`-A` (`--anonymize`) replaces sensitive values consistently across every file in the bundle, so
+it stays readable while masking most identifying details. It is not a guarantee of full redaction:
+internal address ranges survive at the default level, and interface names, indexes, MTUs, and
+flags are never anonymized. Read the bundle before posting it publicly. Two levels are
+available:
+
+| Level | How to select | What it masks |
+| --- | --- | --- |
+| `default` | `-A` / `--anonymize`, or `--anonymize-level default` | Public IP addresses, IPv6 ULA addresses, MAC addresses, and domains other than `netbird.io`, `netbird.cloud`, `netbird.selfhosted`, and `netbird.stage`. IPv4 private, CGNAT, and link-local ranges are kept, and interface names are not anonymized |
+| `strict` | `--anonymize-level strict` (implies `-A`) | The above, plus IPv4 private, CGNAT, and link-local ranges, peer names in front of `netbird.cloud`, `netbird.selfhosted`, and `netbird.stage`, and WireGuard public keys. Labels under `netbird.io` are kept, since it only hosts infrastructure |
+
+Use `strict` when internal addressing or peer naming is itself sensitive. Either way, private
+keys and SSH keys are never included, and the packet capture (`capture.pcap`) is left out of
+anonymized bundles because it holds raw decrypted packets.
+
+`-U` (`--upload-bundle`) uploads the bundle and returns a file key you can paste into the thread
+instead of attaching an archive. Retention is controlled by the upload service; check its policy
+before uploading, and configure cleanup for self-hosted deployments.
+
+For more detail, see [troubleshooting client issues](https://docs.netbird.io/help/troubleshooting-client),
+which explains [what a debug bundle contains](https://docs.netbird.io/help/troubleshooting-client#debug-bundle),
+and the [CLI reference](https://docs.netbird.io/get-started/cli#debug-for).
+
+Intermittent problems are still worth reporting. They just need enough detail to investigate:
+trigger, frequency, timing, timestamps, and any related logs.
+
+For a feature request, describe the problem before the solution: what you are trying to
+accomplish, who is affected and how often, why the current behavior or workaround is not enough,
+and what you would like to see instead.
+
+## What happens after you post
+
+Our team, maintainers, or community members may ask for missing details, link related
+threads, merge duplicates, move your post to a better category, or try to reproduce the problem.
+
+Not every discussion becomes an issue. Some are answered in Q&A, some turn out to be
+configuration problems, and some need more information before engineering can act. A
+well-answered discussion is still a useful outcome.
+
+When a report is confirmed and actionable, a maintainer opens a validated issue linked back to
+the discussion, in whichever repository the fix belongs to. You do not need to know which
+repository that is. Routing is part of triage.
+
+## A note on issues
+
+Issues in this repository are maintainer-curated work items. Every open issue is something a
+maintainer or contributor can pick up and act on. Issues opened without a linked validated
+discussion may be closed and redirected here.
+
+Maintainers can still open issues directly for work found internally, such as regressions caught
+during development, planned maintenance, or release blockers.
+
+## Related reading
+
+- [How to use Discussions, Issues, and Pull Requests](https://github.com/netbirdio/netbird/discussions/6075)
+- [Moving to a discussion-first approach](https://github.com/netbirdio/netbird/discussions/6074)
+- [CONTRIBUTING.md](CONTRIBUTING.md) for opening pull requests
+- [CODE_OF_CONDUCT.md](CODE_OF_CONDUCT.md)

agent-network/README.md

@@ -110,12 +110,12 @@ Two roles delegate Agent Network access without account-admin rights:
   read-only users, groups, peers, and account info (needed to build policies).
   Nothing else in the account.
 - **`usage_viewer`** — the regular User baseline plus read on
-  `agent_network.usage` (the aggregated usage and cost overview) and read-only
-  access to the resources the usage filters resolve against: users, groups,
-  peers, and the provider list (connection config redacted — no upstream URLs
-  or operator-supplied header values). No policies, and no account-wide
-  request-level access logs; like any caller, it still reads its own requests
-  through the self-scoped endpoints below.
+  `agent_network.usage` (the aggregated usage and cost overview) and
+  `agent_network.logs` (the account-wide request-level access logs, which can
+  contain captured prompts), and read-only access to the resources those
+  filters resolve against: users, groups, peers, and the provider list
+  (connection config redacted — no upstream URLs or operator-supplied header
+  values). No policies, guardrails, budgets, or settings.
 
 Every authenticated user, regardless of role, can read the caller-scoped
 self-service endpoint `GET /api/agent-network/agent-config` (the endpoint, providers,

client/android/ssh_client.go

@@ -31,6 +31,8 @@ const (
 // PasswordRequiredMarker tells Java to prompt for a password and retry. It is
 // a string because gomobile flattens errors to their message, so a sentinel
 // value would not survive the binding.
+//
+//nolint:gosec // G101 false positive: a sentinel marker, not a credential
 const PasswordRequiredMarker = "netbird-ssh-password-required"
 
 // HostKeyUnknownMarker tells Java to show the fingerprint and, on confirmation,

client/cmd/debug.go

@@ -23,7 +23,10 @@ import (
 	"github.com/netbirdio/netbird/version"
 )
 
-const errCloseConnection = "Failed to close connection: %v"
+const (
+	errCloseConnection = "Failed to close connection: %v"
+	noUpDownFlag       = "no-updown"
+)
 
 var (
 	logFileCount             uint32
@@ -257,13 +260,14 @@ func runForDuration(cmd *cobra.Command, args []string) error {
 	}
 
 	stateWasDown := stat.Status != string(internal.StatusConnected) && stat.Status != string(internal.StatusConnecting)
+	noUpDown, _ := cmd.Flags().GetBool(noUpDownFlag)
 
 	initialLogLevel, err := client.GetLogLevel(cmd.Context(), &proto.GetLogLevelRequest{})
 	if err != nil {
 		return fmt.Errorf("failed to get log level: %v", status.Convert(err).Message())
 	}
 
-	if stateWasDown {
+	if stateWasDown && !noUpDown {
 		if _, err := client.Up(cmd.Context(), &proto.UpRequest{}); err != nil {
 			cmd.PrintErrf("Failed to bring service up: %v\n", status.Convert(err).Message())
 		} else {
@@ -284,34 +288,20 @@ func runForDuration(cmd *cobra.Command, args []string) error {
 	}
 
 	needsRestoreUp := false
-	if _, err := client.Down(cmd.Context(), &proto.DownRequest{}); err != nil {
-		cmd.PrintErrf("Failed to bring service down: %v\n", status.Convert(err).Message())
+	if noUpDown {
+		enableSyncResponsePersistence(cmd, client)
 	} else {
-		needsRestoreUp = !stateWasDown
-		cmd.Println("netbird down")
+		needsRestoreUp = restartDaemon(cmd, client, stateWasDown)
 	}
 
-	time.Sleep(1 * time.Second)
-
-	// Enable sync response persistence before bringing the service up
-	if _, err := client.SetSyncResponsePersistence(cmd.Context(), &proto.SetSyncResponsePersistenceRequest{
-		Enabled: true,
-	}); err != nil {
-		cmd.PrintErrf("Failed to enable sync response persistence: %v\n", status.Convert(err).Message())
-	}
-
-	if _, err := client.Up(cmd.Context(), &proto.UpRequest{}); err != nil {
-		cmd.PrintErrf("Failed to bring service up: %v\n", status.Convert(err).Message())
-	} else {
-		needsRestoreUp = false
-		cmd.Println("netbird up")
-	}
-
-	time.Sleep(3 * time.Second)
-
 	cpuProfilingStarted := false
 	if _, err := client.StartCPUProfile(cmd.Context(), &proto.StartCPUProfileRequest{}); err != nil {
-		cmd.PrintErrf("Failed to start CPU profiling: %v\n", err)
+		if msg := status.Convert(err).Message(); strings.Contains(msg, "already in progress") {
+			cmd.PrintErrln("CPU profiling is already running (started with `netbird debug cpu start`). " +
+				"It is left running and is included in a bundle created after `netbird debug cpu stop`.")
+		} else {
+			cmd.PrintErrf("Failed to start CPU profiling: %v\n", msg)
+		}
 	} else {
 		cpuProfilingStarted = true
 		defer func() {
@@ -401,7 +391,7 @@ func runForDuration(cmd *cobra.Command, args []string) error {
 		}
 	}
 
-	if stateWasDown {
+	if stateWasDown && !noUpDown {
 		if _, err := client.Down(cmd.Context(), &proto.DownRequest{}); err != nil {
 			cmd.PrintErrf("Failed to restore service down state: %v\n", status.Convert(err).Message())
 		} else {
@@ -458,6 +448,47 @@ func setSyncResponsePersistence(cmd *cobra.Command, args []string) error {
 	return nil
 }
 
+// enableSyncResponsePersistence asks the daemon to keep the latest sync
+// response so the bundle carries the network map. With a running daemon only
+// syncs received after the call are kept.
+func enableSyncResponsePersistence(cmd *cobra.Command, client proto.DaemonServiceClient) {
+	if _, err := client.SetSyncResponsePersistence(cmd.Context(), &proto.SetSyncResponsePersistenceRequest{
+		Enabled: true,
+	}); err != nil {
+		cmd.PrintErrf("Failed to enable sync response persistence: %v\n", status.Convert(err).Message())
+	}
+}
+
+// restartDaemon cycles the daemon down and up with sync response persistence
+// enabled so the bundle carries the network map. It reports whether the
+// daemon was left down although it was running before, so the caller can
+// bring it back up.
+func restartDaemon(cmd *cobra.Command, client proto.DaemonServiceClient, stateWasDown bool) bool {
+	needsRestoreUp := false
+	if _, err := client.Down(cmd.Context(), &proto.DownRequest{}); err != nil {
+		cmd.PrintErrf("Failed to bring service down: %v\n", status.Convert(err).Message())
+	} else {
+		needsRestoreUp = !stateWasDown
+		cmd.Println("netbird down")
+	}
+
+	time.Sleep(1 * time.Second)
+
+	// Enable sync response persistence before bringing the service up
+	enableSyncResponsePersistence(cmd, client)
+
+	if _, err := client.Up(cmd.Context(), &proto.UpRequest{}); err != nil {
+		cmd.PrintErrf("Failed to bring service up: %v\n", status.Convert(err).Message())
+	} else {
+		needsRestoreUp = false
+		cmd.Println("netbird up")
+	}
+
+	time.Sleep(3 * time.Second)
+
+	return needsRestoreUp
+}
+
 func waitForDurationOrCancel(ctx context.Context, duration time.Duration, cmd *cobra.Command) error {
 	ticker := time.NewTicker(1 * time.Second)
 	defer ticker.Stop()
@@ -546,4 +577,5 @@ func init() {
 	forCmd.Flags().StringVar(&uploadBundleURLFlag, "upload-bundle-url", types.DefaultBundleURL, "Service URL to get an URL to upload the debug bundle")
 	forCmd.Flags().BoolVar(&uploadBundleInsecureFlag, "upload-bundle-insecure", false, "Allow uploading to an http or untrusted-TLS upload server (self-hosted); requires root")
 	forCmd.Flags().Bool("capture", false, "Capture packets during the debug duration and include in bundle")
+	forCmd.Flags().Bool(noUpDownFlag, false, "Keep the daemon running instead of bringing it down and up before collecting. The bundle only includes the network map if a sync arrives during the run")
 }

client/cmd/debug_cpu.go

@@ -0,0 +1,83 @@
+package cmd
+
+import (
+	"fmt"
+
+	log "github.com/sirupsen/logrus"
+	"github.com/spf13/cobra"
+	"google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/client/proto"
+)
+
+var debugCPUCmd = &cobra.Command{
+	Use:   "cpu",
+	Short: "Profile the daemon's CPU usage",
+	Long: `Starts and stops CPU profiling in the running daemon without restarting it.
+The profile is included in the next debug bundle as cpu.prof.
+
+Profiling is not time limited: it keeps running, and keeps costing CPU, until
+"netbird debug cpu stop" is run.`,
+}
+
+var debugCPUStartCmd = &cobra.Command{
+	Use:     "start",
+	Short:   "Start CPU profiling in the daemon",
+	Example: "  netbird debug cpu start",
+	Args:    cobra.NoArgs,
+	RunE:    debugCPUStart,
+}
+
+var debugCPUStopCmd = &cobra.Command{
+	Use:   "stop",
+	Short: "Stop CPU profiling in the daemon",
+	Long: `Stops CPU profiling. The captured profile stays in the daemon until the next
+debug bundle is created, which includes it as cpu.prof.`,
+	Example: "  netbird debug cpu stop && netbird debug bundle",
+	Args:    cobra.NoArgs,
+	RunE:    debugCPUStop,
+}
+
+func debugCPUStart(cmd *cobra.Command, _ []string) error {
+	conn, err := getClient(cmd)
+	if err != nil {
+		return err
+	}
+	defer func() {
+		if err := conn.Close(); err != nil {
+			log.Errorf(errCloseConnection, err)
+		}
+	}()
+
+	if _, err := proto.NewDaemonServiceClient(conn).StartCPUProfile(cmd.Context(), &proto.StartCPUProfileRequest{}); err != nil {
+		return fmt.Errorf("start CPU profiling: %v", status.Convert(err).Message())
+	}
+
+	cmd.Println("CPU profiling started and runs until stopped. Run `netbird debug cpu stop` and then `netbird debug bundle` to collect it.")
+	return nil
+}
+
+func debugCPUStop(cmd *cobra.Command, _ []string) error {
+	conn, err := getClient(cmd)
+	if err != nil {
+		return err
+	}
+	defer func() {
+		if err := conn.Close(); err != nil {
+			log.Errorf(errCloseConnection, err)
+		}
+	}()
+
+	if _, err := proto.NewDaemonServiceClient(conn).StopCPUProfile(cmd.Context(), &proto.StopCPUProfileRequest{}); err != nil {
+		return fmt.Errorf("stop CPU profiling: %v", status.Convert(err).Message())
+	}
+
+	cmd.Println("CPU profiling stopped. Run `netbird debug bundle` to include cpu.prof.")
+	return nil
+}
+
+func init() {
+	debugCPUCmd.AddCommand(debugCPUStartCmd)
+	debugCPUCmd.AddCommand(debugCPUStopCmd)
+	debugCmd.AddCommand(debugCPUCmd)
+}

client/cmd/debug_cpu_test.go

@@ -0,0 +1,164 @@
+package cmd
+
+import (
+	"bytes"
+	"context"
+	"os/user"
+	"strings"
+	"testing"
+
+	"github.com/spf13/cobra"
+	"github.com/spf13/pflag"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/profilemanager"
+)
+
+// startDebugTestDaemon starts an in-process daemon with an isolated profile
+// directory and returns the address the CLI should dial.
+func startDebugTestDaemon(t *testing.T) string {
+	t.Helper()
+
+	tempDir := t.TempDir()
+	origDefaultProfileDir := profilemanager.DefaultConfigPathDir
+	origActiveProfileStatePath := profilemanager.ActiveProfileStatePath
+	origConfigDirOverride := profilemanager.ConfigDirOverride
+	origDaemonAddr := daemonAddr
+	t.Cleanup(func() {
+		profilemanager.DefaultConfigPathDir = origDefaultProfileDir
+		profilemanager.ActiveProfileStatePath = origActiveProfileStatePath
+		profilemanager.ConfigDirOverride = origConfigDirOverride
+		daemonAddr = origDaemonAddr
+	})
+
+	profilemanager.DefaultConfigPathDir = tempDir
+	profilemanager.ActiveProfileStatePath = tempDir + "/active_profile.json"
+	profilemanager.ConfigDirOverride = tempDir
+
+	currUser, err := user.Current()
+	require.NoError(t, err)
+	sm := profilemanager.ServiceManager{}
+	created, err := sm.AddProfile("test1", currUser.Username)
+	require.NoError(t, err)
+	require.NoError(t, sm.SetActiveProfileState(&profilemanager.ActiveProfileState{
+		ID:       created.ID,
+		Username: currUser.Username,
+	}))
+
+	ctx, cancel := context.WithCancel(internal.CtxInitState(context.Background()))
+	srv, lis := startClientDaemon(t, ctx, "", tempDir+"/config.json")
+	t.Cleanup(func() {
+		cancel()
+		srv.Stop()
+	})
+
+	return "tcp://" + lis.Addr().String()
+}
+
+// runDebugCmd runs `netbird debug <args>` against the daemon at addr and
+// returns everything the command printed.
+func runDebugCmd(addr string, args ...string) (string, error) {
+	daemonAddr = addr
+	var out bytes.Buffer
+	rootCmd.SetOut(&out)
+	rootCmd.SetErr(&out)
+	rootCmd.SetArgs(append(append([]string{"debug"}, args...), "--daemon-addr", addr, "--log-file", ""))
+	err := rootCmd.Execute()
+	rootCmd.SetOut(nil)
+	rootCmd.SetErr(nil)
+	rootCmd.SetArgs(nil)
+	resetFlags(rootCmd)
+	return out.String(), err
+}
+
+// resetFlags puts every flag of the command and its subcommands back to its
+// default so a value parsed in one run does not leak into the next in-process
+// execution.
+func resetFlags(cmd *cobra.Command) {
+	reset := func(f *pflag.Flag) {
+		// Set appends to a slice flag and would parse the "[a,b]" default
+		// text as elements, so slices are replaced instead.
+		if sv, ok := f.Value.(pflag.SliceValue); ok {
+			var def []string
+			if trimmed := strings.Trim(f.DefValue, "[]"); trimmed != "" {
+				def = strings.Split(trimmed, ",")
+			}
+			_ = sv.Replace(def)
+		} else {
+			_ = f.Value.Set(f.DefValue)
+		}
+		f.Changed = false
+	}
+	cmd.Flags().VisitAll(reset)
+	cmd.PersistentFlags().VisitAll(reset)
+	// Commands pin their writers to the buffer of the run that first used
+	// them, so a later run would print into the old buffer.
+	cmd.SetOut(nil)
+	cmd.SetErr(nil)
+	for _, sub := range cmd.Commands() {
+		resetFlags(sub)
+	}
+}
+
+// TestResetFlagsSliceDefault guards against Set("[]") on slice flags, which
+// stores a literal "[]" element instead of the empty default.
+func TestResetFlagsSliceDefault(t *testing.T) {
+	cmd := &cobra.Command{Use: "x"}
+	var env, withDefault []string
+	cmd.Flags().StringSliceVar(&env, "env", nil, "")
+	cmd.Flags().StringSliceVar(&withDefault, "names", []string{"a", "b"}, "")
+	require.NoError(t, cmd.Flags().Parse([]string{"--env", "K=V", "--names", "c"}))
+
+	resetFlags(cmd)
+
+	assert.Empty(t, env, "slice flag with no default must reset to empty")
+	assert.Equal(t, []string{"a", "b"}, withDefault, "slice flag must reset to its default")
+}
+
+func TestDebugCPUStartStop(t *testing.T) {
+	addr := startDebugTestDaemon(t)
+
+	run := func(args ...string) error {
+		_, err := runDebugCmd(addr, append([]string{"cpu"}, args...)...)
+		return err
+	}
+
+	require.Error(t, run("stop"), "stop without a running profile must fail")
+	require.NoError(t, run("start"))
+	assert.Error(t, run("start"), "second start must be rejected while profiling")
+	require.NoError(t, run("stop"))
+	assert.Error(t, run("stop"), "second stop must be rejected")
+	assert.NoError(t, run("start"), "profiling can be started again after a stop")
+	assert.NoError(t, run("stop"))
+}
+
+// TestDebugForKeepsRunningCPUProfile covers `debug for` started while a
+// profile from `debug cpu start` is running: it must say so, leave the
+// profile alone, and still create the bundle.
+func TestDebugForKeepsRunningCPUProfile(t *testing.T) {
+	addr := startDebugTestDaemon(t)
+
+	_, err := runDebugCmd(addr, "cpu", "start")
+	require.NoError(t, err)
+
+	out, err := runDebugCmd(addr, "for", "1s", "-S=false", "--no-updown")
+	require.NoError(t, err, "output: %s", out)
+	assert.Contains(t, out, "CPU profiling is already running", "the conflict must be explained")
+	assert.NotContains(t, out, "rpc error", "the raw RPC error must not reach the user")
+	assert.Contains(t, out, "Local file:", "the bundle must still be created")
+
+	_, err = runDebugCmd(addr, "cpu", "stop")
+	assert.NoError(t, err, "the profile started by the user must still be running")
+}
+
+func TestDebugForNoUpDown(t *testing.T) {
+	addr := startDebugTestDaemon(t)
+
+	out, err := runDebugCmd(addr, "for", "1s", "-S=false", "--no-updown")
+	require.NoError(t, err, "output: %s", out)
+	assert.NotContains(t, out, "netbird down", "--no-updown must not bring the daemon down")
+	assert.NotContains(t, out, "netbird up", "--no-updown must not bring the daemon up")
+	assert.Contains(t, out, "Local file:", "the bundle must still be created")
+}

client/cmd/service.go

@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"net/http"
 	"runtime"
+	"slices"
 	"strings"
 	"sync"
 
@@ -25,6 +26,30 @@ var serviceCmd = &cobra.Command{
 
 const defaultJSONSocket = "unix:///var/run/netbird-http.sock"
 
+// forbiddenServiceEnvVars are the environment variables the service is never
+// registered with, keyed in upper case since these are Windows names. Each one
+// decides where the daemon resolves something it then uses with the privileges
+// of the account it runs under — LocalSystem on Windows, root elsewhere: the
+// executables it runs (PATH, PATHEXT, COMSPEC, SystemRoot, windir) or the
+// directory it writes temporary files in (TEMP, TMP). The daemon needs none of
+// them, and the utilities it shells out to are resolved by absolute path.
+var forbiddenServiceEnvVars = map[string]struct{}{
+	"PATH":       {},
+	"PATHEXT":    {},
+	"SYSTEMROOT": {},
+	"WINDIR":     {},
+	"COMSPEC":    {},
+	"TEMP":       {},
+	"TMP":        {},
+}
+
+// forbiddenServiceEnvPrefixes are the dynamic-loader families, refused whole
+// rather than by name: LD_PRELOAD, DYLD_INSERT_LIBRARIES and their siblings all
+// reach the loader of the process, the set differs per platform and libc, and
+// new members arrive with new OS releases. Listing them one by one is a list
+// that is wrong the moment it is written.
+var forbiddenServiceEnvPrefixes = []string{"LD_", "DYLD_"}
+
 var (
 	serviceName      string
 	serviceEnvVars   []string
@@ -127,8 +152,33 @@ func parseServiceEnvVars(envVars []string) (map[string]string, error) {
 			return nil, fmt.Errorf("empty environment variable key in: %s", env)
 		}
 
+		if isForbiddenServiceEnvVar(key) {
+			return nil, fmt.Errorf("environment variable %s cannot be set on the service: it decides where the service resolves the executables, libraries or temporary files it uses", key)
+		}
+
 		envMap[key] = value
 	}
 
 	return envMap, nil
 }
+
+// isForbiddenServiceEnvVar reports whether name is one the service must not be
+// registered with.
+//
+// The names are matched case-insensitively only on Windows, where they are the
+// same variable however they are spelled. Elsewhere the environment is
+// case-sensitive, so Path and PATH are two different variables and only the
+// exact spelling is the one the loader reads.
+func isForbiddenServiceEnvVar(name string) bool {
+	if runtime.GOOS == "windows" {
+		name = strings.ToUpper(name)
+	}
+
+	if _, forbidden := forbiddenServiceEnvVars[name]; forbidden {
+		return true
+	}
+
+	return slices.ContainsFunc(forbiddenServiceEnvPrefixes, func(prefix string) bool {
+		return strings.HasPrefix(name, prefix)
+	})
+}

client/cmd/service_params.go

@@ -14,6 +14,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/configs"
 	"github.com/netbirdio/netbird/client/internal/daemonaddr"
+	"github.com/netbirdio/netbird/client/internal/elevate"
 	"github.com/netbirdio/netbird/util"
 )
 
@@ -43,10 +44,33 @@ func serviceParamsPath() string {
 
 // loadServiceParams reads saved service parameters from disk.
 // Returns nil with no error if the file does not exist.
+//
+// The file is read by an elevated install and decides the arguments and the
+// environment of the service it then registers, so it is used only when its
+// ownership and permissions are the ones saveServiceParams leaves behind. That
+// restricted ACL is applied when the file is written, which is not necessarily
+// before it is first read, so this is checked rather than assumed. A file that
+// fails the check is treated as absent, and the install proceeds with its
+// defaults.
 func loadServiceParams() (*serviceParams, error) {
 	path := serviceParamsPath()
 
-	data, err := os.ReadFile(path)
+	// Resolve links first so the checks apply to the file that is actually read.
+	// Since the check covers every directory above it as well, nobody who fails
+	// it can swap the file between here and the read below.
+	resolved, err := filepath.EvalSymlinks(path)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return nil, nil //nolint:nilnil
+		}
+		return nil, fmt.Errorf("resolve service params %s: %w", path, err)
+	}
+
+	if err := elevate.CheckOnlyOwnerWritable(resolved); err != nil {
+		return nil, fmt.Errorf("refusing to read service params from %s: %w", resolved, err)
+	}
+
+	data, err := os.ReadFile(resolved)
 	if err != nil {
 		if os.IsNotExist(err) {
 			return nil, nil //nolint:nilnil
@@ -182,10 +206,16 @@ func applyServiceParams(cmd *cobra.Command, params *serviceParams) {
 // If --service-env was explicitly set to empty, all saved env vars are cleared.
 // If --service-env was not set, saved env vars are used entirely.
 func applyServiceEnvParams(cmd *cobra.Command, params *serviceParams) {
+	// A forbidden name explicitly passed on the command line is an error the
+	// operator is told about, but one restored from a file written by an older
+	// version is dropped: an install that refuses to run would leave the host
+	// without a daemon over a variable nobody is asking for any more.
+	saved := dropForbiddenServiceEnvVars(cmd, params.ServiceEnvVars)
+
 	if !cmd.Flags().Changed("service-env") {
-		if len(params.ServiceEnvVars) > 0 {
+		if len(saved) > 0 {
 			// No explicit env vars: rebuild serviceEnvVars from saved params.
-			serviceEnvVars = envMapToSlice(params.ServiceEnvVars)
+			serviceEnvVars = envMapToSlice(saved)
 		}
 		return
 	}
@@ -204,13 +234,13 @@ func applyServiceEnvParams(cmd *cobra.Command, params *serviceParams) {
 		return
 	}
 
-	if len(params.ServiceEnvVars) == 0 {
+	if len(saved) == 0 {
 		return
 	}
 
 	// Merge saved values underneath explicit ones.
-	merged := make(map[string]string, len(params.ServiceEnvVars)+len(explicit))
-	maps.Copy(merged, params.ServiceEnvVars)
+	merged := make(map[string]string, len(saved)+len(explicit))
+	maps.Copy(merged, saved)
 	maps.Copy(merged, explicit) // explicit wins on conflict
 	serviceEnvVars = envMapToSlice(merged)
 }
@@ -233,6 +263,20 @@ var resetParamsCmd = &cobra.Command{
 	},
 }
 
+// dropForbiddenServiceEnvVars returns the saved entries that may still be
+// registered on the service, reporting every one it leaves behind.
+func dropForbiddenServiceEnvVars(cmd *cobra.Command, saved map[string]string) map[string]string {
+	kept := make(map[string]string, len(saved))
+	for key, value := range saved {
+		if isForbiddenServiceEnvVar(key) {
+			cmd.PrintErrf("Warning: ignoring saved service environment variable %s: it decides where the service resolves the executables, libraries or temporary files it uses\n", key)
+			continue
+		}
+		kept[key] = value
+	}
+	return kept
+}
+
 // envMapToSlice converts a map of env vars to a KEY=VALUE slice.
 func envMapToSlice(m map[string]string) []string {
 	s := make([]string, 0, len(m))

client/cmd/service_params_test.go

@@ -9,6 +9,7 @@ import (
 	"go/token"
 	"os"
 	"path/filepath"
+	"runtime"
 	"strings"
 	"testing"
 
@@ -353,6 +354,59 @@ func TestApplyServiceEnvParams_NotChanged(t *testing.T) {
 	assert.Equal(t, map[string]string{"FROM_SAVED": "val"}, result)
 }
 
+func TestParseServiceEnvVars_RejectsForbiddenNames(t *testing.T) {
+	for _, env := range []string{"PATH=C:\\somewhere", "LD_PRELOAD=/tmp/lib.so", "DYLD_FALLBACK_LIBRARY_PATH=/tmp"} {
+		_, err := parseServiceEnvVars([]string{"KEEP=me", env})
+		require.Errorf(t, err, "%s selects what the service resolves and must be refused", env)
+	}
+}
+
+func TestIsForbiddenServiceEnvVar(t *testing.T) {
+	// The loader families are matched by prefix, so a name nobody has heard of
+	// yet is refused too.
+	for _, name := range []string{
+		"PATH", "PATHEXT", "COMSPEC", "SYSTEMROOT", "WINDIR", "TEMP", "TMP",
+		"LD_PRELOAD", "LD_AUDIT", "DYLD_INSERT_LIBRARIES", "DYLD_FALLBACK_FRAMEWORK_PATH",
+	} {
+		assert.Truef(t, isForbiddenServiceEnvVar(name), "%s must be refused", name)
+	}
+
+	// The prefix must not swallow names that merely start with the same letters.
+	for _, name := range []string{"NB_LOG_LEVEL", "NB_WG_DEBUG", "HTTPS_PROXY", "LDAP_URL", "DYLDX"} {
+		assert.Falsef(t, isForbiddenServiceEnvVar(name), "%s has no reason to be refused", name)
+	}
+
+	// On Windows a variable is the same one however it is spelled; elsewhere
+	// Path and PATH are two variables and only the exact one is read.
+	if runtime.GOOS == "windows" {
+		assert.True(t, isForbiddenServiceEnvVar("Path"))
+		assert.True(t, isForbiddenServiceEnvVar("ld_preload"))
+	} else {
+		assert.False(t, isForbiddenServiceEnvVar("Path"))
+		assert.False(t, isForbiddenServiceEnvVar("ld_preload"))
+	}
+}
+
+func TestApplyServiceEnvParams_DropsForbiddenSavedNames(t *testing.T) {
+	origServiceEnvVars := serviceEnvVars
+	t.Cleanup(func() { serviceEnvVars = origServiceEnvVars })
+
+	serviceEnvVars = nil
+
+	cmd := &cobra.Command{}
+	cmd.Flags().StringSlice("service-env", nil, "")
+
+	saved := &serviceParams{
+		ServiceEnvVars: map[string]string{"PATH": "C:\\attacker", "NB_LOG_FORMAT": "json"},
+	}
+
+	applyServiceEnvParams(cmd, saved)
+
+	result, err := parseServiceEnvVars(serviceEnvVars)
+	require.NoError(t, err, "a saved PATH must be dropped rather than fail the install")
+	assert.Equal(t, map[string]string{"NB_LOG_FORMAT": "json"}, result)
+}
+
 func TestApplyServiceEnvParams_ExplicitEmptyClears(t *testing.T) {
 	origServiceEnvVars := serviceEnvVars
 	t.Cleanup(func() { serviceEnvVars = origServiceEnvVars })

client/cmd/service_params_trust_test.go

@@ -0,0 +1,57 @@
+//go:build !windows && !ios && !android
+
+package cmd
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/client/configs"
+)
+
+// The Windows equivalent of this is the ACL check in
+// elevate.CheckOnlyOwnerWritable, covered by that package's own tests; here the
+// point is that loadServiceParams asks the question at all.
+func TestLoadServiceParams_RefusesWorldWritableFile(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	original := configs.StateDir
+	t.Cleanup(func() { configs.StateDir = original })
+	configs.StateDir = tmpDir
+
+	path := filepath.Join(tmpDir, serviceParamsFile)
+	require.NoError(t, os.WriteFile(path, []byte(`{"log_level":"debug"}`), 0o666))
+	// WriteFile is subject to the umask, so set the bits that matter explicitly.
+	require.NoError(t, os.Chmod(path, 0o666))
+
+	params, err := loadServiceParams()
+	require.Error(t, err, "a service.json anyone can rewrite must not be trusted")
+	assert.Nil(t, params)
+
+	require.NoError(t, os.Chmod(path, 0o600))
+	params, err = loadServiceParams()
+	require.NoError(t, err)
+	require.NotNil(t, params)
+	assert.Equal(t, "debug", params.LogLevel)
+}
+
+func TestLoadServiceParams_RefusesWorldWritableDirectory(t *testing.T) {
+	tmpDir := t.TempDir()
+	stateDir := filepath.Join(tmpDir, "state")
+	require.NoError(t, os.Mkdir(stateDir, 0o777))
+	require.NoError(t, os.Chmod(stateDir, 0o777))
+
+	original := configs.StateDir
+	t.Cleanup(func() { configs.StateDir = original })
+	configs.StateDir = stateDir
+
+	require.NoError(t, os.WriteFile(filepath.Join(stateDir, serviceParamsFile), []byte(`{}`), 0o600))
+
+	params, err := loadServiceParams()
+	require.Error(t, err, "a service.json in a directory anyone can replace entries in must not be trusted")
+	assert.Nil(t, params)
+}

client/firewall/iptables/family_linux.go

@@ -24,6 +24,7 @@ const (
 	tableFilter = "filter"
 	tableNat    = "nat"
 	tableMangle = "mangle"
+	tableRaw    = "raw"
 
 	// chainACLInput is the peer ACL chain that holds installed
 	// peer-filtering rules.
@@ -34,6 +35,7 @@ const (
 	mangleForwardKey chainKey = "MANGLE-FORWARD"
 
 	chainInput       = "INPUT"
+	chainOutput      = "OUTPUT"
 	chainPostrouting = "POSTROUTING"
 	chainPrerouting  = "PREROUTING"
 	chainForward     = "FORWARD"

client/firewall/iptables/manager_linux.go

@@ -25,9 +25,8 @@ type Manager struct {
 
 	wgIface iFaceMapper
 
-	ipv4Client   *iptables.IPTables
-	family4      *family
-	rawSupported bool
+	ipv4Client *iptables.IPTables
+	family4    *family
 
 	// IPv6 counterparts, nil when no v6 overlay
 	ipv6Client *iptables.IPTables
@@ -108,10 +107,6 @@ func (m *Manager) Init(stateManager *statemanager.Manager) error {
 		return err
 	}
 
-	if err := m.initNoTrackChain(); err != nil {
-		log.Warnf("raw table not available, notrack rules will be disabled: %v", err)
-	}
-
 	// Trust after all fatal init steps so a later failure doesn't leave the
 	// interface in firewalld's trusted zone without a corresponding Close.
 	if err := firewalld.TrustInterface(m.wgIface.Name()); err != nil {
@@ -285,10 +280,6 @@ func (m *Manager) Close(stateManager *statemanager.Manager) error {
 
 	var merr *multierror.Error
 
-	if err := m.cleanupNoTrackChain(); err != nil {
-		merr = multierror.Append(merr, fmt.Errorf("cleanup notrack chain: %w", err))
-	}
-
 	if m.hasIPv6() {
 		if err := m.family6.Reset(); err != nil {
 			merr = multierror.Append(merr, fmt.Errorf("reset v6 family: %w", err))
@@ -440,134 +431,6 @@ func (m *Manager) RemoveOutputDNAT(localAddr netip.Addr, protocol firewall.Proto
 	return m.family4.RemoveOutputDNAT(localAddr, protocol, originalPort, translatedPort)
 }
 
-const (
-	chainNameRaw = "NETBIRD-RAW"
-	chainOutput  = "OUTPUT"
-	tableRaw     = "raw"
-)
-
-// SetupEBPFProxyNoTrack creates notrack rules for eBPF proxy loopback traffic.
-// This prevents conntrack from tracking WireGuard proxy traffic on loopback, which
-// can interfere with MASQUERADE rules (e.g., from container runtimes like Podman/netavark).
-//
-// Traffic flows that need NOTRACK:
-//
-//  1. Egress: WireGuard -> fake endpoint (before eBPF rewrite)
-//     src=127.0.0.1:wgPort -> dst=127.0.0.1:fakePort
-//     Matched by: sport=wgPort
-//
-//  2. Egress: Proxy -> WireGuard (via raw socket)
-//     src=127.0.0.1:fakePort -> dst=127.0.0.1:wgPort
-//     Matched by: dport=wgPort
-//
-//  3. Ingress: Packets to WireGuard
-//     dst=127.0.0.1:wgPort
-//     Matched by: dport=wgPort
-//
-//  4. Ingress: Packets to proxy (after eBPF rewrite)
-//     dst=127.0.0.1:proxyPort
-//     Matched by: dport=proxyPort
-//
-// Rules are cleaned up when the firewall manager is closed.
-func (m *Manager) SetupEBPFProxyNoTrack(proxyPort, wgPort uint16) error {
-	m.mutex.Lock()
-	defer m.mutex.Unlock()
-
-	if !m.rawSupported {
-		return fmt.Errorf("raw table not available")
-	}
-
-	wgPortStr := fmt.Sprintf("%d", wgPort)
-	proxyPortStr := fmt.Sprintf("%d", proxyPort)
-
-	// Egress rules: match outgoing loopback UDP packets
-	outputRuleSport := []string{"-o", "lo", "-s", "127.0.0.1", "-d", "127.0.0.1", "-p", "udp", "--sport", wgPortStr, "-j", "NOTRACK"}
-	if err := m.ipv4Client.AppendUnique(tableRaw, chainNameRaw, outputRuleSport...); err != nil {
-		return fmt.Errorf("add output sport notrack rule: %w", err)
-	}
-
-	outputRuleDport := []string{"-o", "lo", "-s", "127.0.0.1", "-d", "127.0.0.1", "-p", "udp", "--dport", wgPortStr, "-j", "NOTRACK"}
-	if err := m.ipv4Client.AppendUnique(tableRaw, chainNameRaw, outputRuleDport...); err != nil {
-		return fmt.Errorf("add output dport notrack rule: %w", err)
-	}
-
-	// Ingress rules: match incoming loopback UDP packets
-	preroutingRuleWg := []string{"-i", "lo", "-s", "127.0.0.1", "-d", "127.0.0.1", "-p", "udp", "--dport", wgPortStr, "-j", "NOTRACK"}
-	if err := m.ipv4Client.AppendUnique(tableRaw, chainNameRaw, preroutingRuleWg...); err != nil {
-		return fmt.Errorf("add prerouting wg notrack rule: %w", err)
-	}
-
-	preroutingRuleProxy := []string{"-i", "lo", "-s", "127.0.0.1", "-d", "127.0.0.1", "-p", "udp", "--dport", proxyPortStr, "-j", "NOTRACK"}
-	if err := m.ipv4Client.AppendUnique(tableRaw, chainNameRaw, preroutingRuleProxy...); err != nil {
-		return fmt.Errorf("add prerouting proxy notrack rule: %w", err)
-	}
-
-	log.Debugf("set up ebpf proxy notrack rules for ports %d,%d", proxyPort, wgPort)
-	return nil
-}
-
-func (m *Manager) initNoTrackChain() error {
-	if err := m.cleanupNoTrackChain(); err != nil {
-		log.Debugf("cleanup notrack chain: %v", err)
-	}
-
-	if err := m.ipv4Client.NewChain(tableRaw, chainNameRaw); err != nil {
-		return fmt.Errorf("create chain: %w", err)
-	}
-
-	jumpRule := []string{"-j", chainNameRaw}
-
-	if err := m.ipv4Client.InsertUnique(tableRaw, chainOutput, 1, jumpRule...); err != nil {
-		if delErr := m.ipv4Client.DeleteChain(tableRaw, chainNameRaw); delErr != nil {
-			log.Debugf("delete orphan chain: %v", delErr)
-		}
-		return fmt.Errorf("add output jump rule: %w", err)
-	}
-
-	if err := m.ipv4Client.InsertUnique(tableRaw, chainPrerouting, 1, jumpRule...); err != nil {
-		if delErr := m.ipv4Client.DeleteIfExists(tableRaw, chainOutput, jumpRule...); delErr != nil {
-			log.Debugf("delete output jump rule: %v", delErr)
-		}
-		if delErr := m.ipv4Client.DeleteChain(tableRaw, chainNameRaw); delErr != nil {
-			log.Debugf("delete orphan chain: %v", delErr)
-		}
-		return fmt.Errorf("add prerouting jump rule: %w", err)
-	}
-
-	m.rawSupported = true
-	return nil
-}
-
-func (m *Manager) cleanupNoTrackChain() error {
-	exists, err := m.ipv4Client.ChainExists(tableRaw, chainNameRaw)
-	if err != nil {
-		if !m.rawSupported {
-			return nil
-		}
-		return fmt.Errorf("check chain exists: %w", err)
-	}
-	if !exists {
-		return nil
-	}
-
-	jumpRule := []string{"-j", chainNameRaw}
-
-	if err := m.ipv4Client.DeleteIfExists(tableRaw, chainOutput, jumpRule...); err != nil {
-		return fmt.Errorf("remove output jump rule: %w", err)
-	}
-
-	if err := m.ipv4Client.DeleteIfExists(tableRaw, chainPrerouting, jumpRule...); err != nil {
-		return fmt.Errorf("remove prerouting jump rule: %w", err)
-	}
-
-	if err := m.ipv4Client.ClearAndDeleteChain(tableRaw, chainNameRaw); err != nil {
-		return fmt.Errorf("clear and delete chain: %w", err)
-	}
-
-	m.rawSupported = false
-	return nil
-}
-
 func getConntrackEstablished() []string {
 	return []string{"-m", "conntrack", "--ctstate", "RELATED,ESTABLISHED", "-j", "ACCEPT"}
 }

client/firewall/manager/firewall.go

@@ -192,10 +192,6 @@ type Manager interface {
 
 	// RemoveOutputDNAT removes an OUTPUT chain DNAT rule.
 	RemoveOutputDNAT(localAddr netip.Addr, protocol Protocol, originalPort, translatedPort uint16) error
-
-	// SetupEBPFProxyNoTrack creates static notrack rules for eBPF proxy loopback traffic.
-	// This prevents conntrack from interfering with WireGuard proxy communication.
-	SetupEBPFProxyNoTrack(proxyPort, wgPort uint16) error
 }
 
 // GenKey builds the rule id for this pair from the given format.

client/firewall/nftables/manager_linux.go

@@ -12,7 +12,6 @@ import (
 	"github.com/google/nftables/expr"
 	"github.com/hashicorp/go-multierror"
 	log "github.com/sirupsen/logrus"
-	"golang.org/x/sys/unix"
 
 	nberrors "github.com/netbirdio/netbird/client/errors"
 	firewall "github.com/netbirdio/netbird/client/firewall/manager"
@@ -55,9 +54,6 @@ type Manager struct {
 	// IPv6 counterpart, nil when no v6 overlay.
 	family6 *family
 
-	notrackOutputChain     *nftables.Chain
-	notrackPreroutingChain *nftables.Chain
-
 	extMonitor *externalChainMonitor
 }
 
@@ -170,10 +166,6 @@ func (m *Manager) initFirewall() (err error) {
 		}
 	}
 
-	if err := m.initNoTrackChains(workTable); err != nil {
-		log.Warnf("raw priority chains not available, notrack rules will be disabled: %v", err)
-	}
-
 	return nil
 }
 
@@ -455,10 +447,6 @@ func (m *Manager) Flush() error {
 		}
 	}
 
-	if err := m.refreshNoTrackChains(); err != nil {
-		log.Errorf("failed to refresh notrack chains: %v", err)
-	}
-
 	return nil
 }
 
@@ -571,176 +559,6 @@ func (m *Manager) RemoveOutputDNAT(localAddr netip.Addr, protocol firewall.Proto
 	return m.family4.RemoveOutputDNAT(localAddr, protocol, originalPort, translatedPort)
 }
 
-const (
-	chainNameRawOutput     = "netbird-raw-out"
-	chainNameRawPrerouting = "netbird-raw-pre"
-)
-
-// SetupEBPFProxyNoTrack creates notrack rules for eBPF proxy loopback traffic.
-// This prevents conntrack from tracking WireGuard proxy traffic on loopback, which
-// can interfere with MASQUERADE rules (e.g., from container runtimes like Podman/netavark).
-//
-// Traffic flows that need NOTRACK:
-//
-//  1. Egress: WireGuard -> fake endpoint (before eBPF rewrite)
-//     src=127.0.0.1:wgPort -> dst=127.0.0.1:fakePort
-//     Matched by: sport=wgPort
-//
-//  2. Egress: Proxy -> WireGuard (via raw socket)
-//     src=127.0.0.1:fakePort -> dst=127.0.0.1:wgPort
-//     Matched by: dport=wgPort
-//
-//  3. Ingress: Packets to WireGuard
-//     dst=127.0.0.1:wgPort
-//     Matched by: dport=wgPort
-//
-//  4. Ingress: Packets to proxy (after eBPF rewrite)
-//     dst=127.0.0.1:proxyPort
-//     Matched by: dport=proxyPort
-//
-// Rules are cleaned up when the firewall manager is closed.
-func (m *Manager) SetupEBPFProxyNoTrack(proxyPort, wgPort uint16) error {
-	m.mutex.Lock()
-	defer m.mutex.Unlock()
-
-	if m.notrackOutputChain == nil || m.notrackPreroutingChain == nil {
-		return fmt.Errorf("notrack chains not initialized")
-	}
-
-	proxyPortBytes := binaryutil.BigEndian.PutUint16(proxyPort)
-	wgPortBytes := binaryutil.BigEndian.PutUint16(wgPort)
-	loopback := []byte{127, 0, 0, 1}
-
-	// Egress rules: match outgoing loopback UDP packets
-	m.rConn.AddRule(&nftables.Rule{
-		Table: m.notrackOutputChain.Table,
-		Chain: m.notrackOutputChain,
-		Exprs: []expr.Any{
-			&expr.Meta{Key: expr.MetaKeyOIFNAME, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: ifname("lo")},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 12, Len: 4}, // saddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 16, Len: 4}, // daddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Meta{Key: expr.MetaKeyL4PROTO, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: []byte{unix.IPPROTO_UDP}},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseTransportHeader, Offset: 0, Len: 2},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: wgPortBytes}, // sport=wgPort
-			&expr.Counter{},
-			&expr.Notrack{},
-		},
-	})
-	m.rConn.AddRule(&nftables.Rule{
-		Table: m.notrackOutputChain.Table,
-		Chain: m.notrackOutputChain,
-		Exprs: []expr.Any{
-			&expr.Meta{Key: expr.MetaKeyOIFNAME, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: ifname("lo")},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 12, Len: 4}, // saddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 16, Len: 4}, // daddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Meta{Key: expr.MetaKeyL4PROTO, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: []byte{unix.IPPROTO_UDP}},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseTransportHeader, Offset: 2, Len: 2},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: wgPortBytes}, // dport=wgPort
-			&expr.Counter{},
-			&expr.Notrack{},
-		},
-	})
-
-	// Ingress rules: match incoming loopback UDP packets
-	m.rConn.AddRule(&nftables.Rule{
-		Table: m.notrackPreroutingChain.Table,
-		Chain: m.notrackPreroutingChain,
-		Exprs: []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: 12, Len: 4}, // saddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 16, Len: 4}, // daddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Meta{Key: expr.MetaKeyL4PROTO, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: []byte{unix.IPPROTO_UDP}},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseTransportHeader, Offset: 2, Len: 2},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: wgPortBytes}, // dport=wgPort
-			&expr.Counter{},
-			&expr.Notrack{},
-		},
-	})
-	m.rConn.AddRule(&nftables.Rule{
-		Table: m.notrackPreroutingChain.Table,
-		Chain: m.notrackPreroutingChain,
-		Exprs: []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: 12, Len: 4}, // saddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseNetworkHeader, Offset: 16, Len: 4}, // daddr
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: loopback},
-			&expr.Meta{Key: expr.MetaKeyL4PROTO, Register: 1},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: []byte{unix.IPPROTO_UDP}},
-			&expr.Payload{DestRegister: 1, Base: expr.PayloadBaseTransportHeader, Offset: 2, Len: 2},
-			&expr.Cmp{Op: expr.CmpOpEq, Register: 1, Data: proxyPortBytes}, // dport=proxyPort
-			&expr.Counter{},
-			&expr.Notrack{},
-		},
-	})
-
-	if err := m.rConn.Flush(); err != nil {
-		return fmt.Errorf("flush notrack rules: %w", err)
-	}
-
-	log.Debugf("set up ebpf proxy notrack rules for ports %d,%d", proxyPort, wgPort)
-	return nil
-}
-
-func (m *Manager) initNoTrackChains(table *nftables.Table) error {
-	m.notrackOutputChain = m.rConn.AddChain(&nftables.Chain{
-		Name:     chainNameRawOutput,
-		Table:    table,
-		Type:     nftables.ChainTypeFilter,
-		Hooknum:  nftables.ChainHookOutput,
-		Priority: nftables.ChainPriorityRaw,
-	})
-
-	m.notrackPreroutingChain = m.rConn.AddChain(&nftables.Chain{
-		Name:     chainNameRawPrerouting,
-		Table:    table,
-		Type:     nftables.ChainTypeFilter,
-		Hooknum:  nftables.ChainHookPrerouting,
-		Priority: nftables.ChainPriorityRaw,
-	})
-
-	if err := m.rConn.Flush(); err != nil {
-		return fmt.Errorf("flush chain creation: %w", err)
-	}
-
-	return nil
-}
-
-func (m *Manager) refreshNoTrackChains() error {
-	chains, err := m.rConn.ListChainsOfTableFamily(nftables.TableFamilyIPv4)
-	if err != nil {
-		return fmt.Errorf("list chains: %w", err)
-	}
-
-	tableName := getTableName()
-	for _, c := range chains {
-		if c.Table.Name != tableName {
-			continue
-		}
-		switch c.Name {
-		case chainNameRawOutput:
-			m.notrackOutputChain = c
-		case chainNameRawPrerouting:
-			m.notrackPreroutingChain = c
-		}
-	}
-
-	return nil
-}
-
 func (m *Manager) createWorkTable() (*nftables.Table, error) {
 	return m.createWorkTableFamily(nftables.TableFamilyIPv4)
 }

client/firewall/nftables/routing_linux.go

@@ -192,7 +192,7 @@ func (r *family) addPostroutingRules() {
 			Data:     binaryutil.NativeEndian.PutUint32(nbnet.PreroutingFwmarkMasquerade),
 		},
 
-		// We need to exclude the loopback interface as this changes the ebpf proxy port
+		// We need to exclude the loopback interface as this changes the wg proxy port
 		&expr.Meta{
 			Key:      expr.MetaKeyOIFNAME,
 			Register: 1,

client/firewall/uspfilter/filter.go

@@ -879,12 +879,6 @@ func (m *Manager) resetState() {
 	}
 }
 
-// SetupEBPFProxyNoTrack is not supported by the userspace firewall: eBPF isn't
-// used in userspace mode, so this should never be called.
-func (m *Manager) SetupEBPFProxyNoTrack(uint16, uint16) error {
-	return errNotSupported
-}
-
 // UpdateSet updates the rule destinations associated with the given set
 // by merging the existing prefixes with the new ones, then deduplicating.
 func (m *Manager) UpdateSet(set firewall.Set, prefixes []netip.Prefix) error {

client/firewall/uspfilter/interface_allower_windows.go

@@ -9,6 +9,7 @@ import (
 	log "github.com/sirupsen/logrus"
 
 	nberrors "github.com/netbirdio/netbird/client/errors"
+	"github.com/netbirdio/netbird/client/internal/wincmd"
 )
 
 type action string
@@ -91,7 +92,7 @@ func manageFirewallRule(ruleName string, action action, extraArgs ...string) err
 	if action == addRule {
 		args = append(args, extraArgs...)
 	}
-	netshCmd := GetSystem32Command("netsh")
+	netshCmd := wincmd.System32("netsh")
 	cmd := exec.Command(netshCmd, args...)
 	cmd.SysProcAttr = &syscall.SysProcAttr{HideWindow: true}
 	return cmd.Run()
@@ -100,7 +101,7 @@ func manageFirewallRule(ruleName string, action action, extraArgs ...string) err
 func isWindowsFirewallReachable() bool {
 	args := []string{"advfirewall", "show", "allprofiles", "state"}
 
-	netshCmd := GetSystem32Command("netsh")
+	netshCmd := wincmd.System32("netsh")
 
 	cmd := exec.Command(netshCmd, args...)
 	cmd.SysProcAttr = &syscall.SysProcAttr{HideWindow: true}
@@ -117,23 +118,10 @@ func isWindowsFirewallReachable() bool {
 func isFirewallRuleActive(ruleName string) bool {
 	args := []string{"advfirewall", "firewall", "show", "rule", "name=" + ruleName}
 
-	netshCmd := GetSystem32Command("netsh")
+	netshCmd := wincmd.System32("netsh")
 
 	cmd := exec.Command(netshCmd, args...)
 	cmd.SysProcAttr = &syscall.SysProcAttr{HideWindow: true}
 	_, err := cmd.Output()
 	return err == nil
 }
-
-// GetSystem32Command checks if a command can be found in the system path and returns it. In case it can't find it
-// in the path it will return the full path of a command assuming C:\windows\system32 as the base path.
-func GetSystem32Command(command string) string {
-	_, err := exec.LookPath(command)
-	if err == nil {
-		return command
-	}
-
-	log.Tracef("Command %s not found in PATH, using C:\\windows\\system32\\%s.exe path", command, command)
-
-	return "C:\\windows\\system32\\" + command + ".exe"
-}

client/iface/configurer/allowedips.go

@@ -0,0 +1,226 @@
+package configurer
+
+import (
+	"net"
+	"net/netip"
+	"slices"
+	"sync"
+
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+)
+
+// allowedIPStore mirrors the allowed IPs configured on each peer of a device.
+//
+// A configurer is the only writer of its device's peer set, so the mirror is authoritative
+// by construction. It spares the paths that have to rewrite one peer's allowed IPs a full
+// device dump just to recover prefixes the process already configured itself.
+//
+// An allowed IP belongs to exactly one peer: configuring a prefix on a peer takes it away
+// from whichever peer held it before, and the configurer leaves that handover to the device
+// rather than removing the prefix from the previous holder itself. The store tracks the
+// owner of each prefix and performs the same handover, so rewriting one peer's list never
+// takes a prefix back from the peer that owns it now.
+//
+// Its own lock guards the map alone, not the device write it accompanies. Consistency
+// between the two rests on the caller serializing every configurer call, which WGIface
+// does with its mutex; two unserialized writers would interleave a device write with the
+// record of a different one.
+//
+// An operator reconfiguring the device out of band, through `wg set` or the UAPI socket,
+// is the one way the mirror can still go stale. A peer missing from it falls back to the
+// device, which reseats that peer's prefixes and their ownership; a peer that is present
+// does not, so one recorded from empty while the device already held prefixes keeps only
+// what was recorded, and the next endpoint removal drops the rest.
+type allowedIPStore struct {
+	mu     sync.RWMutex
+	peers  map[wgtypes.Key][]netip.Prefix
+	owners map[netip.Prefix]wgtypes.Key
+}
+
+func newAllowedIPStore() *allowedIPStore {
+	return &allowedIPStore{
+		peers:  make(map[wgtypes.Key][]netip.Prefix),
+		owners: make(map[netip.Prefix]wgtypes.Key),
+	}
+}
+
+// get returns the prefixes recorded for a peer, and whether the peer is known at all.
+// The caller receives a copy and may retain or modify it freely.
+func (s *allowedIPStore) get(key wgtypes.Key) ([]netip.Prefix, bool) {
+	s.mu.RLock()
+	defer s.mu.RUnlock()
+
+	prefixes, ok := s.peers[key]
+	if !ok {
+		return nil, false
+	}
+	return slices.Clone(prefixes), true
+}
+
+// set replaces the prefixes recorded for a peer.
+func (s *allowedIPStore) set(key wgtypes.Key, prefixes []netip.Prefix) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	k := key
+	s.releaseLocked(k)
+
+	normalized := normalizePrefixes(prefixes)
+	for _, prefix := range normalized {
+		s.claimLocked(k, prefix)
+	}
+	s.peers[k] = normalized
+}
+
+// add records prefixes on a peer without dropping the ones already there, matching the
+// union semantics of a peer update that does not replace its allowed IPs. It records the
+// peer if it is not known yet, so it belongs to the operations that create a peer on the
+// device rather than to the update-only ones.
+func (s *allowedIPStore) add(key wgtypes.Key, prefixes []netip.Prefix) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	s.mergeLocked(key, prefixes)
+}
+
+// addExisting is add for an update-only device operation. Such an operation is a silent
+// no-op when the peer is absent, so recording a peer here would leave the store claiming
+// prefixes the device never took, and the peer would then be recreated by the next endpoint
+// removal, stealing those allowed IPs from the peer that legitimately holds them.
+func (s *allowedIPStore) addExisting(key wgtypes.Key, prefixes []netip.Prefix) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	k := key
+	if _, ok := s.peers[k]; !ok {
+		return
+	}
+	s.mergeLocked(k, prefixes)
+}
+
+// ensure records a peer with no prefixes unless it is already known. A device operation
+// that is not update-only creates the peer when it is absent, so it has to be recorded even
+// when it configures nothing else; otherwise the peer exists on the device while the store
+// treats it as unknown, and a prefix later handed over to it is not accounted for.
+func (s *allowedIPStore) ensure(key wgtypes.Key) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	k := key
+	if _, ok := s.peers[k]; !ok {
+		s.peers[k] = nil
+	}
+}
+
+// forget drops every prefix recorded for a peer.
+func (s *allowedIPStore) forget(key wgtypes.Key) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	k := key
+	s.releaseLocked(k)
+	delete(s.peers, k)
+}
+
+// reset drops every peer, mirroring a device reconfiguration that replaces the peer set.
+func (s *allowedIPStore) reset() {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	s.peers = make(map[wgtypes.Key][]netip.Prefix)
+	s.owners = make(map[netip.Prefix]wgtypes.Key)
+}
+
+// mergeLocked unions normalized prefixes into a peer and transfers their ownership.
+// The caller must hold s.mu for writing.
+func (s *allowedIPStore) mergeLocked(k wgtypes.Key, prefixes []netip.Prefix) {
+	merged := s.peers[k]
+	for _, prefix := range prefixes {
+		prefix = normalizePrefix(prefix)
+		s.claimLocked(k, prefix)
+		if !slices.Contains(merged, prefix) {
+			merged = append(merged, prefix)
+		}
+	}
+	s.peers[k] = merged
+}
+
+// claimLocked hands a prefix over to a peer, taking it from its previous owner the way the
+// device does when the same prefix is configured on a second peer.
+func (s *allowedIPStore) claimLocked(k wgtypes.Key, prefix netip.Prefix) {
+	if owner, ok := s.owners[prefix]; ok && owner != k {
+		s.peers[owner] = slices.DeleteFunc(s.peers[owner], func(p netip.Prefix) bool {
+			return p == prefix
+		})
+	}
+	s.owners[prefix] = k
+}
+
+// releaseLocked drops a peer's claim on every prefix it currently holds.
+func (s *allowedIPStore) releaseLocked(k wgtypes.Key) {
+	for _, prefix := range s.peers[k] {
+		if s.owners[prefix] == k {
+			delete(s.owners, prefix)
+		}
+	}
+}
+
+// normalizePrefix puts a prefix into the form the store recognises it by. It clears the
+// host bits, which a device does on its own, so a caller passing 10.20.0.1/16 still matches
+// the 10.20.0.0/16 read back from the device; and it unmaps a v4-mapped prefix so that it
+// compares equal to, and marshals like, the plain v4 prefix for the same network.
+//
+// Masking comes first because it also decides the address family: only a prefix at least 96
+// bits long keeps the mapped marker through the mask, so a shorter prefix inside the mapped
+// range is a genuine v6 prefix and unmapping it would yield an invalid v4 prefix.
+func normalizePrefix(prefix netip.Prefix) netip.Prefix {
+	masked := prefix.Masked()
+
+	addr := masked.Addr()
+	if !addr.Is4In6() {
+		return masked
+	}
+	return netip.PrefixFrom(addr.Unmap(), masked.Bits()-96)
+}
+
+// normalizePrefixes returns a normalized copy without changing the caller's slice.
+func normalizePrefixes(prefixes []netip.Prefix) []netip.Prefix {
+	normalized := make([]netip.Prefix, len(prefixes))
+	for i, prefix := range prefixes {
+		normalized[i] = normalizePrefix(prefix)
+	}
+	return normalized
+}
+
+// ipNetsToPrefixes converts addresses read back from a device. Unmap keeps a v4-mapped v6
+// address comparable to the plain v4 prefix the configurer was given.
+func ipNetsToPrefixes(ipNets []net.IPNet) []netip.Prefix {
+	prefixes := make([]netip.Prefix, 0, len(ipNets))
+	for _, ipNet := range ipNets {
+		addr, ok := netip.AddrFromSlice(ipNet.IP)
+		if !ok {
+			continue
+		}
+
+		ones, maskBits := ipNet.Mask.Size()
+		// A device may report a v4 prefix as a v4-mapped address. Align the address form with
+		// the mask rather than unmapping on sight: a 32 bit mask always describes v4, while a
+		// 128 bit mask describes v4 only when it covers the mapped prefix, so a genuine v6
+		// prefix inside the mapped range stays v6 instead of being dropped as invalid.
+		if addr.Is4In6() {
+			switch {
+			case maskBits == 32:
+				addr = addr.Unmap()
+			case maskBits == 128 && ones >= 96:
+				addr, ones = addr.Unmap(), ones-96
+			}
+		}
+
+		prefix := netip.PrefixFrom(addr, ones)
+		if !prefix.IsValid() {
+			continue
+		}
+		prefixes = append(prefixes, prefix.Masked())
+	}
+	return prefixes
+}

client/iface/configurer/allowedips_test.go

@@ -0,0 +1,263 @@
+package configurer
+
+import (
+	"net"
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+)
+
+// The store keys on the parsed key, so the tests use two distinct ones rather than names.
+var (
+	testPeer  = wgtypes.Key{1}
+	otherPeer = wgtypes.Key{2}
+)
+
+func TestAllowedIPStoreUnknownPeer(t *testing.T) {
+	s := newAllowedIPStore()
+
+	prefixes, ok := s.get(testPeer)
+	assert.False(t, ok, "an unconfigured peer must be reported as unknown, not as one without prefixes")
+	assert.Nil(t, prefixes, "an unknown peer has no prefixes")
+}
+
+func TestAllowedIPStoreAddUnions(t *testing.T) {
+	s := newAllowedIPStore()
+	overlay := netip.MustParsePrefix("100.64.0.1/32")
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+
+	s.set(testPeer, []netip.Prefix{overlay})
+	// A peer update does not replace allowed IPs, and a repeated prefix must not be doubled.
+	s.add(testPeer, []netip.Prefix{overlay, routed})
+
+	prefixes, ok := s.get(testPeer)
+	require.True(t, ok, "peer must be known after set")
+	assert.Equal(t, []netip.Prefix{overlay, routed}, prefixes, "add must union rather than replace")
+}
+
+func TestAllowedIPStoreGetReturnsCopy(t *testing.T) {
+	s := newAllowedIPStore()
+	overlay := netip.MustParsePrefix("100.64.0.1/32")
+	s.set(testPeer, []netip.Prefix{overlay})
+
+	prefixes, ok := s.get(testPeer)
+	require.True(t, ok, "peer must be known after set")
+	prefixes[0] = netip.MustParsePrefix("0.0.0.0/0")
+
+	stored, _ := s.get(testPeer)
+	assert.Equal(t, []netip.Prefix{overlay}, stored, "a caller mutating the returned slice must not corrupt the store")
+}
+
+func TestAllowedIPStoreForgetAndReset(t *testing.T) {
+	s := newAllowedIPStore()
+	s.set(testPeer, []netip.Prefix{netip.MustParsePrefix("100.64.0.1/32")})
+	s.set(otherPeer, []netip.Prefix{netip.MustParsePrefix("100.64.0.2/32")})
+
+	s.forget(testPeer)
+	_, ok := s.get(testPeer)
+	assert.False(t, ok, "a forgotten peer must be unknown")
+	_, ok = s.get(otherPeer)
+	assert.True(t, ok, "forgetting one peer must not touch the others")
+
+	s.reset()
+	_, ok = s.get(otherPeer)
+	assert.False(t, ok, "reset must drop every peer")
+}
+
+func TestIPNetsToPrefixes(t *testing.T) {
+	tests := []struct {
+		name  string
+		ipNet net.IPNet
+		want  string
+	}{
+		{
+			name:  "v4",
+			ipNet: net.IPNet{IP: net.IP{10, 20, 0, 0}, Mask: net.CIDRMask(16, 32)},
+			want:  "10.20.0.0/16",
+		},
+		{
+			name:  "v4 mapped under a 128 bit mask",
+			ipNet: net.IPNet{IP: net.ParseIP("10.20.0.0"), Mask: net.CIDRMask(112, 128)},
+			want:  "10.20.0.0/16",
+		},
+		{
+			name:  "v6",
+			ipNet: net.IPNet{IP: net.ParseIP("fd00::"), Mask: net.CIDRMask(64, 128)},
+			want:  "fd00::/64",
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			got := ipNetsToPrefixes([]net.IPNet{tc.ipNet})
+			require.Len(t, got, 1, "the address must be converted, not dropped")
+			assert.Equal(t, tc.want, got[0].String(), "converted prefix")
+		})
+	}
+}
+
+func TestIPNetsToPrefixesRoundTrip(t *testing.T) {
+	prefixes := []netip.Prefix{
+		netip.MustParsePrefix("100.64.0.1/32"),
+		netip.MustParsePrefix("10.20.0.0/16"),
+		netip.MustParsePrefix("fd00::/64"),
+	}
+
+	assert.Equal(t, prefixes, ipNetsToPrefixes(prefixesToIPNets(prefixes)),
+		"prefixes handed to a device must come back unchanged")
+}
+
+func TestAllowedIPStoreNormalizesMappedPrefixes(t *testing.T) {
+	s := newAllowedIPStore()
+	v4 := netip.MustParsePrefix("10.20.0.0/16")
+	mapped := netip.PrefixFrom(netip.AddrFrom16(v4.Addr().As16()), 112)
+
+	s.set(testPeer, []netip.Prefix{mapped})
+	// A v4 rule only matches a v4-mapped address once it has been unmapped, so the store must
+	// hold the plain form and recognise the two spellings as the same prefix.
+	s.add(testPeer, []netip.Prefix{v4})
+
+	prefixes, ok := s.get(testPeer)
+	require.True(t, ok, "peer must be known after set")
+	assert.Equal(t, []netip.Prefix{v4}, prefixes, "a mapped prefix must be stored unmapped and not duplicated")
+}
+
+func TestNormalizePrefix(t *testing.T) {
+	v4 := netip.MustParsePrefix("10.20.0.0/16")
+	v6 := netip.MustParsePrefix("fd00::/64")
+
+	assert.Equal(t, v4, normalizePrefix(v4), "a plain v4 prefix is unchanged")
+	assert.Equal(t, v6, normalizePrefix(v6), "a real v6 prefix is unchanged")
+	assert.Equal(t, v4, normalizePrefix(netip.PrefixFrom(netip.AddrFrom16(v4.Addr().As16()), 112)),
+		"a mapped prefix under a 128 bit mask becomes plain v4")
+	// A prefix shorter than /96 inside the mapped range is a genuine v6 prefix. Unmapping it
+	// would pair a v4 address with a v6 sized mask, which is invalid, and the store would then
+	// record a zero prefix that can never recreate the allowed IP.
+	for _, tc := range []string{"::ffff:0:0/64", "::ffff:1.2.3.4/80", "::ffff:1.2.3.4/95"} {
+		got := normalizePrefix(netip.MustParsePrefix(tc))
+		assert.True(t, got.IsValid(), "%s must normalize to a valid prefix", tc)
+		assert.False(t, got.Addr().Is4(), "%s must stay v6", tc)
+	}
+}
+
+func TestAllowedIPStoreAddExistingDoesNotCreate(t *testing.T) {
+	s := newAllowedIPStore()
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+
+	// An update-only device operation on an absent peer is a silent no-op, so nothing may be
+	// recorded for a peer the store does not already know.
+	s.addExisting(testPeer, []netip.Prefix{routed})
+	_, ok := s.get(testPeer)
+	assert.False(t, ok, "addExisting must not record an unknown peer")
+
+	overlay := netip.MustParsePrefix("100.64.0.1/32")
+	s.set(testPeer, []netip.Prefix{overlay})
+	s.addExisting(testPeer, []netip.Prefix{routed})
+
+	prefixes, _ := s.get(testPeer)
+	assert.Equal(t, []netip.Prefix{overlay, routed}, prefixes, "addExisting must union onto a known peer")
+}
+
+func TestAllowedIPStoreHandsPrefixOverToTheNewOwner(t *testing.T) {
+	s := newAllowedIPStore()
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+	other := otherPeer
+
+	s.set(testPeer, []netip.Prefix{netip.MustParsePrefix("100.64.0.1/32"), routed})
+	s.set(other, []netip.Prefix{netip.MustParsePrefix("100.64.0.2/32")})
+
+	// The device takes an allowed IP away from its previous holder when it is configured on
+	// another peer, so the store must do the same rather than list it under both.
+	s.addExisting(other, []netip.Prefix{routed})
+
+	previous, _ := s.get(testPeer)
+	assert.NotContains(t, previous, routed, "the previous owner must lose the prefix")
+	current, _ := s.get(other)
+	assert.Contains(t, current, routed, "the new owner must hold the prefix")
+}
+
+func TestAllowedIPStoreForgetReleasesOwnership(t *testing.T) {
+	s := newAllowedIPStore()
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+
+	s.set(testPeer, []netip.Prefix{routed})
+	s.forget(testPeer)
+	s.set(otherPeer, []netip.Prefix{routed})
+
+	// A forgotten peer must not be resurrected as a key in the peer map by a later claim.
+	_, ok := s.get(testPeer)
+	assert.False(t, ok, "the forgotten peer must stay unknown")
+	current, _ := s.get(otherPeer)
+	assert.Equal(t, []netip.Prefix{routed}, current, "the new owner must hold the prefix")
+}
+
+func TestNormalizePrefixClearsHostBits(t *testing.T) {
+	// A device stores a prefix masked, so a caller passing host bits must still match what a
+	// device fallback seeded, otherwise that prefix could never be removed by value.
+	assert.Equal(t, netip.MustParsePrefix("10.20.0.0/16"),
+		normalizePrefix(netip.MustParsePrefix("10.20.0.1/16")), "host bits must be cleared")
+	assert.Equal(t, netip.MustParsePrefix("fd00::/64"),
+		normalizePrefix(netip.MustParsePrefix("fd00::1/64")), "host bits must be cleared for v6")
+}
+
+func TestIPNetsToPrefixesKeepsV6InTheMappedRange(t *testing.T) {
+	// ::ffff:0:0/64 reads as v4-mapped but is a genuine v6 prefix: unmapping it would leave a
+	// v4 address under a 64 bit mask, which is invalid, and the allowed IP would be dropped.
+	got := ipNetsToPrefixes([]net.IPNet{{
+		IP:   net.ParseIP("::ffff:0:0"),
+		Mask: net.CIDRMask(64, 128),
+	}})
+
+	require.Len(t, got, 1, "the prefix must be converted, not dropped")
+	assert.False(t, got[0].Addr().Is4(), "a v6 prefix in the mapped range must not become v4")
+	assert.Equal(t, 64, got[0].Bits(), "the prefix length must survive the conversion")
+}
+
+func TestPrefixesToIPNetsNormalizes(t *testing.T) {
+	// net.IPNet prints a v4-mapped address as v4 but takes the length from its 16 byte
+	// mask, so an unnormalized ::ffff:10.1.2.3/64 reaches a userspace device as 10.1.2.3/0,
+	// an allowed IP that matches every v4 address.
+	tests := []struct {
+		name  string
+		given string
+		want  string
+	}{
+		{name: "mapped below /96", given: "::ffff:10.1.2.3/64", want: "::/64"},
+		{name: "mapped at /112", given: "::ffff:10.1.2.3/112", want: "10.1.0.0/16"},
+		{name: "host bits are cleared", given: "10.20.0.1/16", want: "10.20.0.0/16"},
+		{name: "v6 is untouched", given: "fd00::1/64", want: "fd00::/64"},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			got := prefixesToIPNets([]netip.Prefix{netip.MustParsePrefix(tc.given)})
+			require.Len(t, got, 1, "the prefix must be converted, not dropped")
+			assert.Equal(t, tc.want, got[0].String(), "what the device is given")
+			assert.NotEqual(t, 0, mustOnes(t, got[0]), "a device must never be given a zero length allowed IP")
+		})
+	}
+}
+
+func mustOnes(t *testing.T, ipNet net.IPNet) int {
+	t.Helper()
+
+	ones, _ := ipNet.Mask.Size()
+	return ones
+}
+
+// TestPrefixesToIPNetsAgreesWithTheStore pins the property the store depends on: what a
+// device is given and what is recorded for it are the same prefix.
+func TestPrefixesToIPNetsAgreesWithTheStore(t *testing.T) {
+	for _, given := range []string{"::ffff:10.1.2.3/64", "::ffff:10.1.2.3/112", "10.20.0.1/16", "fd00::1/64"} {
+		prefix := netip.MustParsePrefix(given)
+
+		toDevice := prefixesToIPNets([]netip.Prefix{prefix})
+		recorded := normalizePrefix(prefix)
+
+		assert.Equal(t, recorded.String(), toDevice[0].String(),
+			"%s must reach the device in the form the store records", given)
+	}
+}

client/iface/configurer/common.go

@@ -19,12 +19,18 @@ func buildPresharedKeyConfig(peerKey wgtypes.Key, psk wgtypes.Key, updateOnly bo
 	}
 }
 
+// prefixesToIPNets converts prefixes on their way to a device. It is the only place that
+// conversion happens, so it also normalizes: the device is then given the same form the
+// store records, and a v4-mapped prefix cannot reach net.IPNet, which prints such an
+// address as v4 while taking the length from its 16 byte mask and so turns
+// ::ffff:10.1.2.3/64 into 10.1.2.3/0 — an allowed IP matching every v4 address.
 func prefixesToIPNets(prefixes []netip.Prefix) []net.IPNet {
 	ipNets := make([]net.IPNet, len(prefixes))
 	for i, prefix := range prefixes {
+		normalized := normalizePrefix(prefix)
 		ipNets[i] = net.IPNet{
-			IP:   prefix.Addr().AsSlice(),                             // Convert netip.Addr to net.IP
-			Mask: net.CIDRMask(prefix.Bits(), prefix.Addr().BitLen()), // Create subnet mask
+			IP:   normalized.Addr().AsSlice(),
+			Mask: net.CIDRMask(normalized.Bits(), normalized.Addr().BitLen()),
 		}
 	}
 	return ipNets

client/iface/configurer/kernel_unix.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"net"
 	"net/netip"
+	"slices"
 	"time"
 
 	log "github.com/sirupsen/logrus"
@@ -18,16 +19,22 @@ import (
 type KernelConfigurer struct {
 	deviceName string
 	statsCache *statsCache
+	allowedIPs *allowedIPStore
 }
 
+// NewKernelConfigurer creates a configurer with an empty allowed IP mirror
+// and a statistics cache for the named kernel device.
 func NewKernelConfigurer(deviceName string) *KernelConfigurer {
 	c := &KernelConfigurer{
 		deviceName: deviceName,
+		allowedIPs: newAllowedIPStore(),
 	}
 	c.statsCache = newStatsCache(statsCacheTTL, c.fetchStats)
 	return c
 }
 
+// ConfigureInterface sets the device key, port and firewall mark, replacing all peers.
+// The allowed IP mirror is reset only after the device accepts the configuration.
 func (c *KernelConfigurer) ConfigureInterface(privateKey string, port int) error {
 	log.Debugf("adding Wireguard private key")
 	key, err := wgtypes.ParseKey(privateKey)
@@ -46,6 +53,8 @@ func (c *KernelConfigurer) ConfigureInterface(privateKey string, port int) error
 	if err != nil {
 		return fmt.Errorf(`received error "%w" while configuring interface %s with port %d`, err, c.deviceName, port)
 	}
+
+	c.allowedIPs.reset()
 	return nil
 }
 
@@ -58,9 +67,20 @@ func (c *KernelConfigurer) SetPresharedKey(peerKey string, psk wgtypes.Key, upda
 	}
 
 	cfg := buildPresharedKeyConfig(parsedPeerKey, psk, updateOnly)
-	return c.configure(cfg)
+	if err := c.configure(cfg); err != nil {
+		return err
+	}
+
+	// Without updateOnly this creates the peer when it is absent, so the store has to
+	// know about it even though no allowed IP was configured.
+	if !updateOnly {
+		c.allowedIPs.ensure(parsedPeerKey)
+	}
+	return nil
 }
 
+// UpdatePeer creates or updates a peer, merging allowed IPs with its existing set.
+// Prefixes assigned to this peer are transferred from their previous owners.
 func (c *KernelConfigurer) UpdatePeer(peerKey string, allowedIps []netip.Prefix, keepAlive time.Duration, endpoint *net.UDPAddr, preSharedKey *wgtypes.Key) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
@@ -83,19 +103,23 @@ func (c *KernelConfigurer) UpdatePeer(peerKey string, allowedIps []netip.Prefix,
 	if err != nil {
 		return fmt.Errorf(`received error "%w" while updating peer on interface %s with settings: allowed ips %s, endpoint %s`, err, c.deviceName, allowedIps, endpoint.String())
 	}
+
+	c.allowedIPs.add(peerKeyParsed, allowedIps)
 	return nil
 }
 
+// RemoveEndpointAddress clears the endpoint of a peer while keeping it configured.
+// Neither the netlink API nor the userspace one can clear an endpoint in place, so the peer
+// is removed and re-added with the allowed IPs it already had.
 func (c *KernelConfigurer) RemoveEndpointAddress(peerKey string) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return err
 	}
 
-	// Get the existing peer to preserve its allowed IPs
-	existingPeer, err := c.getPeer(c.deviceName, peerKey)
+	allowedIPs, err := c.peerAllowedIPs(peerKeyParsed)
 	if err != nil {
-		return fmt.Errorf("get peer: %w", err)
+		return err
 	}
 
 	removePeerCfg := wgtypes.PeerConfig{
@@ -104,26 +128,27 @@ func (c *KernelConfigurer) RemoveEndpointAddress(peerKey string) error {
 	}
 
 	if err := c.configure(wgtypes.Config{Peers: []wgtypes.PeerConfig{removePeerCfg}}); err != nil {
-		return fmt.Errorf(`error removing peer %s from interface %s: %w`, peerKey, c.deviceName, err)
+		return fmt.Errorf("remove peer %s from interface %s: %w", peerKey, c.deviceName, err)
 	}
 
-	//Re-add the peer without the endpoint but same AllowedIPs
 	reAddPeerCfg := wgtypes.PeerConfig{
 		PublicKey:         peerKeyParsed,
-		AllowedIPs:        existingPeer.AllowedIPs,
+		AllowedIPs:        prefixesToIPNets(allowedIPs),
 		ReplaceAllowedIPs: true,
 	}
 
 	if err := c.configure(wgtypes.Config{Peers: []wgtypes.PeerConfig{reAddPeerCfg}}); err != nil {
+		c.allowedIPs.forget(peerKeyParsed)
 		return fmt.Errorf(
-			`error re-adding peer %s to interface %s with allowed IPs %v: %w`,
-			peerKey, c.deviceName, existingPeer.AllowedIPs, err,
+			"re-add peer %s to interface %s with allowed IPs %v: %w",
+			peerKey, c.deviceName, allowedIPs, err,
 		)
 	}
 
 	return nil
 }
 
+// RemovePeer removes a peer and forgets its allowed IPs after a successful device write.
 func (c *KernelConfigurer) RemovePeer(peerKey string) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
@@ -142,15 +167,13 @@ func (c *KernelConfigurer) RemovePeer(peerKey string) error {
 	if err != nil {
 		return fmt.Errorf(`received error "%w" while removing peer %s from interface %s`, err, peerKey, c.deviceName)
 	}
+
+	c.allowedIPs.forget(peerKeyParsed)
 	return nil
 }
 
+// AddAllowedIP adds a prefix to an existing peer; an absent peer is a silent no-op.
 func (c *KernelConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix) error {
-	ipNet := net.IPNet{
-		IP:   allowedIP.Addr().AsSlice(),
-		Mask: net.CIDRMask(allowedIP.Bits(), allowedIP.Addr().BitLen()),
-	}
-
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return err
@@ -159,7 +182,7 @@ func (c *KernelConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix)
 		PublicKey:         peerKeyParsed,
 		UpdateOnly:        true,
 		ReplaceAllowedIPs: false,
-		AllowedIPs:        []net.IPNet{ipNet},
+		AllowedIPs:        prefixesToIPNets([]netip.Prefix{allowedIP}),
 	}
 
 	config := wgtypes.Config{
@@ -169,52 +192,69 @@ func (c *KernelConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix)
 	if err != nil {
 		return fmt.Errorf(`received error "%w" while adding allowed Ip to peer on interface %s with settings: allowed ips %s`, err, c.deviceName, allowedIP)
 	}
+
+	c.allowedIPs.addExisting(peerKeyParsed, []netip.Prefix{allowedIP})
 	return nil
 }
 
+// RemoveAllowedIP removes a prefix while preserving the peer's other allowed IPs.
+// A prefix not assigned to the peer is a no-op.
 func (c *KernelConfigurer) RemoveAllowedIP(peerKey string, allowedIP netip.Prefix) error {
-	ipNet := net.IPNet{
-		IP:   allowedIP.Addr().AsSlice(),
-		Mask: net.CIDRMask(allowedIP.Bits(), allowedIP.Addr().BitLen()),
-	}
-
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return fmt.Errorf("parse peer key: %w", err)
 	}
 
-	existingPeer, err := c.getPeer(c.deviceName, peerKey)
+	currentAllowedIPs, err := c.peerAllowedIPs(peerKeyParsed)
 	if err != nil {
-		return fmt.Errorf("get peer: %w", err)
+		return err
 	}
 
-	newAllowedIPs := existingPeer.AllowedIPs
-
-	for i, existingAllowedIP := range existingPeer.AllowedIPs {
-		if existingAllowedIP.String() == ipNet.String() {
-			newAllowedIPs = append(existingPeer.AllowedIPs[:i], existingPeer.AllowedIPs[i+1:]...) //nolint:gocritic
-			break
-		}
+	idx := slices.Index(currentAllowedIPs, normalizePrefix(allowedIP))
+	if idx < 0 {
+		return nil
 	}
+	newAllowedIPs := slices.Delete(currentAllowedIPs, idx, idx+1)
 
 	peer := wgtypes.PeerConfig{
 		PublicKey:         peerKeyParsed,
 		UpdateOnly:        true,
 		ReplaceAllowedIPs: true,
-		AllowedIPs:        newAllowedIPs,
+		AllowedIPs:        prefixesToIPNets(newAllowedIPs),
 	}
 
 	config := wgtypes.Config{
 		Peers: []wgtypes.PeerConfig{peer},
 	}
-	err = c.configure(config)
-	if err != nil {
+	if err := c.configure(config); err != nil {
 		return fmt.Errorf("remove allowed IP %s on interface %s: %w", allowedIP, c.deviceName, err)
 	}
+
+	c.allowedIPs.set(peerKeyParsed, newAllowedIPs)
 	return nil
 }
 
-func (c *KernelConfigurer) getPeer(ifaceName, peerPubKey string) (wgtypes.Peer, error) {
+// peerAllowedIPs returns the allowed IPs configured for a peer, reading them from the device
+// only for a peer the store has not seen. Dumping the device costs a netlink round trip
+// proportional to the whole network map, and this runs on every relay and ICE transition.
+func (c *KernelConfigurer) peerAllowedIPs(peerKey wgtypes.Key) ([]netip.Prefix, error) {
+	if prefixes, ok := c.allowedIPs.get(peerKey); ok {
+		return prefixes, nil
+	}
+
+	existingPeer, err := c.getPeer(c.deviceName, peerKey)
+	if err != nil {
+		return nil, fmt.Errorf("get peer: %w", err)
+	}
+
+	prefixes := ipNetsToPrefixes(existingPeer.AllowedIPs)
+	c.allowedIPs.set(peerKey, prefixes)
+	return prefixes, nil
+}
+
+// getPeer scans the device for one peer. wgtypes.Key is an array, so the comparison is a
+// plain equality: Key.String would base64 encode into a fresh allocation for every peer.
+func (c *KernelConfigurer) getPeer(ifaceName string, peerPubKey wgtypes.Key) (wgtypes.Peer, error) {
 	wg, err := wgctrl.New()
 	if err != nil {
 		return wgtypes.Peer{}, fmt.Errorf("wgctl: %w", err)
@@ -231,7 +271,7 @@ func (c *KernelConfigurer) getPeer(ifaceName, peerPubKey string) (wgtypes.Peer,
 		return wgtypes.Peer{}, fmt.Errorf("get device %s: %w", ifaceName, err)
 	}
 	for _, peer := range wgDevice.Peers {
-		if peer.PublicKey.String() == peerPubKey {
+		if peer.PublicKey == peerPubKey {
 			return peer, nil
 		}
 	}

client/iface/configurer/usp.go

@@ -8,6 +8,7 @@ import (
 	"net/netip"
 	"os"
 	"runtime"
+	"slices"
 	"strconv"
 	"strings"
 	"time"
@@ -41,31 +42,38 @@ type WGUSPConfigurer struct {
 	deviceName       string
 	activityRecorder *bind.ActivityRecorder
 	statsCache       *statsCache
+	allowedIPs       *allowedIPStore
 
 	uapiListener net.Listener
 }
 
+// NewUSPConfigurer creates a userspace configurer and starts its UAPI listener.
 func NewUSPConfigurer(device *device.Device, deviceName string, activityRecorder *bind.ActivityRecorder) *WGUSPConfigurer {
 	wgCfg := &WGUSPConfigurer{
 		device:           device,
 		deviceName:       deviceName,
 		activityRecorder: activityRecorder,
+		allowedIPs:       newAllowedIPStore(),
 	}
 	wgCfg.statsCache = newStatsCache(statsCacheTTL, wgCfg.fetchStats)
 	wgCfg.startUAPI()
 	return wgCfg
 }
 
+// NewUSPConfigurerNoUAPI creates a userspace configurer without a UAPI listener.
 func NewUSPConfigurerNoUAPI(device *device.Device, deviceName string, activityRecorder *bind.ActivityRecorder) *WGUSPConfigurer {
 	wgCfg := &WGUSPConfigurer{
 		device:           device,
 		deviceName:       deviceName,
 		activityRecorder: activityRecorder,
+		allowedIPs:       newAllowedIPStore(),
 	}
 	wgCfg.statsCache = newStatsCache(statsCacheTTL, wgCfg.fetchStats)
 	return wgCfg
 }
 
+// ConfigureInterface sets the device key, port and firewall mark, replacing all peers.
+// The allowed IP mirror is reset only after the device accepts the configuration.
 func (c *WGUSPConfigurer) ConfigureInterface(privateKey string, port int) error {
 	log.Debugf("adding Wireguard private key")
 	key, err := wgtypes.ParseKey(privateKey)
@@ -80,7 +88,12 @@ func (c *WGUSPConfigurer) ConfigureInterface(privateKey string, port int) error
 		ListenPort:   &port,
 	}
 
-	return c.device.IpcSet(toWgUserspaceString(config))
+	if err := c.device.IpcSet(toWgUserspaceString(config)); err != nil {
+		return err
+	}
+
+	c.allowedIPs.reset()
+	return nil
 }
 
 // SetPresharedKey sets the preshared key for a peer.
@@ -92,14 +105,38 @@ func (c *WGUSPConfigurer) SetPresharedKey(peerKey string, psk wgtypes.Key, updat
 	}
 
 	cfg := buildPresharedKeyConfig(parsedPeerKey, psk, updateOnly)
-	return c.device.IpcSet(toWgUserspaceString(cfg))
+	if err := c.device.IpcSet(toWgUserspaceString(cfg)); err != nil {
+		return err
+	}
+
+	// Without updateOnly this creates the peer when it is absent, so the store has to
+	// know about it even though no allowed IP was configured.
+	if !updateOnly {
+		c.allowedIPs.ensure(parsedPeerKey)
+	}
+	return nil
 }
 
+// UpdatePeer creates or updates a peer, merging allowed IPs with its existing set.
+// It validates the endpoint before writing and records changes after a successful write.
 func (c *WGUSPConfigurer) UpdatePeer(peerKey string, allowedIps []netip.Prefix, keepAlive time.Duration, endpoint *net.UDPAddr, preSharedKey *wgtypes.Key) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return err
 	}
+
+	// Everything that can fail is done before the device is touched, so a failure here
+	// cannot leave the device holding a peer that the activity recorder and the allowed
+	// IP store never learned about.
+	var addrPort netip.AddrPort
+	if endpoint != nil {
+		addr, err := netip.ParseAddr(endpoint.IP.String())
+		if err != nil {
+			return fmt.Errorf("parse endpoint address: %w", err)
+		}
+		addrPort = netip.AddrPortFrom(addr.Unmap(), uint16(endpoint.Port))
+	}
+
 	peer := wgtypes.PeerConfig{
 		PublicKey:         peerKeyParsed,
 		ReplaceAllowedIPs: false,
@@ -119,47 +156,27 @@ func (c *WGUSPConfigurer) UpdatePeer(peerKey string, allowedIps []netip.Prefix,
 	}
 
 	if endpoint != nil {
-		addr, err := netip.ParseAddr(endpoint.IP.String())
-		if err != nil {
-			return fmt.Errorf("failed to parse endpoint address: %w", err)
-		}
-		addrPort := netip.AddrPortFrom(addr.Unmap(), uint16(endpoint.Port))
 		c.activityRecorder.UpsertAddress(peerKey, addrPort)
 	}
+
+	c.allowedIPs.add(peerKeyParsed, allowedIps)
 	return nil
 }
 
+// RemoveEndpointAddress clears the endpoint of a peer while keeping it configured.
+// The UAPI cannot clear an endpoint in place, so the peer is removed and re-added with the
+// allowed IPs it already had.
 func (c *WGUSPConfigurer) RemoveEndpointAddress(peerKey string) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return fmt.Errorf("parse peer key: %w", err)
 	}
 
-	ipcStr, err := c.device.IpcGet()
+	allowedIPs, err := c.peerAllowedIPs(peerKeyParsed)
 	if err != nil {
-		return fmt.Errorf("get IPC config: %w", err)
+		return err
 	}
 
-	// Parse current status to get allowed IPs for the peer
-	stats, err := parseStatus(c.deviceName, ipcStr)
-	if err != nil {
-		return fmt.Errorf("parse IPC config: %w", err)
-	}
-
-	var allowedIPs []net.IPNet
-	found := false
-	for _, peer := range stats.Peers {
-		if peer.PublicKey == peerKey {
-			allowedIPs = peer.AllowedIPs
-			found = true
-			break
-		}
-	}
-	if !found {
-		return fmt.Errorf("peer %s not found", peerKey)
-	}
-
-	// remove the peer from the WireGuard configuration
 	peer := wgtypes.PeerConfig{
 		PublicKey: peerKeyParsed,
 		Remove:    true,
@@ -169,14 +186,13 @@ func (c *WGUSPConfigurer) RemoveEndpointAddress(peerKey string) error {
 		Peers: []wgtypes.PeerConfig{peer},
 	}
 	if ipcErr := c.device.IpcSet(toWgUserspaceString(config)); ipcErr != nil {
-		return fmt.Errorf("failed to remove peer: %s", ipcErr)
+		return fmt.Errorf("remove peer: %w", ipcErr)
 	}
 
-	// Build the peer config
 	peer = wgtypes.PeerConfig{
 		PublicKey:         peerKeyParsed,
 		ReplaceAllowedIPs: true,
-		AllowedIPs:        allowedIPs,
+		AllowedIPs:        prefixesToIPNets(allowedIPs),
 	}
 
 	config = wgtypes.Config{
@@ -184,12 +200,15 @@ func (c *WGUSPConfigurer) RemoveEndpointAddress(peerKey string) error {
 	}
 
 	if err := c.device.IpcSet(toWgUserspaceString(config)); err != nil {
-		return fmt.Errorf("remove endpoint address: %w", err)
+		c.allowedIPs.forget(peerKeyParsed)
+		return fmt.Errorf("re-add peer without endpoint: %w", err)
 	}
 
 	return nil
 }
 
+// RemovePeer removes a peer, then clears its activity and allowed IP records.
+// A failed device write leaves both records intact.
 func (c *WGUSPConfigurer) RemovePeer(peerKey string) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
@@ -204,18 +223,17 @@ func (c *WGUSPConfigurer) RemovePeer(peerKey string) error {
 	config := wgtypes.Config{
 		Peers: []wgtypes.PeerConfig{peer},
 	}
-	ipcErr := c.device.IpcSet(toWgUserspaceString(config))
-
-	c.activityRecorder.Remove(peerKey)
-	return ipcErr
-}
-
-func (c *WGUSPConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix) error {
-	ipNet := net.IPNet{
-		IP:   allowedIP.Addr().AsSlice(),
-		Mask: net.CIDRMask(allowedIP.Bits(), allowedIP.Addr().BitLen()),
+	if ipcErr := c.device.IpcSet(toWgUserspaceString(config)); ipcErr != nil {
+		return ipcErr
 	}
 
+	c.activityRecorder.Remove(peerKey)
+	c.allowedIPs.forget(peerKeyParsed)
+	return nil
+}
+
+// AddAllowedIP adds a prefix to an existing peer; an absent peer is a silent no-op.
+func (c *WGUSPConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix) error {
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
 	if err != nil {
 		return err
@@ -224,79 +242,89 @@ func (c *WGUSPConfigurer) AddAllowedIP(peerKey string, allowedIP netip.Prefix) e
 		PublicKey:         peerKeyParsed,
 		UpdateOnly:        true,
 		ReplaceAllowedIPs: false,
-		AllowedIPs:        []net.IPNet{ipNet},
+		AllowedIPs:        prefixesToIPNets([]netip.Prefix{allowedIP}),
 	}
 
 	config := wgtypes.Config{
 		Peers: []wgtypes.PeerConfig{peer},
 	}
 
-	return c.device.IpcSet(toWgUserspaceString(config))
+	if err := c.device.IpcSet(toWgUserspaceString(config)); err != nil {
+		return err
+	}
+
+	c.allowedIPs.addExisting(peerKeyParsed, []netip.Prefix{allowedIP})
+	return nil
 }
 
+// RemoveAllowedIP removes a prefix while preserving the peer's other allowed IPs.
+// It returns ErrAllowedIPNotFound if the prefix is not assigned to the peer.
 func (c *WGUSPConfigurer) RemoveAllowedIP(peerKey string, allowedIP netip.Prefix) error {
-	ipc, err := c.device.IpcGet()
-	if err != nil {
-		return err
-	}
-
 	peerKeyParsed, err := wgtypes.ParseKey(peerKey)
+	if err != nil {
+		return fmt.Errorf("parse peer key: %w", err)
+	}
+
+	currentAllowedIPs, err := c.peerAllowedIPs(peerKeyParsed)
 	if err != nil {
 		return err
 	}
-	hexKey := hex.EncodeToString(peerKeyParsed[:])
 
-	lines := strings.Split(ipc, "\n")
+	idx := slices.Index(currentAllowedIPs, normalizePrefix(allowedIP))
+	if idx < 0 {
+		return ErrAllowedIPNotFound
+	}
+	newAllowedIPs := slices.Delete(currentAllowedIPs, idx, idx+1)
 
 	peer := wgtypes.PeerConfig{
 		PublicKey:         peerKeyParsed,
 		UpdateOnly:        true,
 		ReplaceAllowedIPs: true,
-		AllowedIPs:        []net.IPNet{},
+		AllowedIPs:        prefixesToIPNets(newAllowedIPs),
 	}
 
-	foundPeer := false
-	removedAllowedIP := false
-	ip := allowedIP.String()
-
-	for _, line := range lines {
-		line = strings.TrimSpace(line)
-
-		// If we're within the details of the found peer and encounter another public key,
-		// this means we're starting another peer's details. So, reset the flag.
-		if strings.HasPrefix(line, "public_key=") && foundPeer {
-			foundPeer = false
-		}
-
-		// Identify the peer with the specific public key
-		if line == fmt.Sprintf("public_key=%s", hexKey) {
-			foundPeer = true
-		}
-
-		// If we're within the details of the found peer and find the specific allowed IP, skip this line
-		if foundPeer && line == "allowed_ip="+ip {
-			removedAllowedIP = true
-			continue
-		}
-
-		// Append the line to the output string
-		if foundPeer && strings.HasPrefix(line, "allowed_ip=") {
-			allowedIPStr := strings.TrimPrefix(line, "allowed_ip=")
-			_, ipNet, err := net.ParseCIDR(allowedIPStr)
-			if err != nil {
-				return err
-			}
-			peer.AllowedIPs = append(peer.AllowedIPs, *ipNet)
-		}
-	}
-
-	if !removedAllowedIP {
-		return ErrAllowedIPNotFound
-	}
 	config := wgtypes.Config{
 		Peers: []wgtypes.PeerConfig{peer},
 	}
-	return c.device.IpcSet(toWgUserspaceString(config))
+	if err := c.device.IpcSet(toWgUserspaceString(config)); err != nil {
+		return fmt.Errorf("remove allowed IP %s: %w", allowedIP, err)
+	}
+
+	c.allowedIPs.set(peerKeyParsed, newAllowedIPs)
+	return nil
+}
+
+// peerAllowedIPs returns the allowed IPs configured for a peer, reading them from the device
+// only for a peer the store has not seen. Reading them back means dumping and parsing the
+// whole device configuration, and this runs on every relay and ICE transition.
+func (c *WGUSPConfigurer) peerAllowedIPs(peerKey wgtypes.Key) ([]netip.Prefix, error) {
+	if prefixes, ok := c.allowedIPs.get(peerKey); ok {
+		return prefixes, nil
+	}
+
+	ipcStr, err := c.device.IpcGet()
+	if err != nil {
+		return nil, fmt.Errorf("get IPC config: %w", err)
+	}
+
+	stats, err := parseStatus(c.deviceName, ipcStr)
+	if err != nil {
+		return nil, fmt.Errorf("parse IPC config: %w", err)
+	}
+
+	// parseStatus reports keys in their textual form, so the comparison needs it once.
+	wanted := peerKey.String()
+	for _, peer := range stats.Peers {
+		if peer.PublicKey != wanted {
+			continue
+		}
+
+		prefixes := ipNetsToPrefixes(peer.AllowedIPs)
+		c.allowedIPs.set(peerKey, prefixes)
+		return prefixes, nil
+	}
+
+	return nil, ErrPeerNotFound
 }
 
 func (c *WGUSPConfigurer) FullStats() (*Stats, error) {

client/iface/configurer/usp_allowedips_test.go

@@ -0,0 +1,318 @@
+package configurer
+
+import (
+	"net"
+	"net/netip"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	wgconn "golang.zx2c4.com/wireguard/conn"
+	wgdevice "golang.zx2c4.com/wireguard/device"
+	"golang.zx2c4.com/wireguard/tun/tuntest"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+
+	"github.com/netbirdio/netbird/client/iface/bind"
+)
+
+// newTestUSPConfigurer builds a configurer over a real wireguard-go device backed by an
+// in-memory TUN. The device stays down, so no socket is opened and no privileges are needed.
+func newTestUSPConfigurer(t *testing.T) *WGUSPConfigurer {
+	t.Helper()
+
+	tun := tuntest.NewChannelTUN()
+	dev := wgdevice.NewDevice(tun.TUN(), wgconn.NewDefaultBind(), wgdevice.NewLogger(wgdevice.LogLevelSilent, ""))
+	t.Cleanup(dev.Close)
+
+	c := NewUSPConfigurerNoUAPI(dev, "wgtest0", bind.NewActivityRecorder())
+
+	key, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err, "generate device private key")
+	require.NoError(t, c.ConfigureInterface(key.String(), 0), "configure test device")
+
+	return c
+}
+
+// seedPeers adds count peers, each with a /32 overlay address, and returns their public keys.
+func seedPeers(t *testing.T, c *WGUSPConfigurer, count int) []string {
+	t.Helper()
+
+	keys := make([]string, 0, count)
+	for i := 0; i < count; i++ {
+		priv, err := wgtypes.GeneratePrivateKey()
+		require.NoError(t, err, "generate peer private key")
+		pub := priv.PublicKey().String()
+
+		addr := netip.PrefixFrom(netip.AddrFrom4([4]byte{100, 64, byte(i >> 8), byte(i)}), 32)
+		require.NoError(t, c.UpdatePeer(pub, []netip.Prefix{addr}, 25*time.Second, nil, nil), "add peer")
+		keys = append(keys, pub)
+	}
+	return keys
+}
+
+func peerAllowedIPs(t *testing.T, c *WGUSPConfigurer, peerKey string) []string {
+	t.Helper()
+
+	stats, err := c.FullStats()
+	require.NoError(t, err, "read device stats")
+
+	for _, p := range stats.Peers {
+		if p.PublicKey != peerKey {
+			continue
+		}
+		got := make([]string, 0, len(p.AllowedIPs))
+		for _, ipNet := range p.AllowedIPs {
+			got = append(got, ipNet.String())
+		}
+		return got
+	}
+	t.Fatalf("peer %s not found on device", peerKey)
+	return nil
+}
+
+// TestRemoveEndpointAddressPreservesRoutedAllowedIPs covers the prefixes the route manager
+// attaches to a routing peer through AddAllowedIP. Those are not known to the peer.Conn that
+// triggers the endpoint removal, so dropping them here would silently blackhole every route
+// behind that peer on each relay or ICE disconnect.
+func TestRemoveEndpointAddressPreservesRoutedAllowedIPs(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	peerKey := seedPeers(t, c, 3)[1]
+
+	routed := []netip.Prefix{
+		netip.MustParsePrefix("10.20.0.0/16"),
+		netip.MustParsePrefix("192.168.7.0/24"),
+	}
+	for _, prefix := range routed {
+		require.NoError(t, c.AddAllowedIP(peerKey, prefix), "add routed prefix")
+	}
+
+	before := peerAllowedIPs(t, c, peerKey)
+	require.Len(t, before, 3, "peer should hold its overlay address plus both routed prefixes")
+
+	require.NoError(t, c.RemoveEndpointAddress(peerKey), "remove endpoint address")
+
+	assert.ElementsMatch(t, before, peerAllowedIPs(t, c, peerKey),
+		"allowed IPs must survive the endpoint removal unchanged")
+}
+
+// TestRemoveEndpointAddressDoesNotScaleWithPeerCount is the regression guard for the actual
+// defect: clearing one peer's endpoint used to dump and parse the whole device, so its cost
+// grew with the size of the network map. On a routing peer with thousands of peers that dump
+// runs on every relay and ICE transition, under the interface lock.
+func TestRemoveEndpointAddressDoesNotScaleWithPeerCount(t *testing.T) {
+	measure := func(peerCount int) float64 {
+		c := newTestUSPConfigurer(t)
+		peerKey := seedPeers(t, c, peerCount)[peerCount/2]
+
+		return testing.AllocsPerRun(5, func() {
+			require.NoError(t, c.RemoveEndpointAddress(peerKey), "remove endpoint address")
+		})
+	}
+
+	small := measure(64)
+	large := measure(1024)
+
+	assert.Less(t, large, small*2,
+		"clearing one endpoint allocated %.0f objects with 1024 peers against %.0f with 64: the cost still scales with the peer count",
+		large, small)
+}
+
+// TestRemoveEndpointAddressFallsBackToDevice covers a peer the store never saw, which is what
+// an out-of-band reconfiguration of the device leaves behind. The device stays the source of
+// truth in that case, so the allowed IPs must still be preserved.
+func TestRemoveEndpointAddressFallsBackToDevice(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	peerKey := seedPeers(t, c, 3)[1]
+	require.NoError(t, c.AddAllowedIP(peerKey, netip.MustParsePrefix("10.20.0.0/16")), "add routed prefix")
+
+	before := peerAllowedIPs(t, c, peerKey)
+	c.allowedIPs.reset()
+
+	require.NoError(t, c.RemoveEndpointAddress(peerKey), "remove endpoint address")
+
+	assert.ElementsMatch(t, before, peerAllowedIPs(t, c, peerKey),
+		"allowed IPs recovered from the device must be preserved")
+
+	recovered, ok := c.allowedIPs.get(mustParseKey(t, peerKey))
+	assert.True(t, ok, "the fallback must seed the store so the next call skips the device dump")
+	assert.Len(t, recovered, 2, "seeded prefixes")
+}
+
+func TestRemoveAllowedIPKeepsTheOtherPrefixes(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	peerKey := seedPeers(t, c, 3)[0]
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+	require.NoError(t, c.AddAllowedIP(peerKey, routed), "add routed prefix")
+	require.NoError(t, c.AddAllowedIP(peerKey, netip.MustParsePrefix("192.168.7.0/24")), "add routed prefix")
+
+	require.NoError(t, c.RemoveAllowedIP(peerKey, routed), "remove routed prefix")
+
+	assert.ElementsMatch(t, []string{"100.64.0.0/32", "192.168.7.0/24"}, peerAllowedIPs(t, c, peerKey),
+		"only the removed prefix should be gone")
+
+	assert.ErrorIs(t, c.RemoveAllowedIP(peerKey, routed), ErrAllowedIPNotFound,
+		"removing a prefix that is no longer configured must be reported")
+}
+
+// TestAddAllowedIPOnAbsentPeerDoesNotResurrectIt covers the lazy connection window documented
+// in #6863: AddAllowedIP is update-only, a silent no-op when the peer is absent, so it must not
+// leave the store claiming prefixes the device never took. RemoveEndpointAddress re-adds a peer
+// without update-only, so a phantom entry would create a peer the device had dropped, and a
+// created peer would steal those allowed IPs from whichever peer legitimately holds them.
+func TestAddAllowedIPOnAbsentPeerDoesNotResurrectIt(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	seedPeers(t, c, 2)
+
+	priv, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err, "generate peer private key")
+	absent := priv.PublicKey().String()
+
+	require.NoError(t, c.AddAllowedIP(absent, netip.MustParsePrefix("10.20.0.0/16")),
+		"update-only add on an absent peer is a silent no-op")
+
+	stats, err := c.FullStats()
+	require.NoError(t, err, "read device stats")
+	require.Len(t, stats.Peers, 2, "the absent peer must not have been created by AddAllowedIP")
+
+	assert.ErrorIs(t, c.RemoveEndpointAddress(absent), ErrPeerNotFound,
+		"clearing the endpoint of a peer the device does not have must fail")
+
+	stats, err = c.FullStats()
+	require.NoError(t, err, "read device stats")
+	assert.Len(t, stats.Peers, 2, "no peer may be created while clearing an endpoint")
+}
+
+// TestRemoveEndpointAddressDoesNotStealAPrefixFromAnotherPeer covers WireGuard's rule that an
+// allowed IP belongs to exactly one peer: configuring a prefix on a peer takes it away from
+// whichever peer held it before. UpdatePeer relies on that rule rather than removing the prefix
+// from the previous holder itself, so a prefix handed over between peers must not come back.
+func TestRemoveEndpointAddressDoesNotStealAPrefixFromAnotherPeer(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	keys := seedPeers(t, c, 2)
+	peerA, peerB := keys[0], keys[1]
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+
+	require.NoError(t, c.AddAllowedIP(peerA, routed), "give the prefix to A")
+	require.Contains(t, peerAllowedIPs(t, c, peerA), routed.String(), "A must hold the prefix")
+
+	// The route moves to B. The device takes it away from A on its own.
+	require.NoError(t, c.AddAllowedIP(peerB, routed), "hand the prefix over to B")
+	require.Contains(t, peerAllowedIPs(t, c, peerB), routed.String(), "B must hold the prefix")
+	require.NotContains(t, peerAllowedIPs(t, c, peerA), routed.String(), "the device must have taken it from A")
+
+	require.NoError(t, c.RemoveEndpointAddress(peerA), "clear A's endpoint")
+
+	assert.NotContains(t, peerAllowedIPs(t, c, peerA), routed.String(),
+		"clearing A's endpoint must not take the prefix back from B")
+	assert.Contains(t, peerAllowedIPs(t, c, peerB), routed.String(),
+		"B must still hold the prefix")
+}
+
+// TestPresharedKeyCreatedPeerTakesPartInPrefixHandover covers a peer created by a preshared
+// key write rather than by a peer update. Rosenpass applies a peer's first key without
+// updateOnly, which creates the peer on the device, so a store that ignored that operation
+// would treat the peer as unknown and would not account for a prefix later handed over to it.
+func TestPresharedKeyCreatedPeerTakesPartInPrefixHandover(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	peerA := seedPeers(t, c, 1)[0]
+	routed := netip.MustParsePrefix("10.20.0.0/16")
+	require.NoError(t, c.AddAllowedIP(peerA, routed), "give the prefix to A")
+
+	priv, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err, "generate peer private key")
+	peerB := priv.PublicKey().String()
+
+	psk, err := wgtypes.GenerateKey()
+	require.NoError(t, err, "generate preshared key")
+	require.NoError(t, c.SetPresharedKey(peerB, psk, false), "a first key creates the peer")
+
+	require.NoError(t, c.AddAllowedIP(peerB, routed), "hand the prefix over to B")
+	require.Contains(t, peerAllowedIPs(t, c, peerB), routed.String(), "B must hold the prefix")
+
+	require.NoError(t, c.RemoveEndpointAddress(peerA), "clear A's endpoint")
+
+	assert.NotContains(t, peerAllowedIPs(t, c, peerA), routed.String(),
+		"clearing A's endpoint must not take the prefix back from B")
+	assert.Contains(t, peerAllowedIPs(t, c, peerB), routed.String(), "B must still hold the prefix")
+}
+
+// TestUpdatePeerDoesNotWidenAMappedPrefixOnTheDevice is the end to end form of the
+// conversion: a v4-mapped prefix must not reach the device as a zero length allowed IP,
+// which would route every v4 address to that peer.
+func TestUpdatePeerDoesNotWidenAMappedPrefixOnTheDevice(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+
+	priv, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err, "generate peer private key")
+	peerKey := priv.PublicKey().String()
+
+	mapped := netip.MustParsePrefix("::ffff:10.1.2.3/112")
+	require.NoError(t, c.UpdatePeer(peerKey, []netip.Prefix{mapped}, 25*time.Second, nil, nil), "add peer")
+
+	onDevice := peerAllowedIPs(t, c, peerKey)
+	assert.NotContains(t, onDevice, "0.0.0.0/0", "the device must not be given a catch-all allowed IP")
+	assert.Equal(t, []string{"10.1.0.0/16"}, onDevice, "the device holds the normalized prefix")
+
+	recorded, ok := c.allowedIPs.get(mustParseKey(t, peerKey))
+	require.True(t, ok, "the peer must be recorded")
+	require.Len(t, recorded, 1, "one prefix recorded")
+	assert.Equal(t, onDevice[0], recorded[0].String(), "device and store must agree")
+}
+
+// TestUpdatePeerWithAnUnusableEndpointTouchesNothing pins the ordering: the endpoint is
+// parsed before the device is configured, so a failure cannot leave the device holding a
+// peer that the store never learned about, with the prefix handover skipped along with it.
+func TestUpdatePeerWithAnUnusableEndpointTouchesNothing(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	seedPeers(t, c, 2)
+
+	priv, err := wgtypes.GeneratePrivateKey()
+	require.NoError(t, err, "generate peer private key")
+	peerKey := priv.PublicKey().String()
+
+	// A three byte address has no textual form netip can parse back.
+	endpoint := &net.UDPAddr{IP: net.IP{1, 2, 3}, Port: 51820}
+	require.Error(t, c.UpdatePeer(peerKey, []netip.Prefix{netip.MustParsePrefix("10.30.0.0/16")},
+		25*time.Second, endpoint, nil), "an unusable endpoint must fail the update")
+
+	stats, err := c.FullStats()
+	require.NoError(t, err, "read device stats")
+	assert.Len(t, stats.Peers, 2, "the peer must not have reached the device")
+
+	_, ok := c.allowedIPs.get(mustParseKey(t, peerKey))
+	assert.False(t, ok, "the peer must not have been recorded either")
+}
+
+// TestRemovePeerKeepsTheRecordWhenTheDeviceRefuses covers a removal that never reached the
+// device. A single peer removal is one write, so a failure leaves the peer on the device
+// exactly as it was, and the record still describes it; dropping it would only force the
+// next caller to read the whole device back for an answer it already had.
+func TestRemovePeerKeepsTheRecordWhenTheDeviceRefuses(t *testing.T) {
+	c := newTestUSPConfigurer(t)
+	peerKey := seedPeers(t, c, 1)[0]
+	require.NoError(t, c.AddAllowedIP(peerKey, netip.MustParsePrefix("10.20.0.0/16")), "add routed prefix")
+
+	before, ok := c.allowedIPs.get(mustParseKey(t, peerKey))
+	require.True(t, ok, "the peer must be recorded before the removal")
+	require.Len(t, before, 2, "overlay address plus routed prefix")
+
+	// A closed device refuses every write, which is the shape of any failed removal.
+	c.device.Close()
+
+	require.Error(t, c.RemovePeer(peerKey), "the removal must report the failure")
+
+	after, ok := c.allowedIPs.get(mustParseKey(t, peerKey))
+	require.True(t, ok, "a peer still on the device must stay recorded")
+	assert.Equal(t, before, after, "the record must describe the peer the device kept")
+}
+
+// mustParseKey turns the textual key the configurer API takes into the form the store
+// keys on.
+func mustParseKey(t *testing.T, key string) wgtypes.Key {
+	t.Helper()
+
+	parsed, err := wgtypes.ParseKey(key)
+	require.NoError(t, err, "parse peer key")
+	return parsed
+}

client/iface/iface.go

@@ -51,7 +51,6 @@ func ValidateMTU(mtu uint16) error {
 
 type wgProxyFactory interface {
 	GetProxy() wgproxy.Proxy
-	GetProxyPort() uint16
 	Free() error
 }
 
@@ -81,12 +80,6 @@ func (w *WGIface) GetProxy() wgproxy.Proxy {
 	return w.wgProxyFactory.GetProxy()
 }
 
-// GetProxyPort returns the proxy port used by the WireGuard proxy.
-// Returns 0 if no proxy port is used (e.g., for userspace WireGuard).
-func (w *WGIface) GetProxyPort() uint16 {
-	return w.wgProxyFactory.GetProxyPort()
-}
-
 // GetBind returns the EndpointManager userspace bind mode.
 func (w *WGIface) GetBind() device.EndpointManager {
 	w.mu.Lock()

client/iface/iface_close_test.go

@@ -52,7 +52,6 @@ func (f *fakeTunDevice) Close() error {
 type fakeProxyFactory struct{}
 
 func (fakeProxyFactory) GetProxy() wgproxy.Proxy { return nil }
-func (fakeProxyFactory) GetProxyPort() uint16    { return 0 }
 func (fakeProxyFactory) Free() error             { return nil }
 
 // TestWGIface_CloseReleasesMutexBeforeTunClose guards against a deadlock

client/iface/iface_destroy_windows.go

@@ -6,27 +6,14 @@ import (
 	"fmt"
 	"os/exec"
 
-	log "github.com/sirupsen/logrus"
+	"github.com/netbirdio/netbird/client/internal/wincmd"
 )
 
 func (w *WGIface) Destroy() error {
-	netshCmd := GetSystem32Command("netsh")
+	netshCmd := wincmd.System32("netsh")
 	out, err := exec.Command(netshCmd, "interface", "set", "interface", w.Name(), "admin=disable").CombinedOutput()
 	if err != nil {
 		return fmt.Errorf("failed to remove interface %s: %w - %s", w.Name(), err, out)
 	}
 	return nil
 }
-
-// GetSystem32Command checks if a command can be found in the system path and returns it. In case it can't find it
-// in the path it will return the full path of a command assuming C:\windows\system32 as the base path.
-func GetSystem32Command(command string) string {
-	_, err := exec.LookPath(command)
-	if err == nil {
-		return command
-	}
-
-	log.Tracef("Command %s not found in PATH, using C:\\windows\\system32\\%s.exe path", command, command)
-
-	return "C:\\windows\\system32\\" + command + ".exe"
-}

client/iface/iface_test.go

@@ -568,11 +568,14 @@ func Test_ConnectPeers(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	// The peers use userspace WireGuard (stdnet transport). A tight busy-loop
-	// here starves the wireguard-go goroutines that process the handshake, so
-	// poll on a ticker instead and yield the CPU between checks. WireGuard also
-	// only retries a lost handshake initiation every REKEY_TIMEOUT (5s), which
-	// is why the overall wait can occasionally stretch to tens of seconds.
+	// On Linux with the kernel module both peers are kernel devices, elsewhere
+	// they run on wireguard-go. A tight busy-loop here would starve the
+	// wireguard-go goroutines that process the handshake, so poll on a ticker
+	// instead and yield the CPU between checks. WireGuard also only retries a
+	// lost handshake initiation every REKEY_TIMEOUT (5s), which is why the
+	// overall wait can occasionally stretch to tens of seconds. Each side sends
+	// its first initiation when its peer is configured, and the first one leaves
+	// before the other device knows the peer, so that one is always wasted.
 	timeout := 30 * time.Second
 	timeoutChannel := time.After(timeout)
 	ticker := time.NewTicker(500 * time.Millisecond)
@@ -590,13 +593,26 @@ func Test_ConnectPeers(t *testing.T) {
 
 		select {
 		case <-timeoutChannel:
-			t.Fatalf("waiting for peer handshake timeout after %s", timeout.String())
+			// The counters tell whether initiations were sent at all, whether they
+			// arrived, and whether only one direction is working.
+			t.Fatalf("waiting for peer handshake timeout after %s\n%s\n%s", timeout.String(),
+				describePeer(peer1ifaceName, peer2Key.PublicKey().String()),
+				describePeer(peer2ifaceName, peer1Key.PublicKey().String()))
 		case <-ticker.C:
 		}
 	}
 
 }
 
+func describePeer(ifaceName, peerPubKey string) string {
+	peer, err := getPeer(ifaceName, peerPubKey)
+	if err != nil {
+		return fmt.Sprintf("%s: peer %s: %v", ifaceName, peerPubKey, err)
+	}
+	return fmt.Sprintf("%s: peer %s endpoint=%v tx=%d rx=%d last_handshake=%v",
+		ifaceName, peerPubKey, peer.Endpoint, peer.TransmitBytes, peer.ReceiveBytes, peer.LastHandshakeTime)
+}
+
 func getPeer(ifaceName, peerPubKey string) (wgtypes.Peer, error) {
 	wg, err := wgctrl.New()
 	if err != nil {

client/iface/wgproxy/ebpf/portlookup.go

@@ -1,32 +0,0 @@
-package ebpf
-
-import (
-	"fmt"
-	"net"
-)
-
-var (
-	portRangeStart = 3128
-	portRangeEnd   = portRangeStart + 100
-)
-
-type portLookup struct {
-}
-
-func (pl portLookup) searchFreePort() (int, error) {
-	for i := portRangeStart; i <= portRangeEnd; i++ {
-		if pl.tryToBind(i) == nil {
-			return i, nil
-		}
-	}
-	return 0, fmt.Errorf("failed to bind free port for eBPF proxy")
-}
-
-func (pl portLookup) tryToBind(port int) error {
-	l, err := net.ListenPacket("udp", fmt.Sprintf(":%d", port))
-	if err != nil {
-		return err
-	}
-	_ = l.Close()
-	return nil
-}

client/iface/wgproxy/ebpf/portlookup_test.go

@@ -1,45 +0,0 @@
-package ebpf
-
-import (
-	"fmt"
-	"net"
-	"testing"
-)
-
-func Test_portLookup_searchFreePort(t *testing.T) {
-	pl := portLookup{}
-	_, err := pl.searchFreePort()
-	if err != nil {
-		t.Fatal(err)
-	}
-}
-
-func Test_portLookup_on_allocated(t *testing.T) {
-	pl := portLookup{}
-
-	portRangeStart = 4128
-	portRangeEnd = portRangeStart + 100
-
-	allocatedPort, err := allocatePort(portRangeStart)
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer allocatedPort.Close()
-
-	fp, err := pl.searchFreePort()
-	if err != nil {
-		t.Fatal(err)
-	}
-
-	if fp != (portRangeStart + 1) {
-		t.Errorf("invalid free port, expected: %d, got: %d", portRangeStart+1, fp)
-	}
-}
-
-func allocatePort(port int) (net.PacketConn, error) {
-	c, err := net.ListenPacket("udp", fmt.Sprintf(":%d", port))
-	if err != nil {
-		return nil, err
-	}
-	return c, err
-}

client/iface/wgproxy/ebpf/proxy.go

@@ -1,243 +0,0 @@
-//go:build linux && !android
-
-package ebpf
-
-import (
-	"context"
-	"fmt"
-	"net"
-	"sync"
-
-	"github.com/hashicorp/go-multierror"
-	"github.com/pion/transport/v3"
-	log "github.com/sirupsen/logrus"
-
-	nberrors "github.com/netbirdio/netbird/client/errors"
-	"github.com/netbirdio/netbird/client/iface/bufsize"
-	"github.com/netbirdio/netbird/client/iface/wgproxy/rawsocket"
-	"github.com/netbirdio/netbird/client/internal/ebpf"
-	ebpfMgr "github.com/netbirdio/netbird/client/internal/ebpf/manager"
-	nbnet "github.com/netbirdio/netbird/client/net"
-)
-
-const (
-	loopbackAddr = "127.0.0.1"
-)
-
-// WGEBPFProxy definition for proxy with EBPF support
-type WGEBPFProxy struct {
-	localWGListenPort int
-	proxyPort         int
-	mtu               uint16
-
-	ebpfManager      ebpfMgr.Manager
-	relayedConnStore map[uint16]net.Conn
-	relayedConnMutex sync.Mutex
-
-	lastUsedPort uint16
-	rawConnIPv4  net.PacketConn
-	rawConnIPv6  net.PacketConn
-	conn         transport.UDPConn
-
-	ctx       context.Context
-	ctxCancel context.CancelFunc
-}
-
-// NewWGEBPFProxy create new WGEBPFProxy instance
-func NewWGEBPFProxy(wgPort int, mtu uint16) *WGEBPFProxy {
-	log.Debugf("instantiate ebpf proxy")
-	wgProxy := &WGEBPFProxy{
-		localWGListenPort: wgPort,
-		mtu:               mtu,
-		ebpfManager:       ebpf.GetEbpfManagerInstance(),
-		relayedConnStore:  make(map[uint16]net.Conn),
-	}
-	return wgProxy
-}
-
-// Listen load ebpf program and listen the proxy
-func (p *WGEBPFProxy) Listen() error {
-	pl := portLookup{}
-	proxyPort, err := pl.searchFreePort()
-	if err != nil {
-		return err
-	}
-	p.proxyPort = proxyPort
-
-	// Prepare IPv4 raw socket (required)
-	p.rawConnIPv4, err = rawsocket.PrepareSenderRawSocketIPv4()
-	if err != nil {
-		return err
-	}
-
-	// Prepare IPv6 raw socket (optional)
-	p.rawConnIPv6, err = rawsocket.PrepareSenderRawSocketIPv6()
-	if err != nil {
-		log.Warnf("failed to prepare IPv6 raw socket, continuing with IPv4 only: %v", err)
-	}
-
-	err = p.ebpfManager.LoadWgProxy(proxyPort, p.localWGListenPort)
-	if err != nil {
-		if closeErr := p.rawConnIPv4.Close(); closeErr != nil {
-			log.Warnf("failed to close IPv4 raw socket: %v", closeErr)
-		}
-		if p.rawConnIPv6 != nil {
-			if closeErr := p.rawConnIPv6.Close(); closeErr != nil {
-				log.Warnf("failed to close IPv6 raw socket: %v", closeErr)
-			}
-		}
-		return err
-	}
-
-	addr := net.UDPAddr{
-		Port: proxyPort,
-		IP:   net.ParseIP(loopbackAddr),
-	}
-
-	p.ctx, p.ctxCancel = context.WithCancel(context.Background())
-
-	conn, err := nbnet.ListenUDP("udp", &addr)
-	if err != nil {
-		if cErr := p.Free(); cErr != nil {
-			log.Errorf("Failed to close the wgproxy: %s", cErr)
-		}
-		return err
-	}
-	p.conn = conn
-
-	go p.proxyToRemote()
-	log.Infof("local wg proxy listening on: %d", proxyPort)
-	return nil
-}
-
-// AddRelayedConn add new relayed connection for the proxy
-func (p *WGEBPFProxy) AddRelayedConn(relayedConn net.Conn) (*net.UDPAddr, error) {
-	wgEndpointPort, err := p.storeRelayedConn(relayedConn)
-	if err != nil {
-		return nil, err
-	}
-
-	log.Infof("relayed conn added to wg proxy store: %s, endpoint port: :%d", relayedConn.RemoteAddr(), wgEndpointPort)
-
-	wgEndpoint := &net.UDPAddr{
-		IP:   net.ParseIP(loopbackAddr),
-		Port: int(wgEndpointPort),
-	}
-	return wgEndpoint, nil
-}
-
-// Free resources except the remoteConns will be keep open.
-func (p *WGEBPFProxy) Free() error {
-	log.Debugf("free up ebpf wg proxy")
-	if p.ctx != nil && p.ctx.Err() != nil {
-		//nolint
-		return nil
-	}
-
-	p.ctxCancel()
-
-	var result *multierror.Error
-	if p.conn != nil {
-		if err := p.conn.Close(); err != nil {
-			result = multierror.Append(result, err)
-		}
-	}
-
-	if err := p.ebpfManager.FreeWGProxy(); err != nil {
-		result = multierror.Append(result, err)
-	}
-
-	if p.rawConnIPv4 != nil {
-		if err := p.rawConnIPv4.Close(); err != nil {
-			result = multierror.Append(result, err)
-		}
-	}
-
-	if p.rawConnIPv6 != nil {
-		if err := p.rawConnIPv6.Close(); err != nil {
-			result = multierror.Append(result, err)
-		}
-	}
-	return nberrors.FormatErrorOrNil(result)
-}
-
-// GetProxyPort returns the proxy listening port.
-func (p *WGEBPFProxy) GetProxyPort() uint16 {
-	return uint16(p.proxyPort)
-}
-
-// proxyToRemote read messages from local WireGuard interface and forward it to remote conn
-// From this go routine has only one instance.
-func (p *WGEBPFProxy) proxyToRemote() {
-	buf := make([]byte, p.mtu+bufsize.WGBufferOverhead)
-	for p.ctx.Err() == nil {
-		if err := p.readAndForwardPacket(buf); err != nil {
-			if p.ctx.Err() != nil {
-				return
-			}
-			log.Errorf("failed to proxy packet to remote conn: %s", err)
-		}
-	}
-}
-
-func (p *WGEBPFProxy) readAndForwardPacket(buf []byte) error {
-	n, addr, err := p.conn.ReadFromUDP(buf)
-	if err != nil {
-		return fmt.Errorf("failed to read UDP packet from WG: %w", err)
-	}
-
-	p.relayedConnMutex.Lock()
-	conn, ok := p.relayedConnStore[uint16(addr.Port)]
-	p.relayedConnMutex.Unlock()
-	if !ok {
-		if p.ctx.Err() == nil {
-			log.Debugf("relayed conn not found by port because conn already has been closed: %d", addr.Port)
-		}
-		return nil
-	}
-
-	if _, err := conn.Write(buf[:n]); err != nil {
-		return fmt.Errorf("forward local WG packet (%d) to remote relayed conn: %w", addr.Port, err)
-	}
-	return nil
-}
-
-func (p *WGEBPFProxy) storeRelayedConn(relayedConn net.Conn) (uint16, error) {
-	p.relayedConnMutex.Lock()
-	defer p.relayedConnMutex.Unlock()
-
-	np, err := p.nextFreePort()
-	if err != nil {
-		return np, err
-	}
-	p.relayedConnStore[np] = relayedConn
-	return np, nil
-}
-
-func (p *WGEBPFProxy) removeRelayedConn(relayedConnID uint16) {
-	p.relayedConnMutex.Lock()
-	defer p.relayedConnMutex.Unlock()
-
-	_, ok := p.relayedConnStore[relayedConnID]
-	if ok {
-		log.Debugf("remove relayed conn from store by port: %d", relayedConnID)
-	}
-	delete(p.relayedConnStore, relayedConnID)
-}
-
-func (p *WGEBPFProxy) nextFreePort() (uint16, error) {
-	if len(p.relayedConnStore) == 65535 {
-		return 0, fmt.Errorf("reached maximum relayed connection numbers")
-	}
-generatePort:
-	if p.lastUsedPort == 65535 {
-		p.lastUsedPort = 1
-	} else {
-		p.lastUsedPort++
-	}
-
-	if _, ok := p.relayedConnStore[p.lastUsedPort]; ok {
-		goto generatePort
-	}
-	return p.lastUsedPort, nil
-}

client/iface/wgproxy/ebpf/proxy_test.go

@@ -1,56 +0,0 @@
-//go:build linux && !android
-
-package ebpf
-
-import (
-	"testing"
-)
-
-func TestWGEBPFProxy_connStore(t *testing.T) {
-	wgProxy := NewWGEBPFProxy(1, 1280)
-
-	p, _ := wgProxy.storeRelayedConn(nil)
-	if p != 1 {
-		t.Errorf("invalid initial port: %d", wgProxy.lastUsedPort)
-	}
-
-	numOfConns := 10
-	for i := 0; i < numOfConns; i++ {
-		p, _ = wgProxy.storeRelayedConn(nil)
-	}
-	if p != uint16(numOfConns)+1 {
-		t.Errorf("invalid last used port: %d, expected: %d", p, numOfConns+1)
-	}
-	if len(wgProxy.relayedConnStore) != numOfConns+1 {
-		t.Errorf("invalid store size: %d, expected: %d", len(wgProxy.relayedConnStore), numOfConns+1)
-	}
-}
-
-func TestWGEBPFProxy_portCalculation_overflow(t *testing.T) {
-	wgProxy := NewWGEBPFProxy(1, 1280)
-
-	_, _ = wgProxy.storeRelayedConn(nil)
-	wgProxy.lastUsedPort = 65535
-	p, _ := wgProxy.storeRelayedConn(nil)
-
-	if len(wgProxy.relayedConnStore) != 2 {
-		t.Errorf("invalid store size: %d, expected: %d", len(wgProxy.relayedConnStore), 2)
-	}
-
-	if p != 2 {
-		t.Errorf("invalid last used port: %d, expected: %d", p, 2)
-	}
-}
-
-func TestWGEBPFProxy_portCalculation_maxConn(t *testing.T) {
-	wgProxy := NewWGEBPFProxy(1, 1280)
-
-	for i := 0; i < 65535; i++ {
-		_, _ = wgProxy.storeRelayedConn(nil)
-	}
-
-	_, err := wgProxy.storeRelayedConn(nil)
-	if err == nil {
-		t.Errorf("invalid relayed conn store calculation")
-	}
-}

client/iface/wgproxy/factory_kernel.go

@@ -8,11 +8,13 @@ import (
 
 	log "github.com/sirupsen/logrus"
 
-	"github.com/netbirdio/netbird/client/iface/wgproxy/ebpf"
+	"github.com/netbirdio/netbird/client/iface/wgproxy/loopback"
 	udpProxy "github.com/netbirdio/netbird/client/iface/wgproxy/udp"
 )
 
 const (
+	envDisableKernelWGProxy = "NB_DISABLE_KERNEL_WG_PROXY"
+	// envDisableEBPFWGProxy is a deprecated alias for envDisableKernelWGProxy.
 	envDisableEBPFWGProxy = "NB_DISABLE_EBPF_WG_PROXY"
 )
 
@@ -20,7 +22,7 @@ type KernelFactory struct {
 	wgPort int
 	mtu    uint16
 
-	ebpfProxy *ebpf.WGEBPFProxy
+	loopbackProxy *loopback.Proxy
 }
 
 func NewKernelFactory(wgPort int, mtu uint16) *KernelFactory {
@@ -29,55 +31,56 @@ func NewKernelFactory(wgPort int, mtu uint16) *KernelFactory {
 		mtu:    mtu,
 	}
 
-	if isEBPFDisabled() {
+	if isKernelProxyDisabled() {
 		log.Infof("WireGuard Proxy Factory will produce UDP proxy")
-		log.Infof("eBPF WireGuard proxy is disabled via %s environment variable", envDisableEBPFWGProxy)
 		return f
 	}
 
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, mtu)
-	if err := ebpfProxy.Listen(); err != nil {
+	loopbackProxy := loopback.NewProxy(wgPort, mtu)
+	if err := loopbackProxy.Listen(); err != nil {
 		log.Infof("WireGuard Proxy Factory will produce UDP proxy")
-		log.Warnf("failed to initialize ebpf proxy, fallback to user space proxy: %s", err)
+		log.Warnf("failed to initialize loopback proxy, fallback to user space proxy: %s", err)
 		return f
 	}
-	log.Infof("WireGuard Proxy Factory will produce eBPF proxy")
-	f.ebpfProxy = ebpfProxy
+	log.Infof("WireGuard Proxy Factory will produce loopback proxy")
+	f.loopbackProxy = loopbackProxy
 	return f
 }
 
 func (w *KernelFactory) GetProxy() Proxy {
-	if w.ebpfProxy == nil {
+	if w.loopbackProxy == nil {
 		return udpProxy.NewWGUDPProxy(w.wgPort, w.mtu)
 	}
 
-	return ebpf.NewProxyWrapper(w.ebpfProxy)
-}
-
-// GetProxyPort returns the eBPF proxy port, or 0 if eBPF is not active.
-func (w *KernelFactory) GetProxyPort() uint16 {
-	if w.ebpfProxy == nil {
-		return 0
-	}
-	return w.ebpfProxy.GetProxyPort()
+	return loopback.NewProxyWrapper(w.loopbackProxy)
 }
 
 func (w *KernelFactory) Free() error {
-	if w.ebpfProxy == nil {
+	if w.loopbackProxy == nil {
 		return nil
 	}
-	return w.ebpfProxy.Free()
+	return w.loopbackProxy.Free()
 }
 
-func isEBPFDisabled() bool {
-	val := os.Getenv(envDisableEBPFWGProxy)
+func isKernelProxyDisabled() bool {
+	env := envDisableKernelWGProxy
+	val := os.Getenv(env)
+	if val == "" {
+		env = envDisableEBPFWGProxy
+		val = os.Getenv(env)
+	}
 	if val == "" {
 		return false
 	}
+
 	disabled, err := strconv.ParseBool(val)
 	if err != nil {
-		log.Warnf("failed to parse %s: %v", envDisableEBPFWGProxy, err)
+		log.Warnf("failed to parse %s: %v", env, err)
 		return false
 	}
+
+	if disabled {
+		log.Infof("kernel WireGuard proxy is disabled via %s", env)
+	}
 	return disabled
 }

client/iface/wgproxy/factory_usp.go

@@ -24,11 +24,6 @@ func (w *USPFactory) GetProxy() Proxy {
 	return proxyBind.NewProxyBind(w.bind, w.mtu)
 }
 
-// GetProxyPort returns 0 as userspace WireGuard doesn't use a separate proxy port.
-func (w *USPFactory) GetProxyPort() uint16 {
-	return 0
-}
-
 func (w *USPFactory) Free() error {
 	return nil
 }

client/iface/wgproxy/loopback/addr.go

@@ -0,0 +1,70 @@
+//go:build linux && !android
+
+package loopback
+
+import (
+	"fmt"
+	"net/netip"
+)
+
+// Peer endpoints live in the upper half of 127.0.0.0/8. Everything in that
+// range is delivered to the loopback device without any address or route being
+// configured, and staying out of 127.0.0.0/9 keeps well-known squatters such as
+// 127.0.0.53 (systemd-resolved) and 127.0.1.1 out of the way.
+const (
+	addrRangeBase   uint32 = 0x7f800000 // 127.128.0.0
+	addrRangeSize   uint32 = 1 << 23    // /9
+	addrRangePrefix        = "127.128.0.0/9"
+)
+
+// allocator hands out one loopback address per relayed connection. The address
+// is the peer's identity: WireGuard sends to it, and the proxy recovers which
+// peer a packet belongs to from the destination address.
+type allocator struct {
+	cursor uint32
+}
+
+// next returns the first free address at or after the cursor, wrapping once.
+// inUse reports whether an address is already handed out.
+func (a *allocator) next(inUse func(netip.Addr) bool) (netip.Addr, error) {
+	for i := uint32(0); i < addrRangeSize; i++ {
+		a.cursor = (a.cursor + 1) % addrRangeSize
+		addr := addrFromOffset(a.cursor)
+		if !addr.IsValid() {
+			continue
+		}
+		if inUse(addr) {
+			continue
+		}
+		return addr, nil
+	}
+	return netip.Addr{}, fmt.Errorf("no free endpoint address in %s", addrRangePrefix)
+}
+
+// addrFromOffset maps an offset in the range to an address, skipping the .0 and
+// .255 hosts. They are unremarkable on loopback, but tools and firewall rules
+// tend to treat them as network and broadcast addresses.
+func addrFromOffset(offset uint32) netip.Addr {
+	last := offset & 0xff
+	if last == 0 || last == 0xff {
+		return netip.Addr{}
+	}
+
+	v := addrRangeBase + offset
+	return netip.AddrFrom4([4]byte{
+		byte(v >> 24),
+		byte(v >> 16),
+		byte(v >> 8),
+		byte(v),
+	})
+}
+
+// inRange reports whether addr is one this proxy could have handed out.
+func inRange(addr netip.Addr) bool {
+	if !addr.Is4() {
+		return false
+	}
+	b := addr.As4()
+	v := uint32(b[0])<<24 | uint32(b[1])<<16 | uint32(b[2])<<8 | uint32(b[3])
+	return v >= addrRangeBase && v < addrRangeBase+addrRangeSize && b[3] != 0 && b[3] != 0xff
+}

client/iface/wgproxy/loopback/addr_test.go

@@ -0,0 +1,114 @@
+//go:build linux && !android
+
+package loopback
+
+import (
+	"net/netip"
+	"testing"
+)
+
+func TestAllocatorHandsOutDistinctAddresses(t *testing.T) {
+	var a allocator
+	taken := make(map[netip.Addr]bool)
+
+	for i := 0; i < 1000; i++ {
+		addr, err := a.next(func(candidate netip.Addr) bool { return taken[candidate] })
+		if err != nil {
+			t.Fatalf("allocate %d: %v", i, err)
+		}
+		if taken[addr] {
+			t.Fatalf("address %s handed out twice", addr)
+		}
+		if !inRange(addr) {
+			t.Fatalf("address %s outside %s", addr, addrRangePrefix)
+		}
+		taken[addr] = true
+	}
+}
+
+func TestAllocatorSkipsNetworkAndBroadcastHosts(t *testing.T) {
+	var a allocator
+	taken := make(map[netip.Addr]bool)
+
+	// enough allocations to walk past a .255/.0 boundary
+	for i := 0; i < 600; i++ {
+		addr, err := a.next(func(candidate netip.Addr) bool { return taken[candidate] })
+		if err != nil {
+			t.Fatalf("allocate %d: %v", i, err)
+		}
+		last := addr.As4()[3]
+		if last == 0 || last == 255 {
+			t.Fatalf("address %s ends in .%d", addr, last)
+		}
+		taken[addr] = true
+	}
+}
+
+func TestAllocatorReusesReleasedAddresses(t *testing.T) {
+	var a allocator
+	taken := make(map[netip.Addr]bool)
+	inUse := func(candidate netip.Addr) bool { return taken[candidate] }
+	alloc := func() netip.Addr {
+		t.Helper()
+		addr, err := a.next(inUse)
+		if err != nil {
+			t.Fatalf("allocate: %v", err)
+		}
+		taken[addr] = true
+		return addr
+	}
+
+	first := alloc()
+	second := alloc()
+	delete(taken, first)
+
+	// The cursor only moves forward, so a released address comes back after a
+	// wrap. Park the cursor near the end of the range instead of allocating
+	// 2^23 addresses: the next call takes the last usable address, and the one
+	// after that wraps past the skipped .255 and .0 hosts to the released one.
+	a.cursor = addrRangeSize - 3
+	last := alloc()
+	if want := netip.MustParseAddr("127.255.255.254"); last != want {
+		t.Fatalf("expected the last usable address %s before the wrap, got %s", want, last)
+	}
+
+	if reused := alloc(); reused != first {
+		t.Fatalf("expected the released address %s after the wrap, got %s", first, reused)
+	}
+
+	// second is still held, so the allocator must step over it.
+	if next := alloc(); next == second {
+		t.Fatalf("allocator handed out %s while it was still in use", second)
+	}
+}
+
+func TestInRange(t *testing.T) {
+	tests := []struct {
+		addr string
+		want bool
+	}{
+		{"127.128.0.1", true},
+		{"127.255.255.254", true},
+		{"127.128.0.0", false},     // network host, never handed out
+		{"127.128.5.255", false},   // broadcast host, never handed out
+		{"127.127.255.255", false}, // below the range, where 127.0.0.53 and friends live
+		{"127.0.0.1", false},
+		{"127.0.0.53", false},
+		{"127.0.1.1", false},
+		{"128.0.0.1", false},
+		{"10.0.0.1", false},
+	}
+
+	for _, tc := range tests {
+		addr := netip.MustParseAddr(tc.addr)
+		if got := inRange(addr); got != tc.want {
+			t.Errorf("inRange(%s) = %v, want %v", tc.addr, got, tc.want)
+		}
+	}
+}
+
+func TestInRangeIgnoresIPv6(t *testing.T) {
+	if inRange(netip.MustParseAddr("::1")) {
+		t.Error("inRange(::1) = true, want false")
+	}
+}

client/iface/wgproxy/loopback/proxy.go

@@ -0,0 +1,291 @@
+//go:build linux && !android
+
+package loopback
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"net/netip"
+	"sync"
+	"syscall"
+
+	"github.com/hashicorp/go-multierror"
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/net/ipv4"
+	"golang.org/x/sys/unix"
+
+	nberrors "github.com/netbirdio/netbird/client/errors"
+	"github.com/netbirdio/netbird/client/iface/bufsize"
+	"github.com/netbirdio/netbird/client/iface/wgproxy/rawsocket"
+)
+
+const (
+	loopbackDevice = "lo"
+
+	portRangeStart = 3128
+	portRangeEnd   = portRangeStart + 100
+)
+
+// Proxy forwards packets between relayed connections and a local kernel
+// WireGuard instance. Every relayed peer gets its own loopback address as its
+// WireGuard endpoint, so a single socket serves all of them: the destination
+// address of an incoming packet identifies the peer.
+type Proxy struct {
+	localWGListenPort int
+	mtu               uint16
+	proxyPort         int
+
+	conn        *net.UDPConn
+	packetConn  *ipv4.PacketConn
+	loIndex     int
+	rawConnIPv4 net.PacketConn
+	rawConnIPv6 net.PacketConn
+
+	relayedConnMutex sync.Mutex
+	relayedConnStore map[netip.Addr]net.Conn
+	addrs            allocator
+
+	ctx       context.Context
+	ctxCancel context.CancelFunc
+}
+
+// NewProxy creates a proxy for the WireGuard instance listening on wgPort.
+func NewProxy(wgPort int, mtu uint16) *Proxy {
+	log.Debugf("instantiate loopback wg proxy")
+	return &Proxy{
+		localWGListenPort: wgPort,
+		mtu:               mtu,
+		relayedConnStore:  make(map[netip.Addr]net.Conn),
+	}
+}
+
+// Listen opens the shared socket and starts forwarding WireGuard packets to the
+// relayed connections.
+func (p *Proxy) Listen() error {
+	rawConnIPv4, err := rawsocket.PrepareSenderRawSocketIPv4()
+	if err != nil {
+		return fmt.Errorf("prepare IPv4 raw socket: %w", err)
+	}
+	p.rawConnIPv4 = rawConnIPv4
+
+	p.rawConnIPv6, err = rawsocket.PrepareSenderRawSocketIPv6()
+	if err != nil {
+		log.Warnf("failed to prepare IPv6 raw socket, continuing with IPv4 only: %v", err)
+	}
+
+	loopback, err := net.InterfaceByName(loopbackDevice)
+	if err != nil {
+		if freeErr := p.Free(); freeErr != nil {
+			log.Errorf("failed to free the wgproxy: %s", freeErr)
+		}
+		return fmt.Errorf("look up %s: %w", loopbackDevice, err)
+	}
+	p.loIndex = loopback.Index
+
+	if err := p.listen(); err != nil {
+		if freeErr := p.Free(); freeErr != nil {
+			log.Errorf("failed to free the wgproxy: %s", freeErr)
+		}
+		return err
+	}
+
+	p.ctx, p.ctxCancel = context.WithCancel(context.Background())
+
+	go p.proxyToRemote()
+	log.Infof("local wg proxy listening on %s:%d", addrRangePrefix, p.proxyPort)
+	return nil
+}
+
+// listen binds the shared socket on the first free port of the range. The bind
+// has to be a wildcard one to receive every peer address in the range, so it is
+// restricted to the loopback device: without that the port would be reachable
+// on every interface.
+func (p *Proxy) listen() error {
+	var lastErr error
+	for port := portRangeStart; port <= portRangeEnd; port++ {
+		err := p.listenOn(port)
+		if err == nil {
+			p.proxyPort = port
+			return nil
+		}
+		lastErr = err
+	}
+	return fmt.Errorf("bind proxy port in range %d-%d: %w", portRangeStart, portRangeEnd, lastErr)
+}
+
+func (p *Proxy) listenOn(proxyPort int) error {
+	lc := net.ListenConfig{
+		Control: func(_, _ string, c syscall.RawConn) error {
+			var sockErr error
+			if err := c.Control(func(fd uintptr) {
+				if err := unix.SetsockoptString(int(fd), unix.SOL_SOCKET, unix.SO_BINDTODEVICE, loopbackDevice); err != nil {
+					sockErr = fmt.Errorf("bind to %s: %w", loopbackDevice, err)
+					return
+				}
+			}); err != nil {
+				return fmt.Errorf("control socket: %w", err)
+			}
+			return sockErr
+		},
+	}
+
+	conn, err := lc.ListenPacket(context.Background(), "udp4", fmt.Sprintf(":%d", proxyPort))
+	if err != nil {
+		return fmt.Errorf("listen on :%d: %w", proxyPort, err)
+	}
+
+	udpConn, ok := conn.(*net.UDPConn)
+	if !ok {
+		if closeErr := conn.Close(); closeErr != nil {
+			log.Errorf("failed to close proxy conn: %s", closeErr)
+		}
+		return fmt.Errorf("unexpected conn type %T", conn)
+	}
+
+	packetConn := ipv4.NewPacketConn(udpConn)
+	// the destination address carries the peer identity, the interface index is
+	// checked on receive as a second line of defense behind SO_BINDTODEVICE
+	if err := packetConn.SetControlMessage(ipv4.FlagDst|ipv4.FlagInterface, true); err != nil {
+		if closeErr := udpConn.Close(); closeErr != nil {
+			log.Errorf("failed to close proxy conn: %s", closeErr)
+		}
+		return fmt.Errorf("request destination address: %w", err)
+	}
+
+	p.conn = udpConn
+	p.packetConn = packetConn
+	return nil
+}
+
+// AddRelayedConn assigns an endpoint address to the relayed connection and
+// returns the address WireGuard should send to, along with the key the
+// connection is stored under.
+func (p *Proxy) AddRelayedConn(relayedConn net.Conn) (*net.UDPAddr, netip.Addr, error) {
+	addr, err := p.storeRelayedConn(relayedConn)
+	if err != nil {
+		return nil, netip.Addr{}, err
+	}
+
+	log.Infof("relayed conn added to wg proxy store: %s, endpoint address: %s", relayedConn.RemoteAddr(), addr)
+
+	return &net.UDPAddr{
+		IP:   addr.AsSlice(),
+		Port: p.proxyPort,
+	}, addr, nil
+}
+
+// Free releases the proxy resources. The relayed connections are left open.
+func (p *Proxy) Free() error {
+	log.Debugf("free up loopback wg proxy")
+	if p.ctx != nil && p.ctx.Err() != nil {
+		//nolint
+		return nil
+	}
+
+	if p.ctxCancel != nil {
+		p.ctxCancel()
+	}
+
+	var result *multierror.Error
+	if p.conn != nil {
+		if err := p.conn.Close(); err != nil {
+			result = multierror.Append(result, err)
+		}
+	}
+
+	if p.rawConnIPv4 != nil {
+		if err := p.rawConnIPv4.Close(); err != nil {
+			result = multierror.Append(result, err)
+		}
+	}
+
+	if p.rawConnIPv6 != nil {
+		if err := p.rawConnIPv6.Close(); err != nil {
+			result = multierror.Append(result, err)
+		}
+	}
+	return nberrors.FormatErrorOrNil(result)
+}
+
+// proxyToRemote reads packets from the local WireGuard instance and forwards
+// them to the relayed connection the destination address belongs to.
+func (p *Proxy) proxyToRemote() {
+	buf := make([]byte, p.mtu+bufsize.WGBufferOverhead)
+	for p.ctx.Err() == nil {
+		if err := p.readAndForwardPacket(buf); err != nil {
+			if p.ctx.Err() != nil {
+				return
+			}
+			log.Errorf("failed to proxy packet to remote conn: %s", err)
+		}
+	}
+}
+
+func (p *Proxy) readAndForwardPacket(buf []byte) error {
+	n, cm, _, err := p.packetConn.ReadFrom(buf)
+	if err != nil {
+		return fmt.Errorf("read UDP packet from WG: %w", err)
+	}
+
+	if cm == nil {
+		return fmt.Errorf("no control message on packet")
+	}
+
+	if cm.IfIndex != p.loIndex {
+		log.Tracef("dropping packet received on interface %d instead of %s", cm.IfIndex, loopbackDevice)
+		return nil
+	}
+
+	dst, ok := netip.AddrFromSlice(cm.Dst.To4())
+	if !ok || !inRange(dst) {
+		log.Tracef("dropping packet for unexpected destination %s", cm.Dst)
+		return nil
+	}
+
+	p.relayedConnMutex.Lock()
+	conn, ok := p.relayedConnStore[dst]
+	p.relayedConnMutex.Unlock()
+	if !ok {
+		if p.ctx.Err() == nil {
+			log.Debugf("relayed conn not found by address because conn already has been closed: %s", dst)
+		}
+		return nil
+	}
+
+	if _, err := conn.Write(buf[:n]); err != nil {
+		return fmt.Errorf("forward local WG packet (%s) to remote relayed conn: %w", dst, err)
+	}
+	return nil
+}
+
+func (p *Proxy) storeRelayedConn(relayedConn net.Conn) (netip.Addr, error) {
+	p.relayedConnMutex.Lock()
+	defer p.relayedConnMutex.Unlock()
+
+	addr, err := p.addrs.next(func(a netip.Addr) bool {
+		_, ok := p.relayedConnStore[a]
+		return ok
+	})
+	if err != nil {
+		return netip.Addr{}, err
+	}
+
+	p.relayedConnStore[addr] = relayedConn
+	return addr, nil
+}
+
+// removeRelayedConn releases an endpoint address. It only removes the entry
+// while it still belongs to relayedConn, so a late release cannot take an
+// address away from the peer it was handed to next.
+func (p *Proxy) removeRelayedConn(addr netip.Addr, relayedConn net.Conn) {
+	p.relayedConnMutex.Lock()
+	defer p.relayedConnMutex.Unlock()
+
+	if stored, ok := p.relayedConnStore[addr]; !ok || stored != relayedConn {
+		return
+	}
+
+	log.Debugf("remove relayed conn from store by address: %s", addr)
+	delete(p.relayedConnStore, addr)
+}

client/iface/wgproxy/loopback/proxy_privileged_test.go

@@ -0,0 +1,196 @@
+//go:build linux && !android && privileged
+
+package loopback
+
+import (
+	"context"
+	"net"
+	"strconv"
+	"testing"
+	"time"
+)
+
+const testWGPort = 51862
+
+// relayEnd stands in for a relayed connection: the proxy writes what it read
+// from WireGuard into it, and the test reads it back out here.
+func relayEnd(t *testing.T) (proxySide net.Conn, testSide *net.UDPConn) {
+	t.Helper()
+
+	testSide, err := net.ListenUDP("udp", &net.UDPAddr{IP: net.ParseIP("127.0.0.1")})
+	if err != nil {
+		t.Fatalf("relay listener: %v", err)
+	}
+	t.Cleanup(func() {
+		if err := testSide.Close(); err != nil {
+			t.Logf("close relay listener: %v", err)
+		}
+	})
+
+	proxySide, err = net.Dial("udp", testSide.LocalAddr().String())
+	if err != nil {
+		t.Fatalf("relay conn: %v", err)
+	}
+	t.Cleanup(func() {
+		if err := proxySide.Close(); err != nil {
+			t.Logf("close relay conn: %v", err)
+		}
+	})
+
+	return proxySide, testSide
+}
+
+// TestProxyDemuxesByDestinationAddress is the core of the design: one socket
+// serves every peer, and the destination address decides which relayed
+// connection a WireGuard packet belongs to.
+func TestProxyDemuxesByDestinationAddress(t *testing.T) {
+	proxy := NewProxy(testWGPort, 1280)
+	if err := proxy.Listen(); err != nil {
+		t.Fatalf("listen: %v", err)
+	}
+	defer func() {
+		if err := proxy.Free(); err != nil {
+			t.Errorf("free proxy: %v", err)
+		}
+	}()
+
+	const peers = 3
+	endpoints := make([]*net.UDPAddr, 0, peers)
+	readers := make([]*net.UDPConn, 0, peers)
+	for i := 0; i < peers; i++ {
+		proxySide, testSide := relayEnd(t)
+		endpoint, _, err := proxy.AddRelayedConn(proxySide)
+		if err != nil {
+			t.Fatalf("add relayed conn %d: %v", i, err)
+		}
+		if endpoint.Port != proxy.proxyPort {
+			t.Errorf("peer %d endpoint port = %d, want the shared proxy port %d", i, endpoint.Port, proxy.proxyPort)
+		}
+		endpoints = append(endpoints, endpoint)
+		readers = append(readers, testSide)
+	}
+
+	// every peer must have its own address, otherwise they are indistinguishable
+	seen := make(map[string]bool, peers)
+	for i, endpoint := range endpoints {
+		if seen[endpoint.IP.String()] {
+			t.Fatalf("peer %d reuses endpoint address %s", i, endpoint.IP)
+		}
+		seen[endpoint.IP.String()] = true
+	}
+
+	wgSock, err := net.ListenUDP("udp", &net.UDPAddr{IP: net.ParseIP("127.0.0.1")})
+	if err != nil {
+		t.Fatalf("wg socket: %v", err)
+	}
+	defer func() {
+		if err := wgSock.Close(); err != nil {
+			t.Logf("close wg socket: %v", err)
+		}
+	}()
+
+	for i, endpoint := range endpoints {
+		payload := []byte{byte(i), 'p', 'k', 't'}
+		if _, err := wgSock.WriteTo(payload, endpoint); err != nil {
+			t.Fatalf("write to peer %d endpoint %s: %v", i, endpoint, err)
+		}
+
+		buf := make([]byte, 1500)
+		if err := readers[i].SetReadDeadline(time.Now().Add(2 * time.Second)); err != nil {
+			t.Fatalf("set read deadline: %v", err)
+		}
+		n, _, err := readers[i].ReadFrom(buf)
+		if err != nil {
+			t.Fatalf("peer %d did not receive its packet: %v", i, err)
+		}
+		if string(buf[:n]) != string(payload) {
+			t.Errorf("peer %d got %q, want %q", i, buf[:n], payload)
+		}
+
+		// no other peer may see it
+		for j, other := range readers {
+			if j == i {
+				continue
+			}
+			if err := other.SetReadDeadline(time.Now().Add(100 * time.Millisecond)); err != nil {
+				t.Fatalf("set read deadline: %v", err)
+			}
+			if _, _, err := other.ReadFrom(buf); err == nil {
+				t.Errorf("packet for peer %d also delivered to peer %d", i, j)
+			}
+		}
+	}
+}
+
+// TestProxyDropsPacketsOutsideTheRange guards the wildcard bind: anything that
+// is not addressed to a handed-out endpoint must not reach a relayed peer.
+func TestProxyDropsPacketsOutsideTheRange(t *testing.T) {
+	proxy := NewProxy(testWGPort+1, 1280)
+	if err := proxy.Listen(); err != nil {
+		t.Fatalf("listen: %v", err)
+	}
+	defer func() {
+		if err := proxy.Free(); err != nil {
+			t.Errorf("free proxy: %v", err)
+		}
+	}()
+
+	proxySide, testSide := relayEnd(t)
+	if _, _, err := proxy.AddRelayedConn(proxySide); err != nil {
+		t.Fatalf("add relayed conn: %v", err)
+	}
+
+	sender, err := net.Dial("udp", net.JoinHostPort("127.0.0.1", strconv.Itoa(proxy.proxyPort)))
+	if err != nil {
+		t.Fatalf("sender: %v", err)
+	}
+	defer func() {
+		if err := sender.Close(); err != nil {
+			t.Logf("close sender: %v", err)
+		}
+	}()
+
+	if _, err := sender.Write([]byte("stray")); err != nil {
+		t.Fatalf("write stray packet: %v", err)
+	}
+
+	buf := make([]byte, 1500)
+	if err := testSide.SetReadDeadline(time.Now().Add(500 * time.Millisecond)); err != nil {
+		t.Fatalf("set read deadline: %v", err)
+	}
+	if _, _, err := testSide.ReadFrom(buf); err == nil {
+		t.Error("packet addressed to 127.0.0.1 was forwarded to a relayed peer")
+	}
+}
+
+// A wrapper that is closed before it starts forwarding still has to give its
+// endpoint address back, otherwise the range leaks an address per attempt.
+func TestClosingBeforeWorkReleasesTheAddress(t *testing.T) {
+	proxy := NewProxy(testWGPort+2, 1280)
+	if err := proxy.Listen(); err != nil {
+		t.Fatalf("listen: %v", err)
+	}
+	defer func() {
+		if err := proxy.Free(); err != nil {
+			t.Errorf("free proxy: %v", err)
+		}
+	}()
+
+	proxySide, _ := relayEnd(t)
+	wrapper := NewProxyWrapper(proxy)
+	if err := wrapper.AddRelayedConn(context.Background(), nil, proxySide); err != nil {
+		t.Fatalf("add relayed conn: %v", err)
+	}
+
+	if got := len(proxy.relayedConnStore); got != 1 {
+		t.Fatalf("store holds %d entries after adding one conn, want 1", got)
+	}
+
+	if err := wrapper.CloseConn(); err != nil {
+		t.Fatalf("close conn: %v", err)
+	}
+
+	if got := len(proxy.relayedConnStore); got != 0 {
+		t.Errorf("store holds %d entries after close, want 0", got)
+	}
+}

client/iface/wgproxy/loopback/wrapper.go

@@ -1,6 +1,6 @@
 //go:build linux && !android
 
-package ebpf
+package loopback
 
 import (
 	"context"
@@ -8,6 +8,7 @@ import (
 	"fmt"
 	"io"
 	"net"
+	"net/netip"
 	"sync"
 
 	"github.com/google/gopacket"
@@ -95,13 +96,14 @@ func NewPacketHeaders(localWGListenPort int, endpoint *net.UDPAddr) (*PacketHead
 
 // ProxyWrapper help to keep the remoteConn instance for net.Conn.Close function call
 type ProxyWrapper struct {
-	wgeBPFProxy *WGEBPFProxy
+	proxy *Proxy
 
 	remoteConn net.Conn
 	ctx        context.Context
 	cancel     context.CancelFunc
 
 	wgRelayedEndpointAddr *net.UDPAddr
+	peerAddr              netip.Addr
 	headers               *PacketHeaders
 	headerCurrentUsed     *PacketHeaders
 	rawConn               net.PacketConn
@@ -113,36 +115,44 @@ type ProxyWrapper struct {
 	closeListener *listener.CloseListener
 }
 
-func NewProxyWrapper(proxy *WGEBPFProxy) *ProxyWrapper {
+func NewProxyWrapper(proxy *Proxy) *ProxyWrapper {
 	return &ProxyWrapper{
-		wgeBPFProxy:   proxy,
+		proxy:         proxy,
 		pausedCond:    sync.NewCond(&sync.Mutex{}),
 		closeListener: listener.NewCloseListener(),
 	}
 }
 
 func (p *ProxyWrapper) AddRelayedConn(ctx context.Context, _ *net.UDPAddr, remoteConn net.Conn) error {
-	addr, err := p.wgeBPFProxy.AddRelayedConn(remoteConn)
+	addr, peerAddr, err := p.proxy.AddRelayedConn(remoteConn)
 	if err != nil {
 		return fmt.Errorf("add relayed conn: %w", err)
 	}
 
-	headers, err := NewPacketHeaders(p.wgeBPFProxy.localWGListenPort, addr)
+	// the endpoint address is otherwise only released by the forwarding
+	// goroutine, which never starts when the setup below fails
+	release := func() { p.proxy.removeRelayedConn(peerAddr, remoteConn) }
+
+	headers, err := NewPacketHeaders(p.proxy.localWGListenPort, addr)
 	if err != nil {
+		release()
 		return fmt.Errorf("create packet sender: %w", err)
 	}
 
 	// Check if required raw connection is available
-	if !headers.isIPv4 && p.wgeBPFProxy.rawConnIPv6 == nil {
+	if !headers.isIPv4 && p.proxy.rawConnIPv6 == nil {
+		release()
 		return errIPv6ConnNotAvailable
 	}
-	if headers.isIPv4 && p.wgeBPFProxy.rawConnIPv4 == nil {
+	if headers.isIPv4 && p.proxy.rawConnIPv4 == nil {
+		release()
 		return errIPv4ConnNotAvailable
 	}
 
 	p.remoteConn = remoteConn
 	p.ctx, p.cancel = context.WithCancel(ctx)
 	p.wgRelayedEndpointAddr = addr
+	p.peerAddr = peerAddr
 	p.headers = headers
 	p.rawConn = p.selectRawConn(headers)
 	return nil
@@ -193,18 +203,18 @@ func (p *ProxyWrapper) RedirectAs(endpoint *net.UDPAddr) {
 		return
 	}
 
-	header, err := NewPacketHeaders(p.wgeBPFProxy.localWGListenPort, endpoint)
+	header, err := NewPacketHeaders(p.proxy.localWGListenPort, endpoint)
 	if err != nil {
 		log.Errorf("failed to create packet headers: %s", err)
 		return
 	}
 
 	// Check if required raw connection is available
-	if !header.isIPv4 && p.wgeBPFProxy.rawConnIPv6 == nil {
+	if !header.isIPv4 && p.proxy.rawConnIPv6 == nil {
 		log.Error(errIPv6ConnNotAvailable)
 		return
 	}
-	if header.isIPv4 && p.wgeBPFProxy.rawConnIPv4 == nil {
+	if header.isIPv4 && p.proxy.rawConnIPv4 == nil {
 		log.Error(errIPv4ConnNotAvailable)
 		return
 	}
@@ -240,6 +250,10 @@ func (p *ProxyWrapper) CloseConn() error {
 
 	p.closeListener.SetCloseListener(nil)
 
+	// releases the endpoint address for a wrapper that was never started, and
+	// is a no-op once the forwarding goroutine has released it
+	p.proxy.removeRelayedConn(p.peerAddr, p.remoteConn)
+
 	p.pausedCond.L.Lock()
 	p.paused = false
 	p.pausedCond.Signal()
@@ -252,9 +266,9 @@ func (p *ProxyWrapper) CloseConn() error {
 }
 
 func (p *ProxyWrapper) proxyToLocal(ctx context.Context) {
-	defer p.wgeBPFProxy.removeRelayedConn(uint16(p.wgRelayedEndpointAddr.Port))
+	defer p.proxy.removeRelayedConn(p.peerAddr, p.remoteConn)
 
-	buf := make([]byte, p.wgeBPFProxy.mtu+bufsize.WGBufferOverhead)
+	buf := make([]byte, p.proxy.mtu+bufsize.WGBufferOverhead)
 	for {
 		n, err := p.readFromRemote(ctx, buf)
 		if err != nil {
@@ -286,7 +300,7 @@ func (p *ProxyWrapper) readFromRemote(ctx context.Context, buf []byte) (int, err
 		}
 		p.closeListener.Notify()
 		if !errors.Is(err, io.EOF) {
-			log.Errorf("failed to read from relayed conn (endpoint: :%d): %s", p.wgRelayedEndpointAddr.Port, err)
+			log.Errorf("failed to read from relayed conn (endpoint: %s): %s", p.wgRelayedEndpointAddr, err)
 		}
 		return 0, err
 	}
@@ -314,7 +328,7 @@ func (p *ProxyWrapper) sendPkg(data []byte, header *PacketHeaders) error {
 
 func (p *ProxyWrapper) selectRawConn(header *PacketHeaders) net.PacketConn {
 	if header.isIPv4 {
-		return p.wgeBPFProxy.rawConnIPv4
+		return p.proxy.rawConnIPv4
 	}
-	return p.wgeBPFProxy.rawConnIPv6
+	return p.proxy.rawConnIPv6
 }

client/iface/wgproxy/proxy_linux_test.go

@@ -9,25 +9,25 @@ import (
 	"github.com/netbirdio/netbird/client/iface/bind"
 	"github.com/netbirdio/netbird/client/iface/wgaddr"
 	bindproxy "github.com/netbirdio/netbird/client/iface/wgproxy/bind"
-	"github.com/netbirdio/netbird/client/iface/wgproxy/ebpf"
+	"github.com/netbirdio/netbird/client/iface/wgproxy/loopback"
 	"github.com/netbirdio/netbird/client/iface/wgproxy/udp"
 )
 
 func seedProxies() ([]proxyInstance, error) {
 	pl := make([]proxyInstance, 0)
 
-	ebpfProxy := ebpf.NewWGEBPFProxy(51831, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		return nil, fmt.Errorf("failed to initialize ebpf proxy: %s", err)
+	loopbackProxy := loopback.NewProxy(51831, 1280)
+	if err := loopbackProxy.Listen(); err != nil {
+		return nil, fmt.Errorf("failed to initialize loopback proxy: %s", err)
 	}
 
-	pEbpf := proxyInstance{
-		name:    "ebpf kernel proxy",
-		proxy:   ebpf.NewProxyWrapper(ebpfProxy),
+	pLoopback := proxyInstance{
+		name:    "loopback kernel proxy",
+		proxy:   loopback.NewProxyWrapper(loopbackProxy),
 		wgPort:  51831,
-		closeFn: ebpfProxy.Free,
+		closeFn: loopbackProxy.Free,
 	}
-	pl = append(pl, pEbpf)
+	pl = append(pl, pLoopback)
 
 	pUDP := proxyInstance{
 		name:    "udp kernel proxy",
@@ -42,18 +42,18 @@ func seedProxies() ([]proxyInstance, error) {
 func seedProxyForProxyCloseByRemoteConn() ([]proxyInstance, error) {
 	pl := make([]proxyInstance, 0)
 
-	ebpfProxy := ebpf.NewWGEBPFProxy(51831, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		return nil, fmt.Errorf("failed to initialize ebpf proxy: %s", err)
+	loopbackProxy := loopback.NewProxy(51831, 1280)
+	if err := loopbackProxy.Listen(); err != nil {
+		return nil, fmt.Errorf("failed to initialize loopback proxy: %s", err)
 	}
 
-	pEbpf := proxyInstance{
-		name:    "ebpf kernel proxy",
-		proxy:   ebpf.NewProxyWrapper(ebpfProxy),
+	pLoopback := proxyInstance{
+		name:    "loopback kernel proxy",
+		proxy:   loopback.NewProxyWrapper(loopbackProxy),
 		wgPort:  51831,
-		closeFn: ebpfProxy.Free,
+		closeFn: loopbackProxy.Free,
 	}
-	pl = append(pl, pEbpf)
+	pl = append(pl, pLoopback)
 
 	pUDP := proxyInstance{
 		name:    "udp kernel proxy",

client/iface/wgproxy/redirect_test.go

@@ -8,7 +8,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/netbirdio/netbird/client/iface/wgproxy/ebpf"
+	"github.com/netbirdio/netbird/client/iface/wgproxy/loopback"
 	"github.com/netbirdio/netbird/client/iface/wgproxy/udp"
 )
 
@@ -198,20 +198,20 @@ func testRedirectAs(t *testing.T, proxy Proxy, wgPort int, nbAddr, p2pEndpoint *
 	}
 }
 
-// TestRedirectAs_eBPF_IPv4 tests RedirectAs with eBPF proxy using IPv4 addresses
-func TestRedirectAs_eBPF_IPv4(t *testing.T) {
+// TestRedirectAs_Loopback_IPv4 tests RedirectAs with the loopback proxy using IPv4 addresses
+func TestRedirectAs_Loopback_IPv4(t *testing.T) {
 	wgPort := 51850
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		t.Fatalf("failed to initialize ebpf proxy: %v", err)
+	loopbackProxy := loopback.NewProxy(wgPort, 1280)
+	if err := loopbackProxy.Listen(); err != nil {
+		t.Fatalf("failed to initialize loopback proxy: %v", err)
 	}
 	defer func() {
-		if err := ebpfProxy.Free(); err != nil {
-			t.Errorf("failed to free ebpf proxy: %v", err)
+		if err := loopbackProxy.Free(); err != nil {
+			t.Errorf("failed to free loopback proxy: %v", err)
 		}
 	}()
 
-	proxy := ebpf.NewProxyWrapper(ebpfProxy)
+	proxy := loopback.NewProxyWrapper(loopbackProxy)
 
 	// NetBird UDP address of the remote peer
 	nbAddr := &net.UDPAddr{
@@ -227,20 +227,20 @@ func TestRedirectAs_eBPF_IPv4(t *testing.T) {
 	testRedirectAs(t, proxy, wgPort, nbAddr, p2pEndpoint)
 }
 
-// TestRedirectAs_eBPF_IPv6 tests RedirectAs with eBPF proxy using IPv6 addresses
-func TestRedirectAs_eBPF_IPv6(t *testing.T) {
+// TestRedirectAs_Loopback_IPv6 tests RedirectAs with the loopback proxy using IPv6 addresses
+func TestRedirectAs_Loopback_IPv6(t *testing.T) {
 	wgPort := 51851
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		t.Fatalf("failed to initialize ebpf proxy: %v", err)
+	loopbackProxy := loopback.NewProxy(wgPort, 1280)
+	if err := loopbackProxy.Listen(); err != nil {
+		t.Fatalf("failed to initialize loopback proxy: %v", err)
 	}
 	defer func() {
-		if err := ebpfProxy.Free(); err != nil {
-			t.Errorf("failed to free ebpf proxy: %v", err)
+		if err := loopbackProxy.Free(); err != nil {
+			t.Errorf("failed to free loopback proxy: %v", err)
 		}
 	}()
 
-	proxy := ebpf.NewProxyWrapper(ebpfProxy)
+	proxy := loopback.NewProxyWrapper(loopbackProxy)
 
 	// NetBird UDP address of the remote peer
 	nbAddr := &net.UDPAddr{
@@ -259,17 +259,17 @@ func TestRedirectAs_eBPF_IPv6(t *testing.T) {
 // TestRedirectAs_Multiple_Switches tests switching between multiple endpoints
 func TestRedirectAs_Multiple_Switches(t *testing.T) {
 	wgPort := 51856
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		t.Fatalf("failed to initialize ebpf proxy: %v", err)
+	loopbackProxy := loopback.NewProxy(wgPort, 1280)
+	if err := loopbackProxy.Listen(); err != nil {
+		t.Fatalf("failed to initialize loopback proxy: %v", err)
 	}
 	defer func() {
-		if err := ebpfProxy.Free(); err != nil {
-			t.Errorf("failed to free ebpf proxy: %v", err)
+		if err := loopbackProxy.Free(); err != nil {
+			t.Errorf("failed to free loopback proxy: %v", err)
 		}
 	}()
 
-	proxy := ebpf.NewProxyWrapper(ebpfProxy)
+	proxy := loopback.NewProxyWrapper(loopbackProxy)
 
 	ctx := context.Background()
 

client/internal/daemonaddr/grpc.go

@@ -0,0 +1,42 @@
+package daemonaddr
+
+import (
+	"os"
+	"strconv"
+
+	log "github.com/sirupsen/logrus"
+)
+
+const (
+	// EnvMaxRecvMsgSize overrides the default gRPC max receive message size for
+	// connections to the daemon. Value is in bytes.
+	EnvMaxRecvMsgSize = "NB_DAEMON_GRPC_MAX_MSG_SIZE"
+
+	// defaultMaxRecvMsgSize is the max gRPC receive message size used for daemon
+	// connections when EnvMaxRecvMsgSize is unset or invalid. It overrides the
+	// gRPC library default of 4 MB, which a detailed status already exceeds on a
+	// network of a few thousand peers.
+	defaultMaxRecvMsgSize = 1024 * 1024 * 16
+)
+
+// MaxRecvMsgSize returns the max gRPC receive message size for daemon connections
+// from the environment, or defaultMaxRecvMsgSize (16 MB) if unset or invalid.
+func MaxRecvMsgSize() int {
+	val := os.Getenv(EnvMaxRecvMsgSize)
+	if val == "" {
+		return defaultMaxRecvMsgSize
+	}
+
+	size, err := strconv.Atoi(val)
+	if err != nil {
+		log.Warnf("invalid %s value %q, using default: %v", EnvMaxRecvMsgSize, val, err)
+		return defaultMaxRecvMsgSize
+	}
+
+	if size <= 0 {
+		log.Warnf("invalid %s value %d, must be positive, using default", EnvMaxRecvMsgSize, size)
+		return defaultMaxRecvMsgSize
+	}
+
+	return size
+}

client/internal/daemonaddr/grpc_test.go

@@ -0,0 +1,112 @@
+package daemonaddr
+
+import (
+	"context"
+	"net"
+	"os"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"google.golang.org/grpc"
+	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/credentials/insecure"
+	"google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/client/proto"
+)
+
+func TestMaxRecvMsgSize(t *testing.T) {
+	tests := []struct {
+		name     string
+		envValue string
+		expected int
+	}{
+		{name: "unset returns default", envValue: "", expected: defaultMaxRecvMsgSize},
+		{name: "non-numeric returns default", envValue: "abc", expected: defaultMaxRecvMsgSize},
+		{name: "negative returns default", envValue: "-1", expected: defaultMaxRecvMsgSize},
+		{name: "zero returns default", envValue: "0", expected: defaultMaxRecvMsgSize},
+		{name: "valid value is used", envValue: "33554432", expected: 33554432},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			// Set first so the previous value is restored on cleanup, then unset to
+			// exercise the absent case.
+			t.Setenv(EnvMaxRecvMsgSize, tc.envValue)
+			if tc.envValue == "" {
+				require.NoError(t, os.Unsetenv(EnvMaxRecvMsgSize), "unset the override")
+			}
+
+			assert.Equal(t, tc.expected, MaxRecvMsgSize(), "max receive message size")
+		})
+	}
+}
+
+// bigStatusServer answers Status with a response larger than gRPC's 4 MB default
+// receive limit, which is what a detailed status on a large network looks like.
+type bigStatusServer struct {
+	proto.UnimplementedDaemonServiceServer
+	payload string
+}
+
+func (s *bigStatusServer) Status(context.Context, *proto.StatusRequest) (*proto.StatusResponse, error) {
+	return &proto.StatusResponse{Status: s.payload}, nil
+}
+
+func startBigStatusServer(t *testing.T, payload string) string {
+	t.Helper()
+
+	listener, err := net.Listen("tcp", "127.0.0.1:0")
+	require.NoError(t, err, "listen on loopback")
+
+	srv := grpc.NewServer()
+	proto.RegisterDaemonServiceServer(srv, &bigStatusServer{payload: payload})
+	go func() {
+		_ = srv.Serve(listener)
+	}()
+	t.Cleanup(srv.Stop)
+
+	return "tcp://" + listener.Addr().String()
+}
+
+func TestDialTargetAcceptsAStatusOverTheGrpcDefault(t *testing.T) {
+	payload := strings.Repeat("x", 5*1024*1024)
+	addr := startBigStatusServer(t, payload)
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	target, opts := DialTarget(addr)
+	conn, err := grpc.NewClient(target, opts...)
+	require.NoError(t, err, "dial the daemon")
+	t.Cleanup(func() { _ = conn.Close() })
+
+	resp, err := proto.NewDaemonServiceClient(conn).Status(ctx, &proto.StatusRequest{})
+	require.NoError(t, err, "a detailed status must not be rejected for its size")
+	assert.Len(t, resp.GetStatus(), len(payload), "the whole response must arrive")
+}
+
+// TestDialTargetRaisesTheDefaultLimit is the negative control: the same response
+// over a connection carrying gRPC's own defaults is refused, which is the failure
+// reported by `netbird status -d` on a large deployment.
+func TestDialTargetRaisesTheDefaultLimit(t *testing.T) {
+	payload := strings.Repeat("x", 5*1024*1024)
+	addr := startBigStatusServer(t, payload)
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	conn, err := grpc.NewClient(
+		strings.TrimPrefix(addr, "tcp://"),
+		grpc.WithTransportCredentials(insecure.NewCredentials()),
+	)
+	require.NoError(t, err, "dial with the library defaults")
+	t.Cleanup(func() { _ = conn.Close() })
+
+	_, err = proto.NewDaemonServiceClient(conn).Status(ctx, &proto.StatusRequest{})
+	require.Error(t, err, "the library default must reject this response")
+	assert.Equal(t, codes.ResourceExhausted, status.Code(err), "gRPC rejects an oversized message")
+}

client/internal/daemonaddr/pipe.go

@@ -36,7 +36,10 @@ const (
 // address. The npipe scheme needs a context dialer because gRPC has no
 // named-pipe resolver; unix and tcp are handled by gRPC itself.
 func DialTarget(addr string) (string, []grpc.DialOption) {
-	opts := []grpc.DialOption{grpc.WithTransportCredentials(insecure.NewCredentials())}
+	opts := []grpc.DialOption{
+		grpc.WithTransportCredentials(insecure.NewCredentials()),
+		grpc.WithDefaultCallOptions(grpc.MaxCallRecvMsgSize(MaxRecvMsgSize())),
+	}
 
 	if name, ok := strings.CutPrefix(addr, pipeScheme); ok {
 		paths := PipePaths(name)

client/internal/dns/host_windows.go

@@ -124,19 +124,9 @@ func newHostManager(wgInterface WGIface) (*registryConfigurator, error) {
 		return nil, err
 	}
 
-	var useGPO bool
-	k, err := registry.OpenKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot, registry.QUERY_VALUE)
-	if err != nil {
-		log.Debugf("failed to open GPO DNS policy root: %v", err)
-	} else {
-		closer(k)
-		useGPO = true
-		log.Infof("detected GPO DNS policy configuration, using policy store")
-	}
-
 	configurator := &registryConfigurator{
 		guid: guid,
-		gpo:  useGPO,
+		gpo:  useGPOPolicyStore(),
 	}
 
 	origNameservers, err := configurator.captureOriginalNameservers()
@@ -576,14 +566,22 @@ func (r *registryConfigurator) setInterfaceRegistryKeyStringValue(key, value str
 	return nil
 }
 
+// deleteInterfaceRegistryKeyProperty removes a value from the interface key.
+// A value that is already gone, or an interface key that is, is not an error:
+// the caller asked for the value not to be there, and a cleanup that runs twice
+// has to reach its later steps on the second run as well.
 func (r *registryConfigurator) deleteInterfaceRegistryKeyProperty(propertyKey string) error {
 	regKey, err := r.getInterfaceRegistryKey()
-	if err != nil {
+	switch {
+	case errors.Is(err, registry.ErrNotExist), errors.Is(err, syscall.ERROR_PATH_NOT_FOUND):
+		log.Debugf("interface key of %s does not exist, nothing to delete %s from", r.guid, propertyKey)
+		return nil
+	case err != nil:
 		return fmt.Errorf("get interface registry key: %w", err)
 	}
 	defer closer(regKey)
 
-	if err := regKey.DeleteValue(propertyKey); err != nil {
+	if err := regKey.DeleteValue(propertyKey); err != nil && !errors.Is(err, registry.ErrNotExist) {
 		return fmt.Errorf("delete registry key %s: %w", propertyKey, err)
 	}
 	return nil
@@ -612,7 +610,12 @@ func (r *registryConfigurator) restoreHostDNS() error {
 
 	go r.flushDNSCache()
 
-	return nil
+	// Last, and only on the way out, once no rule of ours is left: during a
+	// session the store is where the rules of this run live, and emptying it
+	// mid-session would have the next rule recreate it anyway. Propagated so a
+	// failure keeps the shutdown state for the next run to retry, rather than
+	// leaving the store to hold up every rule change from here on.
+	return removeEmptyGPOPolicyStore()
 }
 
 // removeDNSMatchPolicies deletes every NRPT rule this client may have created,
@@ -651,6 +654,73 @@ func (r *registryConfigurator) restoreUncleanShutdownDNS() error {
 	return r.restoreHostDNS()
 }
 
+// useGPOPolicyStore reports whether NRPT rules have to go into the group policy
+// store, and clears an empty one out of the way first.
+//
+// The order is the point. A store left empty by an earlier run would otherwise
+// decide this run too, sending its rules somewhere the resolver only reads when
+// the policy engine next applies DNS client policy. Removing it before the
+// choice is made leaves the local store authoritative for the whole session,
+// including the first one after an upgrade.
+func useGPOPolicyStore() bool {
+	if err := removeEmptyGPOPolicyStore(); err != nil {
+		// Nothing to retry against here: the worst case is the run going
+		// through the group policy store, which is where it would have gone
+		// before this check existed.
+		log.Warnf("%v", err)
+	}
+
+	k, err := registry.OpenKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot, registry.QUERY_VALUE)
+	if err != nil {
+		log.Debugf("failed to open GPO DNS policy root: %v", err)
+		return false
+	}
+	closer(k)
+
+	log.Infof("detected GPO DNS policy configuration, using policy store")
+	return true
+}
+
+// removeEmptyGPOPolicyStore deletes the group policy DnsPolicyConfig key once
+// nothing is left in it. The key survives the deletion of the last rule it
+// held, and the client treats its presence as "group policy configures the
+// NRPT", so an empty one left behind keeps every later run writing rules there.
+// Rules in that store reach the resolver only when the policy engine next
+// applies DNS client policy, and a rule this client writes belongs to no GPO,
+// so nothing schedules that application: both adding and removing a rule are
+// held up by a minute or more, and for a removal that is a catch-all rule
+// resolving every name over an interface that no longer exists. With the store
+// absent the local one is authoritative and a change applies at once.
+//
+// A store that still holds rules, values or subkeys of somebody else's is left
+// alone.
+func removeEmptyGPOPolicyStore() error {
+	k, err := registry.OpenKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot, registry.QUERY_VALUE)
+	switch {
+	case errors.Is(err, registry.ErrNotExist), errors.Is(err, syscall.ERROR_PATH_NOT_FOUND):
+		return nil
+	case err != nil:
+		return fmt.Errorf("open HKEY_LOCAL_MACHINE\\%s: %w", GPODNSPolicyConfigRoot, err)
+	}
+
+	info, err := k.Stat()
+	closer(k)
+	if err != nil {
+		return fmt.Errorf("stat HKEY_LOCAL_MACHINE\\%s: %w", GPODNSPolicyConfigRoot, err)
+	}
+
+	if info.SubKeyCount != 0 || info.ValueCount != 0 {
+		return nil
+	}
+
+	if err := registry.DeleteKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot); err != nil {
+		return fmt.Errorf("delete empty HKEY_LOCAL_MACHINE\\%s: %w", GPODNSPolicyConfigRoot, err)
+	}
+
+	log.Infof("removed the empty GPO DNS policy store, leaving the local one authoritative")
+	return nil
+}
+
 // listNRPTRuleKeys returns the names of our NRPT rule keys under a policy store
 // root. An absent root holds nothing to clean up, which is the normal state of
 // the GPO store on a machine without DNS Client policy.

client/internal/dns/host_windows_test.go

@@ -8,6 +8,8 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"golang.org/x/sys/windows/registry"
+
+	"github.com/netbirdio/netbird/client/internal/winregistry"
 )
 
 // TestNRPTEntriesCleanupOnConfigChange tests that old NRPT entries are properly cleaned up
@@ -405,3 +407,130 @@ func TestNRPTDomainBatching(t *testing.T) {
 		})
 	}
 }
+
+// TestRemoveEmptyGPOPolicyStore verifies that cleanup takes the GPO policy
+// store itself with it once our rules are gone, since the store existing keeps
+// the local one from being applied, and that a store with somebody else's rule
+// in it is left alone.
+func TestRemoveEmptyGPOPolicyStore(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping registry integration test in short mode")
+	}
+
+	t.Cleanup(func() { cleanupRegistryKeys(t) })
+	cleanupRegistryKeys(t)
+
+	testIP := netip.MustParseAddr("100.64.0.1")
+	cfg := &registryConfigurator{gpo: true}
+
+	// a store holding a rule of ours is kept, because the rule is still applied
+	require.NoError(t, cfg.addDNSMatchPolicy([]string{".example.com"}, testIP))
+	exists, err := registryKeyExists(gpoDnsPolicyConfigMatchPath + "-0")
+	require.NoError(t, err)
+	require.True(t, exists, "Should write the rule to the GPO policy store")
+
+	require.NoError(t, removeEmptyGPOPolicyStore())
+	exists, err = registryKeyExists(GPODNSPolicyConfigRoot)
+	require.NoError(t, err)
+	assert.True(t, exists, "Should keep a policy store that still holds a rule")
+
+	// once the rules are gone the store goes with them
+	require.NoError(t, cfg.removeDNSMatchPolicies())
+	require.NoError(t, removeEmptyGPOPolicyStore())
+
+	exists, err = registryKeyExists(GPODNSPolicyConfigRoot)
+	require.NoError(t, err)
+	assert.False(t, exists, "Should remove the GPO policy store once it is empty")
+
+	// A store is not ours to remove while somebody else has a rule in it. The
+	// rule is written volatile like our own: the rules above created the parent
+	// chain volatile, and Windows refuses a stable subkey under a volatile
+	// parent.
+	foreignRule := GPODNSPolicyConfigRoot + `\{2A3B4C5D-6E7F-4041-8283-84858687888A}`
+	foreignKey, _, err := winregistry.CreateVolatileKey(registry.LOCAL_MACHINE, foreignRule, registry.SET_VALUE)
+	require.NoError(t, err, "Should create a foreign GPO rule")
+	foreignKey.Close()
+	t.Cleanup(func() {
+		_ = registry.DeleteKey(registry.LOCAL_MACHINE, foreignRule)
+		_ = registry.DeleteKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot)
+	})
+
+	require.NoError(t, cfg.removeDNSMatchPolicies())
+	require.NoError(t, removeEmptyGPOPolicyStore())
+
+	exists, err = registryKeyExists(foreignRule)
+	require.NoError(t, err)
+	assert.True(t, exists, "Should not remove a foreign rule")
+	exists, err = registryKeyExists(GPODNSPolicyConfigRoot)
+	require.NoError(t, err)
+	assert.True(t, exists, "Should keep a policy store that still holds a foreign rule")
+}
+
+// TestDeleteInterfaceRegistryKeyPropertyTwice verifies that removing a value
+// that is already gone, or one on an interface key that is, reports success.
+// Teardown runs again after a failed cleanup, and the steps that follow this
+// one have to be reached on that second run.
+func TestDeleteInterfaceRegistryKeyPropertyTwice(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping registry integration test in short mode")
+	}
+
+	testGUID := "{12345678-1234-1234-1234-123456789ABC}"
+	interfacePath := InterfaceConfigPath + `\` + testGUID
+	testKey, _, err := registry.CreateKey(registry.LOCAL_MACHINE, interfacePath, registry.SET_VALUE)
+	require.NoError(t, err, "Should create test interface registry key")
+	testKey.Close()
+	t.Cleanup(func() {
+		_ = registry.DeleteKey(registry.LOCAL_MACHINE, interfacePath)
+	})
+
+	cfg := &registryConfigurator{guid: testGUID}
+
+	require.NoError(t, cfg.setInterfaceRegistryKeyStringValue(interfaceConfigSearchListKey, "example.com"))
+	require.NoError(t, cfg.deleteInterfaceRegistryKeyProperty(interfaceConfigSearchListKey))
+	assert.NoError(t, cfg.deleteInterfaceRegistryKeyProperty(interfaceConfigSearchListKey),
+		"Should report success for a value that is already gone")
+
+	// and with the interface key itself gone, as it is once the adapter is
+	require.NoError(t, registry.DeleteKey(registry.LOCAL_MACHINE, interfacePath))
+	assert.NoError(t, cfg.deleteInterfaceRegistryKeyProperty(interfaceConfigSearchListKey),
+		"Should report success when the interface key does not exist")
+}
+
+// TestUseGPOPolicyStoreClearsEmptyStore verifies that the store is cleared
+// before it is consulted, so an empty one left by an earlier run does not send
+// this run's rules to the group policy store. A store somebody else has a rule
+// in still decides where the rules go.
+func TestUseGPOPolicyStoreClearsEmptyStore(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping registry integration test in short mode")
+	}
+
+	t.Cleanup(func() { cleanupRegistryKeys(t) })
+	cleanupRegistryKeys(t)
+
+	// the leftover an earlier run used to keep, which the client read as
+	// "group policy configures the NRPT" for every run after it
+	emptyStore, _, err := winregistry.CreateVolatileKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot, registry.SET_VALUE)
+	require.NoError(t, err, "Should create the GPO policy store")
+	emptyStore.Close()
+
+	assert.False(t, useGPOPolicyStore(), "An empty store should not decide where the rules go")
+	exists, err := registryKeyExists(GPODNSPolicyConfigRoot)
+	require.NoError(t, err)
+	assert.False(t, exists, "Should clear the empty store before consulting it")
+
+	foreignRule := GPODNSPolicyConfigRoot + `\{2A3B4C5D-6E7F-4041-8283-84858687888A}`
+	foreignKey, _, err := winregistry.CreateVolatileKey(registry.LOCAL_MACHINE, foreignRule, registry.SET_VALUE)
+	require.NoError(t, err, "Should create a foreign GPO rule")
+	foreignKey.Close()
+	t.Cleanup(func() {
+		_ = registry.DeleteKey(registry.LOCAL_MACHINE, foreignRule)
+		_ = registry.DeleteKey(registry.LOCAL_MACHINE, GPODNSPolicyConfigRoot)
+	})
+
+	assert.True(t, useGPOPolicyStore(), "A store holding a rule should decide where the rules go")
+	exists, err = registryKeyExists(GPODNSPolicyConfigRoot)
+	require.NoError(t, err)
+	assert.True(t, exists, "Should keep a store that holds a rule")
+}

client/internal/ebpf/ebpf/bpf_bpfeb.go

@@ -1,148 +0,0 @@
-// Code generated by bpf2go; DO NOT EDIT.
-//go:build mips || mips64 || ppc64 || s390x
-
-package ebpf
-
-import (
-	"bytes"
-	_ "embed"
-	"fmt"
-	"io"
-
-	"github.com/cilium/ebpf"
-)
-
-// loadBpf returns the embedded CollectionSpec for bpf.
-func loadBpf() (*ebpf.CollectionSpec, error) {
-	reader := bytes.NewReader(_BpfBytes)
-	spec, err := ebpf.LoadCollectionSpecFromReader(reader)
-	if err != nil {
-		return nil, fmt.Errorf("can't load bpf: %w", err)
-	}
-
-	return spec, err
-}
-
-// loadBpfObjects loads bpf and converts it into a struct.
-//
-// The following types are suitable as obj argument:
-//
-//	*bpfObjects
-//	*bpfPrograms
-//	*bpfMaps
-//
-// See ebpf.CollectionSpec.LoadAndAssign documentation for details.
-func loadBpfObjects(obj interface{}, opts *ebpf.CollectionOptions) error {
-	spec, err := loadBpf()
-	if err != nil {
-		return err
-	}
-
-	return spec.LoadAndAssign(obj, opts)
-}
-
-// bpfSpecs contains maps and programs before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfSpecs struct {
-	bpfProgramSpecs
-	bpfMapSpecs
-	bpfVariableSpecs
-}
-
-// bpfProgramSpecs contains programs before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfProgramSpecs struct {
-	NbXdpProg *ebpf.ProgramSpec `ebpf:"nb_xdp_prog"`
-}
-
-// bpfMapSpecs contains maps before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfMapSpecs struct {
-	NbFeatures           *ebpf.MapSpec `ebpf:"nb_features"`
-	NbWgProxySettingsMap *ebpf.MapSpec `ebpf:"nb_wg_proxy_settings_map"`
-}
-
-// bpfVariableSpecs contains global variables before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfVariableSpecs struct {
-	FlagFeatureWgProxy *ebpf.VariableSpec `ebpf:"flag_feature_wg_proxy"`
-	MapKeyFeatures     *ebpf.VariableSpec `ebpf:"map_key_features"`
-	MapKeyProxyPort    *ebpf.VariableSpec `ebpf:"map_key_proxy_port"`
-	MapKeyWgPort       *ebpf.VariableSpec `ebpf:"map_key_wg_port"`
-	ProxyPort          *ebpf.VariableSpec `ebpf:"proxy_port"`
-	WgPort             *ebpf.VariableSpec `ebpf:"wg_port"`
-}
-
-// bpfObjects contains all objects after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfObjects struct {
-	bpfPrograms
-	bpfMaps
-	bpfVariables
-}
-
-func (o *bpfObjects) Close() error {
-	return _BpfClose(
-		&o.bpfPrograms,
-		&o.bpfMaps,
-	)
-}
-
-// bpfMaps contains all maps after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfMaps struct {
-	NbFeatures           *ebpf.Map `ebpf:"nb_features"`
-	NbWgProxySettingsMap *ebpf.Map `ebpf:"nb_wg_proxy_settings_map"`
-}
-
-func (m *bpfMaps) Close() error {
-	return _BpfClose(
-		m.NbFeatures,
-		m.NbWgProxySettingsMap,
-	)
-}
-
-// bpfVariables contains all global variables after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfVariables struct {
-	FlagFeatureWgProxy *ebpf.Variable `ebpf:"flag_feature_wg_proxy"`
-	MapKeyFeatures     *ebpf.Variable `ebpf:"map_key_features"`
-	MapKeyProxyPort    *ebpf.Variable `ebpf:"map_key_proxy_port"`
-	MapKeyWgPort       *ebpf.Variable `ebpf:"map_key_wg_port"`
-	ProxyPort          *ebpf.Variable `ebpf:"proxy_port"`
-	WgPort             *ebpf.Variable `ebpf:"wg_port"`
-}
-
-// bpfPrograms contains all programs after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfPrograms struct {
-	NbXdpProg *ebpf.Program `ebpf:"nb_xdp_prog"`
-}
-
-func (p *bpfPrograms) Close() error {
-	return _BpfClose(
-		p.NbXdpProg,
-	)
-}
-
-func _BpfClose(closers ...io.Closer) error {
-	for _, closer := range closers {
-		if err := closer.Close(); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-// Do not access this directly.
-//
-//go:embed bpf_bpfeb.o
-var _BpfBytes []byte

client/internal/ebpf/ebpf/bpf_bpfel.go

@@ -1,148 +0,0 @@
-// Code generated by bpf2go; DO NOT EDIT.
-//go:build 386 || amd64 || arm || arm64 || loong64 || mips64le || mipsle || ppc64le || riscv64 || wasm
-
-package ebpf
-
-import (
-	"bytes"
-	_ "embed"
-	"fmt"
-	"io"
-
-	"github.com/cilium/ebpf"
-)
-
-// loadBpf returns the embedded CollectionSpec for bpf.
-func loadBpf() (*ebpf.CollectionSpec, error) {
-	reader := bytes.NewReader(_BpfBytes)
-	spec, err := ebpf.LoadCollectionSpecFromReader(reader)
-	if err != nil {
-		return nil, fmt.Errorf("can't load bpf: %w", err)
-	}
-
-	return spec, err
-}
-
-// loadBpfObjects loads bpf and converts it into a struct.
-//
-// The following types are suitable as obj argument:
-//
-//	*bpfObjects
-//	*bpfPrograms
-//	*bpfMaps
-//
-// See ebpf.CollectionSpec.LoadAndAssign documentation for details.
-func loadBpfObjects(obj interface{}, opts *ebpf.CollectionOptions) error {
-	spec, err := loadBpf()
-	if err != nil {
-		return err
-	}
-
-	return spec.LoadAndAssign(obj, opts)
-}
-
-// bpfSpecs contains maps and programs before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfSpecs struct {
-	bpfProgramSpecs
-	bpfMapSpecs
-	bpfVariableSpecs
-}
-
-// bpfProgramSpecs contains programs before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfProgramSpecs struct {
-	NbXdpProg *ebpf.ProgramSpec `ebpf:"nb_xdp_prog"`
-}
-
-// bpfMapSpecs contains maps before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfMapSpecs struct {
-	NbFeatures           *ebpf.MapSpec `ebpf:"nb_features"`
-	NbWgProxySettingsMap *ebpf.MapSpec `ebpf:"nb_wg_proxy_settings_map"`
-}
-
-// bpfVariableSpecs contains global variables before they are loaded into the kernel.
-//
-// It can be passed ebpf.CollectionSpec.Assign.
-type bpfVariableSpecs struct {
-	FlagFeatureWgProxy *ebpf.VariableSpec `ebpf:"flag_feature_wg_proxy"`
-	MapKeyFeatures     *ebpf.VariableSpec `ebpf:"map_key_features"`
-	MapKeyProxyPort    *ebpf.VariableSpec `ebpf:"map_key_proxy_port"`
-	MapKeyWgPort       *ebpf.VariableSpec `ebpf:"map_key_wg_port"`
-	ProxyPort          *ebpf.VariableSpec `ebpf:"proxy_port"`
-	WgPort             *ebpf.VariableSpec `ebpf:"wg_port"`
-}
-
-// bpfObjects contains all objects after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfObjects struct {
-	bpfPrograms
-	bpfMaps
-	bpfVariables
-}
-
-func (o *bpfObjects) Close() error {
-	return _BpfClose(
-		&o.bpfPrograms,
-		&o.bpfMaps,
-	)
-}
-
-// bpfMaps contains all maps after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfMaps struct {
-	NbFeatures           *ebpf.Map `ebpf:"nb_features"`
-	NbWgProxySettingsMap *ebpf.Map `ebpf:"nb_wg_proxy_settings_map"`
-}
-
-func (m *bpfMaps) Close() error {
-	return _BpfClose(
-		m.NbFeatures,
-		m.NbWgProxySettingsMap,
-	)
-}
-
-// bpfVariables contains all global variables after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfVariables struct {
-	FlagFeatureWgProxy *ebpf.Variable `ebpf:"flag_feature_wg_proxy"`
-	MapKeyFeatures     *ebpf.Variable `ebpf:"map_key_features"`
-	MapKeyProxyPort    *ebpf.Variable `ebpf:"map_key_proxy_port"`
-	MapKeyWgPort       *ebpf.Variable `ebpf:"map_key_wg_port"`
-	ProxyPort          *ebpf.Variable `ebpf:"proxy_port"`
-	WgPort             *ebpf.Variable `ebpf:"wg_port"`
-}
-
-// bpfPrograms contains all programs after they have been loaded into the kernel.
-//
-// It can be passed to loadBpfObjects or ebpf.CollectionSpec.LoadAndAssign.
-type bpfPrograms struct {
-	NbXdpProg *ebpf.Program `ebpf:"nb_xdp_prog"`
-}
-
-func (p *bpfPrograms) Close() error {
-	return _BpfClose(
-		p.NbXdpProg,
-	)
-}
-
-func _BpfClose(closers ...io.Closer) error {
-	for _, closer := range closers {
-		if err := closer.Close(); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-// Do not access this directly.
-//
-//go:embed bpf_bpfel.o
-var _BpfBytes []byte

client/internal/ebpf/ebpf/manager_linux.go

@@ -1,115 +0,0 @@
-package ebpf
-
-import (
-	_ "embed"
-	"net"
-	"sync"
-
-	"github.com/cilium/ebpf/link"
-	"github.com/cilium/ebpf/rlimit"
-	log "github.com/sirupsen/logrus"
-
-	"github.com/netbirdio/netbird/client/internal/ebpf/manager"
-)
-
-const (
-	mapKeyFeatures uint32 = 0
-
-	featureFlagWGProxy = 0b00000001
-)
-
-var (
-	singleton     manager.Manager
-	singletonLock = &sync.Mutex{}
-)
-
-// required packages libbpf-dev, libc6-dev-i386-amd64-cross
-
-// GeneralManager is used to load multiple eBPF programs with a custom check (if then) done in prog.c
-// The manager simply adds a feature (byte) of each program to a map that is shared between the userspace and kernel.
-// When packet arrives, the C code checks for each feature (if it is set) and executes each enabled program (e.g., wg_proxy.c).
-//
-//go:generate go run github.com/cilium/ebpf/cmd/bpf2go -cc clang-14 bpf src/prog.c -- -I /usr/x86_64-linux-gnu/include -include src/bpf_map_def.h
-type GeneralManager struct {
-	lock         sync.Mutex
-	link         link.Link
-	featureFlags uint16
-	bpfObjs      bpfObjects
-}
-
-// GetEbpfManagerInstance return a static eBpf Manager instance
-func GetEbpfManagerInstance() manager.Manager {
-	singletonLock.Lock()
-	defer singletonLock.Unlock()
-	if singleton != nil {
-		return singleton
-	}
-	singleton = &GeneralManager{}
-	return singleton
-}
-
-func (tf *GeneralManager) setFeatureFlag(feature uint16) {
-	tf.featureFlags |= feature
-}
-
-func (tf *GeneralManager) loadXdp() error {
-	if tf.link != nil {
-		return nil
-	}
-	// it required for Docker
-	err := rlimit.RemoveMemlock()
-	if err != nil {
-		return err
-	}
-
-	iFace, err := net.InterfaceByName("lo")
-	if err != nil {
-		return err
-	}
-
-	// load pre-compiled programs into the kernel.
-	err = loadBpfObjects(&tf.bpfObjs, nil)
-	if err != nil {
-		return err
-	}
-
-	tf.link, err = link.AttachXDP(link.XDPOptions{
-		Program:   tf.bpfObjs.NbXdpProg,
-		Interface: iFace.Index,
-	})
-
-	if err != nil {
-		_ = tf.bpfObjs.Close()
-		tf.link = nil
-		return err
-	}
-	return nil
-}
-
-func (tf *GeneralManager) unsetFeatureFlag(feature uint16) error {
-	tf.lock.Lock()
-	defer tf.lock.Unlock()
-	tf.featureFlags &^= feature
-
-	if tf.link == nil {
-		return nil
-	}
-
-	if tf.featureFlags == 0 {
-		return tf.close()
-	}
-
-	return tf.bpfObjs.NbFeatures.Put(mapKeyFeatures, tf.featureFlags)
-}
-
-func (tf *GeneralManager) close() error {
-	log.Debugf("detach ebpf program ")
-	err := tf.bpfObjs.Close()
-	if err != nil {
-		log.Warnf("failed to close eBpf objects: %s", err)
-	}
-
-	err = tf.link.Close()
-	tf.link = nil
-	return err
-}

client/internal/ebpf/ebpf/manager_linux_test.go

@@ -1,31 +0,0 @@
-package ebpf
-
-import (
-	"testing"
-)
-
-func TestManager_setFeatureFlag(t *testing.T) {
-	mgr := GeneralManager{}
-	mgr.setFeatureFlag(featureFlagWGProxy)
-	if mgr.featureFlags != featureFlagWGProxy {
-		t.Errorf("invalid feature state")
-	}
-
-	mgr.setFeatureFlag(featureFlagWGProxy)
-	if mgr.featureFlags != featureFlagWGProxy {
-		t.Errorf("setting a flag twice must be idempotent, got: %d", mgr.featureFlags)
-	}
-}
-
-func TestManager_unsetFeatureFlag(t *testing.T) {
-	mgr := GeneralManager{}
-	mgr.setFeatureFlag(featureFlagWGProxy)
-
-	err := mgr.unsetFeatureFlag(featureFlagWGProxy)
-	if err != nil {
-		t.Errorf("unexpected error: %s", err)
-	}
-	if mgr.featureFlags != 0 {
-		t.Errorf("invalid feature state, expected: %d, got: %d", 0, mgr.featureFlags)
-	}
-}

client/internal/ebpf/ebpf/src/bpf_map_def.h

@@ -1,16 +0,0 @@
-// libbpf 1.0 removed struct bpf_map_def, but the programs here keep the legacy
-// map definitions: they load on kernels built without BTF, which BTF-style
-// (SEC(".maps")) definitions do not. Define the struct ourselves so the
-// programs compile against current libbpf headers.
-#ifndef NB_BPF_MAP_DEF_H
-#define NB_BPF_MAP_DEF_H
-
-struct bpf_map_def {
-	unsigned int type;
-	unsigned int key_size;
-	unsigned int value_size;
-	unsigned int max_entries;
-	unsigned int map_flags;
-};
-
-#endif

client/internal/ebpf/ebpf/src/prog.c

@@ -1,54 +0,0 @@
-#include <stdbool.h>
-#include <linux/if_ether.h> // ETH_P_IP
-#include <linux/udp.h>
-#include <linux/ip.h>
-#include <netinet/in.h>
-#include <linux/bpf.h>
-#include <bpf/bpf_helpers.h>
-#include "wg_proxy.c"
-
-const __u16 flag_feature_wg_proxy = 0b01;
-
-const __u32 map_key_features = 0;
-struct bpf_map_def SEC("maps") nb_features = {
-	.type = BPF_MAP_TYPE_ARRAY,
-	.key_size = sizeof(__u32),
-	.value_size = sizeof(__u16),
-	.max_entries = 10,
-};
-
-SEC("xdp")
-int nb_xdp_prog(struct xdp_md *ctx) {
-    __u16 *features;
-    features = bpf_map_lookup_elem(&nb_features, &map_key_features);
-    if (!features) {
-        return XDP_PASS;
-    }
-
-    void *data     = (void *)(long)ctx->data;
-    void *data_end = (void *)(long)ctx->data_end;
-    struct ethhdr *eth = data;
-    struct iphdr  *ip  = (data + sizeof(struct ethhdr));
-    struct udphdr *udp = (data + sizeof(struct ethhdr) + sizeof(struct iphdr));
-
-    // return early if not enough data
-    if (data + sizeof(struct ethhdr) + sizeof(struct iphdr) + sizeof(struct udphdr) > data_end){
-        return XDP_PASS;
-    }
-
-    // skip non IPv4 packages
-    if (eth->h_proto != htons(ETH_P_IP)) {
-       return XDP_PASS;
-    }
-
-    // skip non UPD packages
-    if (ip->protocol != IPPROTO_UDP) {
-       return XDP_PASS;
-    }
-
-    if (*features & flag_feature_wg_proxy) {
-        xdp_wg_proxy(ip, udp);
-    }
-    return XDP_PASS;
-}
-char _license[] SEC("license") = "GPL";

client/internal/ebpf/ebpf/src/readme.md

@@ -1,27 +0,0 @@
-# XDP programs
-
-`prog.c` is attached to the `lo` device and dispatches to the features enabled in the
-`nb_features` map. The only feature is the WireGuard proxy (`wg_proxy.c`): it rewrites
-loopback UDP sent from the WireGuard listen port so it reaches the userspace relay proxy
-port instead, and swaps the peer endpoint port into the source so the proxy can tell
-peers apart.
-
-Maps use the legacy `struct bpf_map_def` form, defined in `bpf_map_def.h` because libbpf
-1.0 removed it. They load on kernels built without BTF, which BTF-style (`SEC(".maps")`)
-definitions do not.
-
-Regenerate the objects with `go generate ./client/internal/ebpf/ebpf/`; it needs
-`clang-14`. Loading a regenerated object needs root, attaching it needs `bpf_link`
-(kernel >= 5.7), and only one XDP program can own `lo` at a time.
-
-# Debug
-
-The CONFIG_BPF_EVENTS kernel module is required for bpf_printk.
-Apply this code to use bpf_printk
-```
-#define bpf_printk(fmt, ...)                                                   \
-  ({                                                                           \
-    char ____fmt[] = fmt;                                                      \
-    bpf_trace_printk(____fmt, sizeof(____fmt), ##__VA_ARGS__);                 \
-  })
-```

client/internal/ebpf/ebpf/src/wg_proxy.c

@@ -1,60 +0,0 @@
-const __u32 map_key_proxy_port = 0;
-const __u32 map_key_wg_port = 1;
-
-struct bpf_map_def SEC("maps") nb_wg_proxy_settings_map = {
-	.type = BPF_MAP_TYPE_ARRAY,
-	.key_size = sizeof(__u32),
-	.value_size = sizeof(__u16),
-	.max_entries = 10,
-};
-
-__u16 proxy_port = 0;
-__u16 wg_port = 0;
-
-bool read_port_settings() {
-    __u16 *value;
-    value = bpf_map_lookup_elem(&nb_wg_proxy_settings_map, &map_key_proxy_port);
-    if (!value) {
-        return false;
-    }
-
-    proxy_port = *value;
-
-    value = bpf_map_lookup_elem(&nb_wg_proxy_settings_map, &map_key_wg_port);
-    if (!value) {
-        return false;
-    }
-    wg_port = htons(*value);
-
-    return true;
-}
-
-int xdp_wg_proxy(struct iphdr  *ip, struct udphdr *udp) {
-    if (proxy_port == 0 || wg_port == 0) {
-        if (!read_port_settings()){
-            return XDP_PASS;
-        }
-        // bpf_printk("proxy port: %d, wg port: %d", proxy_port, wg_port);
-    }
-
-    // 2130706433 = 127.0.0.1
-    if (ip->daddr != htonl(2130706433)) {
-        return XDP_PASS;
-    }
-
-    if (udp->source != wg_port){
-        return XDP_PASS;
-    }
-
-    __be16 new_src_port = udp->dest;
-    __be16 new_dst_port = htons(proxy_port);
-    udp->dest = new_dst_port;
-    udp->source = new_src_port;
-
-    // The ports are covered by the UDP checksum. This is an IPv4 loopback hop
-    // and the payload is already integrity-protected, so clear the checksum (a
-    // zero UDP checksum means "not computed" for IPv4) rather than leave a
-    // stale value the kernel would drop as UDP_CSUM.
-    udp->check = 0;
-	return XDP_PASS;
-}

client/internal/ebpf/ebpf/wg_proxy_linux.go

@@ -1,41 +0,0 @@
-package ebpf
-
-import log "github.com/sirupsen/logrus"
-
-const (
-	mapKeyProxyPort uint32 = 0
-	mapKeyWgPort    uint32 = 1
-)
-
-func (tf *GeneralManager) LoadWgProxy(proxyPort, wgPort int) error {
-	log.Debugf("load ebpf WG proxy")
-	tf.lock.Lock()
-	defer tf.lock.Unlock()
-
-	err := tf.loadXdp()
-	if err != nil {
-		return err
-	}
-
-	err = tf.bpfObjs.NbWgProxySettingsMap.Put(mapKeyProxyPort, uint16(proxyPort))
-	if err != nil {
-		return err
-	}
-
-	err = tf.bpfObjs.NbWgProxySettingsMap.Put(mapKeyWgPort, uint16(wgPort))
-	if err != nil {
-		return err
-	}
-
-	tf.setFeatureFlag(featureFlagWGProxy)
-	err = tf.bpfObjs.NbFeatures.Put(mapKeyFeatures, tf.featureFlags)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-func (tf *GeneralManager) FreeWGProxy() error {
-	log.Debugf("free ebpf WG proxy")
-	return tf.unsetFeatureFlag(featureFlagWGProxy)
-}

client/internal/ebpf/instantiater_linux.go

@@ -1,15 +0,0 @@
-//go:build !android
-
-package ebpf
-
-import (
-	"github.com/netbirdio/netbird/client/internal/ebpf/ebpf"
-	"github.com/netbirdio/netbird/client/internal/ebpf/manager"
-)
-
-// GetEbpfManagerInstance is a wrapper function. This encapsulation is required because if the code import the internal
-// ebpf package the Go compiler will include the object files. But it is not supported on Android. It can cause instant
-// panic on older Android version.
-func GetEbpfManagerInstance() manager.Manager {
-	return ebpf.GetEbpfManagerInstance()
-}

client/internal/ebpf/instantiater_nonlinux.go

@@ -1,10 +0,0 @@
-//go:build !linux || android
-
-package ebpf
-
-import "github.com/netbirdio/netbird/client/internal/ebpf/manager"
-
-// GetEbpfManagerInstance return error because ebpf is not supported on all os
-func GetEbpfManagerInstance() manager.Manager {
-	panic("unsupported os")
-}

client/internal/ebpf/manager/manager.go

@@ -1,7 +0,0 @@
-package manager
-
-// Manager is used to load multiple eBPF programs. E.g., the WireGuard proxy
-type Manager interface {
-	LoadWgProxy(proxyPort, wgPort int) error
-	FreeWGProxy() error
-}

client/internal/elevate/trusted.go

@@ -6,6 +6,17 @@ import (
 	"path/filepath"
 )
 
+// CheckOnlyOwnerWritable reports an error unless path, and every directory
+// leading to it, is owned by an account that can already act with the privileges
+// the caller holds, and is writable by nobody else.
+//
+// Exported for callers outside elevation that read a file while privileged and
+// then act on what it says: the same question this package asks of an
+// executable, asked of a configuration file.
+func CheckOnlyOwnerWritable(path string) error {
+	return checkOnlyOwnerWritable(path)
+}
+
 // trustedSelf returns the path of this executable, provided it is one we are
 // willing to have run as root.
 //

client/internal/engine.go

@@ -671,10 +671,6 @@ func (e *Engine) Start(netbirdConfig *mgmProto.NetbirdConfig, mgmtURL *url.URL)
 	}
 	e.wgDevice.Store(e.wgInterface.GetWGDevice())
 
-	// Set up notrack rules immediately after proxy is listening to prevent
-	// conntrack entries from being created before the rules are in place
-	e.setupWGProxyNoTrack()
-
 	// Start after interface is up since port may have been resolved from 0 or changed if occupied
 	e.shutdownWg.Add(1)
 	go func() {
@@ -841,23 +837,6 @@ func (e *Engine) initFirewall() error {
 	return nil
 }
 
-// setupWGProxyNoTrack configures connection tracking exclusion for WireGuard proxy traffic.
-// This prevents conntrack/MASQUERADE from affecting loopback traffic between WireGuard and the eBPF proxy.
-func (e *Engine) setupWGProxyNoTrack() {
-	if e.firewall == nil {
-		return
-	}
-
-	proxyPort := e.wgInterface.GetProxyPort()
-	if proxyPort == 0 {
-		return
-	}
-
-	if err := e.firewall.SetupEBPFProxyNoTrack(proxyPort, uint16(e.config.WgPort)); err != nil {
-		log.Warnf("failed to setup ebpf proxy notrack: %v", err)
-	}
-}
-
 func (e *Engine) blockLanAccess() {
 	if e.config.BlockInbound {
 		// no need to set up extra deny rules if inbound is already blocked in general
@@ -1104,7 +1083,11 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 			// back to empty if the FQDN doesn't have the expected shape.
 			dnsName = extractDNSDomainFromFQDN(pc.GetFqdn())
 		}
-		result, err := nbnetworkmap.EnvelopeToNetworkMap(e.ctx, envelope, localKey, dnsName)
+		// With the firewall disabled there is no ACL manager to program, so
+		// RoutesFirewallRules would be built and then dropped. On a peer that
+		// routes many network resources that is the single most expensive
+		// step of the sync.
+		result, err := nbnetworkmap.EnvelopeToNetworkMap(e.ctx, envelope, localKey, dnsName, e.config.DisableFirewall)
 		if err != nil {
 			return fmt.Errorf("decode network map envelope: %w", err)
 		}

client/internal/engine_test.go

@@ -65,7 +65,6 @@ type MockWGIface struct {
 	GetStatsFunc               func() (map[string]configurer.WGStats, error)
 	GetInterfaceGUIDStringFunc func() (string, error)
 	GetProxyFunc               func() wgproxy.Proxy
-	GetProxyPortFunc           func() uint16
 	GetNetFunc                 func() *netstack.Net
 	LastActivitiesFunc         func() map[string]monotime.Time
 }
@@ -162,12 +161,6 @@ func (m *MockWGIface) GetProxy() wgproxy.Proxy {
 	return m.GetProxyFunc()
 }
 
-func (m *MockWGIface) GetProxyPort() uint16 {
-	if m.GetProxyPortFunc != nil {
-		return m.GetProxyPortFunc()
-	}
-	return 0
-}
 
 func (m *MockWGIface) GetNet() *netstack.Net {
 	return m.GetNetFunc()

client/internal/iface_common.go

@@ -28,7 +28,6 @@ type wgIfaceBase interface {
 	Up() (*udpmux.UniversalUDPMuxDefault, error)
 	UpdateAddr(newAddr wgaddr.Address) error
 	GetProxy() wgproxy.Proxy
-	GetProxyPort() uint16
 	UpdatePeer(peerKey string, allowedIps []netip.Prefix, keepAlive time.Duration, endpoint *net.UDPAddr, preSharedKey *wgtypes.Key) error
 	RemoveEndpointAddress(key string) error
 	RemovePeer(peerKey string) error

client/internal/ipcauth/forward_test.go

@@ -38,7 +38,7 @@ func asDaemon(t *testing.T, id Identity) {
 	prevID, prevKnown, prevDelegate := selfIdentity, selfKnown, selfMayDelegate
 	t.Cleanup(func() { selfIdentity, selfKnown, selfMayDelegate = prevID, prevKnown, prevDelegate })
 	selfIdentity, selfKnown = id, true
-	selfMayDelegate = !id.IsPrivileged()
+	selfMayDelegate = mayDelegate(id)
 }
 
 func TestCallerIdentity_DirectConnections(t *testing.T) {

client/internal/ipcauth/identity.go

@@ -18,7 +18,8 @@ import (
 	"google.golang.org/grpc/peer"
 )
 
-// Well-known Windows SIDs that identify a fully privileged principal.
+// Well-known Windows SIDs. Only LocalSystem and BUILTIN\Administrators identify a
+// privileged principal; the service accounts are shared by unrelated services.
 const (
 	sidLocalSystem    = "S-1-5-18"     // NT AUTHORITY\SYSTEM
 	sidLocalService   = "S-1-5-19"     // NT AUTHORITY\LOCAL SERVICE
@@ -67,9 +68,9 @@ func (i Identity) IsWindows() bool {
 // user-to-root boundary.
 //
 // On Windows the decision comes from the caller's token rather than from
-// account names or group RIDs: an elevated token, one of the service accounts
-// the daemon itself may run as, or a token with BUILTIN\Administrators
-// enabled. A UAC-filtered administrator has that group marked deny-only, and
+// account names or group RIDs: an elevated token, the LocalSystem SID, or a
+// token with BUILTIN\Administrators enabled. LocalService and NetworkService
+// are not privileged by SID. A UAC-filtered administrator has that group marked deny-only, and
 // deny-only groups are dropped when the identity is captured, so such a
 // caller is correctly reported as unprivileged. Domain group memberships
 // (Domain Admins and friends) are deliberately not consulted: they say
@@ -83,8 +84,7 @@ func (i Identity) IsPrivileged() bool {
 		return true
 	}
 
-	switch i.SID {
-	case sidLocalSystem, sidLocalService, sidNetworkService:
+	if i.SID == sidLocalSystem {
 		return true
 	}
 

client/internal/ipcauth/identity_test.go

@@ -64,3 +64,58 @@ func TestIdentitySameUser(t *testing.T) {
 		})
 	}
 }
+
+func TestIdentityIsPrivileged(t *testing.T) {
+	tests := []struct {
+		name string
+		id   Identity
+		want bool
+	}{
+		{
+			name: "Root",
+			id:   Identity{UID: 0, GID: 0},
+			want: true,
+		},
+		{
+			name: "Non-root",
+			id:   Identity{UID: 1000, GID: 1000},
+			want: false,
+		},
+		{
+			name: "Local system windows",
+			id:   Identity{SID: sidLocalSystem},
+			want: true,
+		},
+		{
+			name: "Windows elevated",
+			id:   Identity{SID: "S-1-5-21-1927267129-3959769253-3036563910-1001", Elevated: true},
+			want: true,
+		},
+		{
+			name: "Admin group windows",
+			id:   Identity{SID: "S-1-5-21-1927267129-3959769253-3036563910-1001", Groups: []string{sidAdministrators}},
+			want: true,
+		},
+		{
+			name: "Regular user windows",
+			id:   Identity{SID: "S-1-5-21-1927267129-3959769253-3036563910-1001"},
+			want: false,
+		},
+		{
+			name: "Network service windows",
+			id:   Identity{SID: sidNetworkService},
+			want: false,
+		},
+		{
+			name: "Local service windows",
+			id:   Identity{SID: sidLocalService},
+			want: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			assert.Equal(t, tt.want, tt.id.IsPrivileged())
+		})
+	}
+}

client/internal/ipcauth/privileged.go

@@ -45,7 +45,15 @@ func init() {
 	// matching there would let a non-elevated shell of an administrator account
 	// act as an administrator, which is the boundary the token check exists to
 	// keep.
-	selfMayDelegate = !id.IsPrivileged()
+	selfMayDelegate = mayDelegate(id)
+}
+
+// mayDelegate reports whether a daemon running as id may extend its authority to
+// callers sharing its identity. The shared service accounts are excluded: their
+// SID is held by unrelated services, so matching on it would grant them the
+// daemon's authority.
+func mayDelegate(id Identity) bool {
+	return !id.IsPrivileged() && id.SID != sidLocalService && id.SID != sidNetworkService
 }
 
 // IsDaemonSelf reports whether an identity is this very process. The JSON gateway

client/internal/ipcauth/privileged_test.go

@@ -98,7 +98,7 @@ func TestIsPrivilegedCaller_SelfRule(t *testing.T) {
 			t.Cleanup(func() { selfIdentity, selfKnown, selfMayDelegate = prevID, prevKnown, prevDelegate })
 
 			selfIdentity, selfKnown = tt.self, tt.selfKnown
-			selfMayDelegate = tt.selfKnown && !tt.self.IsPrivileged()
+			selfMayDelegate = tt.selfKnown && mayDelegate(tt.self)
 
 			if got := IsPrivilegedCaller(tt.caller); got != tt.want {
 				t.Fatalf("IsPrivilegedCaller(%v) with daemon %v = %t, want %t",
@@ -132,3 +132,28 @@ func TestIsPrivilegedCaller_ThisProcess(t *testing.T) {
 		t.Errorf("an unrelated identity %v was treated as privileged", other)
 	}
 }
+
+// The shared service accounts are held by unrelated services, so a daemon running
+// as one of them must not extend its authority to every process with that SID.
+func TestMayDelegate(t *testing.T) {
+	tests := []struct {
+		name string
+		self Identity
+		want bool
+	}{
+		{name: "unprivileged unix user", self: Identity{UID: 1000}, want: true},
+		{name: "root", self: Identity{UID: 0}, want: false},
+		{name: "unprivileged windows user", self: Identity{SID: "S-1-5-21-1-2-3-1001"}, want: true},
+		{name: "elevated windows user", self: Identity{SID: "S-1-5-21-1-2-3-1001", Elevated: true}, want: false},
+		{name: "local system", self: Identity{SID: sidLocalSystem}, want: false},
+		{name: "local service", self: Identity{SID: sidLocalService}, want: false},
+		{name: "network service", self: Identity{SID: sidNetworkService}, want: false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := mayDelegate(tt.self); got != tt.want {
+				t.Errorf("mayDelegate(%+v) = %v, want %v", tt.self, got, tt.want)
+			}
+		})
+	}
+}

client/internal/peer/conn.go

@@ -176,9 +176,10 @@ type Conn struct {
 	// used to store the remote Rosenpass key for Relayed connection in case of connection update from ice
 	rosenpassRemoteKey []byte
 
-	wgProxyICE   wgproxy.Proxy
-	wgProxyRelay wgproxy.Proxy
-	handshaker   *Handshaker
+	wgProxyICE     wgproxy.Proxy
+	wgProxyRelay   wgproxy.Proxy
+	relayedConnRef *relayClient.Conn
+	handshaker     *Handshaker
 
 	guard *guard.Guard
 	wg    sync.WaitGroup
@@ -625,7 +626,7 @@ func (conn *Conn) onRelayConnectionIsReady(rci RelayConnInfo) {
 	conn.mu.Lock()
 	defer conn.mu.Unlock()
 
-	if conn.ctx.Err() != nil {
+	if conn.ctx.Err() != nil || rci.relayedConn.Context().Err() != nil {
 		if err := rci.relayedConn.Close(); err != nil {
 			conn.Log.Warnf("failed to close unnecessary relayed connection: %v", err)
 		}
@@ -640,7 +641,9 @@ func (conn *Conn) onRelayConnectionIsReady(rci RelayConnInfo) {
 		conn.Log.Errorf("failed to add relayed net.Conn to local proxy: %v", err)
 		return
 	}
-	wgProxy.SetDisconnectListener(conn.onRelayDisconnected)
+	wgProxy.SetDisconnectListener(func() {
+		conn.onRelayDisconnected(rci.relayedConn)
+	})
 
 	conn.dumpState.NewLocalProxy()
 
@@ -648,7 +651,7 @@ func (conn *Conn) onRelayConnectionIsReady(rci RelayConnInfo) {
 
 	if conn.isICEActive() {
 		conn.Log.Debugf("do not switch to relay because current priority is: %s", conn.currentConnPriority.String())
-		conn.setRelayedProxy(wgProxy)
+		conn.setRelayedProxy(wgProxy, rci.relayedConn)
 		conn.statusRelay.SetConnected()
 		conn.updateRelayStatus(rci.relayedConn.RemoteAddr().String(), rci.rosenpassPubKey, pqOK, time.Now())
 		return
@@ -679,15 +682,26 @@ func (conn *Conn) onRelayConnectionIsReady(rci RelayConnInfo) {
 	conn.rosenpassRemoteKey = rci.rosenpassPubKey
 	conn.currentConnPriority = conntype.Relay
 	conn.statusRelay.SetConnected()
-	conn.setRelayedProxy(wgProxy)
+	conn.setRelayedProxy(wgProxy, rci.relayedConn)
 	conn.updateRelayStatus(rci.relayedConn.RemoteAddr().String(), rci.rosenpassPubKey, pqOK, updateTime)
 	conn.Log.Infof("start to communicate with peer via relay")
 	conn.doOnConnected(rci.rosenpassPubKey, rci.rosenpassAddr, updateTime)
 }
 
-func (conn *Conn) onRelayDisconnected() {
+// onRelayDisconnected reports the teardown of a relayed connection. relayedConn
+// names the connection the signal belongs to, so a signal that arrives after
+// its connection was replaced is ignored instead of tearing down its successor.
+// A nil relayedConn means the caller does not track generations and the current
+// connection is always torn down.
+func (conn *Conn) onRelayDisconnected(relayedConn *relayClient.Conn) {
 	conn.mu.Lock()
 	defer conn.mu.Unlock()
+
+	if relayedConn != nil && conn.relayedConnRef != relayedConn {
+		conn.Log.Debugf("ignoring relay disconnect of a superseded connection")
+		return
+	}
+
 	conn.handleRelayDisconnectedLocked()
 }
 
@@ -711,6 +725,7 @@ func (conn *Conn) handleRelayDisconnectedLocked() {
 		_ = conn.wgProxyRelay.CloseConn()
 		conn.wgProxyRelay = nil
 	}
+	conn.relayedConnRef = nil
 
 	changed := conn.statusRelay.Get() != worker.StatusDisconnected
 	if changed {
@@ -1020,13 +1035,14 @@ func (conn *Conn) logTraceConnState() {
 	}
 }
 
-func (conn *Conn) setRelayedProxy(proxy wgproxy.Proxy) {
+func (conn *Conn) setRelayedProxy(proxy wgproxy.Proxy, relayedConn *relayClient.Conn) {
 	if conn.wgProxyRelay != nil {
 		if err := conn.wgProxyRelay.CloseConn(); err != nil {
 			conn.Log.Warnf("failed to close deprecated wg proxy conn: %v", err)
 		}
 	}
 	conn.wgProxyRelay = proxy
+	conn.relayedConnRef = relayedConn
 }
 
 // onWGHandshakeSuccess is called when the first WireGuard handshake is detected

client/internal/peer/handshaker.go

@@ -159,7 +159,7 @@ func (h *Handshaker) notifyListeners(remoteOfferAnswer *OfferAnswer) {
 }
 
 func (h *Handshaker) handleRemoteOffer(remoteOfferAnswer OfferAnswer) {
-	h.log.Infof("received offer, running version %s, remote WireGuard listen port %d, session id: %s, remote ICE supported: %t", remoteOfferAnswer.Version, remoteOfferAnswer.WgListenPort, remoteOfferAnswer.SessionIDString(), remoteOfferAnswer.hasICECredentials())
+	h.log.Infof("received offer, running version %s, remote WireGuard listen port %d, session id: %s, remote ICE supported: %t, relay server: %s, relay IP: %s", remoteOfferAnswer.Version, remoteOfferAnswer.WgListenPort, remoteOfferAnswer.SessionIDString(), remoteOfferAnswer.hasICECredentials(), remoteOfferAnswer.RelaySrvAddress, remoteOfferAnswer.RelaySrvIP)
 	h.onSignalReceived(&remoteOfferAnswer)
 
 	// If we are the controller running the KEM, a responder's offer is handled by
@@ -180,7 +180,7 @@ func (h *Handshaker) handleRemoteOffer(remoteOfferAnswer OfferAnswer) {
 }
 
 func (h *Handshaker) handleRemoteAnswer(remoteOfferAnswer OfferAnswer) {
-	h.log.Infof("received answer, running version %s, remote WireGuard listen port %d, session id: %s, remote ICE supported: %t", remoteOfferAnswer.Version, remoteOfferAnswer.WgListenPort, remoteOfferAnswer.SessionIDString(), remoteOfferAnswer.hasICECredentials())
+	h.log.Infof("received answer, running version %s, remote WireGuard listen port %d, session id: %s, remote ICE supported: %t, relay server: %s, relay IP: %s", remoteOfferAnswer.Version, remoteOfferAnswer.WgListenPort, remoteOfferAnswer.SessionIDString(), remoteOfferAnswer.hasICECredentials(), remoteOfferAnswer.RelaySrvAddress, remoteOfferAnswer.RelaySrvIP)
 	h.onSignalReceived(&remoteOfferAnswer)
 
 	// Feed the KEM answer (derive+store PSK) BEFORE bringing up the connection so the WG
@@ -309,7 +309,7 @@ func (h *Handshaker) sendOffer() error {
 	if h.config.PQ != nil {
 		offer.MlkemPayload, offer.MlkemPort = h.config.PQ.OfferPayload(h.config.Key)
 	}
-	h.log.Debugf("sending offer with serial: %s", offer.SessionIDString())
+	h.log.Debugf("sending offer with serial: %s, relay server: %s, relay IP: %s", offer.SessionIDString(), offer.RelaySrvAddress, offer.RelaySrvIP)
 
 	return h.signaler.SignalOffer(offer, h.config.Key)
 }
@@ -323,7 +323,7 @@ func (h *Handshaker) sendAnswer(remoteOffer *OfferAnswer) error {
 		}
 		answer.MlkemPayload, answer.MlkemPort = h.config.PQ.AnswerPayload(h.config.Key, recvOffer)
 	}
-	h.log.Debugf("sending answer with serial: %s", answer.SessionIDString())
+	h.log.Debugf("sending answer with serial: %s, relay server: %s, relay IP: %s", answer.SessionIDString(), answer.RelaySrvAddress, answer.RelaySrvIP)
 
 	return h.signaler.SignalAnswer(answer, h.config.Key)
 }

client/internal/peer/status.go

@@ -830,8 +830,8 @@ func (d *Status) SetSessionExpiresAt(deadline time.Time) {
 // "none" would blank the UI at the exact moment it should say the session
 // ended.
 func (d *Status) GetSessionExpiresAt() time.Time {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return d.sessionExpiresAt
 }
 

client/internal/peer/worker_ice.go

@@ -121,11 +121,8 @@ func (w *WorkerICE) OnNewOffer(remoteOfferAnswer *OfferAnswer) {
 			}
 		}
 
-		sessionID, err := NewICESessionID()
-		if err != nil {
-			w.log.Errorf("failed to create new session ID: %s", err)
-		}
-		w.sessionID = sessionID
+		// Keep the ID already advertised to the remote. Answers do not get a
+		// reply, so changing it here makes the next offer restart both sides.
 		w.abandonNegotiation()
 	}
 
@@ -205,6 +202,9 @@ func (w *WorkerICE) Close() {
 	w.muxAgent.Lock()
 	defer w.muxAgent.Unlock()
 
+	if w.agent != nil || w.agentConnecting {
+		w.renewSessionID()
+	}
 	if w.agent != nil {
 		w.agentDialerCancel()
 		if err := w.agent.Close(); err != nil {
@@ -366,16 +366,23 @@ func (w *WorkerICE) closeAgent(agent *icemaker.ThreadSafeAgent, cancel context.C
 	// Only the owner of the current session may reset its state: a stale dial
 	// goroutine waking after a newer attempt must not clobber it.
 	if w.agent == agent {
-		sessionID, err := NewICESessionID()
-		if err != nil {
-			w.log.Errorf("failed to create new session ID: %s", err)
-		}
-		w.sessionID = sessionID
+		w.renewSessionID()
 		w.abandonNegotiation()
 	}
 	return sessionChanged
 }
 
+// renewSessionID starts a new local session, so the remote treats our next offer
+// or answer as a restart. Caller holds muxAgent.
+func (w *WorkerICE) renewSessionID() {
+	sessionID, err := NewICESessionID()
+	if err != nil {
+		w.log.Errorf("failed to create new session ID: %s", err)
+		return
+	}
+	w.sessionID = sessionID
+}
+
 // abandonNegotiation drops all recorded ICE session state so the worker treats the
 // next offer as a fresh start instead of a duplicate of a dead negotiation. The
 // agent and agentConnecting flags must change together: leaving one stale wedges

client/internal/peer/worker_ice_session_test.go

@@ -0,0 +1,375 @@
+package peer
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	icemaker "github.com/netbirdio/netbird/client/internal/peer/ice"
+)
+
+func TestWorkerICE_RemoteRestartPreservesAdvertisedSession(t *testing.T) {
+	w := newTestWorkerICE(t)
+	t.Cleanup(w.Close)
+	w.dialFunc = parkDial
+	advertised := w.SessionID()
+	remoteSession := ICESessionID("remote-first")
+	offer := OfferAnswer{
+		IceCredentials: IceCredentials{UFrag: "remoteufrag", Pwd: "remote-password-long-enough"},
+		SessionID:      &remoteSession,
+	}
+	w.OnNewOffer(&offer)
+	require.True(t, w.InProgress(), "the first remote session must start ICE")
+	w.muxAgent.Lock()
+	firstAgent := w.agent
+	w.muxAgent.Unlock()
+
+	// The same callback handles answers. A changed remote ID must not create
+	// an unannounced local ID that makes the remote restart on our next offer.
+	secondSession := ICESessionID("remote-restarted")
+	answer := offer
+	answer.SessionID = &secondSession
+	w.OnNewOffer(&answer)
+	assert.Equal(t, advertised, w.SessionID(), "following a remote restart must keep our advertised ID")
+	w.muxAgent.Lock()
+	secondAgent := w.agent
+	w.muxAgent.Unlock()
+	assert.NotSame(t, firstAgent, secondAgent, "the changed remote session must still rebuild ICE")
+
+	w.OnNewOffer(&answer)
+	w.muxAgent.Lock()
+	defer w.muxAgent.Unlock()
+	assert.Same(t, secondAgent, w.agent, "a repeated answer must keep the replacement agent")
+}
+
+func TestWorkerICE_LocalCloseChangesAdvertisedSession(t *testing.T) {
+	w := newTestWorkerICE(t)
+	dialStarted := make(chan struct{})
+	dialDone := make(chan struct{})
+	w.dialFunc = func(ctx context.Context, _ *icemaker.ThreadSafeAgent, _ *OfferAnswer) (net.Conn, error) {
+		close(dialStarted)
+		defer close(dialDone)
+		<-ctx.Done()
+		return nil, ctx.Err()
+	}
+	session := ICESessionID("remote-session")
+	w.OnNewOffer(&OfferAnswer{
+		IceCredentials: IceCredentials{UFrag: "remoteufrag", Pwd: "remote-password-long-enough"},
+		SessionID:      &session,
+	})
+	<-dialStarted
+	advertised := w.SessionID()
+	w.Close()
+	assert.NotEqual(t, advertised, w.SessionID(), "a local teardown must tell the remote to restart")
+	closedSession := w.SessionID()
+
+	// The abandoned dial goroutine cleans up after Close returned.
+	<-dialDone
+	assert.Never(t, func() bool { return w.SessionID() != closedSession }, 200*time.Millisecond, 10*time.Millisecond,
+		"the late cleanup of a closed negotiation must not restart again")
+	w.Close()
+	assert.Equal(t, closedSession, w.SessionID(), "closing an idle worker must not restart again")
+}
+
+// parkDial stands in for the ICE dial. It never connects and returns once the
+// negotiation is abandoned, so a test decides when a negotiation fails.
+func parkDial(ctx context.Context, _ *icemaker.ThreadSafeAgent, _ *OfferAnswer) (net.Conn, error) {
+	<-ctx.Done()
+	return nil, ctx.Err()
+}
+
+func newTestSessionID(t *testing.T) ICESessionID {
+	t.Helper()
+	sid, err := NewICESessionID()
+	require.NoError(t, err)
+	return sid
+}
+
+// handshakeSide is one end of a simulated signaling exchange.
+type handshakeSide interface {
+	// message builds the offer or answer the side would send now.
+	message() OfferAnswer
+	// receive hands a remote offer or answer to the side's ICE logic.
+	receive(msg OfferAnswer)
+	// teardowns counts negotiations the side tore down to follow a remote restart.
+	teardowns() int
+	// failAgent ends the side's current negotiation as an ICE failure does.
+	failAgent()
+}
+
+// workerSide drives a real WorkerICE.
+type workerSide struct {
+	t        *testing.T
+	w        *WorkerICE
+	replaced int
+}
+
+func newWorkerSide(t *testing.T) *workerSide {
+	t.Helper()
+	w := newTestWorkerICE(t)
+	w.dialFunc = parkDial
+	t.Cleanup(w.Close)
+	return &workerSide{t: t, w: w}
+}
+
+func (s *workerSide) message() OfferAnswer {
+	sid := s.w.SessionID()
+	ufrag, pwd := s.w.GetLocalUserCredentials()
+	return OfferAnswer{IceCredentials: IceCredentials{UFrag: ufrag, Pwd: pwd}, SessionID: &sid}
+}
+
+func (s *workerSide) receive(msg OfferAnswer) {
+	before := s.agent()
+	s.w.OnNewOffer(&msg)
+	if after := s.agent(); before != nil && after != before {
+		s.replaced++
+	}
+}
+
+func (s *workerSide) teardowns() int { return s.replaced }
+
+func (s *workerSide) agent() *icemaker.ThreadSafeAgent {
+	s.w.muxAgent.Lock()
+	defer s.w.muxAgent.Unlock()
+	return s.w.agent
+}
+
+// failAgent runs the cleanup the dial goroutine or the Failed state callback
+// performs when the current negotiation dies.
+func (s *workerSide) failAgent() {
+	s.t.Helper()
+	s.w.muxAgent.Lock()
+	agent, cancel := s.w.agent, s.w.agentDialerCancel
+	s.w.muxAgent.Unlock()
+	require.NotNil(s.t, agent, "failing requires a running negotiation")
+	s.w.closeAgent(agent, cancel)
+}
+
+// legacySide models a remote peer running a release from before this change:
+// when it follows a remote restart it also picks a new session ID of its own,
+// which it announces only with its next offer or answer.
+type legacySide struct {
+	t         *testing.T
+	sessionID ICESessionID
+	remoteID  ICESessionID
+	hasAgent  bool
+	replaced  int
+}
+
+func newLegacySide(t *testing.T) *legacySide {
+	return &legacySide{t: t, sessionID: newTestSessionID(t)}
+}
+
+func (s *legacySide) message() OfferAnswer {
+	sid := s.sessionID
+	return OfferAnswer{
+		IceCredentials: IceCredentials{UFrag: "legacyufrag", Pwd: "legacy-password-long-enough"},
+		SessionID:      &sid,
+	}
+}
+
+func (s *legacySide) receive(msg OfferAnswer) {
+	if msg.SessionID == nil {
+		s.hasAgent = true
+		return
+	}
+	if s.hasAgent {
+		if *msg.SessionID == s.remoteID {
+			return
+		}
+		s.replaced++
+		s.sessionID = newTestSessionID(s.t)
+	}
+	s.hasAgent = true
+	s.remoteID = *msg.SessionID
+}
+
+func (s *legacySide) teardowns() int { return s.replaced }
+
+func (s *legacySide) failAgent() {
+	s.hasAgent = false
+	s.remoteID = ""
+	s.sessionID = newTestSessionID(s.t)
+}
+
+// exchange runs one guard-driven round in the order Handshaker.Listen uses: the
+// answerer handles the offer and answers with the session ID it holds
+// afterwards, and the offerer handles the answer without replying.
+func exchange(offerer, answerer handshakeSide) {
+	answerer.receive(offerer.message())
+	offerer.receive(answerer.message())
+}
+
+// offerPattern decides which side's guard sends the offer in a round.
+type offerPattern struct {
+	name   string
+	picker func(round int, local, remote handshakeSide) (offerer, answerer handshakeSide)
+}
+
+var offerPatterns = []offerPattern{
+	{
+		// A routing peer whose relay is down keeps offering on its own.
+		name: "local peer offers",
+		picker: func(_ int, local, remote handshakeSide) (handshakeSide, handshakeSide) {
+			return local, remote
+		},
+	},
+	{
+		name: "both peers offer",
+		picker: func(round int, local, remote handshakeSide) (handshakeSide, handshakeSide) {
+			if round%2 == 0 {
+				return local, remote
+			}
+			return remote, local
+		},
+	},
+}
+
+// assertSettles runs guard rounds and requires the pair to stop restarting
+// each other: at most maxTeardowns in total, and none once half the rounds ran.
+func assertSettles(t *testing.T, pattern offerPattern, local, remote handshakeSide, maxTeardowns int) {
+	t.Helper()
+	const rounds = 10
+
+	total := func() int { return local.teardowns() + remote.teardowns() }
+	start := total()
+	var halfway int
+	for round := range rounds {
+		if round == rounds/2 {
+			halfway = total()
+		}
+		offerer, answerer := pattern.picker(round, local, remote)
+		exchange(offerer, answerer)
+	}
+
+	assert.LessOrEqual(t, total()-start, maxTeardowns, "the peers must not keep restarting each other")
+	assert.Equal(t, halfway, total(), "the negotiation must be stable in the later rounds")
+}
+
+// establish runs the first offer and answer, so both sides negotiate.
+func establish(t *testing.T, local, remote handshakeSide) {
+	t.Helper()
+	exchange(local, remote)
+	require.Zero(t, local.teardowns()+remote.teardowns(), "the first exchange must not restart anything")
+}
+
+func TestICESession_SettlesAfterAgentFailure(t *testing.T) {
+	sides := []struct {
+		name   string
+		remote func(t *testing.T) handshakeSide
+	}{
+		{name: "current remote", remote: func(t *testing.T) handshakeSide { return newWorkerSide(t) }},
+		{name: "legacy remote", remote: func(t *testing.T) handshakeSide { return newLegacySide(t) }},
+	}
+	failures := []struct {
+		name string
+		fail func(local, remote handshakeSide)
+	}{
+		{name: "remote agent fails", fail: func(_, remote handshakeSide) { remote.failAgent() }},
+		{name: "local agent fails", fail: func(local, _ handshakeSide) { local.failAgent() }},
+		{name: "both agents fail", fail: func(local, remote handshakeSide) {
+			local.failAgent()
+			remote.failAgent()
+		}},
+	}
+
+	for _, side := range sides {
+		for _, failure := range failures {
+			for _, pattern := range offerPatterns {
+				t.Run(fmt.Sprintf("%s/%s/%s", side.name, failure.name, pattern.name), func(t *testing.T) {
+					local := newWorkerSide(t)
+					remote := side.remote(t)
+					establish(t, local, remote)
+
+					failure.fail(local, remote)
+					assertSettles(t, pattern, local, remote, 2)
+				})
+			}
+		}
+	}
+}
+
+// TestICESession_LocalCloseRestartsRemote covers an explicit teardown, as on a
+// WireGuard handshake timeout. The remote must start over as well, or it keeps
+// answering from the negotiation this side just abandoned.
+func TestICESession_LocalCloseRestartsRemote(t *testing.T) {
+	for _, pattern := range offerPatterns {
+		t.Run(pattern.name, func(t *testing.T) {
+			local := newWorkerSide(t)
+			remote := newWorkerSide(t)
+			establish(t, local, remote)
+
+			local.w.Close()
+			assertSettles(t, pattern, local, remote, 1)
+			assert.Equal(t, 1, remote.teardowns(), "the remote must restart its negotiation exactly once")
+		})
+	}
+}
+
+func TestICESession_DuplicateMessagesKeepNegotiation(t *testing.T) {
+	local := newWorkerSide(t)
+	remote := newWorkerSide(t)
+
+	offer := local.message()
+	remote.receive(offer)
+	answer := remote.message()
+	local.receive(answer)
+
+	// Signaling may deliver the same message again, and a peer answers every
+	// offer, including repeats of one it already handled.
+	remote.receive(offer)
+	local.receive(answer)
+	local.receive(remote.message())
+
+	assert.Zero(t, local.teardowns(), "a repeated answer must not restart the negotiation")
+	assert.Zero(t, remote.teardowns(), "a repeated offer must not restart the negotiation")
+}
+
+// TestICESession_RemoteWithoutSessionIDKeepsNegotiation covers remote peers
+// too old to send session IDs: once negotiating, their messages cannot tell a
+// restart from a repeat, so they must not tear anything down.
+func TestICESession_RemoteWithoutSessionIDKeepsNegotiation(t *testing.T) {
+	local := newWorkerSide(t)
+	unversioned := OfferAnswer{IceCredentials: IceCredentials{UFrag: "oldufrag", Pwd: "old-password-long-enough"}}
+
+	local.receive(unversioned)
+	require.NotNil(t, local.agent(), "a message without a session ID must still start ICE")
+	advertised := local.w.SessionID()
+
+	for range 3 {
+		local.receive(unversioned)
+	}
+	assert.Zero(t, local.teardowns(), "messages without a session ID must not restart the negotiation")
+	assert.Equal(t, advertised, local.w.SessionID(), "the advertised session must not change")
+}
+
+// TestWorkerICE_StaleCleanupKeepsAdvertisedSession covers the cleanup of a
+// replaced negotiation finishing late, from its dial goroutine or its Closed
+// state callback. It must neither pick a new session ID, an unannounced local
+// restart, nor disturb the negotiation that replaced it.
+func TestWorkerICE_StaleCleanupKeepsAdvertisedSession(t *testing.T) {
+	local := newWorkerSide(t)
+	remote := newWorkerSide(t)
+	establish(t, local, remote)
+
+	local.w.muxAgent.Lock()
+	oldAgent, oldCancel := local.w.agent, local.w.agentDialerCancel
+	local.w.muxAgent.Unlock()
+
+	remote.failAgent()
+	exchange(local, remote)
+	require.Equal(t, 1, local.teardowns(), "the local side must follow the remote restart")
+	advertised := local.w.SessionID()
+	current := local.agent()
+
+	local.w.closeAgent(oldAgent, oldCancel)
+
+	assert.Equal(t, advertised, local.w.SessionID(), "a stale cleanup must not change the advertised session")
+	assert.Same(t, current, local.agent(), "a stale cleanup must keep the current negotiation")
+	assertSettles(t, offerPatterns[1], local, remote, 0)
+}

client/internal/peer/worker_relay.go

@@ -3,7 +3,6 @@ package peer
 import (
 	"context"
 	"errors"
-	"net"
 	"net/netip"
 	"sync"
 	"sync/atomic"
@@ -14,7 +13,7 @@ import (
 )
 
 type RelayConnInfo struct {
-	relayedConn     net.Conn
+	relayedConn     *relayClient.Conn
 	rosenpassPubKey []byte
 	rosenpassAddr   string
 }
@@ -27,7 +26,7 @@ type WorkerRelay struct {
 	conn         *Conn
 	relayManager *relayClient.Manager
 
-	relayedConn net.Conn
+	relayedConn *relayClient.Conn
 	relayLock   sync.Mutex
 
 	relaySupportedOnRemotePeer atomic.Bool
@@ -80,12 +79,7 @@ func (w *WorkerRelay) OnNewOffer(remoteOfferAnswer *OfferAnswer) {
 	w.relayedConn = relayedConn
 	w.relayLock.Unlock()
 
-	err = w.relayManager.AddCloseListener(srv, w.onRelayClientDisconnected)
-	if err != nil {
-		log.Errorf("failed to add close listener: %s", err)
-		_ = relayedConn.Close()
-		return
-	}
+	go w.watchRelayedConn(relayedConn)
 
 	w.log.Debugf("peer conn opened via Relay: %s", srv)
 	go w.conn.onRelayConnectionIsReady(RelayConnInfo{
@@ -109,12 +103,15 @@ func (w *WorkerRelay) RelayIsSupportedLocally() bool {
 
 func (w *WorkerRelay) CloseConn() {
 	w.relayLock.Lock()
-	defer w.relayLock.Unlock()
-	if w.relayedConn == nil {
+	conn := w.relayedConn
+	w.relayedConn = nil
+	w.relayLock.Unlock()
+
+	if conn == nil {
 		return
 	}
 
-	if err := w.relayedConn.Close(); err != nil {
+	if err := conn.Close(); err != nil {
 		w.log.Warnf("failed to close relay connection: %v", err)
 	}
 }
@@ -133,6 +130,8 @@ func (w *WorkerRelay) preferredRelayServer(myRelayAddress, remoteRelayAddress st
 	return remoteRelayAddress
 }
 
-func (w *WorkerRelay) onRelayClientDisconnected() {
-	go w.conn.onRelayDisconnected()
+func (w *WorkerRelay) watchRelayedConn(relayedConn *relayClient.Conn) {
+	<-relayedConn.Context().Done()
+
+	w.conn.onRelayDisconnected(relayedConn)
 }

client/internal/profilemanager/active_state_test.go

@@ -0,0 +1,76 @@
+package profilemanager
+
+import (
+	"fmt"
+	"sync"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// Regression test: a concurrent Get and Set of the ActiveProfileState will
+// fail on Windows since the write is a temp file renamed over an open file.
+// Windows will refuse to replace a file another handle holds open by default.
+func TestActiveProfileState_ReadsDoNotBreakAConcurrentWrite(t *testing.T) {
+	withTempConfigDir(t, func(configDir string) {
+		withPatchedGlobals(t, configDir, func() {
+			sm := &ServiceManager{}
+			require.NoError(t, sm.CreateDefaultProfile())
+			require.NoError(t, sm.SetActiveProfileStateToDefault())
+
+			const switched = ID("0123456789abcdef0123456789abcdef")
+			const rounds = 50
+
+			var wg sync.WaitGroup
+			errs := make(chan error, 128)
+
+			for i := 0; i < 8; i++ {
+				wg.Add(1)
+				go func() {
+					defer wg.Done()
+					for r := 0; r < rounds; r++ {
+						state, err := sm.GetActiveProfileState()
+						if err != nil {
+							errs <- fmt.Errorf("read: %w", err)
+							return
+						}
+						if state.ID != defaultProfileName && state.ID != switched {
+							errs <- fmt.Errorf("read: active profile is %q, which no writer wrote", state.ID)
+							return
+						}
+					}
+				}()
+			}
+
+			for i := 0; i < 2; i++ {
+				wg.Add(1)
+				go func() {
+					defer wg.Done()
+					for r := 0; r < rounds; r++ {
+						id := switched
+						if r%2 == 0 {
+							id = defaultProfileName
+						}
+						if err := sm.SetActiveProfileState(&ActiveProfileState{ID: id, Username: "testuser"}); err != nil {
+							errs <- fmt.Errorf("switch: %w", err)
+							return
+						}
+					}
+				}()
+			}
+
+			wg.Wait()
+			close(errs)
+
+			for err := range errs {
+				assert.NoError(t, err, "a switch and a read of the active profile state must not collide")
+			}
+
+			state, err := sm.GetActiveProfileState()
+			require.NoError(t, err)
+			assert.Contains(t, []ID{defaultProfileName, switched}, state.ID,
+				"the file holds whichever switch landed last, not a mix of the two")
+		})
+	})
+}

client/internal/wincmd/system32_windows.go

@@ -0,0 +1,30 @@
+// Package wincmd locates the Windows utilities the client shells out to.
+package wincmd
+
+import (
+	"path/filepath"
+
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/sys/windows"
+)
+
+// defaultSystem32Dir is where the system directory is on every supported
+// install, used only when the API that reports it fails.
+const defaultSystem32Dir = `C:\Windows\System32`
+
+// System32 returns the full path of a Windows utility under the system
+// directory.
+//
+// PATH is deliberately not consulted. The daemon runs as LocalSystem with an
+// environment of its own, so whoever can place an entry in that PATH chooses
+// which binary runs with those privileges. The system directory is read from
+// the API rather than from %SystemRoot% for the same reason.
+func System32(command string) string {
+	sysDir, err := windows.GetSystemDirectory()
+	if err != nil {
+		log.Warnf("Failed to locate the Windows system directory, falling back to %s: %v", defaultSystem32Dir, err)
+		sysDir = defaultSystem32Dir
+	}
+
+	return filepath.Join(sysDir, command+".exe")
+}

client/internal/wincmd/system32_windows_test.go

@@ -0,0 +1,31 @@
+package wincmd
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestSystem32IgnoresPATH(t *testing.T) {
+	// A directory holding something that would win a PATH lookup, in front of
+	// everything else: the daemon runs as LocalSystem, so a PATH entry must not
+	// be able to decide what it executes.
+	planted := t.TempDir()
+	require.NoError(t, os.WriteFile(filepath.Join(planted, "netsh.exe"), []byte("not really netsh"), 0o600))
+	t.Setenv("PATH", planted+string(os.PathListSeparator)+os.Getenv("PATH"))
+
+	got := System32("netsh")
+
+	assert.True(t, filepath.IsAbs(got), "the path must be absolute, got %q", got)
+	assert.NotContains(t, got, planted, "a PATH entry must not be consulted")
+	assert.True(t, strings.EqualFold(filepath.Base(got), "netsh.exe"), "unexpected file name in %q", got)
+
+	// The system directory is what Windows reports it to be, not %SystemRoot%,
+	// which the same caller could have set alongside PATH.
+	t.Setenv("SystemRoot", planted)
+	assert.Equal(t, got, System32("netsh"), "%SystemRoot% must not move the lookup")
+}

client/ssh/server/executor_windows.go

@@ -6,7 +6,6 @@ import (
 	"context"
 	"errors"
 	"fmt"
-	"os"
 	"os/exec"
 	"os/user"
 	"strings"
@@ -506,15 +505,37 @@ func userExists(fullUsername, username, domain string) error {
 	return nil
 }
 
-// isLocalUser determines if this is a local user vs domain user
+// isLocalUser reports whether domain refers to this machine rather than to a
+// Windows domain.
 func (pd *PrivilegeDropper) isLocalUser(domain string) bool {
-	hostname, err := os.Hostname()
-	if err != nil {
-		hostname = "localhost"
+	return isLocalDomain(domain, netbiosComputerName)
+}
+
+// isLocalDomain compares against the NetBIOS name because Windows qualifies local
+// accounts with it, and it is the DNS host name truncated to 15 characters.
+// An unknown name falls back to the domain path: treating it as local could
+// authenticate a same named local account instead.
+// https://learn.microsoft.com/en-us/windows/win32/sysinfo/computer-names
+func isLocalDomain(domain string, machineName func() (string, error)) bool {
+	if domain == "" || domain == "." {
+		return true
 	}
 
-	return domain == "" || domain == "." ||
-		strings.EqualFold(domain, hostname)
+	name, err := machineName()
+	if err != nil {
+		log.Debugf("read NetBIOS computer name: %v", err)
+		return false
+	}
+	return strings.EqualFold(domain, name)
+}
+
+func netbiosComputerName() (string, error) {
+	buf := make([]uint16, windows.MAX_COMPUTERNAME_LENGTH+1)
+	size := uint32(len(buf))
+	if err := windows.GetComputerNameEx(windows.ComputerNamePhysicalNetBIOS, &buf[0], &size); err != nil {
+		return "", fmt.Errorf("GetComputerNameEx: %w", err)
+	}
+	return windows.UTF16ToString(buf[:size]), nil
 }
 
 // authenticateLocalUser handles authentication for local users

client/ssh/server/executor_windows_test.go

@@ -0,0 +1,48 @@
+//go:build windows
+
+package server
+
+import (
+	"errors"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.org/x/sys/windows"
+)
+
+// Past 15 characters the DNS host name and the NetBIOS name differ, and Windows
+// qualifies local accounts with the NetBIOS one.
+func TestIsLocalDomain(t *testing.T) {
+	const dnsHostname = "WINTESTMACHINE01XYZ" // 19 characters
+	netbios := dnsHostname[:windows.MAX_COMPUTERNAME_LENGTH]
+	require.NotEqual(t, strings.ToLower(dnsHostname), strings.ToLower(netbios),
+		"a 19 character name must not equal its 15 character truncation")
+
+	name := func() (string, error) { return netbios, nil }
+	unreadable := func() (string, error) { return "", errors.New("name unavailable") }
+
+	tests := []struct {
+		name        string
+		domain      string
+		machineName func() (string, error)
+		want        bool
+	}{
+		{"empty_domain", "", unreadable, true},
+		{"dot_domain", ".", unreadable, true},
+		{"truncated_netbios_name", netbios, name, true},
+		{"netbios_name_lowercase", strings.ToLower(netbios), name, true},
+		{"untruncated_dns_host_name", dnsHostname, name, false},
+		{"real_domain", "CORP", name, false},
+		// Must not resolve to local: that could authenticate the wrong account.
+		{"unreadable_machine_name", netbios, unreadable, false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			assert.Equal(t, tt.want, isLocalDomain(tt.domain, tt.machineName),
+				"classification of domain %q", tt.domain)
+		})
+	}
+}

client/ui/frontend/index.html

@@ -6,7 +6,25 @@
     <title>NetBird</title>
     <style>
       html, body { background: #181A1D; }
+      html:not(.dark), html:not(.dark) body { background: #F3F3F3; }
     </style>
+    <script>
+      // Pre-paint theme guard: apply the last-known theme before first render
+      // to avoid a flash of the wrong theme. ThemeContext keeps the mirror
+      // fresh from the persisted preference and the Go-reported OS appearance.
+      (function () {
+        try {
+          var pref = localStorage.getItem("nb-theme-pref") || "system";
+          var dark;
+          if (pref === "dark") dark = true;
+          else if (pref === "light") dark = false;
+          else dark = window.matchMedia("(prefers-color-scheme: dark)").matches;
+          document.documentElement.classList.toggle("dark", dark);
+        } catch (e) {
+          /* keep the default dark class */
+        }
+      })();
+    </script>
   </head>
   <body>
     <div id="root"></div>

client/ui/frontend/src/app.tsx

@@ -13,6 +13,7 @@ import { SkeletonTheme } from "react-loading-skeleton";
 import "react-loading-skeleton/dist/skeleton.css";
 import { welcome } from "@/lib/welcome";
 import LoginWaitingForBrowserDialog from "@/modules/login/LoginWaitingForBrowserDialog.tsx";
+import { ThemeProvider } from "@/contexts/ThemeContext.tsx";
 import { initI18n } from "@/lib/i18n";
 import { initPlatform } from "@/lib/platform";
 import { initLogForwarding } from "@/lib/logs";
@@ -35,30 +36,38 @@ Promise.all([
 ]).finally(() => {
     ReactDOM.createRoot(document.getElementById("root")!).render(
         <React.StrictMode>
-            <SkeletonTheme baseColor={"#25282d"} highlightColor={"#33373e"}>
-                <HashRouter>
-                    <Routes>
-                        <Route path={"dialog"}>
-                            <Route
-                                path={"browser-login"}
-                                element={<LoginWaitingForBrowserDialog />}
-                            />
-                            <Route path={"install-progress"} element={<UpdateInProgressDialog />} />
-                            <Route
-                                path={"session-expiration"}
-                                element={<SessionExpirationDialog />}
-                            />
-                            <Route path={"welcome"} element={<WelcomeDialog />} />
-                            <Route path={"error"} element={<ErrorDialog />} />
-                        </Route>
-                        <Route element={<AppLayout />}>
-                            <Route index element={<MainPage />} />
-                            <Route path={"settings"} element={<SettingsPage />} />
-                            <Route path={"*"} element={<Navigate to={"/"} replace />} />
-                        </Route>
-                    </Routes>
-                </HashRouter>
-            </SkeletonTheme>
+            <ThemeProvider>
+                <SkeletonTheme
+                    baseColor={"rgb(var(--skeleton-base))"}
+                    highlightColor={"rgb(var(--skeleton-highlight))"}
+                >
+                    <HashRouter>
+                        <Routes>
+                            <Route path={"dialog"}>
+                                <Route
+                                    path={"browser-login"}
+                                    element={<LoginWaitingForBrowserDialog />}
+                                />
+                                <Route
+                                    path={"install-progress"}
+                                    element={<UpdateInProgressDialog />}
+                                />
+                                <Route
+                                    path={"session-expiration"}
+                                    element={<SessionExpirationDialog />}
+                                />
+                                <Route path={"welcome"} element={<WelcomeDialog />} />
+                                <Route path={"error"} element={<ErrorDialog />} />
+                            </Route>
+                            <Route element={<AppLayout />}>
+                                <Route index element={<MainPage />} />
+                                <Route path={"settings"} element={<SettingsPage />} />
+                                <Route path={"*"} element={<Navigate to={"/"} replace />} />
+                            </Route>
+                        </Routes>
+                    </HashRouter>
+                </SkeletonTheme>
+            </ThemeProvider>
         </React.StrictMode>,
     );
 });